Advances in INORGANIC CHEM1STRY
Volume 46
ADVISORY BOARD
1. Bertini
D. M. P. Mingos, FRS
Universita degli Stud; d...
32 downloads
4426 Views
27MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Advances in INORGANIC CHEM1STRY
Volume 46
ADVISORY BOARD
1. Bertini
D. M. P. Mingos, FRS
Universita degli Stud; di Firenze Florence, ltaly
lmperial College of Science, Technology, and Medicine London, United Kingdom
A. H. Cowley, FRS
J. Reedijk
University of Texas Austin, Texas, USA
H. B. Gray California Institute of Technology Pasadena, California, USA
M. L. H. Green, FRS University of Oxford Oxford, United Kingdom
0. Kahn lnsfitut de Chimie de la Matiere Condensee de Bordeaux Pessac, France
Andre
E. Merbach
lnsfitut de Chimie Minerale et Analyfique Universite de Lausanne Lausanne, Switzerland
Leiden University Leiden, The Netherlands
A. M. Sargeson, FRS The Australian National University Canberra, Australia
Y. Sasaki Hokkaido University Sapporo, Japan
D.
F. Shriver
Northwestern University Evanston, Illinois, USA
R. van Eldik Universitat Erlangen-Numberg Erlangen, Germany
K. Wieghardt Max-flanck lnstitut Mulheim, Germany
Advances in
INORGANIC CHEMISTRY Including Bioinorganic Studies EDITED BY A. G. Sykes Department of Chemistry The University of Newcastle Newcastle upon Tyne United Kingdom
VOLUME 46
@
ACADEMIC PRESS
San Diego London Boston New York Sydney Tokyo Toronto
This book is printed on acid-free paper.
@
Copyright 0 1999 by ACADEMIC PRESS All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-1999 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0898-8838/99 $25.00
Academic Press u division of Hurcouri Bruce & Company
525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.apnet.com Academic Press 24-28 Oval Road, London NW I 7DX, UK http://www.hbuk.co.uk/ap/ International Standard Book Number: 0-12-023646-X PRINTED IN THE UNITED STATES OF AMEIUCA 98 99 0 0 0 1 02 0 3 B B 9 8 7 6
5
4
3 2
1
CONTENTS The Octahedral MsY8and M6Y,* Clusters of Group 4 and 5 Transition Metals NICHOLAS PKOKOPLIK AND D. F. SHRIVER 1. Introduction . . 11. Group 6 . 111. Group 5 Metal Halide Clusters . IV. Materials Chemistry Derived from Soluble Metal Halide Clusters References .
. .
. .
.
1 3 24 35 44
Recent Advances in Noble-Gas Chemistry JOHN H. HOLLOWAYA N D ERICG. H o w I. Introduction
.
Recent Review Literature . The Possibility of Argon Chemistry . Krypton Chemistry . Xenon Chemistry . VI. Radon Chemistry . References .
11. 111. TV. V.
. .
.
. . .
. .
51 53 54 55 61 91 93
Coming to Grips with Reactive Intermediates ANTHONY J. DOWNSA N D TIMOTHY M. GREENE Introduction . Reaction Intermediates: Nerve Centers of Chemical Reactions . Experimental Characterization of Reaction Intermediates: Retardation Experimental Characterization of Reaction Intermediates: Time-Resolved Methods V. Experimental Characterization of Reaction Intermediates: Flow and Other Methods . VI. Conclusions . . References
I. 11. 111. IV.
V
. .
101 106 107
.
136
.
155 163 166
,
.
.
vi
CONTENTS
Toward the Construction of Functional Solid-State Supramolecular Metal Complexes Containing Copper(1) and Silver(1) MEGUMU MUNAKATA, LIANGPINGWU, AND TAKAYOSHI KURODA-SOWA I. 11. 111. IV. V.
VI. VII. VIII. IX.
. Introduction Helical Frameworks . S----SContact-Assembled Frameworks . Hexagonal Frameworks and Graphite-like Structures Hydrogen-Bond-Assembled Frameworks . a-a-Interaction-Assembled Frameworks . Diamondoid Frameworks . Other Frameworks Based on Covalent Bonds , Concluding Remarks . . References .
.
. . .
. . . . . . .
174 176 192 204 219 228 240 251 292 293
Manganese Redox Enzymes and Model Systems: Properties, Structures, and Reactivity NEILA. LAW,M. TYLERCAUDLE, AND VINCENT L. PECORARO I. Introduction . 11. The Enzymes . 111. Structural Models IV. Physical Properties V. Reactivity . VI. Conclusion . References .
.
.
. . . . . . .
305 310 343 379 393 424 425
.
441 443 456 470 485 485
Calcium-Binding Proteins BRYANE. FINNAND TORBJORN DRAKENBERG I. 11. 111. IV. V.
Introduction . Intracellular EF-Hand Calcium-Binding Proteins . Calcium-Mediated Membrane-Binding Proteins . Extracellular Calcium-Binding Proteins , Summary . References .
. .
. . .
vii
CONTENTS
Leghemoglobin: Properties and Reactions MICHAEL J.
D A V I E S , CHRISTEL M A T H I E U , A N D
. I. Introduction 11. Structure . 111. Biological Localization . IV. Reactions with Different Molecules . V. Oxidation of Fe(I1) Leghemoglobin . VI. Oxidation of Fe(II1) Leghemoglobin . VII. Reduction of Fe(IV)=O Leghemoglobin VIII. Reduction of Fe(II1) Leghemoglobin . Reactions of Globin-Derived Radicals References .
ALAINPUPPO 495 497 499 501 507 511 519 524 527 538
. . .
Ix.
INDEX . CONTENTS OF PREVIOUS VOLUMES
. .
.
543 555
This Page Intention ally Left Blank
ADVANCES IN INORGANIC CHEMISTRY, VO1.
46
THE OCTAHEDRAL M6Y8 AND M6Y12 CLUSTERS OF GROUP 4 AND 5 TRANSITION METALS NICHOLAS PROKOPUK and D. F. SHRIVER Department of Chemistry, Northwestern University, Evanston, Illinois 60208
I. Introduction 11. Group 6 A. Synthesis of Group 6 Clusters B. Axial Ligand Chemistry C. Inner Ligand Chemistry D. Redox Chemistry and Photophysics of the Group 6 Metal Halide Clusters E. Molecular and Electronic Structure of the Group 6 Metal Halid Clusters 111. Group 5 Metal Halide Clusters A. Synthesis of Cluster Core B. Redox Chemistry of the Group 5 Clusters C. Ligand Substitution D. Electronic and Molecular Structure E. Niobium Iodide Clusters {Nb,I,}"+ IV. Materials Chemistry Derived from Soluble Metal Halide Clusters A. Higher Nuclearity Clusters B. Supported Cluster Materials C. Charge-Transfer Salt Complexes D. Extended Solids E. Chemically Modified Surfaces References
I. Introduction
The early transition metal halide clusters (MfiY8}"+(M = Mo, W; Y = C1, Br, I; M = Nb; Y = I) and (MfiYl2}"+(M = Nb, Ta; Y = C1, Br) are the basis for a diverse solid-state chemistry in which cluster fragments are interconnected by halide bridges. Discrete clusters containing the {MsYs}"+and {MfiYln}"+ units are obtained from these materials by addition of ligands that disrupt the halide bridges. Control of the coordination environment about the metal octahedron in the soluble clusters provides control of the chemical and physical properties 1 Copyright C 1999 bv Academic Press All lights 01 reproduction in any farm reserved os9s-smwY $25 on
2
PROKOPUK AND SHRIVER
of the cluster. In addition, the discrete clusters provide building blocks for the construction of new types of extended cluster arrays and cluster based materials. Octahedral clusters of the electropositive metals, groups 3 to 7, are stabilized by a-donor ligands such as halides, chalcogenides, and alkoxides, but the majority accessible to solution chemistry are the halide complexes. These highly symmetric and aesthetically pleasing {M6Y8}"' and {M6Y12}n+clusters contain a robust core of six metal atoms and either 8 face-capping {M,Y8}"+or 12 edge-bridging {M6Y12}"+ halide ligands Y. An additional 6 ligands X are terminally bound, one to each metal center, completing the structures [M6Y8Y61n-6z and [M6Y12&61n-6Z (Fig. 1).In the present discussion it is convenient to refer t o the charge of only the {M6Y8}"+ or {M6Y,2}"+ unit instead of referring to the entire cluster and its full complement of axial ligands. This notation is especially useful when there is ambiguity as to the type and charge of the ligands occupying the axial sites or when the properties of the cluster core are independent of the axial ligands. Nomenclature developed by Schafer and von Schnering (1) denotes the bridging ligands as inner (superscript i) and the terminal ligands outer (from ausser) or axial (superscript a); thus, [Mo&~,,]~is noted [Mo6Cl!!Ckjl". Soluble octahedral clusters {M6Y8}ntand {M6Y12}n+ with inner halide ligands have been reported for Nb, Ta, Mo, and W. Molybdenum and tungsten are observed predominately in the {M,Y,}"+ geometry, and the group 5 metals are found with the {M6Y12}nt arrangement. One notable exception is the niobium iodide clusters such as "b618(NH2CH3)6]in which the iodide ligands occupy face-capping sites on the Nb, octahedron (2). Solid compounds with {M6Ys}"+or (M6Yl2)"' units and halide ligands have been observed for other metals, including those of the rare earth
0
0
A
B
OM
QY
OX
FIG. 1. Structures of [MQY&l' A and [M6Y1&10B.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
3
Am Cm Rk Cf Es Fm Md No Lr FIG.2. Periodic table of metals (shaded) found in {MM,Y,P'or {M,Y,J"+ geometry.
and late d-block elements (Fig. 2). Extensive intercluster bridging has confined the chemistry of the majority of these compounds to the condensed phases. By contrast, the weak intercluster bridging in the group 5 and 6 derivatives provides solid-state materials that are easily dissolved into molecular cluster species. Subsequent solution chemistry has revealed rich photophysical and redox properties for these compounds. The relationship between the ligation of the group 5 and 6 metal halide clusters and their chemical and physical properties, as well as the subsequent materials chemistry that has evolved from this association, is the subject of this review. The solid-state and solution chemistry of the related zirconium halide clusters {Zr,Y,,}"+, which occur with interstitial atoms occupying the center of the metal octahedron or hydrides ligands bridging the octahedron faces, is relatively new and has been the subject of a number of reviews ( 3 , 4 ) .
11. Group 6
Recent interest in halide clusters containing the {M,Y8}"+ unit stems in part from its structural relationship to the superconducting Chevral phases PbMosQs (Q = S, Se, Te) in which eight chalcogenide ligands Q occupy the face-capping positions ( 5 , 6). Early development of the ligand substitution chemistry of the group 6 clusters had led to recent applications of these compounds to catalysis and materials
4
PROKOPUK AND SHRIVER
chemistry. Molybdenum(I1) chloride was discovered over 100 years ago (71,and the related cluster M O ~ C ~ ~ ( O H ) ~was ( H one ~ O )of~ the first complexes characterized by X-ray diffraction (8).Renewed interest in the substitution chemistry of the inner ligands arises from the promise of generating {MO&g}2- (& = s, Se, Te) units in solution, which are potential precursors to Chevral phases. Also of current interest is the discovery of the long-lived excited state exhibited by the molybdenum and tungsten clusters (9, 10). A. SYNTHESIS OF GROUP6 CLUSTERS Early synthetic procedures for the group 6 metal halide clusters relied on disproportionation reactions of the intermediate halides, MoC13,MoBr3, WCl,, and WBr4, at elevated temperatures to produce the clusters MsXlz(11,12). These reactions suffered from inherently poor yields and the need to synthesize and isolate the reactive intermediate metal halide. McCarley and co-workers ( 1 3 ) introduced the use of sodium tetrahaloaluminate melts to reduce the halides MoX5 and WX, (X = C1, Br) with aluminum at lower temperatures: 6MoC15+ 6Al- Mo6Cllz+ 6AlC1, 6WCls + 8Al+ W&!l,
+ 8AlC1,.
(1) (2)
Near quantitative yields are achieved for the molybdenum halides, but substantially lower yields (50%) are obtained in the synthesis of the tungsten clusters, which require higher temperatures and the use of Vycor reaction vessels. A higher yield low-temperature synthesis was devised by McCarley and co-workers for W6C1,2using iron as the reducing agent (14). The chemical reduction of the higher Mo and W halides provides good yields of the octahedral clusters, but the mechanism is obscure. By contrast, chemical oxidation of zero-valent Mo and W leads t o the bromo and iodo cluster species in poor yields but provides considerable insight into the formation of these cluster compounds. The reaction of the hexacarbonyls Mo(CO)$or W(C0l6 with Iz at moderately low temperatures produces a mixture of metal halide phases (15, 16). In the reaction W(CO), with Iz, lower nuclearity clusters have been isolated as reaction intermediates that lead to w6 species. The tri-, tetra-, and pentanuclear tungsten iodide species are obtained from W(CO)s and I2 by varying reaction times and temperatures, and the
OCTAHEDRAL. CLUSTERS OF TRANSITION METALS
5
following reaction sequence was proposed by Franolic, Long, and Holm:
The lower nuclearity clusters [ M O ~ C I ~were ~ I ~isolated from MoClSin AlCl,/KCl/BiCl,/Bi melts that led to the hexanuclear species ( 17). This is the only intermediate isolated from reaction mixture that pro~ } ~ Solution + chemistry has yielded Moq, Mag, duces the { M O ~ C ~core. and M o ~halide species from mononuclear molybdenum complexes, suggesting that a nucleation process similar to that proposed for the tungsten systems may take place during the formation of hexanuclear molybdenum clusters ( 18-22). Extraction of the cluster from the products of the solid-state reactions with HX yields the hydronium salts of the molecular species (H30),(M&&).6H,0, which are converted into the polymeric material M 6 m & under vacuum at 200°C (12, 23). The structure of M&X$X&, also designated MGX12, consists of a {M6&}4+ core with two terminal axial ligands X2” and four axial ligands bridged to neighboring clusters Xjp2(11)(Fig. 3). Both M a l zand (H3O),(M&&).6H20 may serve as precursors to new molecular species containing the metal halide cluster core (MO&}~’. B. AXIALLIGAND CHEMISTRY The ligand substitution chemistry of the axial ligands of the group 6 clusters is considerably more developed than the substitution chem-
FIG.3. Structure of Mo,C1,C12Cl,, . Reprinted with permission from Ref. (130).
6
PROKOPUK AND SHRIVER
istry of the inner ligands. This is due largely to the robust nature of the {M6Y8}*+core, which requires substantially harsher conditions to displace the inner ligands. The increased lability of the axial ligands enables the coordination environment about the {M6Y8}4t framework to be altered without disrupting the metal octahedron on the Y8 cube.
1. Mechanism of Ligand Exchange Kinetic studies of the ligand exchange reaction
were performed by Sheldon with 36C1-labeled[Mo6C18C1612in aqueous solutions. The rate-determining step was found to be first order with respect to the molybdenum cluster and independent of halide concentration, and hence the aquation step
was proposed as rate limiting (24). Schafer found similar results with the exchange reactions of [Mo6Cl8Cl6I2and [w6c18c1612-with Br- and I- (25).Two significant findings from Schafer’s work are the equivalency of the six axial positions and the rate dependence on the metal itself, which show that reactions of molybdenum clusters proceed faster than those of the analogous tungsten complexes. Preetz substantiated the equivalency of the axial positions for the ligands F-, C1-, Br-, I-, and NCS- by monitoring the 19Fand 15Nand 95MoNMR spectra of the ligand exchange reactions
and
Reaction rates were found to increase with the series X = C1 < Br < I < SCN < F, equilibration of the thiocyanate cluster was achieved in 10 h, and 2-3 days were required for the formation of the fluoride cluster at room temperature (26,271. All the clusters [ M O ~ C ~ ~ F ~ - ~ X I ~ and [Mo&~,(NCS)~-,X,]~including structural isomers Ke., mer or fac of [Mo&18F3X3I2> were identified by NMR. In all cases a statistical
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
7
distribution of all possible compounds, inclusive of the ratios for the structural isomers, was obtained. Importantly, no evidence of isomersteering electronic effects was observed. 2. Synthesis of Molecular {M6Y8}Clusters
Various types of neutral ligands, including nitrogen donors (12,23, 28, 29), oxygen donors (12, 28, 30, 311, phosphines (28, 32, 331, and solvent molecules (28, 34, 351, disrupt the extended MosCllz and WsCllzsolid structure. Under harsher conditions terminal chloride ligands of MosCllzare displaced by neutral donor ligand, L:
For instance, addition of PPh3 to MosCllz yields M o , C ~ , C ~ , ( P Pat ~~)~ room temperature, but it is necessary to reflux MosCllzwith an excess of PPh3 in tetrahydrofuran to produce the triphosphine cluster [ M O ~ C ~ ~ C ~ ~ ((28). P P ~Similar ~ ) ~ I Cexperiments ~ by Walton and coworkers with PEt3 and Mo6ClIzin refluxing ethanol resulted in the reduced cluster M o ~ C ~ , C ~ , ( P Ewhich ~ , ) ~ , contains a {Mo,C~,}~+ core. Magnetic susceptibility measurements conducted on M o , C ~ , C ~ , ( P E ~ ~ ) ~ indicate that the cluster is diamagnetic, a result inconsistent with the expected odd number of electrons in the {Mo,C~~}~+ core. The compound was subsequently reformulated as the reduced species [ M O & ~ ~ ( P E ~ ~which ) ~ ] ~ + ,crystallizes with the counteranion [Mo6Cl8Cl6l2-(321. The introduction of two triphenyl phosphine ligands increases the lability of the remaining chloride ligand in the biphosphine cluster Mo6C18C1,(PPh3),. Both ethanol and pyridine dissolve Mo6Cl12,forming the solvent complexes Mo6C18Cl,(solvent)z,but dissolution of the biphosphine cluster M O ~ C ~ , C ~ , ( Pin P ~these , ) ~ solvents to leads to the species [Mo,C1,C1z(PPh3)2(HOEt)212+ and [MosCl&12(PPhs)A py)z12+ (281. Neutral clusters with the formula Mo6ClsCl4L2can exist as cis or trans isomers. Differences in polarity of the cis and trans structures were employed by Saito and co-workers to separate the isomers of [Mo6C1&1,(P(n-C3H7)3]2by chromatography to produce 3 and 15% yields of the cis and trans respectively ( 3 3 ) .The less polar trans isomers of MO&!~&~~(PR~)~ (R = n-C,H9, n-C,HI1) have also been isolated in greater than 40% yields (Fig. 4). Shifts in the 31PNMR spectra reveal a preference for the formation of the all-trans isomer. DiSalvo and co-workers effected the isomerization of trans-MosC18C14[P(nC5Hl113]2into ~ ~ ~ - M O ~ C ~ ~ C ~ by, immersing [ P ( C ~ Hthe ~ )trans-P(n~~~ C,H,1)3 cluster in a THF solution of P(C,H,), (Fig. 5). A 95% conversion
8
PROKOPUK AND SHRIVER
ii
FIG.4. Structure of tralzs-MofiClsC1,tP(n-C4Hs)r12.
to the cis isomer was obtained (36).Alkylation of the trans phosphine clusters with trialkylaluminum reagents AIR; (R’ = CH, , CzHs, n-C3H7,n-C4H9,and n-C6HI3)leads to the replacement of two axial chloride ligands with R groups, M O & ~ & ~ ~ ( P (R’Iz R , ) ~(Fig. 6). Additional AIR; reacts with M O ~ C ~ ~ C ~ ~ ( P R ,producing ) , ( R ’ ) ~ the trialkyl M O ~ C ~ ~ C ~ ( P R ~and ) ~ (both R ’ ) mer ~ , and fac isomers were shown to be present by 31PNMR. These organometallic derivatives are the only examples in which carbon atoms are directly bound to the metal atoms of the { M o ~ C ~framework. ~}~+ Addition of two equivalents of sodium methoxide to Mo6ClI2in methanol leads to [ M O ~ C ~ ~ C ~ ~ ( O(37). C H Single-crystal ~)~]~X-ray diffraction reveals that the methoxide ligands occupy axial positions in a trans geometry (Fig. 7). However, it is not clear whether the trans
FIG.5. Structure of cis-MofiC1,C1,[P(C,H,),l,.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
9
FIG.6. Structure of all-trans-Mo~Cl,CI,(C2H,),[P(C,H,),I,.
isomer is formed exclusively or only crystallizes preferentially because neither solution NMR nor X-ray powder pattern data on the bulk material was reported. In contrast to the bisphosphine clusters MO,$~~C~,(PR,), , in which both electronic and steric factors may affect the cisltrans ratio of the products, only electronic effects are likely to influence the formation of the trans isomer of [Mo,Cl,Cl,(OCH,),]”. The trans isomer of [ M O ~ C ~ ~ C was ~ ~ Bprepared ~ ~ ] ” from crystalline MoGClHC14(H20)2 in a heterogeneous reaction with Br-. Homogeneous mixtures of Mo6C1RC1,(H20)2and LiBrl[(CfiH5)4AslBrlead to bromide-rich clusters [(CfiH5)4As12[Mo6C18Cll,yBr4.81 after only 30 min (35).
FIG.7. Structure of trans-[M0,CI,C1,(0CH,,),]~
10
PROKOPUK AND S H R I W R
Mixed halo clusters (H30)2M6C&Y6 (M = Mo, W; Y = F, Br, I) have been prepared from the aqueous acids HY and Mo6Cllz(12, 27, 32, 35). It is likely the substitution results from an excess of the acid, HX, rather than the relative affinities of the { M O ~ Ccore ~ ~ for } ~ the ~ halide, X-, ions. Dehydration of (H30)z[M6C18Y6] leads to the mixed ligand bridged systems Mo6C18YzY41a, which are analogous in structure and reactivity to MoGCllZin that addition of 2L produces Mo6Cl8Y4L2. One of the first comparative studies of the mixed halide clusters MosC18YzY41z (Y = C1, Br, and I) revealed that the axial iodide ligands of M O ~ C ~ , are I ~ Imore ~ , ~ susceptible to displacement by nitrobenzene than the terminal bromide or chloride lands of Mo6C18YzY41z (Y = C1, Br) ( 2 3 ) .Dissolution of Mo6C181z141z in nitrobenzene produces and the bromo and chloro derivatives yield bisnitrobenzene complexes M O ~ C ~ , Y ~ ( N ~ ~ C ~ H ~ ) ~ . Strongly basic anions such as hydroxide and alkoxide easily displace the axial chloride ligands of [Mo6C18C1612-;thus, titration of [MO~C~,C~,]~with OH- initially precipitates the tetrahydroxo cluster M O ~ C ~ ~ ( O H ) ~14H20, ( H ~ O )which redissolves as the hexahydroxo cluster [MO&l&OH)6]2-with increasing pH (23). Similarly, when Mo6Cllz is refluxed with sodium alkoxides in alcohol solution, the alkoxide substitutes for all of the axial chloride ligands (35, 3 9 ) . Methoxide, ethoxide, and pentafluorophenoxide clusters [MO~C~,(OR)~]~have also been generated by this method (35, 38). Increased lability of the axial positions has been achieved by introducing weakly coordinating ligands such as triflate (trifluoromethanesulfonate) (40, 41 1, tetrafluoroborate (26, 42-45), nitrate (43, 46, 4 7 ) tosylate (401, trifluoroacetate (40,481,and perchlorate ( 3 4 ) .(Caution: perchlorate complexes are hazardously explosive.) Reactions of the chloro cluster [Mo6C18C1612with the silver salt or the protonated form of the weakly coordinating base B- generate the reactive species [MO&18B6]2-:
Only the axial chloride ligands were abstracted by the silver salts or acids, leaving the inner ligands and the (MO,C~,}~~ core intact. The weakly coordinating ligands are easily displaced by anionic, X-, or neutral ligands, L, generating the hexasubstituted clusters Cotton and Curtis first employed this [Mo6C1&l2- or [Mo~C~,L,]~+. strategy with AgC104 to prepare the first tetracationic clusters with
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
11
the ( M O ~ C ~ ~unit, } ~ ' [ M ~ ~ C l ~ ( s o l v e n t ) ~ l ( C(solvent l O ~ ) ~ = DMF or DMSO) (34). These compounds demonstrate the utility of using weakly coordinating ions to alter the coordination environments about the { M O ~ C ~core, ~ } ~but + they are potentially explosive. The triflate derivative [ M O ~ C ~ ~ ( O S O ~(Fig. C F 81, ~ ~ has ~ I ~been used extensively in our laboratory to prepare new cluster derivatives with neutral ligands not previously observed to bind to the ( M O ~ C ~core ~}~' in all six axial positions. Limits of the substitutional lability of the triflate ligand of [Mo~C~,(OSO,CF~),~~have been explored with 19F NMR (49). Diagnostic shifts in the 19FNMR signal of coordinated and uncoordinated (free) OSOzCF3- may be employed to monitor the displacement reactions. As expected, only one peak is observed in the 19F NMR spectrum of [Mo6C18( OSOzCF3)6]"in noncoordinating methylene chloride. Two peaks are observed with THF, acetonitrile, and acetone, suggesting incomplete substitution of the triflate ligand by solvent molecules. A single peak is observed for [MO&18(OSO$F3)6]2- in both DMSO and methanol, indicating that the triflate ligands are completely displaced in these solvents (49). Titration of [Mo6C18 (OS02CF3)6]2in methylene chloride with phosphine ligands PR3 (R = Et, t-butyl, Ph) generates all possible species [MosCls (OS0zCF3)6-x(PR3),1'2-"'( 0 5 x 5 6) as determined by 31PNMR. Unlike the ligand exchange reactions between [MO&18F612- and [Mo6C18Y612(Y = C1, Br, I, and NCS), in which a statistical distribution of products was formed (26,271, the x = 2 species is disfavored in the phosphine titration reactions. This may be due in part to the polar CH,Cl, solvent, which may disfavor the neutral species. Interestingly species
13 FIG.8. Structure of ~Mo6Cl8(OSO,CF3),]'
12
PROKOPUK AND SHRrVER
with x < 6 persist in the presence of excess of the phosphines P[C(CH3)3]S and P(C2H5)3( 4 9 ) .It is unclear whether steric or electronic factors prevent full substitution of the triflate ligands by the phosphines. The acetonitrile ligated cluster [ M O & ~ ~ ( N C C His ~ )an ~ ]even ~ + more labile cluster than DiSalvo and co-workers prepared the hexaacetonitrile complex as the [SbCl,]- salt by treating [ B U , N ] ~ [ M O , C ~ , ( ~ S O ~ C Fwith ~ ) ~ ] KSH to generate [Bu4N12 [MO~C~~(SH which ) ~ I , reacts with SbCl, in CH3CN to produce [ M O ~ C ~ ~ ( N C C H ~ ) in ~ ] less [ S ~than C ~ ~5% ] ~ yield (50).Some decomposition of the { M O ~ C ~core ~ } ~ 'is . required to provide the sixth chlorine atom for the formation of the [SbCl,]- counteranion. A similar abstraction of inner chloride ligands may have occurred in the alkylation reactions of the phosphine clusters with trialkyl aluminum reagents studied by Saito ( 3 3 ) . The [Mo&~&NCCH&~~+ cation also can be prepared in near quantitative yield by reaction of the azide ligands of [ M O ~ C ~ & Nwith ~ ) ~NO' ] ~ - in acetonitrile; however, the CH3CN ligands are removed under vacuum and the BF4- ions then bridge molybdenum centers to fill the coordination sphere of the {Mo~C~,}~+ units ( 4 9 ) . The methoxide derivatives [MO&~&OCHS),#- and [ M o , C ~ & ~ ~ (OCHS)212provide an alternative strategy for generating new clusters with the { M o & ~ ~ core } ~ + (51-55). The strongly basic OCH,- ions are easily abstracted by Bronsted acids, liberating methanol. Nannelli et al. (54)first demonstrated the lability of the methoxide ligands with metathesis reactions between [Mo~C~,(OCH,)~]~and phenol:
The trans isomer of [Mo,C~,(OCH,A~)~]~(Ar = g-anthracenemethano11 was prepared from trUnS-[MOsC18C14(0CH3)212- (Fig. 91, indicating that little or no isomerization or ligand exchange between clusters occurs (37).This strategy is convenient in that readily available proton sources including alcohols, phenols, thiols, and carboxylic acids can be used, and the only by-products are solvent molecules. Other organic ligands with pK, values comparable to phenol, such as sulfanamides, benzamide, pentafluoroanaline, acetylacetonate, and nitromethane, do not undergo the expected metathesis reactions with [MO&l~(OCH&12-(56). Interestingly, the hexaphenoxide derivative [ M O ~ C ~ ~ ( O C reacts ~ H ~ ) with ~ ] ~ -benzoic acid to produce the benzoate cluster:
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
13
e FIG.9. Structure of trans-lMo,Cl,CI,(OCH,C, ,HqI2lL
However, no reaction is observed between [Mo&!~~(OC~H,),]~and sodium benzoate, suggesting that formation of the alcohol or phenol is the driving force for these reactions (56). C. INNERLIGANDCHEMISTRY 1. Mechanism of Ligand Exchange
Substitution of the inner ligands of {MO&!~~}~+ and {WtiC18}4+ clusters requires more forcing conditions than those used t o exchange axial ligands. Neither boiling aqua regia nor fuming sulfuric acid affect the {Mo6ClXJ4+ core (57). Basic solutions, however, disrupt the cubic array of inner ligands. The inner halide ligands of [ M O ~ C ~ ~ ( O Hand )~I” [Mo,Br,(OH),I2- are susceptible to partial displacement by hydroxide ions, [MO&-,(OH),(OH)~I’~. Kinetic studies of the substitution of the inner chloride and bromide ligands by OH- reveal that the process occurs with a second-order rate law, which is first order with respect to both cluster and the hydroxide ion. Interestingly, substitution of by OH- was found to be autothe bromide ligands of [Mo6BrH(OH),I2 catalytic, suggesting that steric factors play a role in displacement of the bromides (58,59).
2. Inner Ligand Exchange
As implied in the previous discussion, the inner ligands have lower lability than their outer counterparts, but full substitution of the in-
14
PROKOPUK AND SHRTVER
ner chloride ligands by methoxide ions can be effected by boiling solutions of sodium methoxide and MosCllzto dryness. The resulting cluster, Naz[Mos(OCH3),(0CH3)6], which has face-capping methoxide ions, is pyrophoric (39, 60). Briickner et al. prepared mixed inner ligand species by heating the mixed halide clusters MO&l&&/z (X = Br, I) to 400°C (44, 45). The resulting mixtures of clusters were converted to the fluoride derivatives [Bu4NIz[Mos(Y,C1,-,)F6] and analyzed by NMR spectroscopy. Deconvolution of the 1D and 2D 19FNMR spectra revealed a mixture of products and structural isomers, all of which were identified in the NMR spectra. A statistical distribution of products is obtained from A contrasting result was observed when Mo6Cl~BrzBr41z. was heated to produce the z = 3 and 4 species as the most abundant species. Complexes containing iodide ligands in close proximity are favored. Complete exchange of both inner and outer chloride ligands of M&1&1zC14/~ (M = Mo, W)occurs in fused salt mixtures of LWKX (X = Br, I) (12, 61). The similarity between the structure of {MosC18}4+ and the MosQs (Q = chalcogenide) unit of the superconducting Chevral phases has stimulated interest in the substitution chemistry of the inner ligands and, in particular, substitution of the inner chloride ligands by a chalcogenide. A number of mixed halide-chalcogenide clusters {Mo6(Y8-,&)}(X = C1, Br, I; Q = S, Se, Te; z = 1, 3, 4, 6, 8) have been prepared from Mo6C18C1zC141z and the elemental chalcogenide at elevated temperatures (9OOOC) (62-64). The clusters are linked into extended arrays through both inner and outer bridging ligands {Mo6(Y,-,Qz)}(Fig. 10). Diamagnetic and dielectric properties of these solids reveal insulator to metal to superconductor transitions upon increasing chalcogenide content, halide size, and intercluster bridging.
84 FIG.10. Intercluster bridging between {MQYJ”’ clusters. The vertices of the cube represent the inner ligands Y.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
15
Lower temperature methods for introducing chalcogenides into the face-capping positions of MosCllzwere developed in McCarley's laboratory, where MosClle in pyridine was refluxed with NaSH. The resulting soluble mixed-capped cluster (pyH),[Mo,(C1,S)Clsl was obtained in 41% yield (65). Saito used similar conditions followed by extensive chromatographic separation to prepare the mixed halidechalcogenide clusters [Mo6(Y7Q)&I3-(X = Y = C1, Br; Q = S, Se) and the disubstituted species [ M O , ( C ~ ~ S ~ ~ ) C(66, ~ ~ I67). " ' ~ -The chemistry of these interesting chalcogenide clusters has not been pursued. The observation that chromatography of the clusters with a chloride-deficient eluent produces the aqua complex [Mo,(C~,Q)(H,O),~~+ indicates that the axial chloro ligands of the chalcogenide-doped clusters are more labile than those of the all-chloro complex [Mo6C18C1612-. Tungsten clusters with inner halide-chalcogenide ligands have been prepared by the reaction of W6Cllzand AzSe (A = Na, K) (68). When W6Cl12and a n alkali metal chalcogenide are refluxed in toluene with triethyl phosphine, clusters with mixed ligands are produced, [Ws(Se,C18-,)(PEt,),]1t ( z = 7.0, A = Na; z = 6.4, A = K). Slight changes in reaction conditions or chalcogenide source significantly affect the halide/chalcogenide ratio of the substituted product. Full substitution of the inner chloride ligands of both MosCllz and W,Clle by chalcogenides produce molecular clusters with the {M6Q8}n+ cores. When W,Cl,, and elemental sulfur are heated to 300"C, a n uncharacterized cluster results, which reacts with magnesium in the presence of triethylphosphine to generate W6S8(PEt3),in 10% yield (69). Better yields for {MsS8}"+(M = Mo, W) are obtained when M,Cl12,NaSH, and NaOBu are refluxed in pyridine. In these reactions the NaOBu serves as a proton abstractor and the chalcogenide clusters are isolated as the pyridine adducts ( 1 4 , 70, 71). The chemistry and structural variations of the group 6 metal chalcogenide clusters were recently reviewed in this series, and the reader is referred to this publication for further information on these compounds (72).
D. Redox Chemistry and Photophysics of the Group 6 Metal Halide Clusters The robust nature of the (MO,C~~}~+ core in combination with its photophysical and redox properties is ideal for solar energy storage (73). Voltammetry on [Mo6Cl8Cl6I2in acetonitrile reveals two l-electron steps at - 1.53 and 1.56 V (vs SCE), corresponding to generation of the oxidized and reduced species [Mo6C1&!ls]'- and [Mo6Cl8Cl,l3-, respectively ( 74). Electrolysis of [Mo6ClsCl6l2-a t potentials more posi-
16
PROKOPUK AND SHRIVER
tive than that of the 2-/1- couple generates the oxidized cluster [Mo6C18C16]'-,with no evidence for decomposition or any new electroactive species (75). Coulometric measurements are consistent with quantitative one-electron oxidation of the cluster; however, the oxidized species [Mo6C18C16]'with a {MO,C~,)~' core was not isolated (10). Voltammetry experiments on [Mo6Cl8Clsl2in basic chloroaluminate melts reveal a reversible oxidation couple corresponding to generation of [Mo6C1,C161'-. In contrast t o the voltammetry experiments in acetonitrile, the [ M O ~ C ~ ~ C ~ ~ ] ~ ~couple / [ MinO the ~ Cionic ~ , Cmelt ~ ~is]chem~~ ically irreversible, presumably due to dissociation of axial chloride ligands (76). The oxidized clusters [MsY&]l- (M = Mo, W; Y = C1, Br, core have been generated electroI; X = C1, Br, I) with the chemically in organic electrolytes (Table I) (10,77, 78). The tungsten clusters [wsY&]'- (X, Y = C1, Br) have been produced by chemical oxidation of [W6Y&l2- with NO+ (79). Oxidation of WsBr12 with liquid bromine in a net two-electron process, and the extended structure, WSBr8Br4(Br4)2,2, results consisting of {W6Br8}6+ units bridged in a linear array by (Br4Y anions (80, 81).The rearrangement of inner halide ligands to edge-bridging positions takes place with w&l&12c14/2 at elevated temperatures in the presence of chlorine (821. The resulting core {w&l12}6+contains 12 edge-bridging chloride ligands, and 6 axial chlorides complete the structure of W&ll&l6. A n alternative route to w6c1&16 employs octachlorocyclopentene, C5C18,as both the oxidant and the chloride source (83). Oxidation of M o ~ C ~ ~ C ~by, C C,C& ~ ~ ,leads , to M0,$1&13, which can be converted to the tetraethylammonium salt of [Mo6Cl12C1613(83). The clusters w6c1&16 and [Mo6Cl1,Clsl3-are the only known examples of TABLE I REDOX POTENTIALSO
FOR
1.60
1.38 1.14 0.99 0.93 0.80
0.57 0.56
SOME{MBY,P+CLUSTERS
- 1.56
I0 10 10 78 77 78 78 78
All values were obtained from voltammetry experiments in CH,CN and are referenced t o SCE.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
17
group 6 metal halide clusters adopting the {MHY1l}’lt geometry. Conversion of the {M6YR}4+ core to {M,Y,,}”+has only been observed in these reactions, and no examples are known in which the reverse transformation from {M6Y12}”T to {M6Y8}4+ occurs. Clusters containing chalcogenide ligands {Mo6(ClR-2Q2)}f1+ are somewhat more substitutionally labile than the all-halo complexes and undergo oxidation at potentials about 0.87 V less positive than that needed to oxidize the parent perhalo cluster (Table 1). The oxidation potentials of the chalcogenide-substituted clusters change very little with the identity of the chalcogenide, but significant solvent dependence is observed for the chalcogen-containing clusters. The cyclic voltammograms of [ M O , ( C ~ ~ S ) Cin~ ~the ~ ” noncoordinating solvent methylene chloride reveal a quasi-reversible wave corresponding to reduction of the cluster a t 0.14 V vs. the ferrocenelferrocenium couple. In acetonitrile this wave is shifted to 0.28 V, indicating that partial substitution of the axial ligands by CH3CN has occurred. New waves grow in upon changing the potential on acetonitrile solutions of [Mo6(C17S)C161”,suggesting that further changes in the coordination environment of the {Mo,(C17S)}”+core take place upon switching the oxidation state of the cluster (66, 84). In the early 1980s Gray and co-workers discovered an unusually long-lived excited-state lifetime, 180 p s , for the cluster anion [Mo&~,C~,]~-. The emission spectrum is substantially red shifted from the absorption bands (9) (Fig. 11).This is one of the longest observed lifetimes of any purely inorganic compound. Shifts in the near-UV absorption band upon variation of the Y ligands in { M o ~ Y ~and }~+ {W6Y8}4were initially taken as an indication that the transition originates from an (inner) ligand-to-metal charge transfer; however, the
,/-
FIG. 11. Absorption (full) and emission (dashed) spectra of IBu,NIL[Mo,C1&l,,I in CH,Cl,.
18
PROKOPUK AND SHRIVER
independence of the emission spectra of the {Mo&}~+unit on changing the identity of inner ligands suggests that the transition responsible for radiative decay is between purely metal-based orbitals (10,77, 85, 86). The {W6Y8}4+clusters display emission spectra that vary only slightly with changes in the inner and outer ligands. A summary of the photophysical data for various molybdenum and tungsten clusters is given in Table 11. The low rate constants for the red-shifted radiative decay are attributed to a spin- and symmetry-forbidden character of the transition. Also important to the low rate constants is the absence of high-energy vibrational modes capable of deactivating the excited state. Additionally, the halide ligands appear to shield the Mo6 and w6 core and thereby inhibit energy transfer t o solvent molecules. From the photophysical and electrochemical data on [ M O ~ C ~ & ~ ~ ] ~ - , a modified Latimer diagram can be constructed relating the excited cluster [ M o ~ C ~ , C ~to ~ ~the " * oxidized and reduced ground states ( 7 4 ) (Fig. 12). When [Mo6C1&l6I2-*is quenched with an electron donor such as benzoquinone, [ M O ~ C ~ , C ~is~ ] generated. ~Similarly, [Mo6C18C1612-* is quenched by the electron acceptor phenothiazine to produce [M0,C1&1~1'-.The long excited-state lifetime of [Mo6C1&1612-* and the strong reducing and oxidizing nature of the electronically TABLE I1
PHOTOPHYSICAL DATA^ OF {M6Yap+CLUSTERS Cluster
A,,~,,,
805
825
880 814 802 766 758 752 701 698 698
7(ps)
180 140 86 190 110 71 84 2.2 4.4 5.6 15 15 19 16 27 30
k , (103 s-1)
k , (103 s-1)
1.1
4.5
2.1
7.0
1.9 13 10 23 10 8 17 11 11 13
Values are reported with acetonitrile solutions of cluster.
10 650 490 310 93 59 50 89 34 20
Refs. 10, 85 89 89 89 10, 85 89 85 10, 77, 89 85, 89 77, 89 77, 89 77, 84 89 77, 89 77, 89 77, 84
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
pIO,CI,CI,]'~-
- I .J3
t1.30
19
vlo,CI,CI,]'
FIG.12. Latimer diagram of [Mo6C1,C1~I". Excited state energy in eV; electrode potentials estimated vs. SCE in CH,CN.
quenched clusters should be useful for converting solar energy to chemical energy ( 73). The potentials in the excited-state Latimer diagram demonstrate that the annihilation reactions of [Mo6C18C1613and [Mo6C18C161'-liberate enough energy to generate [Mo6Cl8Cl6l"* and electrogenerated chemiluminescence (ECL). This phenomenon has been used extensively to study electron transfer reactions between the excited states of a broad range of the halide clusters (M = Mo, W; X = C1, Br, I; Y = C1, Br, I) in combination with both electron donors and acceptors (78, 87, 88). Chemical oxidation of benzyl alcohol and 2-propanol by electrochemically generated [Mo6ClsC1611-has been demonstrated; however, photochemically generated [MosC18C161'~ is less effective for the oxidation of alcohols (75). The excited cluster [Mo6C18C1612-* is quenched by molecular oxygen through an energy transfer mechanism to generate singlet oxygen, which was identified by reaction with l-methylcyclohexene (891. E. MOLECULAR AND ELECTRONIC STRUCTURE OF THE GROUP 6 METAL HALIDE CLUSTERS 1. Electronic Structure Electronic structure calculations carried out at a variety of levels on these octahedral metal-halide clusters vary in detail (90-104). Our view of the metal-based orbitals responsible for metal-metal bonding in the {M6Y8}4+ unit has evolved very little since Cotton and Haas' semiempirical treatment in 1964 (90). Undoubtedly the importance of symmetry in the calculations has led to the consistent description of the bonding scheme. Only a brief summary emphasizing the relation between the proposed models to the physical and chemical properties of the clusters is presented here. For a more detailed discussion of the bonding numerous reviews are available ( 105-1091.
20
PROKOPUK AND SHRIWR
Orbitals responsible for the metal-metal bonding in {M6Y8}4+ have t2", and eg symmetry and are directed along the edges of alg,tlu,tZg, the M6 framework, generating 12 two-electron bonds. A total of 24 cluster bonding electrons, CBE (CBE number is equal to the number of valence electrons of the metal atoms minus the number of anionic ligands and the charge of the cluster) fill the bonding orbitals, creating "closed shell" diamagnetic compounds (Fig. 13). ESR measurements are consistent with the proposed diamagnetism of [M6Y8&I2(M = Mo, W) clusters and paramagnetism of the oxidized species CM~Y&l*-(10, 79). Some dispute persists as to the size of the highest occupied molecular orbital (HOMO)-lowest unoccupied molecular orbital (LUMO) gap in the closed-shell 24-electron species. This aspect of the molecular orbital diagram is of particular interest for its implications on the photophysical properties of the clusters. Electrochemical measurements are consistent with a large, 2- to 3-eV, separation; however, the exact nature of the excited state remains uncertain. 2. Molecular Structure These highly symmetric compounds provide aesthetically pleasing structures with near perfect Mo6and octahedrons that expand and contract to accommodate different coordination environments. Comprehensive studies of the structures of the halide clusters [Mo6Y&I2-
FIG.13. A molecular orbital diagram for {M,Y,}4+ based on a n Extended Hiickel calculation, adapted from reference (108).
21
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
(Y = C1, Br, I; X = F, C1, Br, I) by Preetz reveals that both electronic and steric effects of the inner and outer ligands influence the Mo-Mo, Mo-Y, and Mo-X bond lengths (42, 43, 110). For a given {Mo,Y,}~+ (Y = C1, Br, I) core, the Mo-Mo bond length increases with the axial ligands series X = F < C1 < Br < I (Table 111). Expansion of the Mo6 octahedron is attributed to changes in the electronegativity of the axial ligands. The 95Mo NMR spectra of the various halide clusters reveal that resonances occur at the lowest field for complexes having six axial fluoride ligands. A high field shift occurs with increasing polarizability of the axial ligands, indicating that the electron density on the metal centers increases with decreasing electronegativity of the outer ligands. Because this increased electron density coincides with an expansion of the Mo6 octahedron, it is reasonable to assume that the added electron density increases occupation of the antibonding orbitals of the Mo6 framework. In addition to increasing the electron density on the metal centers, changes in the identity of the axial ligands X from F to I also result in a reduced HOMO-LUMO gap as the ligand field strength of X decreases. The 95MoNMR shift of the halide clusters varies linearly with the electronegativity of the axial ligands (110), and a similar relationship exists between the 95Mo NMR shift and the Mo-Mo bond length (Fig. 14). TABLE I11 SELECTED BONDLENGTHS" OF SOME LMo6Y8&]*-CLUSTERS Cluster
d(Mo-Mo)
2.593 [Mo,CI,F,I" [MofiC1$ls12 2.602 [ M O ~ C I ~ B ~ ~ ~2.604 ~ ~ [Mo~CI,IJ2.615 [MotiBr,F61*2.618 [MosBrxClsl" 2.636 [MotiBr8Brtil" 2.640 [MotiBr,I,l" 2.644 [Mo&F,I" 2.650 [Mo,I,CI,]'~ 2.655 [MotiIRBrhlZ. 2.670 [Mo&I,I'2.675 IMos(CI$)CIJ 2.609 [MoB(C17Se)ClJ3 2.616 [Mo6(Cl6Se,)C1,1" 2.610 ~
" All distances are in A.
d(Mo-Y)
d(Mo-X)
d(Y-Y)
d(Y-X)
Ref.
2.488 2.469 2.465 2.466 2.622 2.604 2.600 2.598 2.798 2.775 2.775 2.767 2.479" 2.501b 2.513"
1.993 2.420 2.565 2.788 2.012 2.451 2.491 2.815 2.008 2.460 2.616 2.846 2.468 2.471 2.496
3.517 3.489 3.483 3.484 3.710 3.680 3.674 3.673 4.006 3.923 3.924 3.911
3.234 3.523 3.629 3.801 3.305 3.591 3.675 3.854 3.394 3.656 3.764 3.923
42 42 42 42 43 43 43 43
3.559
3.591
'' Chalcogens are disordered over the capping positions.
110
43 43 43 66 66 67
22
PROKOPUK AND SHRIVER
X.
m
r .
I
Average Mo-Mo bond length I Angstroms
FIG.14. Plot of 95Mo NMR shift vs. average Mo-Mo bond length. Data are taken from references (43, 66).
The expansion of the Mo6 octahedron correlates with contraction of the cubic array of the eight inner ligands and shortening of the Mo-Y bonds. The net result of these distortions is a deviation in the planarity of the Mo-Y4 units on the face of the Y8 cube (Fig. 15). The Mo atoms of the clusters [MO,C~~&]~(X = F, C1, Br, I) protrude outside the halide cube with progressively increasing distances as the axial ligand is changed from F through I. At the other extreme, the metal atoms lie within the Is cube and approach the Y4 face as X increases in size for [Mo61&12-. The [Mo6Brs&12-clusters exhibit intermediate structures, with the metal atoms of [Mo6BrsF612inside the halide cube and those of [Mo6BrsBr6I2and [ M O ~ B ~ &outside. ]~With the ex~ ] ~ [- M O ~ B ~ ~ overlap F ~ ] ~ - of , the inner and ception of [ M O & ~ ~ Fand outer ligands induces a strain within the cluster framework, and this strain becomes more pronounced as the size of either X or Y increases. The strain may account for the increased reactivity in clusters with larger ligands (26, 27).
tI
FIG. 15. Distortions of Y8 cube and X, octahedron with changes in X or Y.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
23
Changes in the inner ligands Y of the clusters [M6Y8X,12from C1 to I results in an expanded Ms octahedron and a Y8 cube. Substitution of chalcogenide atoms for the inner halide ligands has only a modest effect on the average Mo-Y and Mo-Mo bond length (Table III), with the chalcogenide ligands distributed among two or more face-capping sites. However, the terminal chloride-molybdenum bond lengths are elongated compared t o the all-halide clusters. These bond length changes have been attributed to decreased electrostatic attraction between the axial chloride ligands and the less positive halidechalcogenide cores { M o ~ Y ~ - , Q ~ ) ~ ~ ~ ~ ’ + . The pseudo-halide derivatives [Mo&YJ- (X = C1, Br; Y = N3 ( 4 6 ) , NCO (471, NCS (X = C1 only) (40, 111, 1121, and NCSe ( 4 7 ) ) have been reported and crystallographically characterized. Some discrepancies exist in the assignment of the disposition of the ambidentate ligands. For instance, both N- and O-bound cyanate are indicated by 14NNMR in tetrachloroethane; however, crystallographic studies are consistent with N-bound cyanate ligands. Selenocyanate is found to be exclusively N-bound by I4N NMR, IR, and X-ray crystallography. Thiocyanate ligands of most [ M O ~ C ~ ~ ( S C N salts )~I~ are - assigned as N-bound to the cluster, with the exception of K2[Mo6C18(SCN)61, which has been identified as S-bound in both the solid state and solution. Interestingly, structures of the cluster anions [Mo6Br8(N3),1”vary with the type of counterion, including relatively noninteracting [Bu4N]+and [Ph3P=N=PPh3]+.The cluster anions of these salts have regular octahedral geometries; however, the Mo-Mo and Mo-Br bonds vary by as much as 8% between the two structures. Several vibrational studies of the [MsY,&12- clusters have been reported (12, 53, 113-117). The structures of these compounds led Hogue and McCarley to predict that extensive mixing of the symmetry coordinates in the normal modes would prevent the application of group-frequency concepts ( 1 2 ) ;however, diagnostic M-M, M-X, M-Y, and X-M-Y stretching frequencies have been identified, and they find use in the characterization of the clusters. Selection rules for the high-symmetry (0,)halide clusters [M6YM&I2-(Y = C1, Br, I; X = F, C1, Br, I) predict a total of ten Raman active (3A,, + 3E, + 4Tz,) and five IR (5T1,) active modes. The three totally symmetric (Alg)bands, which consist of the breathing mode of the M6octahedron, the YMcube, and the terminal X , ligands, are virtually unmixed and thus provide diagnostic features that are easily identified by their totally polarized Raman bands. Decreased symmetry in the environment about the MI; core, owing to site symmetry in the solid or the presence of axial ligands having lower symmetry, does not prevent application of groupfrequency concepts.
24
PROKOPUK AND SHRIVER
Pyridine N-oxide derivatives, M O ~ C ~ , C ~ ~ ( O N C ,have H ~ Rbeen ) ~ , prepared, and the v(Mo-0) band was assigned around 1200-1270 cm-'. The v(M-0) frequencies correlated with the Hammett constant (T of the R group. As the R group becomes more electron withdrawing (or increases in u),the v(Mo-0) stretching frequency decreases. This observation contrasts with trends in the v(M-0) frequency with latetransition metal complexes such as Ni, Co, and Cu, in which the M-0 vibrational energy increases with the electron-withdrawing nature of the R group and indicates that n-bonding is unimportant in the axial ligand-metal bonds (31). Oxidation of the {M6Y8}4tcore removes electrons from the bonding orbitals of the M6framework, causing an expansion of the metal octahedron. Few of these oxidized clusters have been isolated and structurally characterized. The EPR spectrum of [Mo6C1&1,]'~indicates that the compound is distorted along one axis; however, single-crystal X-ray data is not available to confirm this distortion. The EPR spectrum of the oxidized tungsten cluster [W6Br8Br6l1- contrasts with the spectrum of [Mo6C1&1,]'- in that the values are consistent with an undistorted octahedron, which has been verified by the crystal structure of [(Ph3P)2Nl[W6Br8Br61 (10, 79).
111. Group 5 Metal Halide Clusters
Clusters containing the {Ta&112}2t and {Nb6C112}2t cores have been known for nearly as long as those of the group 6 metals, and their structures were first reported by Linus Pauling (118).Despite the long history of the group 5 metal halide clusters {MsY12}"+(M = Nb, Ta) the ligand substitution chemistry is not as well developed as that of the analogous group 6 metal compounds {M6Y8}4+ (M = Mo, W), because of the extensive redox chemistry associated with the {M6Y1,}"+ core. Five different oxidation states have been observed for the Nb and Ta cores, with compounds in the n = 2-4 states being most common. The n = 1 and n = 5 states are electrochemically accessible; however, clusters in these oxidation states have not been isolated. As the average oxidation state changes for the metals in the cluster core (M6Y,,}"', the affinity for the axial ligands is altered. In the more electron-rich (reduced) state, the clusters have higher affinity for IT-acid ligands such as phosphines, and in the more electron-deficient state, the core has a greater affinity for n-donor ligands such as chloride. Thus, the stability of a particular oxidation state is altered by the coordination environment about the {M6Y,,}"+core.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
25
A. SYNTHESIS OF CLUSTER CORE clusters of niobium and tanThe most convenient route to {M6Y12}2+ talum is the conproportionation of the pentahalides NbY5 and TaY5 (Y = C1, Br) with excess of the metal in molten alkali halide: 20NaY + 14MY5+ 16M + 5Na4M,Y18.
(13)
Formation of lower nuclearity compounds can be avoided by using a 10 to 7 ratio of NaY to MY5 (119).The conproportionation reaction is convenient for the synthesis of the group 5 metal clusters because it allows the use of glass reaction vessels as opposed to electron-beamsealed metal tubes, it provides almost quantitative yield of cluster (in excess of go%), and it requires only 24 h for the reaction to reach completion, as opposed to 4-6 days. Work-up of the product in aqueous acid H Y solutions with small amounts of stannous chloride to prevent oxidation produces M6Y14.8H20. Unlike the group 6 polymeric materials MsYlz, M,Y14*8H,O consists of discrete molecules with a {M6Y12}+core coordinated by two halide ligands and four water molecules, and is accurately described as MsH12Y2(Hz0)4.4H20. The soluble tertiary salts &Nb6Y12Yshave been used as precursors to other cluster compounds with the {Nb6YI2)a+ core, with limited success (120, 121). B. REDOX CHEMISTRY OF THE GROUP 5 CLUSTERS Initial studies on the redox properties of the {M6Y12}n+ clusters were reported on poorly defined complexes simply noted as [TasC1,,]2’ or [NbsCIJt (1221. Aqueous cluster solutions were prepared from the sulfate salts Ta6C1,,S04 and Nb6C1,,SO, with various electrolytes. Some combination of water molecules and anions fills the coordination sphere of the {M6Y12}”+cluster. Despite these ambiguities, the oxidized compounds {M6Y12}4.C were obtained with various chemical oxidants by a two-electron transfer process (123): {M6Y12}2+ + 20x -+ {MsY12}4+ + 2Red.
(14)
Subsequent investigations ( 124-126) revealed an equilibrium between the intermediate {MfiY12}3+ state and the {M6Y,2}2+and {M6Y12}4 clusters, +
26
PROKOPUK AND SHRIVER
and a slow disproportionation of the oxidized { T E ~ ~ Y (Y~ =z }C1, ~ ~Br) cluster to mononuclear Ta(V) complexes and the reduced clusters {Ta6Y12}2tand {Ta6Y1~}~'. Cyclic voltammetry studies (127) on {TasBr12}2treveal two reversible one-electron oxidations corresponding to generation of the n = 3 and n = 4 states,
* -e
-e
{Ta6Br1z}2t
te
{Ta6Br12}3+ {Ta6Br12}4t, +e
(16)
suggesting only a minimal structural rearrangement upon changing oxidation state. Contrasting results were obtained by Hussey and coworkers (128, 129) who investigated the cyclic voltammetry response of [TasCllzC16]'6-n)in acetonitrile. Three 1-electron transfer processes were observed, corresponding to generation of the n = 2-5 species. The reduced cluster [Ta6C11zC16]4undergoes facile substitution by an acetonitrile molecule, generating [Ta6CllZClS(CH3CN)l3(Fig. 16), which is detected in the voltammogram along with the liberated chloride ion. Noncoordinating electrolytes in methylene chloride prevent the loss of the chloride ligand, and the redox couples are chemically reversible (Fig. 16). Interestingly, [TasC11S(OSOzCF3)6]2undergoes two reversible one-electron reductions in acetonitrile to generate [TasC11z(0S02CF3)s13-and [TasCl,~(oso,CF,),]~-, with no evidence for
-1.5
0.5 1.0 1.5 Potential in V vs. Ag/AgCI
-1.0 -0.5 0.0
FIG. 16. Cyclic voltammogram of [Bu,Nl,[TasC1,,Clsl in (A) CHZC1, and 0.1 M [Bu,NlBF,. Sweep rate is 0.10V/s.
(B)CH,CN
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
27
ligand displacement on the cyclic voltammetry time scale. Presumably the strong (T- and n-donor properties of the chloride ions weaken the Ta-C1” bonds of [Ta6Cl12C1614-, making the chloride ligands susceptible to displacement. In contrast, the reduced {TasC1,2}2+core of [TasC112(OS02CF3)6]4is stabilized by the weak donor properties of the triflate ligands. In acidic chloroaluminate melts, the redox potentials of the {Ta&112}2+’3+and {Ta&!112}3+14+couples are anodically shifted by one volt compared to values obtained in acetonitrile electrolytes, suggesting that the majority of the axial chloride ligands have been replaced by [AlClJ ions (129).A similar shift in the redox potentials of couples is observed when changing the {Nb6C112}2fi3t and {Nb6C112}3+14 from a basic chloroaluminate melt to an acidic one. In C1--rich molten salts, the highly reduced cluster [Nb6c112c1615with a {Nb6clI2}’+core is accessible (128),but this species is not accessible in acidic melts or acetonitrile (130).At the other extreme, the highly oxidized cluster [Nb6C112Cls]1with a {Nb6C112}5+ core is generated at 1.76 V vs. SCE in acetonitrile electrolytes. +
C. LIGANDSUBSTITUTION The ligand substitution chemistry of the “b6Y12&I‘6-n’- and [Ta6Y12&](6-’1)clusters is complicated by disproportionation reactions and dependence of the ligand affinity on the oxidation state of the cluster. Unless measures are taken to control chemical potentials, unwanted redox reactions can produce mixtures of partially substituted clusters in a mixture of oxidation states. Consequently, no strategies for exchanging all six axial ligands with a wide range of donor groups exist. Instead, a number of less versatile methods have been developed that introduce a limited range of ligands a t the axial positions and provide moderate control of the redox chemistry. The axial ligands of &Nb6YrLY6(A = alkali metal; Y = C1, Br) undergo simple metathesis with azide and thiocyanate. Dissolution of &NbsYI2Y6in alcohol solutions of N,: or NCS- produces the crystalline compounds &NbsYl2& (X = NB, NCS) (131-133). Long reaction times, on the order of weeks, are required when the A4Nb6Y12Y6 clusters are used as precursors to new cluster compounds. Also, the relative affinity of the {M6Y12}‘+ core for different anionic and neutral ligands is unknown and may restrict the types of clusters accessible from &Nb6Y12Y6. The hydrated complexes M6Yl2Y2(H2O),-4H20 (M6Y14) provide more versatile precursors t o new {M6Y12}’+ (M = Nb, Ta) derivatives. The
28
PROKOPUK AND SHRIVER
aqua ligands are slowly displaced by neutral L or anionic X- donor ligands, generating M6YlzYzL4or [M6Y12Y2&l4-(31, 134-1361. The range of ligands amenable to this strategy has not been delineated, but if care is taken to prevent oxidation of the cluster, a variety of 8-acids are likely to bind to the reduced {M6Ylz}2+ core. Saito and coworkers used chromatography to isolate the cis and trans isomers of the phosphine complexes [M6C112C1z(PR3)4] in moderate yields (1337%) (135) (Fig. 17). Also isolated in the separation procedure was the trisubstituted cluster [Ta6C112C13(PEt3)31, indicating that substitution of the water molecules of Ta6Cll,C1,(H20),.4H20is more complex than first proposed. Repeated addition of aliphatic alcohols ROH to M6Y12Yz(H20)4.4Hz0 (M = Nb, Ta; Y = C1, Br) under an oxygen-free atmosphere eventually yields [M6Ylz(ROH)61X2 (137).Interestingly, 'H NMR and conductivity measurements indicate that the axial halide ligands (especially bromide) are displaced by methanol before the axial water ligands (138). Under anaerobic conditions [M6Ylz~OCH3)z(HOCHz)41 can be obtained from methanolic solutions of M6YlzY2(Hz0)4.4Hz0 and two equivalents of sodium methoxide (139). Care must be taken to exclude air from these reaction mixtures to avoid unwanted oxidation of the cluster. In basic solutions the {M6Y12}'+ core is rapidly oxidized upon exposure to air. Presumably the strong u- and 8-donor properties of hydroxide and alkoxide ligands increase electron density on the
Q
FIG.17. Structure of t r a n ~ - T ~ C l ~ , C l ~ ~ P ( C ~ H ~ ) ~ 1 ~ .
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
29
{M6Yl,}2rcore, making the clusters [M6Y,2(OR)s14(R = H, CHd susceptible to oxidation. Titration of M6YIzYa(Hz0)4 with OH- precipitates the neutral complex M,Y,,( OH),(H,O), . 6Hz0under Na (140).Isolation of M6Y,,(OH),(H20)4.6H,0indicates that the axial chloride ligands are displaced in preference to the axial H,O ligands. Subsequent addition of OH- yields &M,Y,,(OH), . In oxygen-containing atmospheres, MfiY12 (OH),(H,O), precipitates and then redissolves as A2MsY,2(OH)fi with increasing pH ( 141 1. The methoxide clusters M,Y,2(OCH3),(HOCH3)4 and A,M,Y,,(OCH,), have been prepared in a similar fashion (139) (Fig. 18). In acid environments the {M6Y12}2+ core of MsY12Y2(H,0)4.4Hz0 is depending on the metal, inner oxidized to either {M6Y,2}3'or {M6Y12}4t to ligand, and acid source. Hydrochloric acid oxidizes TafiC112Cl,(H20)4 Ta6C1,,C1,(H,O), with a {Ta6C11,}4+core (142); however, NbsCllzCl~ core (143, 144). (HzO), is oxidized to [Nb6C1&1,1S- with a {Nb6C112}3+ Synthesis of the two-electron oxidized cluster [Nb6CllzCls12~ requires (143, 145). chlorine gas to oxidize Nb,ClIaCla(H~O), Introduction of the weakly coordinating triflate ion at the axial positions of the {M6y12}"+cluster is accomplished by the reactions [ B U ~ N I J ~ " ~ C+~xsHOSO2CFS ~~C~~] + [Bu4N1~[Ta,C1,,(0S0~CF:~~~1 + 6HCl(g). (17) The resulting cluster is an electron deficient species (146) (Fig. 19). The niobium analog [Bu4Nls[Nb,Cl,a(OS0,CF3)61has also been prepared. Cyclic voltammetry on solutions of [Bu4Nla[Ta6Cllz(OSOzCFB)61 reveals two 1-electron transfers a t 0.89 V and 0.29 V vs. AgIAgC1, corresponding to generation of [Ta,Cl,,(OSOaCF,),]3- and [Ta6Clln (OS02CFS)fi]4-, respectively. These reduction potentials are among the
FIG.18. Structure of [Ta,C1,,(OCH,),JL .
30
PROKOPUK AND SHRIWR
FIG. 19. Structure of [Ta,Cl,,(OSO,CF,),IZ-
most positive observed for the {TasC11z}4+ and {TagC11~)~+ clusters and attest to the poor donor strength of the triflate ion. The strong oxidizO~CF,)~ the ] simple meing character of [ B U ~ N ] ~ [ T ~ , C ~ ~ ~ ( O Sprevents tathesis reactions with anionic ligands X- such as C1-, Br-, I-, NCS-, and CN- that are observed with the analogous [Bu4Nlz[MosCls (oso2cF,),] (40). Instead, a complex redox process accompanies the substitution reactions and reduced clusters [Bu4N13[Ta6Cl12&I(X = C1, Br, I, NCS) are obtained. A slight excess of X- is required for complete substitution of the triflate ligands, and the reaction rate increases with the ligand series C1- < Br- < NCS- < I- < CN-. This trend does not follow the basicity or polarizability of the ligand, so other factors, possibly oxidation potentials of X-,influence the redox/ substitution process. In fact the reduced cluster [Ta&Il2(OSO2CF3),I3is observed spectroscopically as an intermediate. Reduction of [Ta6C11z(OSOzCF3)s]2by C1- (1.29 V vs. Ag/AgCl) would not be predicted from redox potentials; however, refluxing methylene chloride solutions of [ B u , ~ ] z [ ~ a 6 ~ 1 1 z ( ~and ~ ~ z[Bu,NICl ~ F 3 ) 6 yields ] the reduced product [ B U ~ N ] ~ [ T ~ , C ~ ,It ~ Cis~likely ,]. that oxidation of C1- is promoted by loss of Clz due to a slow stream of Nz passed over the reaction mixture. In mixtures of [BU~N],[T~,C~~~(~S~~CF,),I and the more reducing ligands I-, Br-, and NCS-, the reduced cluster core {TasC112}2+ is detected spectroscopically and the proposed reaction/substitution scheme in Fig. 20 is proposed. Surprisingly, the simple metathesis reaction
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
31
FIG.20. Reaction scheme of [TasCl~~(OSOzCF3),12with X
is not observed for any anionic, X-, or neutral, L, ligands including phosphine oxides.
D. ELECTRONIC AND MOLECULAR STRUCTURE Numerous electronic structure calculations on the {MfiY12}n+ cluster have been performed at various levels (90, 91, 93, 94, 96, 98, 104). The molecular orbital structure obtained from extended Huckel calculations indicates that the eight metal-based orbitals responsible for M-M bonding in the {MfiY12}n+ unit have alg,tZg, tlu,and a2,, symmetry (147)(Fig. 21). In contrast to the directionality of the bonding orbitals in {M6Ys}4+ (see Section II.E.l), the bonding orbitals of {M6Ylz}"+are directed along the faces of the metal octahedron, creating 8 threecenter, two-electron bonds. The number of cluster-bonding electrons available to fill the eight bonding orbitals is equal to the number of valence electrons of the metals minus the number of anionic ligands and the charge of the cluster. Thus, [Nb6Cl12C16]4is electron precise, with 16 CBE [30 - 18 - (-4)1, and [Nb6Cll2Cl6l4is electron deficient with 14 CBE. The exact bonding or antibonding character of the a2,,level differs among the various calculations depending on how the contribution of the axial ligands to the molecular orbitals is considered. Oxidation of the {Nb6cll2}" core to {NbfiC112}4+ corresponds to depopulation of the
32
PROKOPUK AND SHRIVER
FIG.21. A molecular orbital diagram for {M,Y,,j"+ based on an extended Hiickel calculation, adapted from reference (108).
a2ulevel; thus, the aZulevel is the HOMO or LUMO of the n = 2 or n = 4 clusters respectively. Magnetic susceptibility measurements and ESR spectra of the clusters with n = 2-4 confirm the assigned aZusymmetry of the HOMO/LUMO (148).Hyperfine splitting of the unpaired electron of [Nb6Y&I3- delocalized over the six niobium centers (spin 9/2) leads to impressively split ESR spectra (Fig. 22). X-ray
l
2.8
.
I
3.0
.
I
3.2
.
1
.
3.4
I
3.6
.
I
3.8
.
I
4.0
.
2
kG FIG.22. EPR spectrum of [Bu4NI~,[NbsCll2BrGI in CHaClaa t room temperature.
OCTAHEDRAL, CLUSTERS OF TRANSITION METALS
33
crystal structure determinations for the cluster series [NbsCl12Clsl(n-6)reveal that the average Nb-Nb bond length decreases from 3.029 to 2.899 A as the oxidation state on the {NbsC112}'L+ core is lowered from n = 4 to n = 2 (149). This trend is consistent with the assignment of the aZulevel as predominately bonding in character for the all-chloro clusters. The high energy of the a2"orbital explains the relative stability of the oxidized clusters {M6Y12}3+ and {M6Y12}4+. The average oxidation state of the metal atoms ranges from 2.33 ( n = 2) to 2.66 ( n = 4). Numerous UV-vis absorption studies on the Nb and Ta clusters have been reported; however, despite the existence of one-to-one correspondence in the spectra of the two metals, there is little agreement on the assignments for the metal-based transitions (135, 150-152). Solutions of both {TasY12}4+ and {Ta6Y,2}3+ are orange-brown in color, and the niobium analogs {Nb6Y,2}4t and {NbsY12}3t are yellow-brown in solution. The reduced states {Nb6Y12}2+ and {TasY~z}~~ are olive green and green, respectively. Attempts to assign the electronic transitions of the clusters have been complicated by the similarity in the spectra of {M6Y12}3+ and {M6Y12}4+ clusters, which makes identifying mixtures of clusters in these oxidation states difficult. Additionally the {MsY12}2-i state is intensely colored and therefore easily obscures the spectra of the oxidized clusters.
E. NIOBIUMIODIDE CLUSTERS {NbsIs}"+ Hexanuclear niobium iodide clusters {Nb61B}f1+ are the lone exception to the usual edge-bridging halide geometry adopted by the group 5 metal halide clusters. With eight face-capping iodide ligands, the {NbsIs}"+core is isostructural with the group 6 metal halide clusters {MO~C~ and ~ }{W6Clg}4+ ~+ (153).The similarity in the metal-ligand configurations permits the use of the molecular orbital structure derived for the group 6 metal halides {M6Y8}"+ to be applied to the {Nb61R}n+ unit. The Nb61B13 is extremely electron deficient, with 19 CBE available for 12 metal-metal bonding orbitals. This unique electronic structure gives rise to the distinctive chemical and physical properties of { N ~ ~ I J J I + . The binary phase Nbs1813is prepared (154) from the decomposition of Nb318at 950°C:
34
PROKOPUK AND SHRIWR
Increased yields are obtained when Niobium metal is added to the reaction mixture to reduce the NbI, by-product: 11NbI4 + 13Nb+ 4Nb&I3.
(20)
The addition of CsI to the reaction mixture produces the reduced cluster CsNb61813, with 20 CBE (155). The electron deficiency of Nb61813and csNb61813 allows these compounds to absorb hydrogen in the solid state at 300°C and atmospheric pressure, generating HNb61813and CsHNb61813( 155, 156). Low-temperature neutron diffraction studies on Nb61813, HNbs1813, and DNb61813 reveal that the hydrogen atom is slightly displaced from the center of gravity of the surrounding Nb6 octahedron (157). The size of the displacement is enough to accommodate multiple hydrogen atoms, suggesting that more than one interstitial hydrogen may reside in the cage in high-temperature phases of HNb61813.Adsorption of hydrogen into the interstitial position is unusual, and this is the only case in which an interstitial atom is implanted in a preexisting M6 octahedron of a metal halide cluster. The uptake of hydrogen by Nb61813increases the number of CBE from 19 to 20, and thus a decrease in the average Nb-Nb bond length is expected. However, single-crystal X-ray diffraction studies of Nbs1813and HNb61813reveal a slight expansion of the Nb6 cage of HNb61813compared t o the intersticontial free cluster (158) (Table Iv). By contrast, Zr6HCl12(EtNH2)6 tains interstitial hydrogen atoms but an average Zr-Zr bond length is -0.3 A shorter than the interstitial free [zr&1&C16I2-, which has a comparable CBE count (both at 13) (159, 160). Low-temperature magnetic susceptibility measurements on Nb61813 and HNb61813 reveal ground spin states of S = 1/2 and S = 0, respecTABLE IV NIOBIUM-NIOBIUM BONDDISTANCES IN SOME {Nb6I8P+CLUSTERS Cluster
CBE
d(Nb-Nb)(AP
Ref.
Average Nb-Nb bond lengths are reported with crystals at room temperature except HNb61813(211 K).
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
35
tively. At higher temperatures the former has a ground state S = 3/2 and the latter S = 1 (161).The change in spin states is not due to simple transitions from the ground to excited state but a coupled electronic and structural phase change (158,161, 162). Heat capacity, X-ray diffraction, and magnetic susceptibility measurements indicate that a second-order or higher phase change occurs at 274 K for Nb61813 and 324 K for HNb61813.The structural deformation, a rotation of one Nb, face of the octahedron by 7", is accompanied by the crossing of electronic levels, which leads to a reduction of the spin degeneracy. This spin-crossover transition is made possible by competition between the effects of a spin flip and changes in the molecular orbital levels that result from the cluster distortion (163). Two discrete molecular compounds containing {Nb618}"+ units were obtained from reactions of Nb61813and NH,R (R = CH, , C,H,) (2).The amines displace all the bridging axial iodide ligands, generating the neutral complexes [Nb618(NH,R)6].These clusters are electron-rich with 22 CBE compared to the 19- and 20-CBE {Nb618}n+ cores obtained from solid-state reactions. Consequently, the increased number of electrons available for cluster bonding orbitals results in shorter Nb-Nb distances (Table IV). Magnetic susceptibility measurements indicate that [Nb618(NH2CH3)6] is diamagnetic. This observation is contrary to the prediction of a half-filled e, HOMO for a 22-CBE species with a {M6Y8}"+ geometry.
IV. Materials Chemistry Derived from Soluble Metal Halide Clusters
A. HIGHER NUCLEARITY CLUSTERS Mixed metal clusters have been generated by reaction of organometallic species with ambidentate cyanide ligands and [Mo6C18 (OSO&F,)6]2- (411. Both [CpMn(C0)2CNl- and CpRu(PPh,),CN displace the triflate ligands of [ M o ~ C ~ ~ ( O ~ O ~ forming C F , ) ~ ]the ~ ~ ,12 metal clusters { M o ~ C ~ ~ [ ( ~ - C N ) M ~ ( C O )and ~ C ~{Mo6C18[(p-CN) ]~}~R U ( P P ~ ~ ) ~ Crespectively. ~ ] ~ } ~ + , Single-crystal X-ray structure determinations on (PPN)2{Mo6C18[(p-CN)Mn(C0)2Cp]6} reveals that the [CpMn(C0)2CNl-fragment is coordinated to the {MO,C~,)~+ core via bridging cyanide ligands (Fig. 23). Interestingly, the electronic spectra of these mixed metal clusters show an extremely intense absorbance that most likely corresponds to a charge transfer band [possibly intervalence charge transfer from the CpMn(C0)2CN- ligandl. The intensity of this band cannot be accounted for from a simple com-
36
PROKOPUK AND SHRIVER
FIG.23. Structure of { M O ~ C ~ ~ [ ( ~ - C N ) C ~ M ~ ( C O ) ~ I ~ } ” .
bination of the absorbencies of the manganese centers and the { M O ~ C ~core; ~ } ~ ’ however, “intensity-stealing” between a Mo6-based d-d“ transition and a charge transfer band may lead to the increased absorption (Fig. 24). Cyclic voltammetry on [ B U ~ N ] ~ { M O ~ C ~ , [ ( ~ - C N ) M ~ (reveals C~)~C~]G) a broad redox wave assigned to the oxidation of the pendant manganese centers. This wave is anodically shifted 0.26 V from the “free” ligand/metal complex and is consistent with reduced electron density on the C ~ M n ( c 0fragment. )~ The breadth of the redox wave is attributed to communication between the manganese and molybdenum atoms through the bridging cyanide ligands. Chemical oxidation of ( M O ~ C ~ , [ ( ~ - C N ) M ~ ( C O ) ~with C ~ ]NO+ , ) ~ - results in clean substitution of one of the carbonyls of the manganese fragment by NO’, generating (Mo6C1,[(~-CN)Mn(CO)(NO)Cp],)4+.
250
300
350
400
450
500
550
600
t
wavelength I nm
FIG. 24. Absorption spectra of (A) [ B U ~ N I [ C ~ M ~ ( C O ) ~ C(B) N I , [Bu,NlZ[Mo6CI, in CH2Clz. (OSOZCFJ)61, and (C) [Bu4NlZ{MosC181(~-CN)CpMn(CO)z16}
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
37
Reaction of NazMo6C18( OCH3)s with the carboxylic acid H02CC5H4 FeCp produces the organometallic derivative Na2M06C18(02CC5H4 ~}~' FeCp), (55). The iron centers are attached to the { M O ~ C ~core through the carboxylate moieties on the Cp ring (Fig. 25). In contrast to voltammograms of {M0~Cl~[(~-CN)MnfC0)~Cpl,f2, cyclic voltammograms of Na2MofiC18(OzCCsH4FeCp)6 in DMF display a single redox wave shifted 0.10 V from that of the acid H02CC5H4FeCp,indicating that the iron centers are independent and equivalent. Cyclic voltin DMSO reveals a chemically ammetry on NazMosC18(02CC5H4FeCp)fi irreversible oxidation of the iron centers. A reduction wave is observed substantially shifted to negative potentials. This chemical irreversibility is attributed to reduction of uncoordinated ferrocenium carboxylate. A decrease in basicity of the carboxylate moiety upon oxidation of the pendant ferrocenyl group should make this oxidized ligand a weaker donor, which is susceptible to dissociation. Presumably, vacant coordination sites of {Mo,C~,}~+ are taken up by solvent molecules.
B. SUPPORTED CLUSTER MATERIALS A number of strategies have been developed to generate composite materials by the incorporation of the { M O ~ C ~unit ~ } ~into ' a host ma-
38
PROKOPUK AND SHRIVER
trix. Coordinating polymers such as poly(4-vinylpyridine) (PVP) provide sites to bind the { M o & ~ ~ units } ~ + directly to the polymer backto bone. Thus, the addition of Mo6Cl12or [Bu4NI2[Mo6C1,(0SO~CFg)61 PVP in ethanol secures the {MO~C~,}~+ unit to the pyridine moieties of the polymer. Because the { M O ~ C ~unit ~ } ~acts + as a cross-linking agent, the nature of the axial ligands on { M O , C ~influences ~}~~ the physical properties of the cluster/polymer composite material ( 164 ). Dissolution of MosCllzin ethanol generates M O , C ~ ~ C ~ ~ ( Ewith ~OH two ) ~ labile , axial sites capable of coordinating pyridine. In contrast [Bu4N12 [MO&1,(OSO&F&] has six sites available for pyridine to bind and thus provides a more effective cross-linker. The number of accessible coordination sites is reflected in the glass transition temperature, T,, of the resulting material. Adsorption of Mo&l12 and [Bu4Nlz[Mo&18 (OS02CF3),]into PVP in a ratio that provides equal donor groups t o cluster binding sites results in composite materials with a higher T, than those derived from Mo6ClI2.Immobilization of the { M O ~ C ~core ~}~+ by the PVP does not drastically alter the emission spectrum of the { M O ~ C ~core; ~ } ~ however, + increasing the number of pyridine groups coordinated to the cluster lowers the excited-state lifetime (164).The photocatalytic properties of these materials have been demonstrated, and singlet oxygen generated by quenching the excited cluster with O2 has been used to oxidize alkenes (165). DiSalvo and co-workers developed a monomer in solvent strategy for preparing monodispersed { M O ~ C ~units , } ~ ~in a polymer matrix in a coordinating sol(166).By dissolving vent with a polymerizable functional group such as N-vinylimidazole (NVI), a monomer unit containing the {MO,C~~}~+ unit is generated, [Mo,C~,(NVI)~](OSO,CF,),. The low solubility of these tetracationic clusters is circumvented by using the monomer unit, in this case NVI, as the solvent. Polymerization of the cluster/monomer solution produces a monodispersed cluster in an organic matrix. In addition t o solvating the cluster cations, the uncoordinated monomer/solvent dilutes the highly cross-linking clusters. By adjusting the cluster concentration in the monomer/solvent, the degree of cross-linking can be controlled. Oxide supports have also been used to immobilize {Mo,Cl8}*+clusters with various degrees of success. Basic and acidic silica adsorb [BU,N]~[MO,C~,(OSO~CF,),] (167).Both electro[ B U ~ N ] ~ [ M O &and ~~C ~~] static and covalent binding of the {M~,cl,}~+ core to the silica surface can be obtained with proper choice of cluster, solvent, and silica treatment. The photophysical and photocatalytic properties of the cluster are retained on the Si02support (168).
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
39
Intercalation of the aqua complex [Mo&~~(H@),]~+ into layered sodium montmorillonite creates a pillared structure (169). The cluster cations increase the d-layer spacing from 12.5 A for sodium montmorillonite to 16.6 A for the intercalated composite material. The driving force for the intercalation process is attributed to ion exchange with the Na' of the montmorillonite; however, the amount of cluster adsorbed by the layered clay is larger than expected based on the sodium content. Similar observations have been made for the intercalation of {Nb6C112}2+'3+ and {Ta6C112}2+ into sodium montmorillonite, which results in an increase in d-layer spacing and adsorption of excess cluster (170). The axial ligands attached to the intercalated group 5 clusters were not unambiguously identified, thus the driving force for intercalation is unclear.
C. CHARGE-TRANSFER SALT COMPLEXES Organic-inorganic charge-transfer salt complexes containing radical organic cations such as tetrathiafulvalenium (TTF)' or tetramethyltetrathiafulvalenium (TMTTF)' and anionic clusters [M6C18C1612and [M6Cl12C16]'6~"'as counterions have been electrocrystallized. Other organic-inorganic salt complexes have been prepared from these organic radicals with smaller anions such as BF;, PF,, ReO;, and SbF, ( 171). Interactions between the organic radical cations, especially overlap of the HOMO of neighboring cations, can produce semiconducting, metallic, or superconducting properties in the molecular salt compounds. Using large cluster anions changes the anion/ cation volume ratio, thereby forcing the cation radicals into new packing arrangements. Also, the use of radical cluster anions creates the possibility of anion-cation spin interactions. Single crystals of the first charge-transfer salt complex containing a hexanuclear metal halide cluster, (TMTTF)2[Mo6C18C16], were prepared from TMTTF and [Et4N]2[Mo6C18C18] (172). The structure of (TMTTF)2[Mo6C18C16] consists of cation dimers with the anionic metal clusters in a distorted CsCl structure (Fig. 26). The CsCl arrangement is likely a consequence of the comparable volumes of the cation $mer and [MO~C~,C~,]~-. Interdimer distances (S...S distance of 4.87 A) are quite large, resulting in an insulating material. As expected no ESR signal could be obtained because the cation radicals are paired. The tetrakis(methy1thio)tetrathiafulvalenium (CH,S),TTF+ salt of [Mo6C18(NCs)6]2-was prepared in the same manner as (TMTTF)z[MosClsCls]and has the same CsC1-based structure with isolated dimers of (CH,S),TTF' (111). However, unlike
40
PROKOPUK AND SHRIVER
FIG.26. Unit cell of [TMTTF12[Mo6C18C161.
(TMTTF)2[Mo6C18C161,the cation dimers of [(CH3SI4TTFl2 [Mo6C1,(NCS),]interact with the axial ligands of the cluster. Sulfursulfur distances separating the (CH3S),TTFt dimer and [Mo&~,(NCS)~]~are 3.793 and 3.755 These cluster-cation interactions stabilize a slipped planar configuration in the cation dimer, which favors a mobile triplet exciton at room temperature. In effect, the triplet state cation dimers are coupled through the [Mo6C1,(NCS),l2-cluster . Charge-transfer salt complexes containing the ESR-active [NbfiCll2Cll2I3cluster with the organic cation radicals TTFt and TMTTF' were prepared by electrocrystallization from CH3CN/CH2CI solutions (130, 173). The TTF' salt, which was isolated as (TTF)2(Bu4N)[Nb6C112Cllz].CH3CN, contains isolated TTF' dimers. At higher temperatures the triplet state of the cation dimer is populated and interacts with the unpaired electron localized on the {NbfiCL12}3' core. In contrast, the TMTTFt salt as well as the selenium analog TMTSF were isolated as (TMTTF),[Nb6ClI2Cll2] .0.5CH2C1, containing a [(TMTTF),I3' cation and neutral TMTTF molecule. The TMTTF units in the [(TMTTF)J3' cation are stacked into infinite layers surrounded by [Nb6C112C11213anions (Fig. 27). The fifth TMTTF molecule is neutral and orthogonal to the cationic TMTTF species. Four partially oxidized TMTTF molecules are present. The HOMO of the organic stack is 5/8 filled (5 electrons in 4 orbitals), and metallic properties are expected. Conductivity measurements, however, indicate that the hybrid material is semiconducting, with band gaps of 0.16 and 0.20 eV for the TMTTF and TMTSF, respectively. Inequivalency in
A.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
41
p
FIG. 27. Unit cell of [TMTTFl,[Nb,Cl,,Cl,]; two clusters a t corners of the unit cell are omitted for clarity. Reprinted with permission from Ref. (130).
the structures of the stacked cations is invoked to explain the electron localization and the resulting semiconducting properties. A single ESR signal is obtained for both (TMTTF),[Nb6Cl1,C1,,1 and (TMTSF)S[NbsC112C112], which indicates mixing of the localized spins of the cluster and cation.
D. EXTENDED SOLIDS The most notable example of a covalently linked extended solid generated from a soluble precursor is Prussian blue (174). A number of strategies for assembling discrete mononuclear complexes into ex-
42
PROKOPUK AND SHRIVER
tended solids have been adapted to the formation of an extended metal halide cluster network. Interactions between counter cations and the ligands of the metal halide clusters often result in solids with bonding in one, two, or three dimensions. Hydrogen bonding between ligands with donor or acceptor groups and water molecules in the crystal lattice also provides intercluster linkage. Alkali metal salts of [M6Y&12- and [MsY12&l'6-n'often contain cation-cluster interactions that fill the coordination sphere of the alkali metal cation. When ligands from neighboring clusters coordinate to a common cation, an extended network results. A recent example is in the ferrocenecarboxylate derivative, Na2[Mo6Cls(02CCSH4Fecp)6], which the ferrocenyl moieties are connected to the { M O ~ C ~core ~}~+ through the carboxylate groups (55).The multidentate properties of the carboxylate group enable the ligand to coordinate to the cluster and the sodium counterions. Two carboxylate ligands from one cluster are linked to a sodium counterion, and a third ligand from an adjacent cluster anion plus a methanol molecule fill the coordination sphere of the cation. These interactions create a one-dimensional array of [ M O ~ C ~ ~ ( O ~ C C S H ~anions F ~ C ~(Fig. ) ~ ]28). ~Crystals of the pseudohalide complexes AM6Y12& (A = K, Rb, Cs; Y = C1, Br; X = NB,NCS) produced from aqueous solutions incorporate water molecules in the unit cell (131-133). These water molecules as well as the inner halide and the pseudohalide ligands coordinate the alkali cations, resulting in an extended solid with a threedimensional array. Interestingly both the a- and P-nitrogens of the azide ligand bind the cations, whereas only the sulfur atoms of the thiocyanate groups coordinate to the alkali metals. In addition to cation bridges, water molecules connect the azide ligands of adjacent clusters via hydrogen bonding.
FIG.28. Display of the linear chain of {Mo&&}~+ units connected through bridging carboxylate ligands and sodium ions of Naz[Mo6Cls(OzCC6H4Fecp)61. The ferrocenyl moieties of the carboxylate groups are omitted for clarity.
43
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
Extensive hydrogen bonding has been observed in many structures containing {M6Y12}n+units with OH-, HzO, or CHBOHligands in the axial sites. The axial ligands act as both proton donors and acceptors. Hydrogen bonding interactions between the axial ligands and solvent molecules of th‘e crystal lattice create three-dimensional cluster systems. A few structures containing extended networks of clusters linked via hydrogen bonds are listed in Table V. It should be noted that the hydrogens were not located in the structures but inferred from short donor-acceptor distances. Two compounds, (Me,N) (175) and [Nb6Brlz(HzO)61[HgBr41 12Hz0(1761, pos[Ta6Cllz(Hz0)61Br4 sess hydrogen bonds between water donors and bromide acceptors. DC electrical measurements indicate that [Nb6Br12(Hz0)61[HgBr41. 12H20has semiconducting properties; however, the origin of this conductivity is not completely understood.
E. CHEMICALLY MODIFIED SURFACES The affinity of the cluster anions [ B U , N ] ~ “ ~ ~ C ~ & ] ~ =-C1, ( X Br, I) for gold and silver surfaces is determined by the identity of the axial ligand X (177). Only [Nb6CllzBr613adsorbs in a monolayer fashion on gold as determined by quartz crystal microgravimetry, X-ray photoelectron spectroscopy, and cyclic voltammetry. Redox potentials for “b6Cl12Br613-adsorbed to gold are anodically shifted 0.15 V relative to the cluster in solution. This shift is consistent with the transfer of electron density from the cluster to the metal surface. By contrast no TABLE V HYDROGEN-BONDED CLUSTER COMPLEXES Cluster
H Bonding
Ref. 175 176 176 179 179 180 181 181 181
Bromide ligands from both the cluster cation and the anion partake in hydrogen bonding. Cation and anion are involved in hydrogen bonding.
44
PROKOPUK AND SHRIVER
sign of niobium or chlorine is observed in the X P S spectrum of gold surfaces that were exposed to [Bu~N]3[Nb6C112C~613~. At the other extreme, solutions of [Bu,N]3[Nb6C11zC161 completely dissolve gold films. The increased reactivity of the clusters with the heavier halides parallels the interaction of gold with the free halide ions., where C1- adsorbs the weakest and I- the strongest (178). Of the three clusters only [Nb6C1,2C16]3-forms a monolayer on silver surfaces. Both [Nb6CllzBr6]3and [Nb&llz16]3- dissolve silver. These findings demonstrate that halide ligands should be versatile anchors for immobilizing many types of metal complexes to metal surfaces. Pyridine groups immobilized on gold surfaces through strong goldsulfur interactions displace the triflate ligands of [Ta6Cl12 (OS02CF3)fi14and result in a immobilized monolayer of cluster on the gold surface. XPS reveals that an average of four triflate ions are replaced by the surface-bound pyridines. Dilution of the pyridine surface with coadsorbed thiolphenol, a noncoordinating moiety, reduces the amount of cluster adsorbed to the surface and the average number of triflate ligands displaced by surface pyridine. The number of triflate ligands remaining on the surface-bound cluster may be adjusted with the thiolphenol/mercaptopyridine ratio. The different coordination environments about the {Ta&112)"t core in the resulting material have been identified with differential pulse voltammetry. As the number of coordinating pyridine ligands coordinated to the {TafiC112}'1+ core decreases, the oxidation potentials of the clusters are cathodically shifted. The shift reflects the n--acid character of the pyridine ligand.
ACKNOWLEDGEMENTS We appreciate the many contributions of former graduate students and senior collaborators, who have contributed to our research in the area of inorganic metal cluster chemistry: Laura Robinson, Dean Johnston, Joseph Hupp, and Donald Ellis. Our research in this area was supported by the National Science Foundation, inorganic chemistry program and the NSF supported Materials Research Center a t Northwestern University.
REFERENCES
1. Schafer, H.; Schnering, H.-G. v. Angew. Chem. Znt. Ed. Engl. 1964,76, 833. 2. Stollmaier, F.; Simon, A.Znorg. Chem. 1985,24, 168. 3. Ziebarth, R. P.; Corbett, J . D. Acc. Chem. Res. 1989,22, 256.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
45
4. Rogel, F.; Zhang, J.; Payne, M. W.; Corbctt, J. D. Adu. Chem. Ser. 1990, 226, 369. 5. Chevrel, R.; Hirrien, M.; Sergent, M. Polyhedron 1986, 5, 87. 6. Hughbanks, T.; Hoffmann, R. J. J. A m . Chenz. Soc. 1983, 105, 1150. 7. Blomstrand, W. J . Prakt. Chem. 1859, 77, 88. 8. Brosset, C. Ark. Kemi. Mineral. Geol. 1943,20A, 1. 9. Maverick, A. W.; Gray, H. B. J. Am. Chem. Soc. 1981, 203, 1298. 10. Maverick, A. W.; Najdzionek, J . S.; MacKenzie, D.; Nocera, D. G.; Gray, H. €3. J . Am. Chem. Soc. 1983, 205, 1878. 11. Schafer, H.; Schnering, H.-G. v.; Tillack, J.; Kuhnen. F.;Wohrle, H.; Baumann, H. 2. Anorg. Allg. Chem. 1967,353, 281. 12. Hogue, R. D.; McCarley, R. E. Inoig. Chenr. 1970, 9, 1354. 23. Dorman, W. C.; McCarley, R. E. Inorg. Chenr. 1974, 13, 491. 14. Zhang, X.; McCarley, R. E. Inorg. Chem. 1995, 34, 2678. 15. Drobot, D. V.; Mikhailova, L. G.; Bol'shakiv, K. A.; Sbitnev, V. L. Russ. J . Inorg. Chem. 1978,23, 643. 16. Franolic, J . D.; Long, J. R.; Holm, R. H. J . A m . Chem. Soc. 1995, 117, 8139. 17. Jiidden, K.; Schafer, H. 2. Anorg. Allg. Chem. 1977,430, 5. 28. Stensvad, S.; Helland, B. J.; Babich, M. W.; Jacobson, R. A,; McCarley, R. E. J. Am. Chem. Soc. 1978, 100, 6257. 19. Aufdembrink, B. A,; McCarley, R. E. J . Am. Chem. Soc. 1986, 208, 2474. 20. Zietlow, T. C.; Gray, H. B. Iiiorg. Chein. 1986, 25, 631. 21. Cotton, F. A,; Poli, R. J . Am. Chem. Snc. 1988, 220, 830. 22. Burini, A.; Cotton, F. A,; Czuchajowska. J. Pdyhedron 1991, 10, 2145. 23. Sheldon, J. C. J. Ch,em. Soc. 1960, 1007. 24. Sheldon, J. C. J. Chem. Soc. 1960, 3106. 25. Lessmeister, P. v.; Schafer, H. 2. Anorg. Allg. C h e m 1975, 427, 171. 26. Harder, K.; Peters, G.; Preetz, W. 2. Anorg. Allg. Chem. 1991, 598/599, 139. 27. Preetz, W.; Braack, P.; Harder, K.; Peters, G. 2. Anorg. Allg. Chern. 1992, 612, 7. 28. Fergusson, J. E.; Robinson, B. H.; Wilkins. C. J. J. Chem. Soc. A. 1967, 486. 29. Carmichael, W. M.; Edwards, D. A. J. Inorg. Nucl. Chem. 1967,29, 1535. 30. Sheldon, J. C. J. Chem. Soc. 1961, 750. 32. Field, R. A,; Kepert, D. L.; Taylor, D. Inorg. Ch.rm. Acta 1970, 4, 113. 32. Hamer, A. D.; Smith, T. J.; Walton, R. A. Inorg. Ch,em. 1976, 15, 1014. 33. Saito, T.; Nishida, M.; Yamagnta, T.; Yamagata, Y.; Yamaguchi, Y. Inorg. Chem. 1986,25, 1111. 34. Cotton, F. A.; Curtis, N. F. Inorg. Ch,etn. 1965, 4, 241. 35. Schafer, H.; Plautz, H.; Abel, H.-J.; Lademann, D. 2. Anorg. Allg. Chem. 1985, 526, 168. 36. Ehrlich, G. M.; Deng, H.; Hill, L. I.; Steigerwald, M. L.; Squattrito, P. J.; DiSalvo, F. J. Inorg. Chem. 1995, 34, 2480. 37. Perchenek, N.; Simon, A. 2. Anorg. Allg. Chem. 1993, 619, 98. 38. Schafer, H.; Brendel, C.; Henkcl, G.; Krebs, B. 2. Anorg. Allg. Chem. 1982, 491, 275. 39. Nannelli, P.; Block, B. P. Inorg. C h e m 1968, 7, 2423. 40. Johnston, D. H.; Gaswick, D. C.; Lonergan, M. C.; Stern, C. L.; Shriver, D. F. Inorg. Chenr. 1992, 31, 1869. 41. Johnston, D. H.; Stern, C. L.; Shliver, D. F. Inorg. Chem. 1993,32, 5170. 42. Preetz, W.; Harder, K.; Schnering, H. G. v.; Niche, G.; Peters, K. J. Alloy. Comp. 1992, 283, 413.
46
PROKOPUK AND SHRJYER
43. Preetz, W.; Dublitz, D.; Schnering, H. G. v.; Safimannshausen, J. 2. Anorg. Allg. Chem. 1994,620, 234. 44. Briickner, P.; Peters, G.; Preetz, W. 2. Anorg. Allg. Chem. 1993, 619, 551. 45. Briickner, P.; Peters, G.; Preetz, W. 2. Anorg. Allg. Chem. 1993,619, 1920. 46. Bublitz, D.; Preetz, W.; Simsek, M. K. 2. Anorg. Allg. Chem. 1997, 623, 1. 47. Simsek, M. K.; Preetz, W. 2. Anorg. Allg. Chem. 1997, 623, 515. 48. Harder, K.; Preetz, W. 2. Anorg. Allg. Chem. 1992,612, 97. 49. Johnston, D. H. Ph.D. Thesis, Northwestern University, Evanston, 1993. 50. Ehrlich, G. M.; Warren, C. J.; Haushalter, R. C.; DiSalvo, F. J. Znorg. Chem. 1996, 34, 4284. 51. Perchenek, N.; Simon, A. 2.Anorg. Allg. Chem. 1993, 619, 103. 52. Perchenek, N.; Simon, A. Acta Crystallogr. 1991, C47, 2354. 53. Schoonover, J. R.; Zietlow, T. C.; Clark, D. L.; Heppert, J . A,; Chisholm, M. H.; Gray, H. B.; Sattelberger, A. P.; Woodruff, W. H. Znorg. Chem. 1996,35, 6606. 54. Nannelli, P.; Block, B. P. Znorg. Chem. 1969,8, 1767. 55. Prokopuk, N.; Shriver, D. F. Znorg. Chem. 1997,36, 5609. 56. Prokopuk, N.; Kennedy, V. 0.; Stern, C. L.; Shriver, D. V. Unpublished results. 57. Sheldon, J . C. Nature 1959, 184, 1210. 58. Sheldon, J. C. J. Chem. SOC.1964, 1287. 59. Sheldon, J. C. J. Chem. SOC.1963, 4183. 60. Chisholm, M. H.; Heppert, J . A.; Huffman, J. C. Polyhedron 1984,3, 475. 61. Sheldon, J. C. J. Chem. SOC.1962, 410. 62. Perrin, C.; Sergent, M.; Traon, F. L.; Traon, A. L. J. Solid State Chem. 1978, 25, 197. 63. Pilet, J. C.; Traon, F. L.; Traon, A. L.; Perrin, C.; Perrin, A.; Leduc, L.; Sergent, M. Surf Sci. 1985, 156, 359. 64. Perrin, A.; Perrin, C.; Sergent, M. J. Less-Common Met. 1988, 137, 241. 65. Michel, J . B.; McCarley, R. E. Znorg. Chem. 1982,21, 1864. 66. Ebihara, M.; Toriumi, K.; Saito, K. Inorg. Chem. 1988,27, 13. 67. Ebihara, M.; Toriumi, K.; Sasaki, Y.; Saito, K. Gazz. Chim. Ztal. 1995, 125, 87. 68. Xie, X.; McCarley, R. E. Znorg. Chem. 1997,36, 4011. 69. Saito, T.; Yoshikawa, A.; Yamagata, T.; Imoto, H.; Unoura, K. Znorg. Chem. 1989, 28, 3588. 70. Hilsenbeck, S. J.; Young, V. G.; McCarley, R. E. Znorg. Chem. 1994,33, 1822. 71. Ehrlich, G. M.; Warren, C. J.; Vennos, D. A.; Ho, D. M.; Haushalter, R. C.; DiSalvo, F. J. Znorg. Chem. 1995,34, 4454. 72. Saito, T. Adv. Znorg. Chem. 1997, 44, 45. 73. Gray, H. B.; Maverick, A. W. Science 1981,214, 1201. 74. Nocera, D. G.; Gray, H. B. J. Am. Chem. SOC.1984, 106, 824. 75. Jackson, J . A.; Mussell, R. D.; Nocera, D. G. Znorg. Chem. 1993,32, 4643. 76. Barnard, P. A.; Sun, I.-W.; Hussey, C. L. Znorg. Chem. 1990,29, 3670. 77. Zietlow, T. C.; Nocera, D. G.; Gray, H. B. Inorg. Chem. 1986,25, 1351. 78. Mussell, R. D.; Nocera, D. G. Znorg. Chem. 1990,29, 3711. 79. Zietlow, T. C.; Schaefer, W. P.; Sadeghi, B.; Nocera, D. G.; Gray, H. B. Inorg. Chem. 1986,25, 2198. 80. Schafer, H.; Siepmann, R. 2. Anorg. Allg. Chem. 1968,357, 273. 81. Siepmann, R.; Schnering, H. G. v. 2.Anorg. Allg. Chem. 1968,357, 289. 82. Siepmann, R.; Schnering, H. G. v.; Schafer, H. Angew. Chem. Int. Ed. Engl. 1967, 6, 637. 83. Kepert, D. L.; Marshall, R. E.; Taylor, D. J. C. S. Dalton Trans. 1974, 506.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
47
84. Ebihara, M.; Isobe, K.; Sasaki, Y.; Saito, K. Znorg. Chem. 1992,31, 1644. 85. Zietlow, T. C.; Hopkins, M. D.; Gray, H. B. J. Solid State Chem. 1985,57, 112. 86. Zietlow, T. C.; Schaefer, W. P.; Sadeghi, B.; Hua, N.; Gray, H. B. Inorg. Chem. 1986,25, 2195. 87. Mussell, R. D.; Nocera, D. G. Polyhedron 1986, 5, 47. 88. Mussell, R. D.; Nocera, D. G. J . Am. Chem. SOC.1988, 110, 2764. 89. Jackson, J . A.; Turro, C.; Newsham, M. D.; Nocera, D. G. J . Phys. Chem. 1990, 94, 4500. 90. Cotton, F. A,; Haas, T. E. Inorg. Chem. 1964, 3, 10. 91. Kettle, S. F. A. Theor. Chim. Acta 1965,3, 211. 92. Guggenberger, J.; Sleight, A. W. Inorg. Chem. 1969,8, 2041. 93. Voronovich, N. S.; Korol’kov, D. V. Zh. Strukt. Khim. 1971, 13, 458. 94. Voronovich, N. S.; Korol’kov, D. V. Zh. Strukt. Khim. 1971,12, 613. 95. Miiller, H. 2. Phys. Chem. (Leipzig) 1972,249, 1. 96. Wirsich, J . Theor. Chim. Acta 1974, 34, 67. 97. Cotton, F. A,; Stanley, G. G. Chem. Phys. Lett. 1978,58, 450. 98. Bursten, B. E.; Cotton, F. A,; Stanley, G. G. Isr. J. Chem. 1980, 19, 132. 99. Seifert, G.; GroBmann, G.; Miiller, H. J. Mol. Struct. 1980, 64, 93. 100. Hughbanks, T.; Hoffmann, R. J. Am. Chem. SOC.1983,105, 1150. 101. Wooley, R. G. Inorg. Chem. 1985,24, 3519. 102. Hughbanks, T. Inorg. Chem. 1986,25, 1492. 103. Robinson, L. M.; Bain, R. L.; Shriver, D. F.; Ellis, D. E. Inorg. Chem. 1995, 34, 5588. 104. Lin, Z.; Williams, I. D. Polyhedron 1996, 15, 3277. 105. Cotton, F. A. Acc. Chem. Res. 1969,2, 240. 106. Corbett, J. D. Acc. Chem. Res. 1981, 14, 239. 107. Simon, A. Angew. Chem. Znt. Ed. Engl. 1988,27, 159. 108. Hughbanks, T. Prog. Solid State Chem. 1989, 19, 329. 109. Cotton, F. A,; Hughbanks, T.; Runyan Jr., C. E.; Wojtczak, W. A. In “Early Transition Metal Clusters with pi-Donor Ligands”; Chisholm, M. H., Ed.; VCH Publishers, Inc.: New York, 1995. 110. Briickner, P.; Preetz, W.; Piinjer, M. 2. Anorg. Allg. Chem. 1997, 623, 8. 111. Guirauden, A,; Johannsen, I.; Batail, P.; Coulon, C. Znorg. Chem. 1993,32, 2446. 112. Weissenhorn, R. G. Z. v. Anorg. Allg. Chem. 1976,426, 159. 113. Clark, R. J. H.; Kepert, D. L.; Nyholm, R. S.; Rodley, G. A. Spectrochim. Acta 1966,22, 1697. 114. Cotton, F. A.; Wing, R. M.; Zimmerman, R. A. Znorg. Chem. 1967,6, 11. 115. Zelverte, A,; Mancour, S.; Caillet, P. Spectrochim. Acta 1986,42A, 837. 116. Mancour, S.; Potel, M.; Caillet, P. J. Mol. Struct. 1987, 162, 1. 117. Mancour, S.; Caillet, P.; Jouan, M.; Dao, N. Q. J . Ramun Spect. 1987, 18. 118. Vaughan, P. A,; Sturdivant, J . H.; Pauling, L. J . Am. Chem. SOC.1950, 72, 5477. 119. Koknat, F. W.; Parsons, J . A.; Vongvusharintra, A. Znorg. Chem. 1974,13, 1699. 120. Recheweg, 0.; Meyer, H.-J. 2. Krist. 1996, 211, 396. 121. Ueno, F.; Simon, A. Actu Crystallogr. 1985, C41, 308. 122. McCarley, R. E.; Hughes, B. G.; Cotton, F. A,; Zimmerman, R. Znorg. Chem. 1965, 4, 1491. 123. Espenson, J . H.; McCarley, R. E. J . Am. Chem. SOC.1966, 88, 1063. 124. Espenson, J. H. Znorg. Chem. 1968,4, 631. 125. Espenson, J. H.; Boone, D. J. Inorg. Chem. 1968,4, 636. 126. Eisenbraun, R.; Schafer, H. 2.Anorg. Allg. Chem. 1985,530, 222.
48 127. 128. 129. 130.
PROKOPUK AND SHRIVER
Cooke, N. E.; Kuwana, T.; Espenson, J. Inorg. Chem. 1971,10, 1081. Quigley, R.; Barnard, P. A.; Hussey, C. L.; Seddon, R. Inorg. Chem. 1992,31, 1255. Hussey, C. L.; Quigley, R.; Seddon, K. R. Inorg. Chem. 1995,34, 370. Penicaud, A.; Batail, P.; Coulon, C.; Canadell, E.; Perrin, C. Chem. Muter. 1990, 2, 123.
131. Meyer, H.-J. 2. Anorg. Allg. Chem. 1995, 621, 921.
132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142.
Reckeweg, 0.; Meyer, H.-J. Z. Naturforsch 1995, 50b, 1377. Reckeweg, 0.; Meyer, H.-J. Z. Anorg. Allg. Chem. 1996,622, 411. Klendworth, D. D.; Walton, R. A. Inorg. Chem. 1981,20, 1151. Imoto, H.; Hayakawa, S.; Morita, N.; Saito, T. Znorg. Chem. 1990,29, 2007. Koknat, F. W.; McCarley, R. E. Inorg. Chem. 1972,11, 812. Kashta, A,; BrniEeviC, N.; McCarley, R. E. Polyhedron 1991, 10, 2031. BrniEeviC, N.; Planinic, P.; BaBic, I.; McCarley, R. E.; Rutar, V.; Xie, X. Inorg. Chem. 1993,32, 3786. BrniEeviC, N.; MuStoviC, F.; McCarley, R. E. Inorg. Chem. 1988,27, 4532. BrniEeviC, N.; Schafer, H. Z. Anorg. Allg. Chem. 1978,441, 219. BrniEeviC, N.; MesariC, S.; Schafer, H. Croat. Chem. Acta 1984, 57, 529. Hughes, B. G.; Meyer, J . L.; Fleming, P. B.; McCarley, R. E. Inorg. Chem. 1970,
9, 1343. 143. Fleming, P. B.; Dougherty, T. A,; McCarley, R. E. J. Am. Chem. Soc. 1967.89, 159. 144. Koknat, F. W.; McCarley, R. E. Inorg. Chem. 1974, 13, 295. 145. Field, R. A.; Kepert, K. L.; Robinson, B. W.; White, A. H. J. C. S. Dalton Trans. 1973, 1858. 146. Kennedy, V. 0.; Stern, C. L.; Shriver, D. F. Inorg. Chem. 1994,33, 5967. 147. Bond, M. R.; Hughbanks, T. Inorg. Chem. 1992,31, 5015. 148. Mackay, R. A,; Schneider, R. F. Inorg. Chem. 1967,6, 549. 149. Perrin, C.; Ihmai'ne, S.; Sergent, M. New J. Chem. 1988, 12, 321. 150. Schneider, R. F.; Mackay, R. A. J. Chem. Phys. 1968,48, 843. 151. Fleming, P. B.; McCarley, R. E. Inorg. Chem. 1970,9, 1347. 152. Robbins, D. J.; Thomson, A. J. J. C. S. Dalton Trans. 1972, 2350. 153. Bateman, L. R.; Blount, J. F.; Dahl, L. F. J. Am. Chem. Soc. 1966,88, 1082. 154. Simon, A.; Schnering, H.-G.; Schafer, H. Z. Anorg. Allg. Chem. 1967,355, 295. 155. Imoto, H.; Corbett, J. D. Inorg. Chem. 1980, 19, 1241. 156. Simon, A. 2. Anorg. Allg. Chem. 1967,355, 311. 157. Fitch, A. N.; Barrett, S. A,; Fender, B. E. F.; Simon, A. J. C. S. Dalton Trans. 1984, 501. 158. Imoto, H.; Simon, A. Inorg. Chem. 1982,21, 308. 159. Rogel, F.; Corbett, J . D. J . Am. Chem. Soc. 1990, 112, 8198. 160. Chen, L.; Cotton, F. A. Inorg. Chem. 1996,35, 7364. 161. Finley, J . J.; Camley, R. E.; Vogel, E. E.; Zevin, V.; Gmelin, E. Phys. Reu. B 1981, 24, 1323. 162. Finley, J . J.; Nohl, H.; Vogel, E. E.; Imoto, H.; Camley, R. E.; Zevin, V.; Andersen, 0. K.; Simon, A. Phys. Reu. Lett. 1981,46, 1472. 163. Nohl, H.; Anderson, 0. K. Znt. Phys. Conf Ser. 1980, No. 55, 61. 264. Robinson, L. M.; Shriver, D. F. Coord. Chem. Reu. 1996,37, 119. 165. Jackson, J. A.; Newsham, M. D.; Worsham, C.; Nocera, D. G. Chem. Muter. 1996, 8, 558. 166. Golden, J. H.; Deng, H.; DiSalvo, F. J.; Frechet, J. M.; Thompson, P. M. Science 1995,268, 1463. 167. Robinson, L. M.; Lu, H.; Hupp, J. T.; Shriver, D. F. Chem. Muter. 1995, 7, 43.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181.
49
Robinson, L. M. Ph. D. Thesis, Northwestern Universit,y, Evanston, 1995. Christiano, S. P.; Pinnavaia, T. J. J . Solid State Chem. 1986, 64, 232. Christiano, S. P.; Wang, J.; Pinnavaia, T. J. Inorg. Chem. 1985,24, 1222. Ouahab, L. Chem. Muter. 1997,9, 1909. Ouahab, L.; Batail, P.; Perrin, C.; Garrigou-Lagrange, C. Muter. Res. Bull. 1986, 21, 1223. Penicaud, A,; Batail, P.; Davidson, P.; Levelut, A.-M.; Coulon, C.; Perrin, C. Chern. Muter. 1990, 2, 117. Keggin, J. F.; Miles, F. D. Nature 1936, 137, 577. BrniEevic, N.; Ruzic-Toros, Z.; Kojic-Prodic, B. J . C. S . Dalton Trans. 1985, 455. Vojnovic, M.; Antolic, S.; Kojic-Prodic, B.; Brnicevic, N.; Miljak, M.; Aviani, I. 2. Anorg. Allg. Chem. 1997, 623, 1247. Prokopuk, N.; Shriver, D. F. Chem. Muter. 1998, 10, 10. Deakin, M. R.; Li, T. T.; Melroy, 0. R. J. Electroanal. Chem. 1988,243, 343. BrniEeviC, N.; Nothig-Hus, D.; Kojic-Prodic, B.; Ruzic-Toros, Z.; DaniloviC, Z.; McCarley, R. E. Inorg. Chem. 1992,31, 3924. BrniEevic, N.; McCarley, R. E.; Hilsenbeck, S.; Kojic-Prodic, B.; Acta Crystallogr. 1991, C47, 315. Beck, U.; Simon, A,; Sirac, S.; BrniEeviC, N.; Z . Anorg. Allg. Chem. 1997, 623, 59.
This Page Intention ally Left Blank
ADVANCES IN INORGANIC CHEMISTRY, VOL.
46
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY JOHN H. HOLLOWAY and ERIC G. HOPE Department of Chemistty, The University of Leicester, Leicester LEI 7RH, United Kingdom
I. Introduction 11. Recent Review Literature 111. The Possibility of Argon Chemistry Krypton Chemistry V. Xenon Chemistry A. Core Chemistry and Recent Advances B. Reactions of the Xenon Fluorides VI . Radon Chemistry References
rv.
I. Introduction
Until "XePtF," was reported in 1962 (I),the only compounds containing noble-gas elements in combination with other elements were weakly bonded species of two main types: (i) gaseous cationic or excited-state species observed by spectroscopic means and (ii)species in which the noble-gas atoms are adsorbed, enclathrated, or encapsulated on or within the lattices of other molecules. The former includes mostly excited-state diatomic molecules containing like or unlike noble-gas atoms, or diatomics containing oxygen, nitrogen, or metal atoms in combination with a noble-gas atom, a wealth of van der Waals-bonded dimers and trimers, diatomic cationic species containing like or unlike noble gases, or noble-gas hydrides. These were briefly reviewed in 1968 (2). The latter consists mainly of clathrates in which the noble-gas atoms lie at the interstices of crystalline cages formed by water, organic molecules, or double hydrates involving both organic and water molecules [e.g., species such as R.2Ng.17H20 (R = CH,COOH, CH2C12,CHC1, or CCll, and Ng = Ar, Kr, or Xe)l (2-41, or where noble-gas encapsulation in zeolites occurs ( 4 ) . Subsequent to the discovery of the xenon-platinum hexafluoride 51 Copyright c) 1999 by Academic P r a s All rights of reproduction In a n y form reserved 0898-8838/99 525 00
52
HOLLOWAY AND HOPE
complex, simple binary fluorides of xenon (XeF,, XeF,, and XeF,), krypton (KrF2),and radon (probably RnF2),have been prepared (2-41, and the general chemistry of these species has been well established and developed. Compounds isolable at ordinary temperatures have been obtained only for the heavier noble gases, and the only stable compounds that can be made by direct synthesis are the fluorides. Most of the compounds that have been obtained consist of the noble gas combined with the most electronegative elements, fluorine and oxygen. However, compounds in which the noble gases are bonded to nitrogen, carbon, or chlorine have been characterized, and there is a growing and interesting chemistry evolving around some of these species. The chemistry of xenon is undoubtedly the most extensive, xenon occurring in oxidation states +2 to +8, and stable species have been isolated for each of these. For krypton and radon, the compounds found so far are all in oxidation state +2 and, until quite recently, all of the known species involved only krypton-to-fluorine or radon-tofluorine bonds. The most sensitive detection methods have provided no definitive evidence for a stable argon compound. The inert status for helium and neon remains unchallenged. The chemistry of xenon stems from the reactions of the binary fluorides and falls into five main categories: 1. reactions in which the xenon fluoride acts as a fluorinating agent, 2. hydrolysis reactions yielding oxide fluorides, oxides, and oxygencontaining salts, 3. fluoridelanion metathesis reactions between the fluoride and an anhydrous acid, 4. reactions with fluoride ion acceptors, yielding fluorocations, and 5. reactions with fluoride ion donors, yielding fluoroanions. Krypton chemistry parallels that of xenon but is, in general, more limited. The bonds between krypton and other elements, even the highly electronegative fluorine and oxygen atoms, are weaker than those with xenon. The outcome is that although krypton-fluoride cationic species are known, Kr-0, Kr-N, and Kr-C bonds are difficult to obtain and, where they exist, are unstable. On the other hand, the weakness of the Kr-F bond, both in neutral KrF, and in the cations [KrFI' and [Kr,F,I+, provides a means of carrying out useful low-temperature fluorinations, which have been exploited in producing novel species that are often inaccessible using elemental fluorine. Radon apparently forms a difluoride and some cationic complexes. However, the evidence is based exclusively on radiochemical tracer
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
53
experiments because radon has no stable isotopes, and the work is made even more difficult by radon’s intense radioactivity, which not only makes it hazardous to manipulate but also leads to decomposition of the reagents involved. A wealth of reviews covering various aspects of noble-gas chemistry were written up to the end of 1979 and some of those written in English are listed in reference ( 5 ) .In addition to the books ( 2 ,3 ) and an extensive and authoritative book chapter ( 4 )already mentioned, three other books appeared in the early days of noble-gas chemistry, and these are listed in reference ( 6 ) .There is also a valuable and authoritative bibliography covering the literature in the period 1962 to 1977 (7). Because the main features of noble-gas chemistry are now well documented to the extent that they appear in all undergraduate inorganic textbooks, this review concerns itself only with the new chemistry that has evolved since 1979, and an effort has been made to pick out the main features of this work and set it within the context of the chemistry established up until that time.
II. Recent Review Literature
Since 1979, there have been a number of reviews of note. Among them is an interesting summary of the history of the discovery of noble-gas chemistry in which the reasons for the failure of Yost and Kaye t o obtain noble-gas compounds in the 1930s are examined. The article also speculates whether there were others who envisioned noble-gas reactivity and discusses work that took place before Bartlett’s key experiments (8).Other general summaries of noble-gas chemistry were published in celebration of the 25th anniversary of the discovery of chemical activity of xenon, a t the 100th anniversary of the discovery of fluorine (9) and very recently in a new “Encyclopaedia of Inorganic Chemistry” (10).A review covering aspects of the chemistry of xenon, and especially its reactions with pentafluoro-oxoselenate and pentafluoro-oxotellurate groups, was published a t the beginning of the review period ( l l ) ,and a detailed and comprehensive academic review of cationic and anionic complexes of the noble gases was published in 1984 (12).An important development in noble-gas chemistry over the last 15 years has been the synthesis of compounds containing stable bonds between noble gases and elements other than fluorine and oxygen. This was foreseen in an article in 1982, which traced
54
HOLLOWAY AND HOPE
then-recent developments and discussed Xe-N bonding as well as the fragmentary evidence for Xe-C and Xe-B bonds (13). An authoritative review on the chemistry of radon, which includes the formation of clathrate compounds, the simple and complex fluorides, ionic radon in solution, redox properties, and the various unsuccessful efforts to prepare an oxide and halides of radon other than the fluoride, was published by the most significant contributor to this field in 1983 (14). A review article on van der Waals complexes of metal-atom noblegas complexes was published in 1995 (15).
111. The Possibility of Argon Chemistry
The existence of krypton and xenon compounds inevitably raises the question as to whether noble-gas chemistry might be extended to argon. The increasing range of krypton compounds that have been synthesized over the last 10 years, and in particular the recent observation of Kr-0 and Kr-N species, will undoubtedly further stimulate interest in this area. Although some years ago ab initio calculations on [HeFI' and [NeF]+ indicated that the ground states of these molecules are unstable, those for [ArF]' implied a stable ground state and suggested that a likely complex to be synthesized would be [ArFI+[PtF,I- (16). Evidence from photoelectron spectroscopy did indeed produce evidence of [ArFI+ and provided a measure of its dissociation energy {D,([ArFI+)2 1.655 eV) but confirmed instability for the helium and neon analogs (17). More recently new ab initio calculations at the MP4C3DTQY6-311G(2df, 2pd)//MP2/6-31G(d,p) level predict that the best counterions for the stabilization of [ArF]+appear to be [AuFJ or [SbFJ (18). High-level calculations also predict that [HC=N-ArFl+ should be stable with a binding energy of 160 kJ.mol-' which is closely similar to that of the recently observed krypton analog [HC= N-KrFl+ (157 k J . mol-') (19).Interesting work, in which electronegativity has been redefined and adjustments to the well-established electronegativity scales have been made, has demonstrated close agreement between the four scales (Allred and Rochow, Sanderson, Mulliken, and Pauling) and that the values for argon and krypton are similar (20).All of this suggests that more effort to prepare an argon fluoride complex might be worthwhile.
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
55
IV. Krypton Chemistry
The chemistry of krypton is less extensive than that of xenon, and the well-established characteristics, summarized in detail in the book chapters in 1973 by Bartlett and Sladky ( 4 ) and in 1984 by Selig and Holloway (12) are dealt with in outline here, but the new developments since 1979 are referenced in detail. The only stable binary fluoride of krypton is the difluoride, KrF2, and all the known chemistry derives from this molecule. Early reports of the preparation of KrF4have not been substantiated, and no simple stable oxides or oxide fluorides have been isolated. However, small amounts of the violet free radical, KrF, have been observed following y-irradiation of KrFz (2-4). There has also now been direct observation of [KrO,]' ( n = 1,2) and [&OH]' by Fourier-transform ion cyclotron resonance (FT-ICR) mass spectrometry, for which ab initio theoretical calculations suggest that the two species are covalently bonded with bond dissociation enthalpies of 237.7 and 169.5 k J . mol-' respectively (211. The volatile, colorless crystalline difluoride was prepared initially by subjecting mixtures of krypton and fluorine at low temperature and pressure to an electric discharge or by irradiation of liquid krypton/fluorine mixtures with UV light (22). The latter is a more rapid and convenient way of preparing KrFz than the earlier discharge method, and it has since been found that laser as well as UV photolysis is also effective (23). These photolytic methods are particularly useful for preparing small quantities (22,231. Because molecular fluorine is not significantly dissociated until 500-600°C, the preparation of thermally unstable compounds such as KrFz by a thermal method did not seem to be an option. However, in 1976, Soviet workers reported the preparation of KrFz using catalytic-thermal generation of atomic fluorine from molecular fluorine (24). The nonequilibrium conditions necessary were created by generating atomic fluorine on the surface of a nickel wire at 600-700°C and allowing the atomic fluorine to react with krypton in a low-temperature zone close to the reactor wall a t -196°C. The rate of production is high (-6 g/h) compared with other preparative methods. The reaction method and operating conditions of another high-yield, thermal catalytic method has also recently been discussed (25). Krypton difluoride has now been shown to exist in two crystallographic forms, a low-temperature ( - 196°C) phase, referred to as a-KrFz, and a high-temperature (-78°C) phase, /3-KrF2(23). Study of the vibrational spectra has shown that the X-ray structure reported
56
HOLLOWAY AND HOPE
in 1972 (261, which contains two linear KrFz molecules per tetragonal unit cell (space group P4,lmnrn) aligned in planes perpendicular to the tetrad axis, is that of P-KrF,. Raman data suggests that the a-KrFpcrystallizes in the same space group as XeF, (271, but a definitive conclusion must await an X-ray structural investigation of the krypton species at low temperature. The adducts of KrF, are analogous to those of XeF, and until quite recently were limited to cationic derivatives, [KrFl', [KrFl' .xKrF2, or [Kr2F3]+. KrFz, of Lewis acid pentafluorides (12,281 or the [KrpF31+ related but somewhat less powerfully Lewis acidic metal oxide tetrafluorides (12,28,29). More recent reports have included the synthesis of [KrF]'[MF,J (M = As, Sb) in anhydrous HF (30), the preparation of [K~F]'[AS~F~~]for the first time, and the observation of two solidstate forms of [KrFl+[SbpFJ (23), providing only the second example of polymorphism in krypton complex chemistry {cf. [KrFl+[AsF61(31I}. The cationic complex of krypton difluoride, [Kr,F31+.xKrFz. [BF41-,incorporating the [BF4]-anion, has been prepared for the first time from the reaction of KrFz with BF3at -40°C in HF. The presence of HF to help bring about the ionization was shown to be essential in this preparation because the complex is not produced a t temperatures as low as -78°C when KrFz and BF3 are brought directly together (30). The use of anhydrous HF has also provided a means of synthesizing two new metal fluoride adducts, 2KrFzsMnF4 and KrFz.MnF4, from the combination of KrFz and MnF, in HF solvent (32). The 2 : 1 adduct decomposes in dynamic vacuum at -45"C, yielding KrFz and the 1: 1 adduct, which itself is stable up to -25°C. Because neither species provided a Raman spectrum, their exact nature is uncertain, but it Seems likely that they are molecular adducts rather than complexes with cationidanionic character. In general, the component molecules of all of the cationidanionic complexes are linked together by fluorine bridges, which have a considerable degree of ionic character. However, Raman spectra of the complexes derived from the metal oxide tetrafluorides indicate that they are best formulated as essentially covalent structures (12). The spectra of the [KrpF31'cation correlate well with those associated with the analogous xenon species, but additional peaks, which have no equivalents in the xenon spectra, indicate that, unlike [XezF3It,the krypton cation is unsymmetrical (Fig. 1)with one short and strong bond, %,-Fa, a weaker and longer bond, Krb-F,, and two bridging bonds, Krh----Fhand Kr,----Fb,which are also of different lengths and strengths (31). Alternatively, the cation may be regarded as a dis-
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
57
FIG.1. The structure of IKr2F1]'
torted KrFz molecule fluorine bridged to a [KrFI' cation, [KrFz. KrF]' (33). As in the case of xenon, one of the most interesting developments in krypton chemistry has been the identification of bonds between noble-gas atoms and elements other than fluorine. The first example of such a species containing a Kr-0 bond, Kr(OTeF5In,has been prepared by the reaction of krypton difluoride with B(OTeF& a t -90 to -112°C in S02C1F solvent (341, the thermally unstable product decomposing to species analogous to those found in the decomposition of Xe(OTeF&. The high solubility of B(OTeF& relative t o KrFz in SOzCIF at the low temperatures required for the stabilization of Kr(OTeF& is thought to serve to maintain B(OTeF& in excess and to prevent the formation of FKr(OTeF5),which was not observed ( 3 4 ) . Perhaps even more remarkable is the extension of krypton chemistry to include Kr-N-bonded species in the cations [HCrN-KrF]' ( 3 5 )and [R,C=N-KrFl+ (Rf = CF:<,C2F5,n-C,F,) (36).The first to be prepared, the fluoro(hydrocyano)krypton(II) cation, [HC=N-KrFl', was obtained from the low-temperature reaction of KrFy with [HC~NH]+[AsFJin H F or BrF5 solvent rather than by direct combination of [KrFl+[AsF,]- with HC=N in H F or BrF5 (which was the route used in the preparation of the xenon analog) because, not only is [KrF]' strongly oxidizing toward HC=N and BrF5, it also tends to undergo autoredox reactions in both solvents (35).Warming of [HC= N-Kr-F]+[AsF,l~ in BrF5 solution to -58 to -55°C for several hours produced little evidence of decomposition, but brief warmings above -50°C led to rapid formation of [NF,]', CF4, and CF3H. However, when the solid complex is warmed above -5O"C, krypton, NF3, and CF, gases are rapidly evolved. Extension of the HCN work to include other nitrogen bases that are oxidatively resistant and have first adiabatic ionization potentials close to or exceeding that of [KrFI' has led to the preparation of three other examples of cationic Kr-Nbonded species. Derived from perfluoroalkyl nitriles, these species, [R&=N-KrFI+[AsFJ (Rf = CF3, C2Fs,n-C3F7),have been prepared by reaction of R&=N/AsF5 mixtures with KrF2 in BrF5 a t tempera-
58
HOLLOWAY AND HOPE
tures between -57 and -61"C, and have been compared with their xenon analogs, which have also been made (36). Inevitably, the discovery of the krypton-nitrogen-bonded species has led to the testing of a number of theoretical models. Nonrelativistic quantum chemical calculations, including electron correlation effects, have been done for the ground-state [HC=N--KrF]' cation, and the computed geometrical structure, stability toward dissociation, and harmonic vibrational spectrum agree closely with the experimental data (37, 38). The calculations suggest a Kr-F bond stronger than that in &Fa. The fact that the xenon analog was also made prompted a comparison of the experimental properties of the bonds formed between nitrogen and fluorine to the noble gas atoms, which suggested that the exceptional ability of the [NgFI+(Ng = Kr, Xe) ions to act as Lewis acids is related to the presence of holes in the valence shell charge concentrations of the krypton and xenon atoms, which expose their cores. The study also provided reason to believe that the mechanism of formation of the Ng-N bonds in the adducts is similar to that in the formation of hydrogen bonds (39). The ability of noble-gas fluorides to bring about oxidation and/or fluorination is in the order XeFz = covalent Xe(I1) derivatives < XeF4 < XeFs < &Fa. Also, it is well known that the cationic derivatives, [XeFl.', [XezF31+,[XeF31t, [XeFJ+, [Xe2F1J+,and [KrFI' or [Kr2F3]+have a higher fluorinating ability than their neutral parent compounds (13).Clearly, from this comparison, it is evident that KrFz and its cationic derivatives are the most powerful noble-gas fluorinating agents. Indeed, thermodynamic data indicate that reactions involving KrFz are about 50.2 kJ.mo1-I more exothermic than those with elemental fluorine. This has two consequences: the first is that such reactions have t o be undertaken with great care; the second is that these reagents offer the potential to synthesize other novel highoxidation-state species (13). This important characteristic has been exploited for some time. For example, KrFz is known to fluorinate xenon to XeFs and iodine to IF7(331, whereas the [KrF]' cation reacts spontaneously with oxygen to yield [O,]' and xenon to give [XeFl' (31).The cations, [KrFl' and [Kr2F31+, oxidize BrF, to IBrFJ' [40,411. The fact that [KrFl' takes BrF5 to [BrF,]' but fails to take XeOF, to [XeOFJ' (42) has been explained in terms of the [KrFI' attacking BrF5 at the bromine because of the nucleophilic attraction of its nonbonding lone pair whereas the most nucleophilic part of the XeOF4 molecule is its oxygen atom. Thus, an intermediate XeF4-OF might be expected (43). This is given further credence by the observation of a yellow explosive intermediate in reactions of [KrF]' with
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
59
XeOF, (44). It is also noteworthy that XeOF, is not fluorinated by PtFG (45). The successful oxidation of BrF5 to [BrF,]’ led to investigations on the oxidative fluorination of metals and metal fluorides. The first example was the successful oxidation of elemental gold to gold(V) (42), the treatment of Au powder in anhydrous HF giving a pale yellow solid, [KrFl+[AuFGI-,which is sparingly soluble in H F to give a pale yellow solution: 7KrFz + 2Au
20°C
aHF
2[KrFI+[AuFJ
+ 5Kr.
(1)
Pyrolysis of the krypton complex at between 60 and 65°C gave gold pentafluoride, AuF5: [KrF]+[AuFJ
60-65°C 8h
A u F ~+ Kr
+ F,.
(2)
Since then, the oxidative fluorinating abilities of KrF, have been used to prepare CoF3 (46), MnF4 (321, and the lanthanum tetrafluorides, LnF4 (Ln = Ce, Pr, or Tb) (471, and their heptafluoro complexes, Cs3LnF7(Ln = Ce, Pr, Nd, Tb, Dy, or Tm) (47, 48). Cobalt trifluoride and the lanthanum compounds can also be prepared using xenon fluorides, except for the thermally unstable PrF4,which can only be obtained using KrF, (47). Further careful investigation of the reaction of KrFz with LnOz (Ln = Ce, Tb, or Pr) in 2 : 1 or 3 : 1 ratios has shown that, here again, LnF, is produced, but reaction at a 1 : 1 ratio gives a product with a composition close to that of LnOFz that has been shown not t o be a mixture of Ln02 and LnF4 (49).Another species in oxidation state four that has been prepared is the hexafluoronickelate(1V) anion, “iF612-, as the [Xe2F,J$[NiF61z-salt, by the reaction of NiFz, XeF6, and KrF, in anhydrous HF (50). Reaction of KrF, with [XeF51+[AuF,I-in anhydrous H F below 273 K gives [XeF5]+[AuF6]-(51). A number of high-oxidation-state fluorides and oxide fluorides have also been prepared using KrFz. For example, Cr02F2can be converted t o CrOF4 (29) [rather than CrOF3 which is the product of reaction with XeF, (5211 and has been shown to form a stable 1 : 1 adduct with KrF2 which has a n essentially covalent structure with a Kr----F----Cr bridge (Fig. 2). A notable achievement has been the synthesis of AgF3 (53)by precipitation from anhydrous HF solutions of [AgF41- salts by addition of fluoro acids such as BF,, PF,, or AsF5. It was also shown
60
HOLLOWAY AND HOPE
Kr
F'
FIG.2. The structure of CrOF4. KrFz.
that previous claims were in fact due to Ag(II)Ag(III)F, (53).There has also been considerable interest in the derivation of high-oxidation-state fluorides from a number of the actinide oxides or fluorides. Reactions of the oxides and tetrafluorides of uranium, neptunium, and americium, and the oxfluorides of neptunium, with KrF, in BrF5 or HF were reported in the mid 1970s and early 1980s (53-55), and brief mention has been made of the preparation of UF6 from UF4 (56), the fluorination of NpOF4 to NpFs (571, and the oxidation of PuF3 to PuF, (58) using KrF, at room temperature or lower. More recently (591, it has been shown that not only uranium and neptunium oxides and low-valent fluorides but also plutonium tetrafluoride can be converted to their hexafluorides at ambient temperature using KrF2vapor or KrF, in anhydrous HF. However, efforts to prepare AmFs from AmO, in anhydrous HF failed, in contrast to the earlier report (55). The highest oxidation-state species synthesized recently using KrF, in anhydrous HF have been TcOF,, made by reaction with Tc02F3 (60), and the new osmium(VII1) oxide fluoride, cis-OsF402,prepared by the fluorination of O s 0 4 (61).The latter was previously erroneously characterized as OsF60 (62). Another interesting aspect of the oxidative fluorinating ability of krypton difluoride is its ability to produce coordinatively saturated fluorocations. By 1974 three such species were known, namely, [NF41+,[ClFsl', and [BrF,]', the last being produced using [KrF]' or [Kr,F,I+ (41, 42). It was later shown that the [NF4]+salts could be prepared using the [KrFl' cation, KrF, reacting with NF3 in the presence of the Lewis acids SbF,, NbF5, PtF5, TiF4, and BF, to yield [NF41'[SbFsl-, [NF41'[NbF61-, [NF4]'[PtF6]-, [NF,I,"TiFsI2-, and [NF41+[BF41respectively (63).In more recent studies, the syntheses of [NF,I'[SbF,I-, [NF41'[AsF61- (301,and [NF41+[BFJ (25, 30) have been studied under different conditions. The fluorination of NF, by [KrF]'[SbFsl- in anhydrous HF has been shown to proceed quantitatively at temperatures as low as -31"C, indicating an ionic two-electron mechanism. It was shown that CIFBand BrF5 could also be con-
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
61
verted to their complex fluorocations by [KrFl'. In contrast, PtFs was shown to be capable of oxidizing only NFB and ClF, but not BrF5. Efforts to fluorinate OF2, CF3NF2, and [C1F401- with [KrFI' salts, however, were not successful (30). An unusual reaction of KrF, has been the observation that KrF, decomposes on zeolites at 233 K, fluorinating the zeolite and encapsulating the krypton released. A proportion of the latter is retained even upon heating to 773 K, demonstrating the potential for R5Krrecovery using KrF,/zeolite systems (64).As might be expected, the fluorination of C,,-fullerene by KrF, is extreme, yielding species ranging from C,,F,, to the hyperfluorinated species C6OF78, as determined by mass spectrometry. An interesting observation in these experiments, however, is that in comparison with other fluorination methods, the concentration of oxygen-containing derivatives is low (65).
V. Xenon Chemistry
A. CORECHEMISTRY AND RECENT ADVANCES With the exception of XeF, which is obtained as a n unstable free radical (2-4), there is no evidence for the occurrence of xenon compounds in odd-numbered oxidation states. The only species that can be synthesized directly from the elements are the three fluorides XeF:,, XeF, , and XeFs . Early reports of the preparation of XeFHhave not been substantiated. There are a variety of methods for the preparation of XeF2 (2-4), and recent additions include oxidation of xenon using the blue solutions of AgF, in H F (66)and a new thermal catalytic synthesis (25). The latter is related to that recently devised for the preparation of gram quantities of KrF2( 2 4 )and has the potential to become one of the better methods for producing XeF,. However, this route in the hands of other workers appears to yield rather pure XeFs (671, so the conditions need further study and appraisal. In the meantime, the best methods to make the difluoride are by heating an excess of xenon with fluorine (2-4) or by irradiating Xe/F2 mixtures with ultraviolet light or sunlight (2-4, 68). Xenon tetrafluoride is the most difficult xenon fluoride to prepare because, even under optimum conditions (heating xenon and fluorine under pressure in a 1: 5 ratio), equilibrium concentrations of XeF2 and XeF, occur (2-4). The claim that XeF,, free from contamination by XeFz or XeF6, can be prepared by irradiation of a gaseous mixture of xenon and fluorine (molar ratio
62
HOLLOWAY AND HOPE
2 1:2) in a reactor with walls coated with NiF2 (69)is probably not valid. However, the thermal combination of xenon and fluorine in a 1:20 ratio of 350°C for two days followed by one day at 50°C in the presence of sodium fluoride yields mainly the involatile complex, Naz[XeF8][Eq. (311 but with traces of XeFz and XeF, . After removal of the volatile mixture of XeF2 and XeF4, the Na2[XeF81is decomposed a t 350°C to give XeF4 [Eq. (4)l (70):
Xe + 3F2+ 2NaF
350°C (2 d) 50°C (1 d)
Naz[XeFsl
350°C
Naz[XeF81
XeF,
+ F2+ 2NaF.
(3) (4)
Another route to high-purity XeF4that has been discovered recently involves a modification of the thermal catalytic method (24)in which the fluorinating agent is 02Fzrather than Fz (67). The best preparative route to XeFs remains the thermal combination of xenon and fluorine in a 1:20 ratio at high pressure (2-4). However, the thermal catalytic method using elemental fluorine also seems to yield a pure product, albeit in a lower yield (67). All of the xenon fluorides are colorless solids except the hexafluoride, which is yellow in the liquid and gaseous states. The di- and tetrafluorides have low volatility and the expected linear (Dmh)and square planar (D.,h) structures respectively in both the solid and vapor states. The bonding in these structures has been accounted for using a three-center, four-electron bond model similar to that employed to explain the structure of C02.Although the structures and bonding for XeFz and XeF4 have been widely accepted for many years, physicochemical studies continue t o be reported and refinements made. For XeF2,new Raman spectra in the solid state and vapor phase at 325 K have been reported, and force constants have been evaluated using a SVFF approximation (71).Raman study of the molecular dynamics of XeF2 in acetonitrile (721, hydrogen fluoride, and bromine pentafluoride (73) (which are the most widely used solvents in noble-gas fluoride chemistry) at various concentrations has been carried out, and vibrational and rotational correlation functions, as well as the characteristic times, have been calculated. The structure of XeFs in the solid, liquid and vapor states continues to be discussed in detail. Out of all the known hexafluorides, it has the highest boiling point, suggesting that, unlike other hexafluorides, it might be polymeric in the liquid and solid states. Some of the uncertainty has arisen because of difficulties in manipulating the com-
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
63
pound. Its powerful fluorinating ability leads to it damaging container and window materials and producing lower-xenon fluorides, which contaminate the specimen. Another difficulty is its ready hydrolysis to the dangerously explosive solid, Xe03. In the vapor state, it has been clear for some time that an octahedral structure for this molecule could not have the lowest energy (74-76). The VSEPR model (74, 75) predicted that an extra lone pair should be stereochemically active and distort the molecule to lower symmetry. Experimental and theoretical work has demonstrated that this is indeed the case, with the molecule having local minima giving CBvand C2, geometries via a pseudorotation mechanism in which the lone pair moves over the octahedral faces and edges (77, 78). Theoretical work in 1987 provided additional information concerning Oh, C2", and Csv structures (791, and more recent experimental work, using high-resolution photoelectron spectroscopy to resolve the ligand field splittings, yielded bond angles for the CSvstructure that agreed with the theoretical work (80). These studies also confirmed earlier work that indicated that the dipole moment is very small. However, both pieces of work referred to the Cpvstructure as a local minimum, and no theoretical vibrational frequency analysis was presented to support this. Now, the Oh, Czv, and C3, structures have been studied with a significantly larger basis set, and harmonic vibrational frequencies for each have been determined and analyzed (81). At the SCF level of theory, the Oh and Cpv structures have been shown not to be minima but to be transition states whose imaginary vibrational frequencies lead to the C13" structure. Nothing new has emerged recently about the structure of XeF, in the liquid state and in solution, and the current state of knowledge was well reviewed by Seppelt and Lentz in 1982 (13). In both the molten state and in solution, monomeric XeFs is in equilibrium with tetrameric ([XeF5]+F-)4 but at low temperatures the monomer disappears. The 25 fluorines of the tetramer are involved in a complex scrambling mechanism, which is perhaps related to that in Rh,(CO),, , although it is clear that no xenon-xenon bonds equivalent to the metal-metal bonds in the rhodium compound occur. In the solid state, at least four phases exist, of which, because of lack of X-ray intensities, absorption problems, and disorder, only the cubic structure, which contains 144 XeF6 molecules in the unit cell arranged as 24 tetramers and 8 hexamers, has been determined. In each, the fluorines attached to each xenon form square pyramidal configurations, with the remaining fluorine atoms contributing to a twofold or threefold bridge linking adjacent xenons. Because the bridging
64
HOLLOWAY AND HOPE
F- ions are not close to the fourfold axis of the [XeF51+groups, it has been suggested that lone pairs of electrons may occupy this position. Thermochemical data on the xenon fluorides continues to accumulate. New melting and boiling points of XeF, at 136 2 5°C and 338°C respectively, and an enthalpy of fusion of 14.04 2 2.09 kJ.mol-l, have been obtained, and disproportion of XeFz to XeF, and Xe has been observed above 350°C (82). A new determination of the heat of formation of XeF, , obtained by the combustion of germanium in the difluoride, is in close agreement with an earlier value obtained by reaction of XeFz with PF3(83). Existing thermodynamic data have been used to calculate successive one-electron potentials for the reduction of XeFa (84). From the beginning of noble-gas chemistry there has been continued interest in theoretical and spectroscopic aspects. The more recent include an ESCA study of XeF, to high statistics in which the satellites observed on the low-energy side of the xenon 3d and 4d primary photolines have been attributed to monopole transitions from the occupied MOs of XeFz to the antibonding and/or Rydberg-state orbitals (85). The nature of the satellite lines has been identified using theoretical SCF-Xa-SW calculations. This has been followed up by X-ray photoelectron spectroscopy and theoretical studies on XeF, and XeFs in which the Xe 4d and Xe 3d spectra have been reported. However, the main center of focus in this paper was on KrFz (86). Other theoretical work that has been carried out includes SCF-MO studies on XeFz and XeF, (87) and the application of the multiple-scattered-wave-SCF-X, method to estimate ionization potentials and charge distribution in XeF,, XeF4, XeFs, and XeC1,. The role of 5d orbitals, especially their importance in XeF, and XeFs binding, has been discussed (88). Many attempts have been made to understand theoretically the electronic structures of XeF, ( n = 1, 2, 4, 6) species, the majority having been made since the mid-1970s (79, 81, 89-93). Generally, these have utilized the conventional Hartree-Fock molecular-orbital approach and in some cases correlation effects have been included (81, 89). Relativistic effects have been ignored despite the fact that, due to the high nuclear charge of xenon, they should be high. The spectra have been studied by different techniques (e.g., 75, 80, 94). Now, allelectron Dirac-Fock and Dirac-Fock-Breit calculations for the series have been caried out, and the role of relativistic effects on bond lengths and dissociation energy has been reported (95).
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
65
B. REACTIONS OF THE XENON FLUORIDES With the exception of the reactions of xenon with the powerful fluorinating agent PtF, and its relatives, all of the chemistry of xenon has been derived from reactions of the binary fluorides. 1. Hydrolysis Reactions Yielding Oxide Fluorides, Oxides, and
Oxygen-Containing Salts All of the known oxides, oxide fluorides and a number of oxygencontaining salts stem from hydrolysis reactions of XeF, or XeF,. It has been well established for some time that no stable oxo-species can be obtained from the difluoride, XeF2. Stable in acid or neutral solution, it decomposes instantaneously in base with liberation of xenon, oxygen, and hydrogen fluoride. The tetrafluoride, on the other hand, is instantly hydrolyzed by water and, depending on the conditions, up to one third of the xenon may be retained in solution. There is also evidence that some XeF2 may be liberated. There has been little recent work in this area, but the most recent (96)suggests that the initial product may be XeOFz and that this decomposes to give XeFz if the water supply is limited. In the presence of larger amounts of water, XeOz and XeO may be produced and in turn can produce Xe03. The reaction of XeF, with water is difficult to control and can result in explosions. However, the reaction can be controlled by passing dry nitrogen over crystalline XeF, a t room temperature to sweep the vapor into water, where the following reaction occurs: XeFs + 3H20+ XeOB+ 6HF.
(5)
Hence, the aqueous hydrolysis of both XeF, and XeFs results in the formation of the trigonal pyramidal trioxide, Xe09,in solution. Aqueous solutions of the trioxide, known as “xenic” acid, are fairly stable and can be used as powerful but kinetically slow oxidizing agents. These might have found wide application were it not for the fact that solid XeOB,which is readily obtained from these solutions on evaporation, is a violent and sensitive explosive. The addition of alkali to XeO, solutions yields “xenate” ions [HXeOJ. Although salts of this species can be isolated, solutions containing [HXe041- disproportionate to give the “perxenate” ion, [XeO6I4-,and xenon: 2[HXe041- + 2[OH]- .--, LXe0614-+ Xe
+ O2 + H20.
(6)
66
HOLLOWAY AND HOPE
Within the time frame of this review, an improved synthesis of perxenate solutions has been devised in which XeFs is dissolved in excess of HOPOFz under carefully controlled conditions, the excess of HOPOFz, HF, and OPF3 is removed, and the Xe03 remaining is dissolved slowly in 4 M NaOH at 0°C (90). Both reactions have been carried out on the 6- to 7-mmol scale without explosions occurring. The addition of solid BazXe06 to cold concentrated sulfuric acid yields a second xenon oxide, XeO,, as an unstable and explosive gas. Not surprisingly, this has been little studied, but infrared spectroscopy and electron diffraction have shown that it has the expected tetrahedral geometry. The controlled hydrolysis of XeFs with water in a 1:1 molar ratio produces the colorless, stable, volatile liquid XeOF4, which has a square pyramidal structure: XeF,
+ HzO+ XeOF4 + 2HF.
(7)
Efforts to produce XeOzFzby the addition of a further stoichiometric quantity of water have resulted only in the production of XeO,, and no evidence for a dioxide difluoride intermediate has ever been obtained. Of the known oxide fluorides, XeOF,, XeOF4, XeOzF2,and XeO3FZ (2-4, 96), the most stable is XeOF4.The monoxide difluoride, XeOF,, mentioned earlier, is unstable but can be obtained from the reaction of XeF, with water at -80°C. The dioxide difluoride is obtained by mixing XeO, with XeOF4 at dry-ice temperature and fractionally distilled from the resulting mixture of XeOzF2,XeOF4, and XeFz. Although thermodynamically unstable with respect to XeF, and Oz, it is sufficiently stable to be held in well-fluorinated Kel-F containers for several days. The trioxide difluoride can be obtained by reaction of XeFs with either Xe04 or Na4[Xe06]but is contaminated with other xenon species, principally XeOF4 (3, 4, 96). The reactions of xenon fluorides and oxide species have been systematized in order of descending acidity, XeFs > XeOZF4> XeO3FZ> XeO, > XeOF4 > XeF, > XeOzFz> XeO, = XeFz (97). The most recent advances in this area include new syntheses of XeOF4 by the reaction of NaN03 (98) [Eq. (€311 or POF3 (99) with a very slight excess of XeFs. In the case of the former, an excess of NaN03 has to be avoided to suppress the secondary reaction (98)[Eq. (9)1, whereas in the latter, excess of POF, may lead to generation of explosive XeO, (99):
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
+ FNO, NaN03 + XeOF, + NaF + XeOzF2+ FNO, . NaN03 + XeFG+ NaF + XeOF,
67 (8)
(9)
Reaction of CsN03 with a large excess of XeF6 at temperatures above the melting point of XeF, proceeds quantitatively according to Eqs. (10) and (11)to yield XeOF, (100). However, if in Eq. (10) the excess of XeF, is reduced to less than loo%, some of the CsF can react with XeOF, to yield Cs+[XeOF,I-. For this and other reasons, substitution of CsN03 by NaN03 is advantageous (98,100). The reaction of CsN03 with a n excess of XeOF, yields FNOz and XeOzFzin quantitative yield with CsF and CsXeOF6 as by-products. Provided a n excess of XeOF, is used, a better new and safe synthesis for Xe02F2is the reaction of XeOF, with N205(100): CsN03 + XeF, -+ CsF + XeOF, CsF + XeF, -+ CsXeF;..
+ FNO,
( 10)
(11)
New chemistry of XeOF4 and XeOzF2includes the preparation of a series of novel pentafluoro-orthotellurate derivatives (1011. Reaction of Xe02Fzwith B(OTeF& according to Eq. (12) has yielded the previously unknown compound OzXe(OTeF5),:
However, with excess of Xe02F2,a mixture of O,Xe(OTeF&, 0,XeF (OTeF,) and XeOzFzwas produced. There has been very little new information on xenic and perxenic acids and their salts since the late 1960s. More recent information includes study of the spectra and protolytic properties of aqueous perxenic acid in the pH range -0.2 to 13, which yielded pK values at 25°C as follows:
The enthalpy of dissociation of [H&!O6]- at 25°C is 0 2 5 kJ.mol-' and that of [H,xeo612- is 3 % 2 kJ.mol-1 (102). The participation of 5d orbitals in the bonding of high-valent oxygen compounds, including [HXeOJ and cXe0614-,has also been investigated by examination of the Kl and I(pl X-ray emission lines in these ions (88).
68
HOLLOWAY AND HOPE
2. Reactions with Fluoride-Ion Donors and Acceptors
A major part of the chemistry of xenon fluorides and oxide fluorides is associated with their reactions with fluoride donors and acceptors to produce a variety of cations and anions. The known species are summarized in Scheme 1. In addition, there are a number of more unusual entities, notably the dixenon cation [Xezl+and [XeO3X1- (X = F, C1, or Br). The preparations, structures, and chemistry of these species was comprehensively reviewed in 1984 (121, so only developments since 1979 are covered in this present account. It is important to remember that the formulation of many of these species as salts is an oversimplification. In general, most are essentially fluorine bridged but with a high degree of ionic character. Also, there are a number of adducts, such as XeFz.XeF4 and XeF2.1F5, whose overall formulas are suggestive of donor-acceptor species, which are simple molecular adducts. a. Xenon Cations The transient green coloration observed in the early days of noble-gas chemistry was identified in 1978 as being due to the dixenon cation (103, 104). The cation can be produced either by oxidation of gaseous xenon using [O,]' or by reduction of [XeFI' with water or other reducing agents such as lead or mercury. The green species is stable in SbF, indefinitely at room temperature under an atmosphere of xenon gas (103-105). Now, almost 20 years after the initial characterization of [Xez]+(103),a structural characterizahas shown that the Xe-Xe bond is weak and tion of [Xezl+[Sb4FzJlonger, at 3.087(1) A, than any other main-group element bond observed so far (Fig. 3). It has also been shown that the presence of HF in the SbF5 is essential (i.e., no reaction of [XeF]'[SbzFJ with Xe XeF2 XeF, XeOFz XeF6
-
[XeF]' ; [XezF3]+ [XeF3]+ [XeOFd
[XeF5]+ ; [Xe2F,,I'
SCHEME 1. Major cationic and anionic derivatives of xenon fluorides and oxide fluorides.
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
69
FIG.3. The structure of [Xe,l'[Sb,F,,I-. IReprinted with permission from (106); copyright 1997 VCH Verlagsgesellschaft mbH. I
occurs in the presence of HF-free SbF5),and it is speculated that the ion IXe(HF)21z+ is the active oxidizing agent in the Xe/XeFz/SbF,/HF system (106). One piece of recent chemistry of IXe,l+ has been its reaction with bromide and iodide ions in the presence of xenon gas, which yields the excited complexes XeBr": and XeI* respectively (107). Xenon difluoride has an extensive fluoride-ion donor chemistry, which is exemplified by its reaction with a large range of transition metal pentafluorides to give adducts formulated as [XeFl'[MzFI,l-, [XeFI'[MF,I-, and IXezFyl+[MF,Iin which the cations and anions are linked by weak fluorine bridging bonds, F-Xe----F----M.The V-shaped [Xe2F31'cation is symmetrical, two [XeFI' units being linked by a fluorine bridge. Related complexes incorporating the weaker Lewis acids, MOF, (M = W or Mo), are similar but have less ionic character. Essentially, no new chemistry has been derived in this area since publication of the review by Selig and Holloway in 1984 (12). However, it has been shown that [N,FI-[AsF,I- will oxidize xenon to [XeFl' (108). On the theoretical front, the electronic structures and bonding in the [Xe2F31'cation and the unknown but isoelectronic and isostructural neutral species XeIF3 (i.e., [FXeFXeF]' and FXeFIF respectively) have been calculated using ab initio molecular orbital theory with polarized split-valence basis sets (109). The [XezF31' cation is known to be bent at the central fluorine, but the calculated structure is liqear and the bonds from the xenons to the central fluorine are 0.09 A shorter than the experimentally determined values. The calculated vibrational frequencies are in reasonable agreement with
70
HOLLOWAY AND HOPE
experimental values after scaling. The [XezF31fcation is a good example of a 512,6e hypervalent bond; for XeIF3,the bonding is less delocalized but the components of the six-electron (5c, 6e) hypervalent bond are still present. In accord with enthalpy of ionization data, xenon tetrafluoride forms complexes with only the strongest fluoride-ion acceptors, but XeFs has a much more extensive and complicated chemistry, forming 1: 2, 1:1, 2 : 1, and 4 : 1 adducts with a variety of tri-, tetra- and pentafluorides that incorporate [XeFJ' or [Xe2F11]+ cations. Solutions of the hexafluoride in anhydrous HF have a higher conductivity than those of XeFz and XeF4 because of the occurrence of both of these cations in solution (12): (DLeF51tF-)4+ nHF G 2[XezFlll++ [(HF),I[XezFlllt+ nHF + 2[XeF51' + [(HF),Fl-.
(13) (14)
Work that was not covered in the 1984 review (12) or has appeared since includes preparation of 1: 1 complexes with LnF3 (Ln = La, Pr, or Nd), which, upon thermal decomposition, gave rise to the adducts 2LnF3 XeF, and 3LnF3* XeFs. The stabilities of the 1: 1 and 2 : 1complexes appeared to decrease with increasing atomic number of the lanthanide (110). Reactions with tin and lead tetrafluorides have given the adducts XeFs.4MF4,3XeFs.4MF4,4XeF6.MF4(M = Pb, Sn), and 4XeFs.3SnF4 for the first time (111).The 1 : 4 adducts contain [XeFJ' cations and polymeric anions, and the 4 : 1 complexes contain [Xe2F11]+cations. The complexes with 4 : 3 and 3 : 4 stoichiometries are novel and are the first xenon hexafluoride complexes that contain both [XezFllltand [XeFJ+ cations. More recent work from the same group on the 4XeF6.NiF4species is the first to have shown definitively the presence of two [XezFlll+cations in the same compound. The complex, prepared from the reaction of NiFz, KrFz,and XeFs in (50).It has anhydrous HF, is therefore formulated as Dre2Flll~[NiFslZalso been shown that displacement of BrF, from BrF3.AuF3with XeFs gives [XeF,lt[AuF,l-, which is isostructural with [XeF51'[AgF41-, the structure of which contains double layers of [XeFJ' and layers of [AgF41-with all the layers parallel to the ab-plane. The greater basicity of [AuF4]-relative to that of [AgF41-implies that the ligand charge in the gold anion should be higher than that in the silver salt. This appears to be confirmed by the difference in the unit cells of the two salts and is consistent with observed differences in the basicity and oxidizability of the anions (51).
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
71
Work on species incorporating metal pentafluorides includes data on 2XeFs-VF5,which had been reported earlier, and the novel complexes XeFs.nVFS( n = 1 or 2). This Raman study suggested that ionic character decreases from essentially ionic to molecular in the order 2XeF6.VF,, XeFs.VF,, XeFs. 2VF5(112). More recently, the first BiF, complex, the white diamagnetic solid [XeF,I'[BiF,I-, has been obtained. This exhibits a strong fluorine-bridge interaction between the cation and anion, which distorts the anion from Oh symmetry (113). Further detailed studies on the interpretation and assignment of vibrational spectra have been reported by two independent workers (114,115).
Adducts of XeFz with [XeF,]' salts were first examined in 1971 (1161,but to obtain a firmer understanding of the factors that determine adduct formation and to investigate the possibility of lower fusion temperatures, the phase diagram for the XeF2/[XeFS]'[AsF61-system has been studied and representative structures of the XeFz: [XeF,I'[AsF,I- products (1: 2, 1: 1, and 2 : 1) have been carried out by X-ray single-crystal methods (117).Whether the complexes are polymeric or nonpolymeric is correlated with the coordination number of the [XeF,]' cations in the complexes, and this is determined by the extent of ionization of the XeFz species. There has been significant interest in the problems associated with heptacoordination in recent years [see (118)references 1-111, and although [XeF71+has not been synthesized, its possible existence is of major interest in this context. Recent ab initio calculations employing effective core potentials and density functional theory calculations have been used for analysis of the isoelectronic series [TeF,]-, IF7,and [XeF71+.These show that all three possess a pentagonal bipyramidal equilibrium geometry and that [XeF7]' is a stable structure. However, it is conceded that the synthesis of such a species is likely to be very difficult because its calculated oxidized strength is closely similar t o that of [KrFl' (118). Xenon oxide tetrafluoride is a poorer fluoride ion donor than XeFz and XeFs, and therefore complexes only with the stronger Lewis acids. Thus, although [XeOF31t[SbFsl- and [XeOF,I+[Sb,F,,I- have been known for some time (121, the first single-crystal structure determination of the [XeOFJ' cation was obtained on the [XeOF,]' [SbFJ salt more recently. As expected, the isolated cation has a geometry that is consistent with VSEPR theory. Oxygen-17 and -18 enrichment of the [XeOF3]+cation in HF and SbF5 solvents has allowed comprehensive investigation of the 170and Iz9Xenuclear magnetic resonance of the cation (119). Xenon dioxide difluoride yields
72
HOLLOWAY AND HOPE
[Xe0~Fl'[SbF~l-and [XeOzFlt[SbzF1ll-,but these too are somewhat unstable, the cation decomposing slowly in SbF, according to Eq. (15) (12): SbF,
[XeO,F]++[Xe-Fl' rt
+ 0,.
(15)
However, the reaction of XeOzFzwith excess of AsF5 in anhydrous HF has been shown t o yield the [XeOzFlt[AsFsI-species, which has a dissociation pressure of 7 torr at 23°C. Under vacuum or when exposed to a laser beam, this loses AsF6to give [FOzXeFXeOzFl'[AsFsl-. This same species can also be prepared by reaction of XeOzFzwith AsF, in a 2 : 1 molar ratio in anhydrous HF. A detailed Raman study of the two adducts has shown that, whereas the [FO2XeFXeOzF1' [AsFJ species has minimal interaction between the anion and cation, the [XeOzFlt[AsF6]-species is significantly fluorine bridged ( 120). b. Anions Derived from Xenon Fluorides and Oxide Fluorides Fluoride ion acceptor behavior among the xenon fluorides has, until the late 1980s, been limited to XeFs. Reactions of the alkali metal fluorides with XeF, yield MzXeFs(M = Cs, Rb, K, or Na) and M'XeF7(M' = Cs or Rb), which lose XeFs when heated. The thermal stability of the adducts increases with increasing relative molecular mass, and the decomposition temperature of the sodium complex, which is below 1OO"C, provides a means of separating XeFs from XeFz and XeF4 (see Section V,A). The related NOF and NOzF adducts have also been prepared, and the NOF derivative [NOld[XeFJ- was shown some years ago to contain a slightly distorted square antiprismatic [XeFs12-anion (12). More recently, it has been demonstrated that XeF, also reacts with [NF41+[HFzIin anhydrous HF to give [NF,I+[XeF,I-, which uncomdergoes laser photolysis at 488 nm t o yield the [NF41~[XeF,I2plex. Vibrational spectroscopic data on these two complexes, along with those of the previously prepared cesium salts, and evidence for the existence of NaXeF, have been obtained (121). The [NF41+CXeF71salt is by far the most energetic of the known [NF,]' salts, and a DSC study of its decomposition has provided a means of calculating its heat of formation (AH? = -490.7 kJ.mol-') (122). The most important and most recent work in this area, however, is the X-ray structural analyses of Cs'[XeF71 -, [N021+[XezF131(123), and Csz+[XeF8l2-. 4BrF5 (124). Prepared from the reaction of XeFs with CsF in BrF5 at 4"C, the anion in Cs+[XeF,] is a capped octahedron, a geometry strictly enforced by the symmetry constraints of the cubic lattice.
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
73
This, of course, is at variance with most other seven-coordinate maingroup compounds, which are pentagonal bipyramidal. The bond from xenon to the capping fluorine is long (210 pm) compared with the average of the rest of the xenon-fluorine bonds (Xe-Foh = 195 pm). This may be a consequence of interaction between Xe-F,,, and the nonbonding electron pair, or it may arise from interaction with three cesium cations (the other fluorines have contacts to only two cations). The anion in [NOzl+[XezFJ can be considered as an adduct of an [XeF7]-anion with a discrete XeFfimolecule (Fig. 4). This is the first structure in which distinct XeFs as opposed to [XeFJ'F- units have been observed. The [XeF71- adopts a capped trigonal prismatic arrangement in which the Xe-F,,, bond is the shortest and the lone pair appears to be between the two Xe-F bonds trans to Xe-F,,, . The XeFs part of the adduct has CZvsymmetry, with two short, two intermediate, and two long Xe-F bonds, and so is significantly distorted from the regular octahedral structures exhibited by the isoelectronic [IF,]and [SeF,I2- anions (123).In the BrF,-solvated Csz'[XeF8I2-complex ( 124), the anion assumes the near-regular square antiprismatic structure found in (N0)2'[XeF81" (125). This structure is related to that of [IF&, but the Xe-F bonds are longer. This can be explained in terms
I I
I I
Q-
F21
F2d
F23
6
3 F 1 5 F13
\
\
1 \ \ \
FIG. 4. The structure of [N021+[Xe2F,31-, showing the disposition of the [XeF,I- and discrete XeF6 molecule and the INO,]'. [Reprinted with permission from (123); copyright 1996 VCH Verlagsgesellschaft mbH.1
74
HOLLOWAY AND HOPE
of shielding by the centrosymmetric nonbonding electron pair or by increased polarity of the Xe-F bonds due to the increased negative charge. The first evidence of possible fluoride-ion-acceptor behavior of XeF, was from reports of the synthesis of Mi[XeF612-(M = Cs, Rb, K, Na) salts from the reaction of XeF, with MF (126-128). The characterization of these species as [XeFs12-salts was based on vibrational spectra (2281, which were later shown to be incompatible with those for an octahedral species (129). Furthermore, the Raman spectrum attributed to the cesium salt was found to be identical to that due to a species observed during the laser photolysis of Cs+[XeFJ and tentatively assigned to Cs~lXeFs12(1211. These discrepancies prompted an investigation of the reactions of XeF4 with N(CH3),F, which is a powerful reagent in the synthesis of complex fluoroanions, and a reexamination of the XeF4/MF systems, which has led to the identification and full characterization of [XeFJ for the first time (130).Direct reaction of excess of XeF, with the alkali-metal fluorides, MF (M = Cs, Rb, K, Na) at 190°C and reaction of equimolar amounts of XeF4 with N(CH3),F in CH3N at -40°C or of XeF4 with excess of FNO at 0°C gave the 1: 1 complexes. Detonation of [N(CH3)41+[XeF5]in CH3CN when cooled with liquid nitrogen can occur, and [NOI+[XeF.J is unstable, with a dissociation pressure of 10 torr at O"C, but the alkali metal salts are stable, as was found in the earlier work (126-128). All of the salts were characterized by vibrational, 19F,and 129XeNMR spectroscopy, and an X-ray single-crystal structure determination was carried out on [N(CH3)41+[XeF51-.The structure contains planar [XeF51- anions of D5h symmetry in which all six atoms are coplanar. The results from the structure and from a normal coordinate analysis indicate a greater rigidity of the XeF5 plane than that in the fluxional IF7, which is due t o repulsion from the xenon free electron pairs (130).A study of the mean amplitudes of vibrations of the [XeFJ ion has also confirmed that the lone pairs exert repulsion, which forces bond lengthening (131). Complexes of XeOF4 with CsF, RbF, KF', and NOF were first prepared in the 1960s (12).The alkali-metal compounds were prepared by treating the fluorides with an excess of XeOF4 and pumping to constant weight under appropriate conditions (for example, see Scheme 2). The structures of the 3 : 1 complexes were shown by X-ray crystallography to contain the [F(XeOF,)J anion, which consists of three essentially undistorted XeOF4 units solvating an F- anion (132, 133). On the basis of vibrational spectroscopic data (1331, the [XeOFJ anion in Cs+[XeOF,]- was assigned a low-symmetry C, struc-
75
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
- oocvac.
CsF + XeOF, (excess)
20OC vac.
CsF.XeOF,
-XeOF,
2CsF.XeOF4
-XeOF,
1I
-XeOF, -125OC vac -270°C vac. 3CsF.XeOF4 -3CsF.2XeOF4 -XeOF,
c
SCHEME 2. The reaction of XeOF4 with CsF.
ture. However, recently two independent studies have described reinvestigations of this anion with different cations. Based on vibrational spectroscopic data and ab initio calculations, the [N(CH,),]' salt has been shown to have a pentagonal pyramidal ((3,")structure (1341,and a n X-ray single-crystal structure on the [NO]' salt also indicates a similar pentagonal pyramidal arrangement for the anion with the oxygen apical and in which xenon is coordinatively unsaturated and forms two weak interactions with adjacent anions (135). A new synthesis of Cs+[XeO2FJ, by recrystallization from a CsF/ Cs+[XeOF51-/Xe02F2mixture, and a preparation of a n unstable adfrom the reaction of XeOzF2 with duct, [N02]+[Xe02F3.nXe02F21-, F N 0 2 , have been reported (100). The two oxygen atoms in [XeO,F,Iare cis and not trans to each other. The Raman spectrum previously attributed to Cs+[Xe02F3]- has been shown to be due to Cs' [Xe02F3.nXeF21-. c. Fluoride I Anion Metathesis Reactions between Xenon Fluorides and a n Anhydrous Acid Reactions involving the replacement of fluorine in xenon fluorides with other groups began with studies of the reactions of XeFs with anhydrous oxygen-containing acids ( 3 , 4 , 9 , 1 3 ) :
XeF2 + HL+ F-Xe-L XeF2 + 2HL+ L-Xe-L
+ HF + 2HF.
( 16)
(17)
In all of the earlier cases the linkage to the xenon was via the electronegative oxygen (L = OCOCH,, OCOCF,, O N O p , OPOF2, OS02F, OSeF,, OTeF5, OC103, or OIF40). Only those containing the highly electronegative pseudo-halogens, -OSeF, and -OTeF, , are stable under ordinary conditions. The -0S0,F and -OIF,O derivatives decompose readily, and the rest are explosive. Recently, reaction of xenon bis(trifluor0acetate) with trifluoromethanesulfonic acid (triflic acid) has given the novel and highly reactive unsymmetrical xenon-
76
HOLLOWAY AND HOPE
0x0 species CF3COOXeOS02CFS(136). Most new work in this area, however, has been concerned with species containing the OTeF, or OSeF, groups. The high electronegativity of the OTeF5 group means that Xe(OTeF& is chemically similar to XeF2, and interesting reactions with a variety of fluoro-olefins and fluorocarbon iodides have provided useful pathways to a variety of substituted fluorocations (137-139). More recently, further work in this area has shown that Xe(OTeF5)2reacts with an excess of perfluorobutadiene to yield mostly TeF50CF2CF=CFCF20TeF5with a small trace of TeF50CF2CF (OTeF5)CF=CF2.These are the first examples of -OTeF5-substituted unsaturated fluorocarbons (140). The products of a similar reaction with perfluorocyclohexene were found to be temperature dependent. T ~ obtained F ~ ) ~ as expected, but At 95°C a high yield of C - C ~ F ~ ~ ( Owas at 115°C the main product was the mono -OTeF5-substituted comthat, at the higher temperature, fluopound C - C ~ F ~ ~ O, indicating T~F, rination of one of the C = C carbons occurs (140). Reaction of Xe(OSeF5)2with CF2=CF2,CFBCF=CF2,CF,=CFCl, and CF2=CFH has shown that the OSeF5 group cannot be added to olefinic double bonds as cleanly as its tellurium analog. As well as modest yield addition of two OSeF5 groups t o the double bond, higher yields of c2F5 OSeF5, n-CgF70SeF5,and SeF50CF,COC1 are also formed. In the case of c-C5F8only monosubstitution occurred (141). The ability of Xe(OTeF& to introduce two OTeF5 groups oxidatively has previously been exploited in the synthesis of neutral high-oxidation-state -OTeF5 derivatives from an appropriate neutral low-oxidation-state derivative. For example, Te(OTeF& can be prepared from Te(OTeF& (142), I(OTeF& from I(OTeF& (1431, and ORe(OTeF,), from ORe(OTeF,), (144).This same approach has now been used for the synthesis of -OTeF5-substituted anionic species in the form of the weakly coordinated tetraalkylammonium pnicogen anion series, [M(OTeF,)J (M = As, Sb, or Bi). The arsenic and bismuth anions were prepared by the reaction of the M(OTeF& derivative (M = As or Bi) with a small excess of [N(CH,),]+[OTeFJ in S02C1F, but the instability of Sb(OTeF5)5made this a n inappropriate precursor for [Sb (OTeF,),]-. Consequently, a two-step process in which [Sb(OTeF5)J was first prepared and then oxidized to [Sb(OTeF,)J with Xe(OTeF5)2 was used (145). An intriguing extension of the chemistry of xenon(I1) pentafluoroorthotellurates has been a study of the ability of the OTeF, group to participate in nonredox metathetical fluorinations and to be replaced by a suitably strong protonic acid. Thus, it has been shown that dissolution of [XeOTeF5]+[AsFs]-in BrF5 results in the formation
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
77
of the fluorine-bridged cations [FXeFXeOTeFJ' and [XeF2.BrOF,]' (146).
The reaction of XeF, with bidentate ligands has been regarded as a particularly interesting challenge, and in 1989 it was shown that it reacts with cz~-(H0)~TeF, to form a polymer (Xe02TeF4),1 that is stable to around 80°C. The reaction of cis-(HO)2TeF, with [Xe2F,l+IAsF6] gives a material with the composition [FXeOTeF40Xe]+[AsF6].Recrystallization of the latter in anhydrous H F gave the HF-solvolyzed species HF[HOTeF,OXe]'[AsF,], for which an X-ray single-crystal structure was obtained. This shows that the complex is closest to approaching a purely ionic formulation of any xenon complex so far isolated. The Xe----F-AsF5 contact is 245.8 pm compared with 234.0 pm in [XeFI+[Sb,F,,]- (147). The use of B(OTeF5)to prepare -OTeF, derivatives of many binary fluorides has inevitably impacted on xenon chemistry. In 1978 Xe(OTeF,), was reported (148), and in 1979 Xe(OTeF,), was prepared for the first time by the reaction of XeF, with B(OTeF,), in perfluoron-pentane at -40°C and the XeOF4 derivative XeO(OTeF& was prepared by a similar route, or by hydrolysis of the xenon hexakis{pentafluorooxotellurate(V1)) (149). All three were discussed in more detail two years later (150). More recently a series of mixed fluorol pentafluoro-orthotellurate derivatives of XeOF, and Xe02F2have been prepared. The novel compound 02Xe(OTeF& was prepared according to Eq. (18):
With excess of XeOzFy,a mixture of 02Xe(OTeF&, 02XeF(OTeF5),and XeO,F, is produced (1511. The structures of all of these species and of the related Xe(OSeF& have been of interest since they were first prepared because the ligands are almost completely unable to form bridging in either solid or liquid structures. It has been anticipated, therefore, that xenon derivatives of these ligands will exhibit the expected geometry of the central atom without influence from weaker interactions. The structure of Xe(OTeF& , although based only on X-ray powder data, has been shown to be a true derivative of XeF, (146). More recently, a single-crystal X-ray study of Xe(OTeF5I4has demonstrated that this, too, is as expected, the xenon being surrounded by four equidistant oxygen atoms in a square planar arrangement. The struc, formed according to the reaction represented by ture of 02Xe(OTeF5)2 Eq. (18),shows a very distorted tetrahedral or pseudo-trigonal bipyra-
78
HOLLOWAY AND HOPE
midal arrangement about the xenon, which is not consistent with VSEPR predictions (152). Stable oxygen-bonded xenon(I1) and xenon(1V) derivatives of the highly electronegative O=IF40 group have also been prepared. The generally convenient route t o such species by direct interaction of the xenon fluoride with the acid, in this case HOIOF,, was not used because it is unsafe; instead, the IOzF3dimer was used in insertion reactions of the type shown in Eqs. (19) or (20) in SOzCIF,CFC13,HF, or BrF5 solvents, or an acid displacement reaction between Xe(OTeF& and HOIOF, in CFC1, solvent [Eq. (21)l was used:
+ (IOZF3),+ FXeOIF40 + IOF, + &02 FXeOIF40 + (I02F3),+ Xe(OIF,O), + IOF3 + $0, Xe(OTeF5)2+ BHOIOF, + Xe(OIF,O), + 2HOTeF5 XeF,
(19) (20) (21)
In solution, the products of Eqs. (19) and (20) are mixtures of cis and trans isomers, but the solid obtained from Eq. (21) is the cis, cisXe(OIOF,), compound (153, 154). The reaction of XeF, with I0,F3 in a 1:2 ratio in CFCI, solution produced F3XeOIOF4(153).Efforts to bring about ammonolysis of XeF, and XeF4 have resulted in the formation of xenon, nitrogen, HF, and ammonium fluoride. Reaction of excess of ammonia and XeF6 yields the same products, whereas reaction between ammonia and an excess of XeFs gives an explosive, white solid (96). The first claim for the successful synthesis of a xenon-nitrogen bond was made in 1974 (155).Initial skepticism, based on concerns that the species might actually involve xenon-oxygen rather than xenon-nitrogen bonds, was removed in the early 1980s when the Xe-Nbonded configuration was definitively demonstrated by a single-crystal structure analysis of FXe[N(S02F)21carried out at 218 K (Fig. 5) (156). In the meantime, a second Xe-N-bonded species had been proposed (157),and another pawith the formulation [{(FS02)2NXe}2Fl+[AsFslper, which outlined further details of these two species and a new compound Xe[N(SO,F),I2, was published (158). The species FXe" (SO,F,),I and Xe[N(S02F)212 were prepared by the low-temperature reactions of XeF, with HN(SO,F), whereas the complex salt [{(FSO,), NXe}2F]t[AsF61-was prepared by reaction of FXe[N(SO,F),I with AsF5 ( 158).The compounds were characterized by multinuclear NMR spectroscopy (19F and 129Xe)(1581, and the nature of Xe[N(SO2F),I2was
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
79
FIG. 5 . Perspective drawing of FXeN(S0,F) showing bond lengths. [Reprinted with permission from (156);copyright 1982 the American Chemical Society.]
subsequently confirmed by multinuclear magnetic resonance spectroscopy of the 15N-enrichedcompound (159). More recently, the fluoride-ion-donor properties of FXe[N(S02F)2] have been exploited in the synthesis of adducts with stoichiometries, FXe[N(S02F)21.AsF5,FXe[N(S02F)21.3SbF5, and 2{FXe[N(S02F)2]}. AsF5 (160). All of the compounds were enriched to 30% with 15N and studied by multinuclear magnetic resonance (15N, lYF,IZ9Xe)and Raman spectroscopy, which has shown that the adducts can best be described in terms of predominantly ionic formulations, CXeN(S02F)21+ [AsFJ, [XeN(S02F)21+[Sb3F161-, and [F(XeN(S02F)2}l+[AsFsl-.The was determined from crystal structure of the [XeN(S02F)21+[Sb3Fls1three-dimensional X-ray data a t - 64°C and contains discrete [XeN (S02F)21+[Sb3F161~ species (Fig. 6) in which the cation is Xe-N-bonded. The Xe-N bond length [2.02(1) A] is somewhat shorter than that in FXe[N(S02F)21[2.00(3) and the cation and anion are linked by a weak covalent &-fluorine bridge bond in which the Xe----Fbond length is 2.457(8) A (160). Xenon-nitrogen chemistry has been further extended and a new class of Xe-N bonded compounds obtained by the reaction of Lewis acidic [XeFl+ with the lone pair of a series of nitriles. The first of this new series were obtained by the reaction of the appropriate nitriles RCGN (R = H, CH3,CH2F,C2H5,C2F5,CxF7,C6F5)with either [XeF]'
A],
80
HOLLOWAY AND HOPE
FIG.6. Perspective drawing of the [XeN(SO,F),l' [Sb3F,J structural unit. [Reprinted with permission from (160); copyright 1986 the American Chemical Society.]
[AsFs]- or [Xe2F31t[AsFsl-in anhydrous HF at low temperature (1611. Subsequently, other novel xenon-nitrogen-bonded cations have been observed by the reaction o f R&=N (RF = CF3, CzF5,n-C,F,) in a similar way ( 3 5 )and by the reaction of equimolar quantitites of [XeFI' [AsF,I- with the perfluoropyridines, 4-RC5F4N(R = F or CF,) in anhydrous HF at -30 to -20°C or at -30°C in BrF6 solutions (162). The cations, [C5F5N-XeFl and [4-CF3C5F,N-XeFl+, are stable at - 30°C and can be isolated from BrF, solution as the [AsFsl- salts. Low-temperature Raman and multinuclear magnetic resonance (14N,I9F,129Xe) studies suggest that the cations are planar, with the xenon atom coordinated to the aromatic ring through the nitrogen lone pair (Fig. 7). +
FfiF F
N I
Xe I
F FIG.7
F
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
81
These nitrogen bases are all oxidatively resistant and have first adiabatic ionization potentials close to the electron affinities of [Xe-F] ' [see reference ( 3 5 ) and Section IVI. The estimated electron affinity of KeF]' is 10.9 eV and the first ionization potential of s-trifluorotriazine, s-C3F3N3,is 11.50 eV, which led to the idea that the [s-C3F3N2NXeFl' cation should exist. Reaction of the liquid triazine with [XeF]' [AsF,I- at room temperature, followed by removal of excess of the s-trifluorotriazine under vacuum, left a white powder that is stable indefinitely at room temperature, for which multinuclear magnetic resonance (19F,lZ9Xe)and Raman spectroscopic studies indicate clearly that it is the Xe-N-bonded cation [s-C3F3N2N-XeFl'(Fig. 8) (35). The most recent attempt to extend xenon-nitrogen chemistry has been an investigation of the reactions of XeF2 with HN3, NaN,, and NaOCN in solution in HzO, anhydrous HF, or S02C1F. Although no stable xenon-nitrogen compounds were obtained, on the basis of product distribution, FXe(N3)and FXe(NC0) have been postulated as intermediates and ab initio calculations have shown that both possess stable minima (163). Since noble-gas chemistry was initiated in 1962, there has been interest in the possibility of making a xenon-carbon bond. Experiments in 1979 appeared t o have provided a solution. When xenon difluoride was bled into the tail of a trifluoromethyl radical plasma, a volatile, waxy white solid was produced and trapped a t -196°C. The reported properties of this material are consistent with the formulation, Xe(CF3)2( I 6 4 ) ,but until this is independently confirmed, some doubt remains about its authenticity. In 1989, two groups working independently prepared pentafluorophenylxenon borates by nucleophilic displacement of fluorine in XeFz using B(C,FsI3 as an aryl-transfer reagent. The resulting colorless solid, which has a stable xenon-carbon bond, was characterized in solution by lZVXe and 19FNMR [e.g., Eq. (22)l (165, 166): F
Xe
I
F FIG.8
82
-
XeFz+ B(C,jF5)3
MeCN
HOLLOWAY AND HOPE
[C,F5Xe]+[B(C6F5)3F]+ BF3.MeCN + BF;.
(22)
Reactions with Te(C,F,), and C6F51yielded [(C,F,),TeI and [(C6F5),II+ respectively (165), and the hydrolysis and thermal stability of [C,F5Xelt were also reported (166). Very soon afterward more spectroscopic information was provided, and the Xe-C bond length was obtained in a n X-ray structural determination on [MeCN----XeC,F51t[(C,F5)zBF,I-(Fig. 9)(167). Since then, several arylxenon derivatives have been prepared via exchange reactions of XeF, and triarylboranes. The salts vary in thermal stability; the 2,6-difluoroaryl species are more stable than the monofluoroaryl derivatves, and the parachlorophenyl derivative decomposes at about 0°C (168).Reactions of arylxenon species, and especially [CsF5Xelt,have become a n area of intense activity. With PhX, [C6F5Xel+[AsF61-reacts to give isomeric mixtures of biphenyls, XCsH4CsF5(X = Me, F, CF3, NOz, CN) (169). Reaction of the same salt with cesium pentafluorobenzoate in water [Eq. (2311 gives an unexpectedly stable covalent acyloxyxenon(I1) derivative, the X-ray crystal structure of which indicates a linear F5C6-Xe-0 arrangement with dimerization through oxygen bridging (170): +
[CsF5Xelf[AsFJ + Cs[0,CC,F51 + C6F5Xe02CC6F5.
(23)
The first alkenylxenon(I1) derivative has been obtained by the reaction of [C6F5Xelt[AsFsl-with xenon difluoride in anhydrous HF, the reaction proceeding in two steps giving first the (heptafluoro-1,4-cyclo-
FIG.9. The structure of the [MeCN-Xe-C,FJ+ cation showing the Xe-ligand bond distances. [Reprinted with permission from (167); copyright 1989 VCH Verlagsgesellschaft mbH.1
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
83
hexadien- 1-yl)xenon(II)derivative and then the (nonafluorocyclohexene-1-yl)xenon(II) salt (171):
Reactions of [C6F5Xe]+[AsF6](i) with halide ions in acetonitrile to give aromatic and coupled aromatic molecules (172) and (ii) with pyridines of different basicities, where coordination weakens the Xe-C bond and promotes C6F5radical formation, have been described (173). As part of this work the adduct [Xe(C,F,)(NC,H,F,-2,6)l+[AsFslhas been prepared and characterized crystallographically (I731. Xenon difluoride in aqueous HF acts as a strong electrophilic oxygenation system, and reaction of both [C,F,Xe]+[AsF6]- and [C6F,Xel'[AsF61with this system has yielded new examples of the transformation of an organic moiety bonded to xenon without cleavage of the Xe-C bond ( 174 1:
F
/"
Reactions of [C6F9Xe]'[AsF61-with halide ions have been shown to be solvent dependent. With iodide and bromide, in both acetonitrile and anhydrous HF, the xenon is displaced by the halide to give CsFJ(X = Br or I). Chloride and fluoride do not react in anhydrous HF, but in acetonitrile the fluoride ions initiate the formation of alkenyl radicals, which abstract hydrogen from the solvent to give C6FSH,whereas chloride shows intermediate behavior, giving both CGF,Hand C6F,C1 (175). The structure of [(CsH3F~-2,6)Xe]+[BF41in the solid state is an
84
HOLLOWAY AND HOPE
ion pair with an Xe----Fdistance of 279.3 pm. This is about 70 pm shorter than the van der Waals distance (176). Anion exchange of this salt with trimethylsilyltriflate gives the related triflate salt as expected (177). Perhaps the most exciting chemistry to emerge so far from investigation of arylxenon derivatives is a preliminary report of the reaction of [C6F,Xe]+with chloride ions in acetonitrile a t below -30°C, which has led to the synthesis and structural characterization of [(CsFSXe)2C1]'[AsF6]-and the NMR spectroscopic identification of the more thermally labile compound C ~ F S X ~both C ~ , containing Xe-Cl bonds (178 1. Other interesting chemistry includes electrophilic substitution of fluorobenzenes or trifluoromethylbenzenes using the unsymmetrical xenon-oxo species mentioned earlier (136) to give the arylxenon trifluoromethylsulfonates as colorless solids (179 ). Subsequently, CF3COOXeOS02CF3has been used in a variety of reactions with a range of benzene derivatives containing electron-withdrawing groups such a s -F, -CF3, -C1, and -NOz t o yield a number of arylxenon trifluoromethylsulfonates. These have been characterized mainly by NMR, mass, and vibrational spectroscopy; however, the structure of [Xe(2,6-FzC6H3)1 +[OS02CF31-has been determined by single-crystal X-ray diffraction methods, which have shown that the arylxenon is weakly coordinated by one oxygen atom of the anion (180). In a related area, interest in the isolation of alkynyl iodonium salts has led to the synthesis of analogous isoelectronic xenon(I1) derivatives. These have been obtained by reaction of XeFz with ButC=CLi or RC=CSiMe3 in BF3.OEt2 (181): R-CrC-SiMe3
+ BF3.OEt2+ XeFz+
[R-C~C-XeI+[BF,I+ Me3SiF. (27)
As well as chemical reactions, some physical measurements have been made. I n particular, cyclic voltametric measurements have yielded reduction potentials for a series of arylxenon(I1) tetrafluoroborate salts that indicate that these are stronger oxidizers than the comparable aryliodonium salts (182). 3. Reactions i n Which the Xenon Fluoride Acts as a Fluorinating Agent
As outlined in Section IV, p. 55, the xenon fluorides and their cationic derivatives are capable of oxidation andlor fluorination of many
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
85
other species. The oxidizinglfluorinating strength is in the sequence, XeF2 = covalent Xe(I1) derivatives < XeF, < XeF6, and their cations are even stronger than the neutral species. It is also known that, in the case of XeF2, the addition of anhydrous H F enhances its fluorinating ability, and the mechanism has been discussed. Before 1977 there was already a vast literature on the fluorination of organic fluorides, especially using XeF,, and this topic will not be covered here. However, a comprehensive and authoritative review up to 1977 [see R. Filler in reference (511 and articles within a n excellent recent book (183)cover this area well, and the reader is referred to these. Xenon difluoride and the other xenon fluorides were established as good oxidative fluorinating agents in the early days of noble-gas chemistry (2-4, 13). Soluble in water, the difluoride was found to be stable in neutral or acid solution and sufficiently strongly oxidizing to convert chloride to chlorine, chromium(II1) t o chromium(IV), and bromate, [BrOJ, to perbromate, [BrOJ (131. In direct combination with other species or in reactions in nonaqueous solution, it reacts with main-group and transition-metal compounds and is incorporated into intercalation compounds. Because of the difficulty and danger associated with the manipulation of XeF, and XeF6, the oxidative fluorination chemistry of these is less closely studied. Work carried out up to and including 1980 has been usefully summarized (13),and only chemistry carried out since then is covered here.
a. Reactions with Main-Group Species There has been extensive application of XeFLas a reagent in main-group chemistry recently. It has been used to cleave the B-0 bonds in B(OR), to give FXeOR (R = SO,CF, or S02C,F,) and Xe(OR), (R = COCF,%or COC,F,) (184). Aryl trimethylsilanes are readily fluorodesilylated with XeF2 in hexafluorobenzene a t room temperature to give aryl fluorides, the reaction occurring via a radical mechanism (185), and vinyl stannanes are fluorinated to fluoroalkenes with XeFz and silver hexafluorophosphate or trifluoromethanesulfonate, the latter giving the faster conversion (186). The silicon and germanium compounds C,F,MX, (M = Si, Ge; X = C1, Br) undergo halogen exchange with XeF2 without C-M cleavage. However, if X is fluorine or alkyl, XeFz in the presence of BFa.OEt, or with [XeFl+[NbF,l- yields the heptafluorocyclohexadienyl derivative C6F7MX3by the addition of two fluorine atoms to the pentafluorophenyl ring (187). The surface chemistry of silicon has also been studied by exposing it to XeF2 vapor and examining the chemisorption of silicon at room temperature and - 150°C using XPS and AES (188).
86
HOLLOWAY AND HOPE
Studies on reactions of main group alkyl and aryl derivatives have been going on for some years. Recent studies have shown that reaction of XeFz with MePhzP and MePhzAsin CH3CNproduces MePhzPFz and MePhzAsFzrespectively, and with 2-MeCsH4P(NEt&, 2-MeC6H4 PF,(NEt2)zis obtained (189). Related reactions of XeF, with Me3X (X = N, P, As, or Sb) all proceed smoothly, the rate reflecting the basicity of the substrate, Me3X.Where X is P, As, or Sb, the difluoride Me3XF2is formed. Where X is N, cleavage of a C-H bond is the main reaction (190). Xenon difluoride also oxidizes phosphorus(II1) acid derivatives to give either the phosphorus(V) monomer or the ionic salt respectively (1911,
and M(C,F,), (M = As, Sb) (192) and (2,6-FzC6H3),Bi(193) to the respecdifluorides M(C,F,),F, (M = As, Sb) and (2,6-FzC6H3)3BiFz tively. Although protonation of HOF by reaction of AsF5 or SbF, with water and HF was not successful, the fluorination of water with [XeFI+[MFJ (M = As, Sb) produced pale reddish [H20Fl+[MF61salts almost quantitatively. These can be stored for about 2 h at room temperature without decomposition, but under vacuum release HOF, HF, and MF5(194). A wide variety of reactions of XeF2 and [XeF]' with low-valent sulfur species have been studied. Thus, an electrophilic fluorination reagent capable of fluorinating carbanions in moderate yield has been observed in the reaction of XeF, with sulfur(I1) derivatives such as MezS (195).Reactions of Me$ (and MezO)with XeF, in trichlorofluoromethane result in cleavage of C-H bonds to form CHzF derivatives as the main reaction pathway. In the absence of a C-H bond, for example if neat (CF3)ZS is used, oxidation to (CF3)2SF2occurs (190). The reaction of MezS with XeF, in nonacidic medium gives rise t o a product that is best formulated as [Me,SCHzSMel+[F(HF>,l-.In HF solution at -23"C, however, the product is [Me2SFlt (196).A similar reaction has also been observed in the combination of (C&?s)aSwith [XeFl+[MFJ (M = As, Sb) in HF, which yields deep violet crystals of An X-ray crystal structure determination of the l(C,F,),SFl'[MFJ. antimony derivative has shown that the cations and anions are connected via interionic fluorine bridges to produce an infinite chain. An-
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
87
other feature is a short S-F bond distance [158.4(3) pm] (197). The conversion of CH3S-- to CH,FS- in methionine and methionylglycine derivatives has also been achieved using XeFa in solution in MeCN a t or below room temperature (198). Oxidation of MeSCN or S(CN), t o [Me(NC>SFl+and [(NC),SFl+ respectively using [XeFI+[MFJ (M = As, Sb) has also been reported (199). The oxidative fluorination of (CF,),C=NSNCO by XeF, gives rise to two isomeric sulfur(IV1 compounds, (CF3),CF -N =S(F)NCO and ( CF3I2CFN=S =NC(0)F. The third possible isomeric product of 1,3 addition, (CF,),C=N-S(F)= NC(O)F, was not observed (200). The oxidative fluorination of sulfur(IV) compounds such as diphenyl sulfoxide has been shown to occur under mild conditions with XeF, and catalytic amounts of chloride ion. In the Ph$O case the product is Ph,S(O)F, , which is formed in essentially quantitative yield within a few minutes. The chloride ion reacts with XeF, to produce fluoride ion, and a mechanism is proposed that involves fluorosulfur(1V) anions and fluorosulfur(V) radicals (2011. Reactions of MezSe with XeF2 in trichlorofluoromethane proceed smoothly to give MezSeFz(190). Oxidative coupling of R2Se2(R = Ph, biphenyl, alkyl) to unsymmetrical disubstituted acetylenes and cycloalkynes using xenon difluoride has been shown to yield the vicinal (E)-fluoroalkenyl selenides in high yield. Alternatively, the PhSeF equivalent can be formed from PhSeSiMe:i and XeF, (202). Although PhSeCl and PhSeBr are commercially available sources of selenic electrophiles, PhSeF, which is too labile to be isolated, can be generated in situ from PhSeSePh with XeF, (203). This has been used to generate the novel phosphaalkenes from phosphaacetylenes. Reaction of XeFz with Ph,Te gives PhzTeF2(189).In related studies PhTeF3 or PhTeTePh gave PhTeF5, PhzTeFz or PhsTeFBgave PhzTeF,, Ph3TeF or Ph3TeCl gave Ph3TeF3 (with small amounts of Ph3TeF,Cl in the case of Ph3TeC1),and Ph4Te gave Ph,TeF2 (204). The latter has subsequently been characterized (205). In a study on the synthesis of tetrakis(perfluoroalky1) tellurium species, Te(Rf)4,the compound Te(CF3), was obtained from the reaction of Te(CF3),ClZwith Cd(CF,), . Reaction of this compound with fluoride ion acceptors gives rise to the complex cation [Te(CF:3)31i,and with certain fluorides the complex anion [Te(CF3l4F1-is formed. However, with XeF, it appears to be oxidized to (CF3I4TeF2(206).The related compound R,TeF, (R = CH3)was prepared similarly by the reaction of TeR4 with XeF2 (207). As part of a study of the Lewis acid behavior of [Te(OTeF,),l toward [OTeF,l-, higher concentrations of the [FTe(OTeF,),]- anion in solution in
88
HOLLOWAY AND HOPE
SOzCIF were generated by the interaction of XeF, with [N(CHs),l+ [Te(OTeF,),I- at -50°C according to Eq. (30) (51): 2[Te(OTeF5)J
+ XeFz
S0,ClF -50°C
2[FTe(OTeF5),I-+ Xe(OTeF&. (30)
Liquid bromine reacts with [XeF]+[MF6]-(M = As, Sb) at room temperature in a complex reaction sequence that gives BrS+[MF6]-,
and its crystal structure has been determined (208). The reaction between HI and [XeFIt[AsF6]-has been shown to give as the final products of reaction. No experiXe, HF, and [I,12'[AsFfi1~ mental evidence for [XeIl' was found in the reaction products, and theoretical calculations show that such a species is bound only in the 'C singlet state and, after interconversion into the unbound triplet state, would immediately dissociate into xenon ('S) and I'(3P) (209). In synthetic and structural studies of a number of anions the [IF,]-, [BrFJ, and [IF4]- salts were synthesized according to Eqs. (32) to (34) (210):
CsBr + XeF,
Cs+[BrF,]- + Xe
The chemistry of the interaction of XeFz with alkyl and aryl iodides has been further extended by investigations of the reactions with CF3CH21,3,5-ClzC6H31,and 2-CF3CfiH41. In each case the alkyl(ary1) iodine difluoride was obtained (189). In related reactions between MeX (X = C1, Br, or I) and XeFz, the reactions proceed smoothly, the rate of reaction reflecting the basicity of the substrate, and in the case of MeI, MeIFz is formed. With MeCl and MeBr, however, MeF is the product (190). Although difluoroiodo compounds have been used as
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
89
synthetic intermediates, none had been isolated until treatment of 4-iodotricyclene with excess of XeFz in CC1, produced the difluoro derivative [Eq. (35)]as a pale yellow, waxy solid. This is stable in air for a few hours and indefinitely in solution under a n inert atmosphere (2111:
I’
IF,
b. Reactions with Transition-Metal, Lanthanide, a n d Actinide Species The reaction of chromyl fluoride with XeFz a t temperatures up to 278°C is a high-yield route to CrOF3 that is superior to other routes (52). Reaction with CoFz, CoClZ, or metallic cobalt have all been shown to yield CoF,, and DTA data have provided information on the temperatures of initiation of the reactions established ( 4 6 ) .The mild fluorinating ability of XeFz in solution has been exploited in the synthesis of Ir(CO)C1z(PEt3)z(P’F4), whose I9F and NMR spectra clearly showed the presence of the Ir-P’F, group (212):
It was also found to fluorinate the five-coordinate carbonyl complex of Ir(I), [Ir(C0)3(PEt3)z]+, to yield the first fluoroacyl derivative, Ir(CO)2 F(COF)(PEt,), (213). Xenon difluoride also reacts with [Pt(NH3),1C12 in the solid state to give t r a n ~ - [ P t ( N H ~ ) ~ F ~but l C lin ~ , solution in MeCN this is converted to t r a n ~ - [ P t ( N H ~ ) ~ C(214 l ~ ] ).F ~ The analogous oxidation of [Pt(py)4]2+[BF41f gives the crystallographically characterized [PtFz(py)412+[BF41; (215). It was established in the 1970s that XeFz in solution was a valuable means of introducing fluorine oxidatively into transition carbonyl compounds to produce carbonyl fluorides (216, 21 7). Recently, there has been renewed activity in this area, supported by the improvements in multinuclear Fourier-Transform NMR techniques. Among
90
HOLLOWAY AND HOPE
the more recent papers is a report of a reaction of XeF, with [Re2(CO),,] in perfluoro-l,3-dimethylcyclohexanethat concludes that the product is Re(C0)3F2(218)even though the properties of the compound produced are clearly those of Re(CO)SF.ReF5 (217). Considerable progress has been made with oxidative fluorination of ruthenium and osmium carbonyls and their derivatives. Thus, [Ru3(CO),,I with XeF, in anhydrous H F gave, in solution, cis-[RuFz(CO),l as the main product along with small amounts of [RuF(CO)J+, [ R U ~ F ~ ( C O ) ~ I , [{RuF(CO),}~(~-F)I',mer- and fuc-[RuF3(CO)J, [{RuF~(CO)~},I, and [{RuF,(CO)3}(p-F){Ru(C0)5}l +,all of which have been characterized by 13C,I9F and 13C-{19F}NMR spectroscopies. Removal of the HF solvent yielded tetrameric [{RuF,(CO)~}~] (219). Similarly, the equivalent reaction with [ O S ~ ( C O )produced, ~~] in solution, a range of mono- and dinuclear osmium carbonyl fluoride species, with c i s - [ 0 ~ ( C O ) ~ Fas~ l the major product and [Os(CO),F]+,[ O S ~ ( C O ) ~ F and ~ I ,[ O S ~ ( C O ) ~ Fas ~]+ minor products. The main product is stable in solution, but removal of the solvent results in CO loss to give the fluoride-bridged tetrameric analog to the ruthenium tricarbonyl difluoride, [{OS(CO)~F,}],(220). The mechanism of the oxidation of low-valent ruthenium and osmium complexes has also been elucidated by a study of the reaction of XeF, with [M(CO),(PPh3),] (M = Ru, 0s) in dichloromethane (221). Fluoroacyl complexes of ruthenium and osmium, [M(COF)(C0)2 F(PPh3)21(M = Ru, 0s) were identified as intermediates in the oxidative addition of XeF, in solution, and the mechanism, probed by addition of BF3, has been shown to involve oxidation of the metal center by [XeFI' to give a monofluorinated cation. Subsequent nucleophilic attack by F- at the coordinated CO with CO elimination gives, finally, [OC-6-131[M(CO),F2(PPh3)21. In the ruthenium case, the ligand arrangement has been confirmed by a single-crystal X-ray crystallographic study (2211. Controlled sequential reaction of [Ir4(C0>121 with XeF2 in anhydrous H F a t low temperature yielded, initially, fac- and mer-[Ir(CO)3FS], and, ultimately, IrF, . Removal of the solvent at the initial stage of the reaction results in the precipitation of f~c-[1r(CO)~F~l, which contains predominantly u-bonded carbonyl ligands with high CO-stretching frequencies (222). Xenon difluoride has been shown to react with organomercury compounds, RzHg (R = PhC=C, p-MeOC6H4,p-Me2NC6H4,p-Et02CC6H4, and PhCH2), with cleavage of the C-Hg bond t o give RHgF or RHgFI HgF, mixtures. When R = benzyl, the RHgF formed undergoes fast demercurization. From the composition of the reaction mixtures it has been established that the reactions involve free radicals. The absence
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
91
of fluorinated organomercury products is evidence that XeF, reacts more readily with the C-Hg bond than with C-H or C-C bonds. With HgX, (X = C1, Br, I) under mild conditions, HgF2, halogen gas, and xenon are produced (223). The xenon fluorides, XeF2, XeF,, and XeF6, have all been used in preparations of some lanthanum tetrafluorides, LnF4 (Ln = Ce, Pr, or Tb) and their heptafluoro complexes, Cs3LnF7(Ln = Ce, Pr, Tb, Dy, or Tm) (224). Detailed examination of the reaction of CssLnCl6 with XeF, a t 100-400°C has shown that CsyLnF7(Ln = Ce, Pr, Tb, Nd, or Dy) species are formed, but for E r and Yb under the same conditions only the lanthanum(II1) species CssLnFs (Ln = E r or Yb) are produced (224). In the actinide field, stopped-flow spectroscopy has been used to study the oxidation of plutonium(II1) to plutonium(1V) in 1M HClO, by XeFz. Two moles of Pu(II1)are oxidized per mole of XeF, consumed, and there is no evidence for oxidation to PuW) or Pu(V1). Rate constants and activation parameters have been obtained and a mechanism postulated that invokes a sequence of 2 one-electron oxidation steps (225). The oxidation of americium by a range of xenoncontaining oxidants in KHCO,, KHCO, and K2C0,, and in K&O, solutions, has been studied (226,227).Xenon difluoride and xenon trioxide convert AmlIII) to Am(IV) in solution (2261, whereas NaJXeOJ oxidizes Am(II1) to Am(IV), A m W , and Am(V1) mixtures unIess the perxenate is in excess, when Am(V1) is produced (226,227). The trioxide, Xe03, reacts with Am(II1) only under UV irradiation (227).
VI. Radon Chemistry
The chemistry of radon (11) was outlined by Stein in 1983 (14). Since then, evidence for radon in a higher oxidation state (RnF, or RnF6) and a radon oxide (RnO,) has been claimed and disputed, and the ions, [HRnOJ', [HRnOJ, and [RnOSFl have been reported. Soon after the initial discovery of xenon compounds in 1962, a lowvolatility fluoride was detected in the product of the combination of radon with fluorine a t 400°C (226). It was also shown that 'L22Rnis oxidized by chlorine and bromine fluorides, IF7 (227-230) or [NiFsl" (227-229) in H F to give stable solutions of radon fluoride, and that radon can be collected as a solid by oxidation of the gas with liquid BrF3 or the solid complexes [C1F2]+[SbFsl-,[BrFJ'[SbFJ, [BrF,]' [Sb2F113-,[IF41:'+3[SbF61-, and [BrF2I'[BiFJ (231-232). Electromigration studies suggested that in the solution studies the radon is pres-
92
HOLLOWAY AND HOPE
ent as [RnF]' or [RnI2+(228, 229). Such ionic formulations were also favored in relativistic quantum mechanical calculations (233). This has been borne out further by more recent chemistry in which solutions of cationic radon have been prepared by oxidation of elemental radon with halogen fluorides and shown t o displace hydrogen, sodium, and potassium in solid materials such as Nafion resins (H+ and K' forms), potassium hexafluorophosphate, and sodium hexafluoroantimonate. It has also been shown that radon in this state can be quantitatively recovered by ion exchange and eluted with bromine trifluoride in sulfuryl chloride (234). By comparison of the properties of the products from all of these reactions with the known properties of the fluoride and fluoride complexes of krypton and xenon, it is almost certain that radon forms a difluoride, RnFz, and derivatives such as [RnFI+[SbFJ, [RnF]+[TaF6]-,and [RnF]+[BiF6]-,which are direct analogs of krypton and xenon species (235). This has been further strengthened by the observation of the reaction of radon with solid [02]+[SbF61-a t room temperature, which yields gaseous oxygen and an involatile radon compound (236). Comparison with xenon chemistry suggests the reaction
This was further substantiated by reactions of radon with fluoronitrogen salts and halogen-fluoride metal salts, and the development of a method of collecting radon from air using [o21+[SbF6l-or [IFJ+[SbFJ (237). Hydrolysis of radon fluoride with water yields radon, oxygen, and HF. This parallels the reactions of KrFz and XeFz and is in marked contrast to the reactions of XeF4 and XeF6, in which some xenon is retained in solution as Xe(V1). It also coprecipitates with XeFz from halogen fluoride solutions and complexes with XeFzbut not XeF4(238, 239) all of which points to the fluoride being RnFz. Claims by Russian workers that a higher fluoride of radon, RnF4 or RnF6, can be prepared in tracer experiments by heating radon, xenon, fluorine, bromine pentafluoride, and either sodium fluoride or nickel fluoride, and converted to RnOs by hydrolysis (240) appeared to others (235) to be due to the precipitation of radon as a solid complex, which is probably [RnF]2+[NiF612-. However, the precipitation of CsXeOBF from aqueous solutions results in the coprecipitation of radon, and this has been taken by the Russian group as confirmation that RnOB is the product of hydrolysis of the fluoride formed (241). Furthermore,
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
93
it has been suggested that the failure of Stein to observe the highoxidation-state [RnO3F1-species was probably due t o high F- concentration (242).More recently, ultracentrifugation of hydrolyzed radon solutions, coprecipitation studies, and kinetic data for the decomposition of the solution species have been interpreted in terms of it being RnO, (243). Recent electromigration studies have also been used to suggest that in acidic aqueous solutions (pH > 5) cationic, [HRnOJ’, and anionic, [HRnOJ, forms of radon are produced, the validity of the electromigration method having been established using xenon(V1) (244).
REFERENCES 1. Bartlett, N. Proc. Chem. Soc. 1962,218. 2. Holloway, J. H. “Noble-Gas Chemistry.” Methuen, London, 1968. 3. Horn, H.-G.; Hein, H. In “Gmelins Handbuck der Anorganischen Chemie”; Buschbeck, K.-C., Lippert, W., and Slawisch, A,, Eds.; Ergiinzungswerk zur 8. Auflage Band 1, Edelgasverbindungen, Verlag Chemie, Weinheim, 1970. 4. Bartlett, N.; Sladky, F. 0. In “Comprehensive Inorganic Chemistry”; TrotmanDickenson, A. F., Ed.; Vol. 1, pp. 213-330. Pergamon Press, Oxford, 1973. 5. ( a ) Hyman, H. H. Science 1964, 145, 773; tb) Holloway, J. H. In “Progress in Inorganic Chemistry”; Cotton, F. A,, Ed.; Vol. 6. pp. 241-269, Interscience Publishers, New York, 1964; (c) Malm, J. G.; Selig, H.; Jortner, J.; Rice, S. A. Chem. Rev. 1965,65,199; (d) Moody, G. J . ; Thomas, J . D. R. Rev. Pure Appl. Chern. 1966, 16, 1; (e) Sladky, F. 0. In “MTP International Review of Science”; Gutman, V., Ed.; Inorg. Chem., Ser. One, Butterworths, London, 1972; (f) Jha, N. K. R.I.C. Reu. 1971,4,147; (g) Stein, L. Yale Scientific Magazine 1970,44,(8), 2; (h) Malm, J . G.; Appelman, E. H. Atomic Energy Rev. 1969, 7 (3), 3; (i) Filler, R. Isr. J . Chem. 1978,17,1. 6. (a) Hyman, H. H. (ed.)“Noble Gas Compounds,” University of Chicago Press, Chicago, 1963; (b) Moody, G. J.; Thomas, J . D. R. “Noble Gases and their Compounds,” Pergamon Press, Oxford, 1964; (c) Claassen, H. H. “The Noble Gases,” D.C. Heath and Co., Boston, 1966. 7. Hawkins, D. T.; Falconer, W. E.; Bartlett, N. “Noble Gas Compounds, a Bibliography, 1962-1977,” Plenum Press, New York, 1978. 8. Laslo, P.; Schrobilgen, G. J. Angew. Chem., Int. Ed. Engl. 1988,27,479. 9. ( a ) Holloway, J. H. Chem.. Br. 1987,23,658; (b) Holloway, J. H. In “Fluorine, the First 100 Years”; Banks, R. E., Sharp, D. W. A,, and Tatlow, J . C., Eds.; Elsevier Sequoia, Lausanne, 1986. 10. Zemva, B. In “Encyclopaedia of Inorganic Chemistry”; King, R. B., Ed.; Vol. 5, pp. 2660-2680, John Wiley & Sons, Chichester, 1994. 11. Seppelt, K. Accounts Chem. Res. 1979,12, 211. 12. Selig, H.; Holloway, J. H. In “Topics in Current Chemistry”; Boschke, F. L., Ed.; Vol. 124, pp. 33-90, Springer-Verlag, Berlin, 1984. 13. Seppelt, K.; Lentz. D. Prog. Inorg. Chem. 1982,29,167. 14. Stein, L. Radiochim. Acta 1983,32,163.
94
HOLLOWAY AND HOPE
15. Breckenridge, W. H.; Jouret, C.; Soep, B. In “Advances in Metal and Semiconductor Clusters”; Duncan, M. A., Ed.; Vol. 111, JAI Press, Greenwich, 1995. 16. Liebman, J. F.; Allen, L. C. J . Am. Chem. Soc. 1970, 92, 3539. 17. Berkowitz, J.; Chupka, W. A. Chem. Phys. Lett. 1970, 7, 447. 18. Frenking, G.; Koch, W.; Deakyne, C. A.; Liebman, J. F.; Bartlett, N. J . Am. Chem. Sac. 1989, 111, 31. 19. Wong, M. W.; Radom, L. J. Chem. SOC.,Chem. Commun. 1989,719. 20. Allen, L. C.; Huheey, J . E. J. Inorg. Nucl. Chem. 1980,42, 1523. 21. Filippi, A.; Troiani, A.; Speranza, M. Chem. Phys. Lett. 1997,278, 202. 22. Slivnik, J.; Smalc, A.; Lutar, K.; Zemva, B.; Frlec, B. J. FZuorzne Chem. 1975, 5, 273. 23. Al-Mukhtar, M.; Holloway, J. H.; Hope, E. G. J. Fiuorine Chem. 1992, 59, 1. 24. Bezmel’nitsyn, V. N.; Legasov, V. A,; Chaivanov, B. B. Dokl. Chem. (Engl. Transl.) 1977,235, 365. 25. Dang, H.; Huag, L.; Xu, R.; Yang, R.; Zhang, B.; Wang, D.; Lu, Z. Wuji Huarue Xuebao 1996,12, 18 (Chem. Abstr. 1996,125, 24976). 26. Burbank, R. D.; Falconer, W. E.; Sunder, W. A. Science 1972, 178, 1285. 27. (a) Siegel, S.; Gebert, E. J . Am. Chem. SOC.1963,83, 240; (b) Levy, H. A.; Agron, P. A. J. Am. Chem. SOC.1963, 85, 241; (c) Brown, G. M.; Levy, H. A. J. Physi. Radium 1964,25, 469. 28. Gillespie, R. J.; Martin, S.; Schrobilgen, G. J. J. Chem. Soc., Dalton Trans. 1980, 1898. 29. Christe, K. 0.;Wilson, W. W.; Bougon, R. A. Inorg. Chem. 1986, 25, 2163. 30. Christe, K. 0.;Wilson, W. W.; Wilson, R. D. Znorg. Chem. 1984,23, 2058. 31. Gillespie, R. J.; Schrobilgen, G. J. Inorg. Chem. 1976, 15, 22. 32. Lutar, K.; Jesih, A,; Zemva, B. Polyhedron. 1988, 7, 1217. 33. Frlec, B.; Holloway, J . H. J. Chem. Soc., Chem. Commun. 1973, 370. 34. Sanders, J. C. P.; Schrobilgen, G. J . J. Chem. SOC.,Chem. Conzmun. 1989, 1576. 35. Schrobilgen, G. J. J. Chem. Soc., Chem. Commun. 1988,863. 36. Schrobilgen, G. J. J. Chem. Soc., Chem. Commun. 1988, 1506. 37. Hillier, I. H.; Vincent, M. A. J. Chem. Soc., Chem. Commun. 1989,30. 38. Koch, W. J. Chem. Soc., Chem. Commun. 1989, 215. 39. MacDougall, P.; Schrobilgen, G. J.; Bader, R. W. F. Inorg. Chem. 1989,28, 763. 40. Gillespie, R. J.; Schrobilgen, G. J. J. Chem. Sac., Chem. Commun. 1974, 90. 41. Gillespie, R. J.; Schrobilgen, G. J. Inorg. Chem. 1974, 13, 1230. 42. Holloway, J. H.; Schrobilgen, G. J . J . Chem. Soc., Chem. Commun. 1975,623. 43. Liebman, J. J. Fluorine Chem. 1977,9, 147. 44. Bartlett, N.; Seppelt, K, unpublished results [see reference (13)1. 45. Christe, K. 0.; Wilson, R. D. J. Fluorine Chem. 1976, 7, 356. 46. Nikulin, V. V.; Popov, A. I.; Zaitseva, I. G.; Korobov, M. V.; Kiselev, Yu. M.; Siderov, L. N. Russ. J. Inorg. Chem. (Engl. Transl.) 1984,29, 21. 47. Spitzin, V. I.; Martynenko, L. I.; Kisel, W.; Ju. M. 2. Anorg. Allg. Chem. 1982, 495, 39. 48. Kiselev, Yu. M.; Goryachenkov, S. A,; Martynenko, L. I. Russ. J. Inorg. Chem. (Engl. Transl.) 1984,29, 38. 49. Kiselev, Yu. M.; Sokolov, V. B. Russ. J. Inorg. Chem. (Engl. Transl.)1984,29, 493. 50. Jesih, A.; Lutar, K.; Leban, I.; Zemva, B. Inorg. Chem. 1989, 2911. 51. Lutar, K.; Jesih, A.; Leban, I.; Zemva, B.; Bartlett, N. Inorg. Chern. 1989,28, 3467. 52. McHughes, M.; Willet, R. D.; Davis, H. B.; Gard, G. L. Inorg. Chem. 1986,25, 426.
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
95
53. Zemva, B.; Lutar, K.; Jesih, A,; Casteel Jr., W. J.; Wilkinson, A. P.; Cox, D. E.; Van Dreele, R. B.; Borrmann, H.; Bartlett, N. J. Am. Chem. SOC. 1991, 113, 4192. 54. (a) Drobyshevskii, Yu. V.; Serik, V. G.; Sokolov, V. B. Dokl. Chem. (Engl. Transl.) 1975,225, 1079; (b) Drobyshevskii, Yu. V.; Serik, V. G.; Sokolov, V. B.; Tul’skii, M. N. Sou. Radiochem. (Engl. Transl.) 1978,20, 200. 55. Drobyshevskii, Yu. V., Prusakov, V. N.; Serik, V. F.; Sokolov, V. B. Sou. Radiochem. (Engl. Transl.) 1980,22, 591. 56. Bacher, W.; Jacob, E. Chem. -Ztg 1982, 106, 177. 57. Peacock, R. D.; Edelstein, N. J . Znorg. Nucl. Chem. 1976,38, 771. 58. Brown, D.; Berry, J. A.; Holloway, J . H. J . Actinides 1981, 11, 81. 59. Asprey, L. B.; Eller, P. G.; Kinkead, S . A. Inorg. Chem. 1986,25, 670. 60. LeBlond, N.; Schrobilgen, G. J . J . Chem. Soc., Chem. Commun. 1996,2479. 61. Christe, K. 0.;Bougon, R. J . Chem. Soc., Chem. Commun. 1992, 1056. 62. Bougon, R., J. Fluorine Chem. 1991,53, 419. 63. Artyukov, A. A.; Khoroshev, S. S. Koord. Khim. 1977,3, 1478. 64. ( a ) Sakurai, T.; Takahashi, A,; Fujisawa, G. J. Nucl. Sci. Technol. 1982, 19, 255; (b) Sakurai, T.; Takahashi, A.; Fujisawa, G. J. Fluorine Chem. 1982,20, 683. 65. Boltalina, 0. V.; Abdul-Sada, A. K.; Taylor, R. J. Chem. SOC., Perkin Trans. 2. 1995,981. 66. Zemva, B.; Hagiwara, R.; Casteel, J r . , W. J.; Lutar K.; Jesih, A,; Bartlett, N. J . Am. Chem. SOC. 1990, 112, 4846. 67. Nielsen, J . B.; Kinkead, S . A.; Purson, 3 . D.; Eller, P. G. Inorg. Chem. 1990, 29, 1779. 68. Smalc, A,; Lutar, K. In “Inorganic Synthesis”; Grimes, R. N., Ed.; Val. 29, pp. 1-4, Wiley-Interscience, New York, 1992. 69. Lutar, K.; Smalc, A.; Slivnik, J . Vestn. Slov. Kem. Drus. 1979,26, 435. 70. Lutar, K.; Smalc, A,; Zemva, B. In “Inorganic Synthesis”; Grimes, R. N., Ed.; Val. 29, pp. 4-6, Wiley-Interscience, New York, 1992. 71. Brassington, N. J.; Edwards, H. G . M. J . Mol. Struct. 1987, 162, 69. 72. Nabiev, Sh. Sh.; Ostroukhova, I. I. Spectrochim,. Acta. 1993, 49A, 1527. 73. Nabiev, Sh. Sh.; Ostroukhova, I. I. Spectrochim. Acta. 1993,49A, 1537. 74. Gillespie, R. J. J . Chem. Ed. 1963,40, 295. 75. Pitzer, K. S.; Bernstein, L. S. J . Chem. Phys. 1975, 63, 3849. 76. Brocas, J.; Rusu, C. Int. J. Quantum Chem. 1982,22, 331. 77. Strauss, H. L. Ann. Reu. Phys. Chem. 1983, 34, 301. 78. Christe, K. 0.; Wilson, W. W. Inorg. Chem. 1989, 28, 3275. 79. Klobukowski, M.; Huzinaga, S.; Seijo, L.; Barandiaran, Z. Theo. Chim. Acta. 1987, 71, 237. 80. Cutler, J . N.; Bancroft, G. M.; Bozek, J . D.: Tan, K. H.; Schrobilgen, G. J. J. Am. Ch,em. SOC. 1991, 113, 9127. 81. Crawford, T. D.; Springer, K. W.; Schaefer 111, H. F. J . Chem. Phys. 1995, 102, 3307. 82. Kiselev, Yu. M.; Goryachenkov, S . A.; Russ. J. Inorg. Chem., (Engl. Transl.) 1983, 28, 9. 83. Leonidov, V. Ya.; Timoleev, I. V.; Kmelev, Yu. M. Dokl. Akad. N a u k S S S R 1979, 248, 1375. 84. Goncharov, A. A.; Kozlov, Yu. N.; Purmal, A. P., Zh. Fiz. Khznz. 1979,53, 2685. 85. Tse, J . S.; Bristow, D. J.; Bancroft, G. M.; Schrobilgen, G. .J. Znorg. Chem. 1979, 18, 1766.
96
HOLLOWAY AND HOPE
86. Bancroft, G. M.; Bristow, D. J.; Tse, J . S.; Schrobilgen, G. J . Inorg. Chem. 1983, 22, 2673. 87. Bartell, M. J.; Rothman, M. J.; Ewig, C. S.; Van Wazer, J. R. J. Chem. Phys. 1980, 73, 367. 88. Scheire, L.; Phariseau, P.; Nuyts, R.; Smith, A. E., Smith, Jr., V. H. Physica A (Amsterdam) 1980, 101, 22. 89. Molkanov, L. I.; Grushko, Yu. S.; Mishin, Ya. K.; Isupov, V. K. Zh. Eksp. Teor. Fiz. 1980, 78, 467. 90. Bagus, P. S.; Liu, B.; Liskow, D. H.; Schaefer, 111, H. F. J . Am. Chem. SOC.1975, 97, 7216. 91. Krauss, M.; Liu, B. Chem. Phys. Lett. 1976,44, 257. 92. Hay, P. J.; Dunning, T. H. J. Chem. Phys. 1977,66, 1306; (b) Dunning, T. H.; Hay, P. J. J. Chem. Phys., 1978, 69, 134; (c) Hay, P. J.; Dunning, T. H. J . Chem. Phys. 1978,69, 2209. 93. (a) Malli, G. L.; Styszynski, J.; da Silva, A. B. F. Int. J. Quantum Chem. 1995,55, 213; (b) Styszynski, J.; Malli, G, L. Int. J. Quantum Chem. 1996,55, 227. 94. Bancroft, G. M.; Malmquist, P.-A.; Svensson, S.; Basilier, E.; Gelius, U.; Siegbahn, K. Inorg. Chem. 1978,17, 1595. 95. Styszynski, J.; Cao, X.; Malli, G. L.; Visscher, L. J. Comput. Chem. 1997, 18, 601. 96. Huston, J. L. Inorg. Chem. 1982,21, 685. 97. Foropoulus, Jr., J.; DesMarteau, D. D. Inorg. Chem. 1982,21, 2503. 98. Christe, K. 0.; Wilson, W. W. Inorg. Chem. 1988,27, 1296. 99. Nielsen, J. B.; Kinkead, S. A.; Eller, P. G. Inorg. Chem. 1990,29, 3621. 100. Christe, K. 0.;Wilson, W. W. Inorg. Chem. 1988,27, 3763. 101. Schumacher, G. A.; Schrobilgen, G. J. Inorg. Chem. 1984,23, 2923. 102. Klaning, U. K.; Appelman, E. H. Inorg. Chem. 1988,27, 3760. 103. Stein, L.; Norris, J. R.; Downs, A. J.; Minihan, A. R. J. Chem. SOC.,Chem., Comnun. 1978,502. 104. Stein, L.; Henderson, W. W. J . Am. Chem. SOC.1980, 102, 2856. 105. Stein, L. J. Fluorine Chem. 1982,20, 65. 106. Drews, T.; Seppelt, K. Angew. Chem., Int. Ed. Engl. 1997, 36, 273. 107. Mezyk, S. P.; Cooper, R.; Sherwell, J. J . Phys. Chem. 1993,97, 9413. 108. Christe, K. 0.; Wilson, R. D.; Bau, R.; Sukumar, S.; Dixon, D. A. J. Am. Chem. SOC.1991,113, 3795. 109. Dixon, D. A.; Arduengo, A. J.; Farnham, W. B. Inorg. Chem. 1989,28, 4589. 110. Misra, S. N. Indian J. Chem., Sect. A 1979, 18, 530. 111. Zemva, B.; Jesih, A. J . Fluorine Chem. 1984,24, 281. 112. Jesih, A.; Zemva, B.; Slivnik, J. J. Fluorine Chem. 1982, 19, 221. 113. Druzina, B.; Zemva, B. J. Fluorine Chem. 1988,29, 309. 114. (a) MiliCev, S. Vib. Spectrosc. 1995,8, 309; (b)MiliCev, S. Mikrochim. Acta [Suppl.] 1997, 14, 539. 115. Nabiev, Sh. Sh. Zh. Neorg. Khim. 1995, 40, 2016. 116. Bartlett, N.; Wechsberg, M. Z. 2.Anorg. Allg. Chem. 1971, 385, 5. 117. Zemva, B.; Jesih, A.; Templeton, D. H.; Zalkin, A.; Cheetham, A. K.; Bartlett, N. J. Am. Chem. Soc. 1987,109, 7420. 118. Christe, K. 0.;Dixon, D. A.; Sanders, J . C. P.; Schrobilgen, G. J.; Wilson, W. W. J. Am. Chem. SOC.1993,115, 9461. 119. Mercier, H. P. A.; Sanders, J . C. P.; Schrobilgen, G. J.; Tsai, S. S. Inorg. Chem. 1993,32, 386. 120. Christe, K. 0.;Wilson, W. W. Inorg. Chem. 1988,27, 2714.
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
97
121. Christe, K. 0 . ;Wilson, W. W. Inorg. Chem. 1982,21, 4113. 122. Christe, K. 0.; Wilson, W. W.; Wilson, R. D.; Bougon, R.; Bui Huy, T. J . Fluorine Ch,em. 1983,23, 399. 123. Ellern, A.; Mahjoub, A.-R.;Seppelt, K. Angew. Chem., Int. Ed. Engl. 1996,35, 1123. 124. Adam, S.; Ellern, A.; Seppelt, K. Chem. Eur. J. 1996,2, 398. 125. Peterson, S. W.; Holloway, J . H.; Coyle, B. A,; Williams, J . M. Science 1971, 173,
1238. 126. IOselev, Yu. M.; Goryachenkov, S. A , ;Martynenko, L. I.; Spitsyn, V. I. Dokl. Acud. Nuuk SSSR 1984,278, 881. 127. Kiselev, Yu. M.; Fadeeva, N. E.; Popov, A. I.; Korobov, M. V.; Nikulin, V. V.; Spiptsyn, V. I. Dokl. Akud. Nuuk SSSR 1987,295, 378. 128. Spitzin, V. I.; Kiselev, Yu. M.; Fadeeva, N. E.; Popov, A. I.; Tchumaevsky, N. A. 2. Anorg. Allg. Chem. 1988, 559, 171. 129. Weidlein, J.; Muller, U.; Dehnicke, K. “Schwingungsspektroscopie.” Geog. Thieme
Verlag, Stuttgart, 1982. 130. Christe, K. 0.; Curtis, E. C.; Dixon, D. A,; Mercier, H. P.; Sanders, J. C. P.; Schrobilgen, G. J. J. Am. Chem. SOC.1991, 113, 3351. 131. Baran, E. J . J . Mot. Struct. 1992,271, 327. 132. Schrobilgen, G. J.; Martin-Rovet, D.; Charpin, P.; Lance, M. J. J. Chem. SOC., Chem. Commun. 1980, 894. 133. Holloway, J. H.; KauEiE, V., Martin-Rovet, D.; Russell, D. R.; Schrobilgen, G. J . ; Selig, H. lnorg. Chem. 1985,24, 678. 134. Christe, K. 0 . ;Dixon, D. A.; Sanders, J . C. P.; Schrobilgen, G. J., Tsai, S. S.; Wilson, W. W. Inorg. Chem. 1995,34. 1868. 135. Ellern, A.; Seppelt, K. Angew. Chem., Int. Ed. Engl. 1995, 34, 1586. 136. Naumann, D.; Tyrra, W.; Gnann, R.; Pfolk, D.; Gilles, T.; Tebbe, K.-F. 2. Anorg. Allg. Chem. 1997, 623, 1821. 137. Schack, C. J.; Christe, K. 0. J. Fluorine Chem. 1984,24, 467. 138. Schack, C. J., Christe, K. 0. J. Fluorine Chem. 1984,26, 19. 139. Schack, C. J., Christe, K. 0. J. Fluorine Chem. 1985,27, 53. 140. Schack, C. J.; Christe, K. 0. J. Fluorine Chem. 1988,39, 153. 141. Schack, C . J.; Christe, K. 0. J . Fluorine Chem. 1988,39, 163. 142. Lentz, D.; Pritzkow, H.; Seppelt, K. Inorg. Chem. 1978,17, 1926. 143. Lentz, D.; Seppelt, K. 2. Anorg. Allg. Chem. 1980, 460, 5. 144. Turowsky, L.; Seppelt, K. 2. Anorg. Allg. Chem. 1990,590, 37. 145. Mercier, H. P. A,; Sanders, J. C. P.; Schrobilgen, G. J . J . Am. Chem. SOC.1994, 116, 2921. 146. Keller, N.; Schrobilgen, G. J. Inorg. Chem. 1981,20, 2118. 147. Turowsky, L.; Seppelt, K. Inorg. Chem. 1990,29, 3226. 148. Lentz, D.; Seppelt, K. Angew Chem., Int. Ed. Engl. 1978, 17, 356. 149. Lentz, D.; Seppelt, K. Aizgew. Chem., Int. Ed. Engl. 1979, 18, 66. 150. Jacob, E.; Lentz, D.; Seppelt, K.; Simon, A. 2. Anorg. Chem. 1981, 472, 7. 151. Mercier, H. P. A.; Sanders, J. C. P.; Schrobilgen, G . J. Inorg. Chem. 1995,34, 5261. 152. Turowsky, L.; Seppelt, K. 2. Anorg. Allg. Chem. 1992, 609, 153. 153. Syvret, R. G.; Schrobilgen, G. J. J . Chem. SOC.,Chem. Commun. 1985, 1529. 154. Syvret, R. G.; Schrobilgen, G. J. Inorg. Chern. 1989, 28, 1564. 155. LeBlond, R. D.; DesMarteau, D. D. J . Chem. Soc., Chem. Commun. 1974, 555. 156. Sawyer, J. F.; Schrobilgen, G . J.; Sutherland, S. J. Inorg. Chem. 1982.21, 4064. 157. DesMarteau, D. D. J . Am. Chem. Soc. 1978, 100, 6270.
98
HOLLOWAY AND HOPE
158. DesMarteau, D. D.; LeBlond, R. D.; Hossain, S. F.; Nothe, D. J. Am. Chem. SOC. 1981, lo.?. 7734. 159. SchumP(,ier, G. A.; Schrobilgen, G. J. Znorg. Chem. 1983,22, 2178. 160. Faggiani, R.; Kennepohl, D. K.; Lock, C. J. L.; Schrobilgen, G. J. Znorg. Chem. 1986,25, 563. 161. Emara, A, A. A.; Schrobilgen, G. J. J. Chem. SOC.,Chem. Commun. 1987, 1644. 162. Emara, A. A. A,; Schrobilgen, G. J . J. Chem. SOC.,Chem. Commun. 1988, 257. 163. Schulz, A.; Klapotke, T. M. Znorg. Chem. 1997,36, 1929. 164. Turbini, L. J.; Aikman, R. E.; Lagow, R. J. J. Am. Chem. SOC.1979,101, 5833. 165. Naumann, D.; Tyrra, W. J. Chem. SOC.,Chem. Commun. 1989,47. 166. Frohn, H. J.; Jakobs, St. J. Chem. SOC.,Chem. Commun. 1989,625. 167. Frohn, H. J.; Jakobs, St.; Heukel. G. Angew. Chem., Znt. Ed. Engl. 1989,28, 1506. 168. (a) Frohn, H. J.; Rossbach, Chr. 2. Anorg. Allg. Chem. 1993, 619, 1672; (b) Naumann, D.; Butler, H.; Gnann, R.; Tyrra, W. Znorg. Chem. 1993, 32, 861; (c) Naumann, D.; Tyrra, W.; Pfolk, D. 2. Anorg. Allg. Chem. 1994, 620, 987; (d) Butler, H.; Naumann, D.; Tyrra, W. Eur. J. Solid State Chem. 1992,29, 739. 169. Frohn, H. J.; Klose, A,; Bardin, V. V. J. Fluorine Chem. 1993, 64, 201. 170. Frohn, H. J.; Klose, A.; Heukel, G. Angew. Chem., Znt. Ed. Engl. 1993, 32, 99. 171. Frohn, H. J.; Bardin, V. V. J. Chem. SOC.,Chem. Commun. 1993, 1072. 172. Frohn, H. J.; Klose, A,; Bardin, V. V.; Kruppa, A. J.; Leshina, T. V. J. Fluorine Chem. 1995, 70, 147. 173. Frohn, H. J.; Schroer, T.; Heukel, G. 2. Naturforsch. Ted B 1995,50, 1799. 174. Frohn, H. J.; Bardin, V. V. 2. Naturforsch. Teil B 1996, 51, 1011. 175. Frohn, H. J.; Bardin, V. V. 2. Anorg. Allg. Chem. 1996, 622, 2031. 176. Gilles, T.; Gnann, R.; Naumann, D.; Tebbe, K.-F. Acta Crystallogr. Sect. C. 1994, 50, 411. 177. Naumann, D.; Gnann, R.; Padelidakis, V.; Tyrra, W. J. Fluorine Chem. 1995, 72, 79. 178. Frohn, H. J.; Schroer, T. Abstract BA(2) C-4 in Abstracts of 15th International Symposium on Fluorine Chemistry, Vancouver, Canada, 2-7 August, 1997. 179. Naumann, D.; Tyrra, W.; Gnann, R.; Pfolk, D. J. Chem. Soc., Chem. Commun. 1994,2651. 180. Naumann, D.; Tyrra, W.; Gnann, R.; Pfolk, D.; Gilles, T.; Tebbe, K.-F. 2. Anorg. Allg. Chem. 1997, 623, 1821. 181. Zhdankin, V. V.; Stang, P. J.; Zefirov, N. S. J. Chem. Soc., Chem. Commun. 1992,578. 182. Naumann, D.; Gnann, R.; Ignat‘ev, N.; Datsenko, S. 2. Anorg. Allg. Chem. 1995, 621, 411. 183. Hudlicky, M.; Pavlath, A. E. “Chemistry of Organic Fluorine Compounds 11.” American Chemical Society, Washington, DC, 1995. 184. Cremer-Lober, B.; Butler, H.; Naumann, D.; Tyrra, W. 2. Anorg. Allg. Chem. 1992, 607, 34. 185. Lothian, A. P.; Ramsden, C. A. Synlett 1993, 753. 186. Tius, M. A.; Kawakami, J . K. Synlett 1993, 207. 187. Bardin, V. V.; Frohn, H. J . J. Fluorine Chem. 1993,60, 141. 188. Chuang, T. J. J. Appl. Phys. 1980,51, 2614. 189. Alan, K.; Janzen, A. F. J. Fluorine Chem. 1987,36, 179. 190. Forster, A. M.; Downs, A. J. Polyhedron 1985,4, 1625. 191. Lermontov, S. A.; Popov, A. V.; Zavorin, S. I.; Sukhojenko, I. I.; Kuryleva, N. V.; Martynov, I. V.; Zefirov, N. S.; Stang, P. J . J. Fluorine Chem. 1994,66, 233.
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY
99
192. Naumann, D.; Nowicki, G.; Sassen, K.-J. 2. An,org. Allg. Chem. 1997, 623, 1183. 193. Lewe, T.; Naumann D.; Nowicki, G.; Schneider, H.; Tyrra, W. 2. Anorg. Allg. Ch,em. 1997, 623, 122. 194. Minkwitz, R.; Nowicki, G. Angew Chem., Znt. Ed. Engl. 1990,29, 688. 195. Patrick, T. B.; Nadji, S. J . Fluorine Chem. 1988, 39, 415. 196. Forster, A. M.; Downs, A. J. J. Chem. Sac., Dalton Trans. 1984, 2827. 197. Minkwitz, R.; Nowicki, G.; Preut, H. 2. Anorg. Allg. Chem. 1992, 611, 23. 198. Janzen, A. F.; Wang, P. M. C . ; Lenire, A. E. J . Fluorine Chem. 1983,22, 557. 199. Minkwitz, R.; Nowicki, J.; Jahnkow, B.; Koch, M. 2. Anorg. Allg. Chem. 1991, 596, 77. 200. Steinbeisser, H.; Mews, R. J. Fluorine Chem. 1981, 17, 505. 201. Janzen, A. F.; Ou, X. J . Fluorine Chem. 1995, 71, 205. 202. Poleschner, H.; Heydenreich, M.; Spindler K.; Haufe, G. Synthesis 1994, 1043. 203. Laahi, K. K.; Fiedler, W.; Regitz, M. J. Chem. Soc., Chem. Commun. 1997, 1641. 204. Alam, K.; Janzen, A. F. J . Fluorine Chem. 1985,27, 467. 205. Minoura, M.; Sagami, T.; Akiba, K.-Y.; Modrakowski, C.; Sudau, A,; Seppelt, K.; Wallenhauer, S. Angew. Chem., Int. Ed. Engl. 1996,35, 2660. 206. Naumann, D.; Butler, H.; Fischer, J.; Hanke, J.; Mogias, J.; Wilkes, B. 2. Anorg. Allg. Chem. 1992, 608, 69. 207. Ahmed, L.; Morrison, J . A. J. Am. Chem. SOC.1990, 112, 7411. 208. Hartl, H.; Nowicki, J.; Minkwitz, R. Angew Chern., Int. Ed. Engl. 1991, 30, 328. 209. Klapotke, T. M. Heteroatom Chem. 1997,8, 473. 210. Zhang, X.; Seppelt, K. 2. Anorg. Allg. Chern. 1997, 623, 491. 211. Bradley, G. W.; Holloway, J . H.; Koh, H. J.; Morris, D. G.; Watson, P. J . J. Chem. SOC., Perkin Trans. 1 1992,3001. 212. Ebsworth, E. A. V.; Holloway, J . H.; Pilkington, N. J.; Rankin, D. W. H. Angew. Chem., Int. Ed. Engl. 1984,23, 630. 213. Ebsworth, E. A. V.; Robertson, N.; Yellowlees, L. J. J . Chem. SOC., Dalton Trans. 1993, 1031. 214. Zheligovskaya, N. N.; Kiselev, Yu. M.; Krasovskaya, E. P. Koord. Khim. 1980, 6, 1080. 215. Drews, H. H.; Preetz, W. 2. Anorg. Allg. Chem. 1997, 623, 509. 216. (a) Bruce, D. M.; Holloway, J . H.; Russell, D. R. J . Chern. SOC., Chem. Cornmun. 1973, 321; (b) Hewitt, A. J.; Holloway, J . H.; Peacock, R. D.; Raynor, J . B.; Wilson, I. L. J. Chem. Soc., Dalton Trans. 1976, 579; (c) Bruce, D. M.; Holloway, J . H.; Russell, D. R. J. Chem. Soc., Dalton Trans. 1978, 1627. 217. Bruce, D. M.; Hewitt, A. J.; Holloway, J. H., Peacock, R. D.; Wilson, I. L. J . Chem. Sac., Dalton Trans. 1976, 2230. 218. Misra, S. N. Indian J . Chem. 1979, 178, 101. 219. Coleman, K. S.; Holloway, J . H.; Hope, E. G. J. Chern. SOC., Dalton Trans. 1997, 1713. 220. Brewer, S. A,; Holloway, J . H.; Hope, E. G. J . Chem. Soc., Dalton Trans. 1994, 1067. 221. Brewer, S. A,; Coleman, K. S.; Fawcett, J.; Holloway, J. H.; Hope, E. G.; Russell, D. R.; Watson, P. G. J. Chem. Soc., Dalton Trans. 1995, 1073. 222. Brewer, S. A.; Brisdon, A. K.; Holloway, J. H.; Hope, E. G.; Peck, L. A.; Watson, P. G. J. Chem. Soc., Dalton Trans. 1995, 2943. 223. Butin, K. P.; Kiselev, Yu. M.; Magdesieva, T. V.; Reutov, 0. A,, J . Organomet. Chem. 1982,235, 127.
100
HOLLOWAY AND HOPE
224. Kiselev, Yu. M.; Goryachenkov, S. A.; Martynenko, L. I. Russ. J . Znorg. Chem. (Engl. Transl.) 1984,29, 38. 225. Cook, R. L.; Woods, M.; Sullivan, J. C.; Appelman, E. H. Znorg. Chem. 1989, 28, 3349. 226. Fields, P. R.; Stein, L.; Zirin, M. H. J. Am. Chem. SOC.1962,84, 4164. 227. Stein, L. J. Am. Chem. Soc. 1969,91, 5396. 228. Stein, L. Yale Scientific Magazine 1970, 44, (8), 2. 229. Stein, L. Science 1970, 168, 362. 230. Stein, L. U.S.P. 3 660 300, 197. 231. Stein, L. Science 1972, 175, 1463. 232. Stein, L. J. Znorg. Nucl. Chem. 1973,35, 39. 233. Pitzer, K. S. J . Chem. Soc., Chem. Commun. 1975, 760. 234. Stein, L. J. Chem. Soc., Chem. Commun. 1985, 1631. 235. Stein, L. Znorg. Chem. 1984,23, 3670. 236. Stein, L. Nature (London) 1973,243, 30. 237. Stein, L.; Hohorst, F. A. Enuiron. Sci. Technol. 1982, 16, (7), 419. 238. Nefedov, V. D.; Torpova, M. A.; Avrorin, V. V.; Dudkin, B. N. Radiokhimiya 1971, 13, 916. 239. Avrorin, V. V.; Nefedov, V. D.; Toropova, M. A. Radiokhimiya 1974, 16, (21, 261. 240. Avrorin, V. V.; Krasikova, R. N.; Nefedov, V. D., Toropova, M. A. Radiokhimiya 1981, 23, (6), 879. 241. Avrorin, V. V.; Krasikova, R. N.; Nefedov, V. D.; Toropova, M. A. Radiokhimiya 1986,27, (4), 511. 242. Avrorin, V. V.; Krasikova, R. N.; Nefedov, V. D.; Toropova, M. A. Radiokhimiya 1987,29, (31, 426. 243. Avrorin, V. V.; Krasikova, R. N.; Nefedov, V. D.; Toropova, M. A. Radiokhimiya 1989,31, (21, 124. 244. Avrorin, V. V.; Krasikova, R. N.; Nefedov, V. D.; Toropova, M. A. Radiokhimiya 1989,31, (61, 63.
ADVANCES IN INORGANIC CHEMISTRY, VOL. 46
COMING TO GRIPS WITH REACTIVE INTERMEDIATES ANTHONY J. DOWNS and TIMOTHY M. GREENE Inorganic Chemistry Laboratory, University of Oxford, Oxford OX1 3QR, United Kingdom
It i s a capital mistake to theorise before one has data. Inseizsibly one begins to twist facts to suit theories, instead of theories to suit facts. A. Conan Doyle, “A Scandal in Bohemia”
I. Introduction II. Reaction Intermediates: Nerve Centers of Chemical Reactions 111. Experimental Characterization of Reaction Intermediates: Retardation A. Gas-Phase Studies a t Low Pressure B. Gas-Phase Studies in Supersonic Jets C. Trapping and Matrix Isolation D. Solution Studies a t Low Temperatures IV. Experimental Characterization of Reaction Intermediates: Time-Resolved Methods V. Experimental Characterization of Reaction Intermediates: Flow and Other Methods VI. Conclusions References
I . Introduction
Chemistry is the science of change. It is ironic then that, although we have become so knowledgeable about the properties of reactants and products-about the beginnings and ends of chemical journeys-we remain more often than not unsure of just how those journeys are made. Our modern armory of sophisticated spectroscopic and diffraction methods provides every expectation of a direct hit on the ‘‘fixed” objects represented by reactants and products, but is seldom 101 Copyright ID1999 by Academic Presa All rights of reproductiun in any form reserved
O R ~ X - H ~ ~$25 R Moo ~
102
DOWNS AND GREENE
equipped to target the shadowy, moving entity presented by the process of their interconversion. How the reaction rate varies with concentration, temperature, and other conditions may give vital clues but rarely establishes with any certainty the mechanism of what may be, at least superficially, quite a simple homogeneous reaction. Elaborating the mechanism of a reaction is an altogether more complicated undertaking than determining the structures of long-lived reagents or products in that we are seeking ultimately a detailed description of the way in which the structure and bonding of the reagents change with time in each of the several individual acts that normally make up a chemical change. In general terms, a complete account of any such mechanism is going to require knowledge of no less than four aspects (1-3): 1. subdivision of the reaction into its individual steps or equilibria; 2. characterization of each intermediate species in terms of its composition, structure, energy states, and lifetime; 3. a description of the transition state appropriate t o each elementary reaction step, again with reference to composition, stereochemistry, and energetics; and 4. a complete specification of the processes leading to and from each transition state in relation to the geometries and energy levels (mainly electronic and vibrational) of the reactants, intermediates, and products in their ground and excited states. With reactions occurring as they usually do in solution, the role of the solvent at each stage needs to be realistically assessed. The history of kinetic studies is littered with the wrecks of mechanistic hypotheses that ultimately foundered on the unsuspected reefs of solvent mediation. It can come as no surprise, then, that a comprehensive mechanistic description is still well beyond the reach of present techniques, both practical and theoretical, except in a few cases, and those mostly confined to the simplest of systems. More often than not we are obliged to piece together a mechanism that is consistent with all the available facts, both kinetic and nonkinetic; very rarely do we have compelling and unambiguous evidence of one particular mechanism. Nothing could be much simpler, it might be thought, than the exchange that the neutral metal hexacarbonyl CI-(CO)~undergoes with free CO in both the solution and gas phases, particularly as it proceeds in accordance with a rate law of the form
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
103
The solvent having little effect on the activation parameters, it is reasonable to assume that a dissociative mechanism is a t work, with rate-determining fission of a Cr-CO bond being followed by rapid attack a t the resulting pentacoordinated intermediate Cr(CO), ( 4 - 6 ) . Early kinetic studies on substitution of CO by another two-electron ligand L, Cr(CO16+ L
-
Cr(C0)SL + CO
(1)
e.g., L = amine, phosphine, or phosphite, appeared also ( 4 - 6 ) to point to a simple dissociative mechanism, again with Cr(CO), as a common intermediate. However, investigations extending over a much wider concentration range have now shown that a two-term rate law is commonly applicable, uzz.
The activation parameters AH$ and AS$ characterizing the kl term are very similar to those associated with CO exchange, whereas AH2$ for the k 2 term is significantly smaller than AHl$, and AS2$, being negative, differs not only in magnitude but also in sign from A S l $ The most plausible interpretation of these and related experiments is that substitution occurs via two competing pathways, one involving a dissociative mechanism closely akin to that envisaged for CO exchange (see Fig. 11, and the other a “dissociative interchange” (Id)
Reaction coordinate
-
FIG.1. Hypothetical free energy profile for the substitution reaction Cr(CO)6+ L + Cr(CO)SL + CO on the assumption that it proceeds via a dissociative mechanism.
104
DOWNS AND GREENE
process in which there is relatively little bond-making in the transition state. How then are we to picture this Id process? Perhaps L forms encounter complexes in which it occupies an outer-sphere site of a solvated Cr(CO)5fragment. There are other possible explanations, though, and we are now on the "tossing sea" of conjecture, however well informed, rather than the firm ground of incontestable result or inference. Still more problematic is the apparent migratory insertion of nitric oxide into transition metal-carbon bonds, an important reaction in metal nitrosyl complexes and one that may be relevant to biochemical reactions (7). On the evidence of isotopic labeling and kinetic experiments, the insertion of NO into the Co-CH3 bond of the (cyclopentadieny1)cobalt complex ( v5-C5H5)Co(NO)(CH3), which occurs in Reaction ( 3 ) ,
is intramolecular, and the experimental results are consistent with, but do not prove, a mechanism involving methyl migration, followed by phosphine addition (see Scheme 1, lower path) (8).This has the attraction of resembling the mechanism generally agreed to hold for CO insertion in a reaction such as MII(CO)~CH~ + CO + Mn(CO)5[C(0)CH31
(4)
and which is a critical step in many important carbon-carbon bondforming processes mediated by homogeneous transition-metal cata-
04"
R
L-PR,
SCHEME 1. Alternative pathways for the migratory insertion reaction of (q5-C5H6)Co(NO)(CH3)in the presence of a phosphine (8).
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
105
lysts. Still, the worm of doubt may turn. For is not NO a more versatile ligand than CO, and can it not, for example, function as either a 3e- ligand or a le- ligand? Thus, addition of a phosphine molecule accommodated by a change in the ligation of the NO to give a bent Co-N-0 unit offers another reaction pathway, as shown by the upper route in Scheme 1. There is persuasive evidence that such a change affords a low-energy channel in other reactions (81,and it is not obvious why it should not do so to effect NO insertion. In the absence of any definitive experimental information on this score, quantum chemical calculations have been carried out to explore each of the two possible pathways ( 8 ) . The geometries and energies of the reactants, intermediates, transition states, and products have been determined on the basis of Density Functional Theory (DFT), with the results illustrated in Fig. 2. Hence it emerges that the mechanism in which NO insertion precedes phosphine association incurs activation barriers not exceeding 84 kJ.mol-' and is therefore favored over the mechanism involving insertion after phosphine addition, which is opposed by appreciably higher barriers (80-218 kJ.mol-'1. This pleasing endorsement by theory of the mechanism deduced by experiment may seem to settle the issue, but the theory has not been able to budget for the effects of solvation. Still, we cannot be entirely sure of our ground, and about this particular reaction the last word has surely not been said. The preceding examples of CO exchange and NO insertion reactions have to do with systems that have attracted considerable experimental and theoretical attention. Accordingly, although there are fundamental questions still to be answered, some mechanistic features at least are not in doubt. For most of the myriad chemical reactions that
FIG. 2. The calculated profiles of the potential energy surfaces (at the DFT-B3LYP level) for the insertion reaction ( q5-C5H5)Co(NO)(CH3) + phosphine ($-C5H,)Co(phosphine)[N(0)CHplon the assumption (a) of an intramolecular rate-determining mechanism, or (b) of an associative mechanism (8). --.f
106
DOWNS AND GREENE
are known to occur, though, we have neither the facts nor the theory to argue the mechanistic case, and analogy and intuition may well be the sole guiding principles. Hence we may be spared the great tragedy of science-that is, the “slaying of a beautiful hypothesis by an ugly fact”-but, without a secure understanding of mechanism, we are hostages to experience, empiricism, and prejudice when it comes to devising the most efficient, safe, and reliable way of engineering a desired chemical change.
II. Reaction Intermediates: Nerve Centers of Chemical Reactions
Crucial to any problem of mechanism must be the part played by any intermediate compound, for example (i) CI-(CO)~ in the exchange reactions of Cr(CO)5L (L = CO or other ligands) and (ii) either in Reaction (3). (C5H5)Co[N(0)CH3] or (C5H5)Co(phosphine)(NO)(CH3) Indeed, were it possible to detect, identify, and characterize the relevant intermediates, the primary issue of which path is taken in Reaction (3) (see Scheme 1) could be settled beyond peradventure. More generally, the intermediates may be open-shell molecules; they may be radicals; they may involve elements in unusual oxidation states or with unusual coordination geometries or partners. However, until we know what these intermediates are and the part they play, we cannot begin to comprehend the mechanism of any given reaction. As a start, we need to be sure of not only the identities but also the molecular and electronic structures of the intermediates. Such properties cannot necessarily be deduced from the properties of stable molecules and may well challenge conventional principles of bonding. This is all very well, but how are we to cope with intermediates that have but a fleeting existence under normal conditions? At a vapor pressure of 1 torr and in the presence of CO, for example, Cr(COI5is found to have a lifetime in the order of one-millionth of a second at room temperature (9). To counter the all too brief stay normally enjoyed by a molecule such as Cr(CO)5,the experimenter has three main options. One way is to slow down or inhibit the reactions disposing of the molecule. Another is to record a spectroscopic autograph, in absorption or emission, within the lifetime of the transient and so monitor its temporal fate (as in “flash photolysis”). The third is to resort to a flow method, depending on continuous generation of the intermediate in a fluid traveling at a uniform rate along a tube; under the right conditions, decay of the intermediate gives rise to a steady-state concentration
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
107
that varies as a function of the distance traveled along the tube, and that may be analyzed by essentially “static” spectroscopic measurements. Each method has its strengths and weaknesses. None is able individually to deliver all the information we are likely to require for a proper appreciation of how the intermediate fits into the overall mechanism of a given reaction. If we are to realize that ambition, we must be looking therefore to exploit not one but several methods. To gain a better idea of the three strategies for characterizing short-lived intermediates-namely, retardation, time-resolved, or flow methods-we turn next to a brief survey of how they work in practice and of just what light they can or cannot shed on a particular intermediate. How mechanistic studies have been advanced in this way will be illustrated in the course of the account.
HI. Experimental Characterization of Reaction Intermediates: Retardation
There are various ways of slowing down or even suppressing totally the reactions of a molecule. In principle, we may limit its collisions with other molecules by confining it to the gas phase and there minimizing its pressure; we may minimize the thermal energy to which it has access by lowering its temperature; or we may trap it with a suitable substrate. The best policy in practice may well be either simultaneously or in separate experiments to apply more than one of these restraints.
A. GAS-PHASE STUDIES AT Low PRESSURE The gas phase at low pressures is the best medium for detailed structural and spectroscopic studies of very simple, robustly bound molecules such as AIH, SiO, or C10 (10). Generated typically by thermal or discharge reactions, these may survive a t partial pressures high enough and for times long enough to be interrogated by their electronic emission, laser-induced fluorescence, microwave, or infrared absorption spectra, such studies being greatly facilitated by the development of tunable lasers and double resonance techniques. Hence an immense effort has been invested over the years in measuring and analyzing the electronic spectra of stable and unstable diatomic molecules, to give a rich return on information about the bond lengths and vibrational and electronic properties that characterize not only their electronic ground states but also their excited states.
108
DOWNS AND GREENE
A good example is the molecule SiO, a cosmic precursor to silicon oxide chemistry as we now know it on earth (11).In more recent times, too, this molecule has assumed considerable significance because of its relevance to oxidation reactions taking place at the surfaces of silicon wafers and to the creation of antireflection coatings on these and other solid-state devices. Unlike its more familiar counterpart carbon monoxide, SiO is normally quick to aggregate and disproportionate [Reaction (5)] at temperatures below 1000°C: 2SiO
Si + SiO,.
(5)
Stability is not a problem at high temperatures, though, and Reaction (5) can be reversed to produce SiO vapor at a concentration sufficient for microwave absorption measurements by heating a homogeneous mixture of silicon and silicon dioxide to 1350°C in a ceramic reaction vessel, which also serves as the spectroscopic sample chamber (12). Alternatively, excited SiO molecules, produced by the action of a discharge on a gaseous mixture including, say, Sic&and O,, can be characterized by their electronic emission spectrum. It has taken 10 to 20 years to unravel all the features of the electronic spectrum extending from the near- to the vacuum-ultraviolet and assign the different band systems, but we now have a relatively full knowledge of the molecule in its electronic ground and numerous excited states ( 1 3 ) . Rather easier to decipher is the microwave spectrum of the molecule in its electronic ground state, whence a wide range of vibrational and rotational states have been accurately detailed (111. These and other studies give every confidence that we are dealing with a z*Si160 molecule having the following equilibrium parameters: re = 150.9739 pm, we = 1241.56 cm-', and pe = 3.0882 D (13). Some of the microwave lines have served as distinctive signatures by which SiO has been recognized in astronomical sources, and SiO masers have relayed important information about both the conditions in circumstellar shells and the mechanisms of mass loss. Solid boron nitride is a ceramic material of some consequence that can be formed at high temperatures by the reaction of boron atoms with Nz or NH3. A likely intermediary in its formation is another high-temperature molecule BN, first observed through its electronic emission spectrum nearly 60 years ago but the properties of which have emerged only slowly (14). Identification of the several low-lying electronic states of the molecule is a major problem and only very recently has it become clear that the ground state is not (as with the isoelectronic Cz molecule) but 311 (15).The A311,-X311itransition,
lz+
109
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
the equivalent of the celebrated Swann bands of C,, can be excited in emission by the action of a microwave discharge on traces of BCls and N2 entrained in a helium gas stream. Analysis of the well-resolved rotational lines of the 0-0 and 1-0 bands yields the following properties for the ground state of llB1*N:re = 132.9 pm and o,= 1514.6 cm-’ (13, 16). The molecule C10 is another reactive species that has attracted still more attention. It plays a significant role in the destruction of stratospheric ozone through catalytic cycles involving heterogeneous and homogeneous reactions, as in Eq. (6) ( 1 7 ) :
c1 f 0:3+ c10 + 0 2
I
(6)
C10+0-+C1+0~.
Current wisdom can be summarized in the flow chart of Fig. 3. More specifically, the C10 radical is involved in several possible kinetic mechanisms linking global release of chlorofluorocarbons (CFCs) to the Antarctic ozone loss during each austral spring ( 18).An accurate characterization of the spectroscopic and other properties of C10 is therefore vital, not least as a prelude to tracing the crucial correlation between CFCs, C10, and 0,. C10 is short-lived a t the high molecular concentrations characteristic of the condensed phases but can be generated a t low pressures in the gas phase, for example by the action of
nv
I
I
I
LIU
I
FIG. 3. Chemical cycles affecting the formation and decay of chlorine oxide trace species in the earth’s atmosphere (reproduced with permission from Wayne, R. P. “Chemistry of Atmospheres,” 2nd ed., p. 137, Clarendon Press: Oxford, 1991).
110
DOWNS AND GREENE
a microwave discharge or a short burst of radiation on a mixture of Clz and 02.On the evidence of a detailed analysis of the A211i-X211i band system as measured both in absorption and in emission at high resolution, of the infrared spectrum also measured at high resolution, and of the microwave spectrum, the 35Cl'60molecule has a 211 ground state with re = 156.960 pm, we = 853.724 cm-l, pe = 1.24 D, and D;= 265.4 k J * mol-l(I3, 19). Of its reactivity, though, the properties give the barest hint. Simple polyatomic molecules such as BH2 (201, BH, (211, and HBNH (22) may be no less amenable to high-resolution spectroscopic studies. Commonly invoked as an intermediate in reactions of boron hydrides, the BH3 molecule is made elusive by its rapid dimerization to diborane. It is produced, nevertheless, by the action of a dc discharge on a gas stream composed of diborane and helium flowing through a multipass absorption cell; here it can be recognised by its vz (a;) and v3 ( e ' ) vibrational bands as recorded with a high-resolution FT-IR spectrometer (21). Detailed analysis of the rotational structure associated with the bands fixes a regular planar DBhstructure for BH, in its ground electronic state with r&B-H) = 119.001 pm. Another discharge reaction, this time involving diborane and ammonia, generates the species HBNH with a lifetime of a few hundred ms under normal conditions. That this is a linear molecule is made clear by the rotational structure of the v3 band (corresponding to the BN stretching fundamental) exposed by an IR diode laser spectrometer (22);the observed rotational constants imply a BN bond length (123.81 pm) appreciably shorter than those in H3BNH3,HzBNHz,and even BN itself (165.76, 140.3, and 132.9 pm, respectively). In principle, electron diffraction is another tool capable of giving precise information about the structure of a gaseous intermediate, even when this is quite a complicated molecule (23). In practice, though, there are numerous complications liable to militate against electron diffraction as the sole agency of characterization. Chief among these is the low resolution of the electron-scattering pattern of a molecule. As a result it is seldom easy to discover much about the structure-sometimes even the identity-of a specific molecule in a gaseous mixture. To resolve an otherwise severely underdetermined problem, we need independent testimony about the molecules making up the mixture, for example by reference to the mass, rotational, or vibrational spectrum of the sample or to the results of quantum chemical calculations (24). A good example of this strategy is provided by a careful study of the thermal decomposition of trimethylamine vapor (25).Mass spectrometric analysis of the vapor stream shows that de-
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
111
composition proceeds at 515°C primarily in accordance with Eq. (7) to give methane and the short-lived compound CH3-N=CH2:
Analysis of the electron-diffraction pattern of the reaction mixture implies the following mole fractions: (CH&N, 0.63; CHI, 0.18; and CH3-N=CH2, 0.19. The structure of the CH3-N=CH2 molecule has then been determined by a joint analysis of the electron-diffraction pattern and the rotational constants (derived from the microwave spectrum), giving parameters [rZ(N=C)= 127.9 pm, rZ(N-C) = 145.8 pm, and LC-N=C = 116.6"l very close to those calcylated by ab znitio methods. Similarly, pyrolysis of propylenimine, HNCH2CHCH3, a t 470°C results in complete decomposition to give the short-lived species syn- and antz-CH3-NH-CH=CH2 and truns-CH,-N=CHCH, (26);geometrical structures and mole fractions of all three products have been deduced by analyzing the electron diffraction pattern of the vapor in concert with the properties implied by quantum chemical calculations. How the proportions of the transients vary with temperature sheds significant light on the reaction path taken by the thermal rearrangement. Life is altogether simpler with only one molecular species in the vapor, and electron diffraction is well established as a primary method of characterizing high-temperature molecules such as MgC12 (27) or C,, (28). In general, however, the optimum conditions can be met only for a relatively long-lived intermediate such as S2F4.This is formed by the dimerization of SF, on the winding reaction path that starts with SClz and a metal fluoride and affords the best known route to SF4, represented mainly by the overall reaction, 3SC12 + 4MF + SF4 + S&1, (e.g., M
=
+ 4MC1.
(8)
Na or K)
The labile molecule has required the combined forces of electron diffraction, microwave, vibrational, and "F NMR spectroscopy, and ab initio calculations before yielding the secrets of its remarkable, unsymmetrical structure, I (29);this can be viewed as a trigonal bipyramid centered on one sulfur atom with a lone pair of electrons, the SF group, and one fluorine atom occupying the equatorial sites. Electrondiffraction studies have also contributed to our understanding of how
112
DOWNS AND GREENE
I
F
1
gaseous organometallic compounds can, through their decomposition, furnish solid materials with useful electronic or other properties. For example, thermal decomposition of t-butylgallium sulfide, [(But)GaS14, results in the growth of a new cubic phase of Gas. Not only is this material ideal for the electronic passivation of GaAs surfaces, it has a large band gap, which makes it suitable as the insulating “gate” layer in certain transistor devices. What is striking about the structure of the gaseous precursor, 11, as determined by electron diffraction, is the presence of a distorted cubane-like Gads4core with a t-butyl group bound to each gallium atom and T symmetry overall (30).Irradiation of the precursor with the output of a UV excimer laser gives rise to the photofragments (But),Ga4S4(where x = 0-31, which can be detected by their time-of-flight mass spectra following ionization by a second UV excimer laser. It appears, then, that fragmentation proceeds primarily by loss of the organic substituents with retention of the Ga4S4core, clearly suggesting that growth of cubic Gas depends on the oligomerization of preformed Ga4S4fragments.
Problems of analysis apart, experiments of this sort can at their best provide a remarkably full picture of a gaseous molecule, at least with regard to its physical character (10). By deliberately seeking to
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
113
frustrate any chemical reactions, however, we are naturally denying ourselves any detailed chemical and, in particular, kinetic information. There are other limitations, too. The conditions are far removed in terms of molecular concentrations, as well as environment, from those normally prevailing in the condensed phases. For reasons of rotational congestion, moreover, high-resolution spectroscopic measurements cannot usefully be applied to molecules containing more than a few atoms, and for reasons of pressure and temperature such experiments may lose sight of molecules incorporating weak bonds. In fact, weakly bound molecules commonly feature as the initial products when two reagent molecules interact. These problems lead naturally t o the question, How can the power and precision of modern high-resolution spectroscopic methods be brought to bear on larger andlor weakly bound transients? Fast freezing in one form or another is the obvious response.
B. GAS-PHASE STUDIES IN SUPERSONIC JETS When a noble gas seeded with a small amount of other molecules expands through a nozzle into a vacuum, rapid adiabatic cooling occurs, provided that the mean free path of the molecules in the unexpanded gas is short compared with the nozzle diameter. Within only a few nozzle diameters (usually ca. 5 mm), the molecules in the jet of gas emerging from the nozzle achieve a very narrow distribution of speeds in the jet direction, clustered around the supersonic value of ca. 5 x lo4 cm . s-l, and have nearly zero relative speed in directions transverse to this. Consequently, within 5 to 10 p s of emerging from the nozzle into the vacuum, the molecules are in an essentially collisionless state and have a velocity distribution in the transverse direction corresponding to temperatures of a few Kelvin. A similar effective temperature is also achieved through cooling of the rotational and vibrational degrees of freedom. If we can arrange for the intermediate to be formed within the noble-gas stream just before it emerges from the nozzle, freezing times in the order of microseconds can be attained, and thereafter the intermediate is effectively isolated, having no mechanism, either unimolecular or bimolecular, by which to react or decompose. One experimental arrangement, illustrated in Fig. 4, consists of two concentric, coterminal tubes (31-33). One reagent admixed with an excess of noble gas is delivered in short pulses via the outer tube while the second reagent (typically also diluted with noble gas) flows continuously through the central tube. The two reagents
114
DOWNS AND GREENE
Series 9 solenoid valve
Teflon outer tube pulse of F2-Ar from solenoid valve 0.3 mrn 1.d. glass capillary
FIG.4. The fast-mixing nozzle used to observe the rotational spectrum of the gaseous complex H,N----F,[reproduced with permission from (33),p. 1121.
meet and mix in the roughly cylindrical interface between the concentric flows as they simultaneously expand. Any intermediate formed here rapidly undergoes collisionless expansion, is then effectively frozen, and may be investigated by such techniques as microwave, infrared diode-laser, laser-induced fluorescence, or photoionization spectroscopy. For characterizing a dipolar molecule in its electronic ground state, few methods are more instructive than pulsed-nozzle Fourier-transform microwave spectroscopy (32). As illustrated schematically in Fig. 5, a short pulse of microwave radiation directed at the gas pulse excites a rotational transition in the species of interest; subsequently the rotationally excited molecules reemit radiation, which is detected. This technique provides a remarkably sensitive probe for transients, the properties of which can be specified with all the precision and detail peculiar to rotational spectroscopy only microseconds after their production. In relation to a weakly bound adduct A----Bformed by two molecular reagents A and B, for example, we may draw on the rotational spectrum to determine such salient molecular properties as symmetry, radial and angular geometry, the intermolecular stretching force constant and internal dynamics, the electric charge distribution, and the electric dipole and quadrupole moments of A----B(see Table I). The mechanism of the addition of molecular chlorine to an alkene
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
115
Pulsed nozzle
Fabry-Plrot mirrors
FIG. 5. Pulsed-nozzle FT microwave measurements. A molecule-radiation interaction occurs when the gas pulse is between mirrors forming a Fabry-Perot cavity. If the transient molecule has a rotational transition of frequency v,,, falling within the narrow band of frequencies carried into the cavity by a short pulse (ca. 1p s ) of monochromatic radiation of frequency v, rotational excitation leads to a macroscopic electric polarization of the gas. This electric polarization decays only slowly (half-life T2ii: 100 p s ) compared with the relatively intense exciting pulse (half-life in the cavity T~ = 0.1 ps). If detection is delayed until ca. 2 ps after the polarization, the exciting pulse has diminished in intensity by a factor of ca. lo6 but the spontaneous coherent emission from the polarized gas is just beginning. This weak emission can then be detected in the absence of background radiation with high sensitivity. For technical reasons, the molecular emission at v, is mixed with some of the exciting radiation v and detected as a signal proportional to the amplitude of the oscillating electric vector a t the beat frequency 0 - I),,, as a function of time, as in NMR spectroscopy; Fourier transformation leads to the frequency spectrum [reproduced with permission from (31),p. 5631.
has been the subject of much investigation and discussion ( 3 4 ) .In a fairly polar solvent and in the dark, the reaction has been shown to be of first order in both components, indicating that the entity responsible for electrophilic attack on the alkene is the dihalogen molecule. There is a general consensus that the mechanism, as represented by Eq. (9), involves a pre-equilibrium molecular association of the alkene and C12to give the adduct 111: alkene
+ C12+ alkene----C12+ alkeneC1' + C1-. I11
Iv
(9)
116
DOWNS AND GREENE
TABLE I MEASURABLE SPECTROSCOPIC AND MOLECULAR PROPERTIES OF GASEOUS A-B MOLECULES (32) ACCESSIBLE THROUGH GROUND-STATE ROTATIONAL SPECTRA Spectroscopic parameter
Molecular property of A----B
Form of the spectrum
Symmetry
Rotational constants, -40, &, GI
Quantitative determination of radial and angular geometry; nature of intermolecular binding Quadratic force constant for A----Bstretching, k,, Relate to electric field gradient qi, a t a nucleus X having a nonzero electric quadrupole moment Electric dipole moment of A....B Molecular g factor and electric quadrupole moment of A----B
Centrifugal distortion constant, DJor AJ Hyperfine coupling constants, x,, and D,, Stark effect Zeeman effect
Comment/example Spectral pattern varies for linear, symmetric-top, and asymmetric-top species, e.g., F2----NH3 is a symmetric top (35). (34) and e.g., CzH4----C12 F,----NH,(35)
DJis roughly proportional t o k,,. qig depends on the detailed
electric charge distribution within A----B. (36) e.g., HC’sNN----H”6Cl
e.g., Ar----HCI(37)
This intermediate then ionizes in the rate-controlling step to give a cation, IV,which reacts rapidly with Cl-. For many years, however, the precise nature of the pre-equilibrium complex I11 has been a matter of speculation. It is clearly desirable to establish with certainty and precision the existence of any such complex. Mixing ethene and molecular chlorine in an apparatus of the type described gives rise to a new ground-state rotational spectrum attributable to the complex CzH4----Cl2. Analysis of the spectrum indicates that this complex has the Czugeometry illustrated in Fig. 6(a), with the Clz molecule lying along the C2 axis of ethene that is perpendicular to the molecular plane. The weakness of the interaction between the molecules is made abundantly clear by the following features (34): the rotational constant A, is only slightly larger than the constant C, of free ethene; the inner C1 atom is relatively remote from the center of the ethene C=C bond (estimated distance 312.8 pm); at 5.9 N m-l the intermolecular stretching force constant, k,,,deduced from the centrifugal distortion
-
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
117
constant is quite minuscule; and the nuclear quadrupole coupling constants xgg(35Cl) undergo quite minor changes, implying that the cylindrical symmetry of the electric charge distribution of Clz is only slightly perturbed on complex formation. Hence, there can be little doubt that the intermediate is formally a b7i.w complex of the weak, outer type, according to Mulliken's classification. More complicated and less easily controlled is the reaction between ammonia and fluorine gases, which under normal conditions of temperature and pressure results in a spectacular flame and the formation of NF, in small yield. Despite the prodigious reactivity of molecular fluorine, it has proved possible through the device of fast mixing and cooling of the gases in a supersonic jet to detect and characterize the weakly bound prereactive intermediate H3N----F2on the basis of its ground-state rotational spectrum (35). The results attest to the C3rstructure illustrated in Fig. 6(b) with a long intermolecular contact, r(N----F),of 270.8 pm and a stretching force constant k,,of no more than 4.7 N.m-'. By contrast, the corresponding complex H,N----ClFfeatures an appreciably stronger intermolecular bond, with r(N----Cl)= 237.6 pm and k,,= 34.3 Nem-', pointing to a small but significant transfer of charge in the sense H3N----Cl+F(38). It is not just loosely bound complexes that can be specified with profit by spectroscopic analysis of a supersonic jet. Of this there is no more thrilling example than the transient organometallic radical VCH (39),formed when vanadium atoms generated by laser ablation are entrained in a pulse of high-pressure helium containing 5-10% CH, prior to expansion to give a supersonic, cooled molecular beam. High-resolution studies of the molecular fluorescence near 800 nm excited by a tunable probe laser reveal extensive vibrational and
I
c'
P-------
FIG.6. Structures of the gaseous prereactive intermediates (a) C2H4----C12 and (b) H:jN----F,.
118
DOWNS AND GREENE
rotational detail. Hyperfine patterns characteristic of a nucleus with spin I = 7/2 confirm that the carrier contains a vanadium atom; the subband structures are those of a linear molecule; the integer rotational quantum numbers point to an even number of unpaired electrons; and the presence of hydrogen is borne out by experiments with CD4, which give rise to the corresponding bands of VCD. The ground state of the molecule can be identified unambiguously as 3A, and the associated rotational constants imply the following dimensions for the substitution structure: rs(V=C) = 170.25 and r,(C-H) = 108.0 pm. Hence, we now have substantive experimental data for an exemplar of the simplest possible type of metal carbyne complex, making clear inter alia that the V=C stretching mode is at 838 cm-', and not near 1300 cm-l as has been suggested for related species. Even more challenging are some of the neutral and charged species that are created and destroyed within an electric discharge. Yet, by subjecting a pulsed molecular beam to the combined actions of supersonic expansion and a corona discharge (on the downstream side of the nozzle), it has been possible to detect and study short-lived intermediates as varied as OH, NB , and CH3(40). Optical spectroscopy has been the traditional source of information, but major contributions to our limited knowledge of the highly complex chemistry of plasmas have come recently from mass, microwave, and infrared laser spectroscopies. Lifetime is not the only problem, for even within the discharge the concentrations of the relevant species are often very small. This has stimulated the development of special modulation techniques to improve both selectivity and sensitivity (40). Use of an ac discharge gives, for example, Doppler shifts of spectroscopic frequencies due to molecular ions, which vary with the discharge frequency vf, and state populations of any transient species, which vary with twice this frequency. When the spectroscopic detection system is referenced to the frequency ur or 2 u f , the resulting demodulated signal consists only of what are velocity-modulated or population-modulated lines, respectively. Transients such as OH and CH3are no mere scientific curiositieswill-o'-the-wisps peculiar t o the elaborate combination of discharge, supersonic expansion, and spectroscopic subterfuge. Some of these molecules play vital roles in regimes seemingly quite remote from plasma chemistry. For example, the OH radical is crucially significant as an oxidizing initiator in combustion and in the troposphere, where its natural concentration is liable to be seriously augmented as a re-
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
119
sult of pollution (17). Two lines of investigation must suffice to demonstrate the scope of modern discharge experiments. 1. The molecules HCCCS and HCCCCS have been characterized by their microwave spectra following their generation by the action of a pulsed discharge on a supersonic molecular beam of argon containing traces of C2H2and CS2(411. The microwave frequency, swept in small steps of a few MHz, is fed to a Fabry-Perot cavity, which is maintained at the resonance frequency by synchronous length adjustment so as to retain the high Q value of the cavity. This feature, combined with cooling t o rotational temperatures near 1-2 K, enhances greatly the sensitivity of the experiment. The resulting high-resolution spectra are noteworthy for their disclosure that both HCCCS and HCCCCS, in common with HCCS, are linear in their electronic ground states (TI) and so contrast with molecules in the analogous family HC,,O ( n = 2-41, all of which are terminated by angular C-C-H units (41).That molecules of both series are created in discharge reactions is of more than terrestrial note, beceuse they are potential players of some significance on the interstellar scene. 2. Silane discharges are of practical use as sources of thin-surface films. Infrared diode laser spectroscopy with Zeeman modulation has been exploited to detect and analyze the bending fundamental v2 of the pyramidal SiH3 radical (42), which appears t o be a primary component of silane plasmas and is likely therefore to be a key intermediate in chemical vapor deposition processes originating in silane. Analysis of the results gives r(Si-H) = 146.8 pm and LH-Si-H = 110.5'. Other discharge products include SiHz and the novel molecule SieHz, which is predominant in a low-pressure silane/argon plasma on the evidence of mass spectrometric measurements. With a low-power plasma cooled at liquid nitrogen temperature, Si2Hecan be detected by its submillimeter-wave rotational spectrum. The rotational lines are unaffected by a confining magnetic field and so cannot arise from a paramagnetic or ionic species. Moreover, they correspond to the pattern expected for a near-prolate symmetric top with ( B + C)/2 = ca. 7.2 GHz, a value close to what would be expected for a molecule containing two Si atoms, and adjacent lines exhibit a 3 : 1 intensity alternation consistent with the nuclear spin statistics associated with two equivalent protons. Putting this and other evidence together leads to the double hydrogen-bridged butterfly structure depicted in Fig. 7(a), quite unlike the familiar linear configuration of CzH2( 4 3 ) .Confirmation that this is indeed the most stable conformer of SizH2comes from
120
DOWNS AND GREENE
FIG.7. Structures of the molecule SizHzpresent in a silane/argon plasma (a) i n its ground electronic state and (b) in an excited state. Dimensions are those determined by analysis of the high-resolution rotational spectra (43, 45).
ab initio quantum chemical calculations (441,which also predict, however, the existence of two other stable conformers, including one having the monobridged structure shown in Fig. 7(b) and lying only about 36 kJ-mol-' above the global minimum. It has been a triumph for theory that such a conformer should now have been detected and characterized experimentally following closer scrutiny of the millimeter- and submillimeter-wave rotational spectrum associated with the products of the silane/argon plasma ( 4 5 ) .The rotational constants of the normal and perdeuterated versions of the molecule afford an ro structure with the dimensions indicated in Fig. 7(b) and in tellingly close agreement with those predicted by the ab initio calculations. Interestingly, the molecule, which is estimated to have a lifetime not exceeding 10 ms, displays the shortest Si-Si bond to be observed to date. Nor is this the end of the tale, for the heteronuclear species SiCHz, produced by striking an electric discharge in a high-pressure argon pulse seeded with tetramethylsilane vapor, has for its ground state the silylidene structure Si=CH2. The molecule has been characterized not only by its microwave spectrum but also by its distinctive, intense laser-induced fluorescence spectrum (461, which consists of
COMING TO GRIPS WITH REACTM? INTERMEDIATES
121
15-20 vibronic bands in the 300- to 342-nm region; the rotational structure of the bands, clearly defined at high resolution, features a central subband that is weaker than the outer ones, a necessary consequence of the nuclear statistical weights if there are two equivalent hydrogen atoms. The rotational constants of SicHz and SiCDz have been used to deduce the following partial substitution parameters for the Czu ground-state structure: rs(C-H) = 109.9 and rsfC=Si) = 170.6 pm, and LH-C-H = 114.4". Than SizHz and SiCH2 we could have no better demonstration of how remote open-shell reactive intermediates may be in their molecular and electronic structures from the familiar world of long-lived, closed-shell molecules with their apparent obedience to conventional bonding principles. Chemical studies of plasmas are still only in their infancy, and more new and exotic intermediates will surely come to light. Whatever we may make of their structures and bonding, they will have to be accommodated by any mechanistic proposal that seeks realistically to account for the reactions taking place in the discharge (47). Then felt I like some watcher of the skies When a new planet swims into his ken . .
C. TRAPPING AND
MATRIX
.
ISOLATION
One way of catching and identifying a reactive transient is to intercept it with a reactive substrate so as to produce a known product that has a distinctive spectroscopic signature. For example, hydrogen atoms formed in the course of a reaction may be effectively scavenged by carbon monoxide to form the well-authenticated radical HCO, which can be identified unambiguously under appropriate conditions by its infrared or ESR spectrum (48).In a similar vein, ground-state sulfur atoms may be scavenged by dioxygen and detected indirectly by the infrared absorption or UV emission spectrum of the SO, molecules thus generated; by contrast, sulfur atoms in the excited 'D state insert characteristically into the C-H bonds of alkanes (48). Some of the first evidence pointing to the formation of the trihydrides of aluminium and gallium came similarly from trapping experiments, this time with trimethylamine to form the known, relatively stable adducts (Me3N),,MH3(M = Al or Ga; n = 1 or 2) ( 4 9 ) .With molecules as with atoms, it may be possible on the basis of suitable trapping
122
DOWNS AND GREENE
experiments to determine whether the intermediate is formed in its electronic ground state or in an excited state. “Singlet” dioxygen, that is, the 0, molecule in its excited lh, electronic state, results from the stoichiometric oxidation of hydrogen peroxide by hypochlorous acid: HOOH + HOCl + lo2+ H20 + HC1.
(10)
It has a lifetime of only 2 ps in aqueous solution, but it may be intercepted and thereby distinguished from ground-state dioxygen by its facile addition to a conjugated molecule such as butadiene (50):
An alternative and more direct strategy involves matrix isolation (51-53). Here the trick is to catch and hold the intermediate by embedding it in a rigid, inert host such as solid argon (the matrix) at low temperatures (4-50 K). Hence, our fugitive is at once isolated and effectively immobilized. To some extent the experiments imitate nature, for the mineral lapis lazuli owes its beautiful blue color to the highly reactive radical anion S , , which is entrapped in an aluminosilicate matrix. They also improve upon nature with the choice of a transparent, weakly interacting host such as a frozen noble gas at low temperatures, in place of the polar aluminosilicate, which is opaque to broad regions of electromagnetic radiation. However short-lived the transient may be under normal conditions, its lifetime can be extended almost indefinitely by matrix isolation, and we can appeal to a variety of spectroscopic techniques-notably vibrational, electronic, and ESR methods-to follow what is going on in the matrix and to detect, identify, and characterize the trapped species. The method has great advantages. Spectroscopic measurements can be made at leisure ai., therefore in detail, and several different methods can be applied to interrogate a particular sample. Moreover, the infrared spectrum, which has been the principal agent of detection and analysis of matrix samples, is usually comparatively simple, with sharp linelike features marking the different transitions (and reflecting the conditions of low temperature, relatively isotropic, weakly interacting environment, and inhibition of rotation). Not only can vibrational frequencies meaningfully be measured to -+ 0.1 cm-
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
123
in many cases, giving values close to those of the gaseous molecule, but the linelike character of the bands bids fair to the study of isotopic effects, frequently the most compelling means of determining the stoichiometry and geometry of the trapped molecule. On the other hand, there is also a price to be paid. Thus, inhibition of rotation means that we lose precise information about the dimensions, vibrationh-otation, distortion, and charge distribution of the isolated molecule. In addition, various complications can arise from the weak but significant perturbations imposed on the trapped species by the cage it inhabits, which is likely to vary in size, geometry, and composition within a given matrix; as a result, specific spectroscopic transitions do not always appear as sharp bands but sometimes as multiplets or relatively broad bands. The combination of rapid quenching with the rigidity of the resulting condensate is also liable to lead to the trapping of different conformers of a particular molecule. In practice no less than principle, the performance of a typical matrix-isolation experiment is quite simple. As illustrated in Fig. 8, a gas mixture is deposited under controlled, high-vacuum conditions on a cold surface the nature of which is dictated by the type of spectroscopic measurement to be made on the sample (51-53). The species of interest may be formed in the gas phase (by pyrolysis or in a discharge) and then quenched rapidly with a n excess of the matrix gas; sometimes it is formed during the cocondensation of two reagents; commonly it is generated in situ, usually by irradiation of an appropriate matrix-isolated precursor. As noted previously, identification of a novel or unfamiliar captive molecule is accomplished more often than not by reference to its infrared spectrum, with the following steps being typical of the way in which the experiment might proceed. 1. First, the number of distinct species present in the matrix has to be established; this is done by allocating the infrared absorptions to the appropriate species on the evidence of how the spectrum varies with changes in the experimental conditions. 2. We need then to assess the number, frequencies, and intensities of all the absorptions associated with a given molecule before drawing on the evidence of distinctive features due to specific groups, analogies with known compounds, or the vibrational selection rules in order to formulate its likely identity. 3. Evidence of identity is greatly strengthened if it is possible to devise a n alternative route to the molecule in question starting from a different precursor. This strategy is well illustrated by the 16-electron
124
DOWNS AND GREENE
Inlet for gaseous sample
Refrigerant
Inlet for matrix gas
(4-20 K)
High-vacuum pumping system
Valve
Photolysing radiation Quartz window
Infrared spectrometer beam
matrix support (C s I window)
Sample inlet
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
hv
125
Ar matrix
FIG.9. Different routes leading to the unsaturated intermediate tungstenocene, (q5C5H5)2W[reproduced with permission from Downs, A. J.; Hawkins, M. Adv. Infrared and Raman Spectrosc. 1983, 10, 61.
intermediate “tungstenocene,” ( $-C6H5),W ( 5 4 ) , which has been shown by its infrared spectrum to be the common product of the matrix reactions set out in Fig. 9. 4. Annealing or photolysis of the matrix may also be instructive for the evidence of reactions such as ( 5 5 )
2BH3
Ar matnx 27
K
BPHB.
(12)
FIG.8. ( a ) Schematic representation of a typical matrix-isolation assembly. (b) Schematic plan of a matrix-isolation assembly suitable for infrared (transmission) measurements and for photolysis experiments [reproduced with permission from Almond, M. J.; Downs, A. J. Adv. Spectrosc. 1989, 17, 31.
126
DOWNS AND GREENE
5. Most revealing of all is usually the response of the spectrum to changes in the isotopic composition of the molecule. If the sample is prepared in such a way as to produce a mixture of isotopomers, then it is often possible to deduce unequivocally the stoichiometry and geometry of the molecule from the number, frequencies, and relative intensities of the infrared absorptions. With sufficient information, we may even be in a position t o estimate one or more interbond angles, if these are not already fixed by the symmetry of the molecule. In the case of metal carbonyls, the vibrational problem can be simplified by concentrating on the restricted region of the infrared spectrum near 2000 cm-I, corresponding to C-0 stretching vibrations, v(C-01, which can be successfully analyzed independently of the other fundamental vibrations through the use of an energy-factored force field. For each of several possible models of a particular metal carbonyl fragment M(CO), , it is then comparatively easy to estimate the frequencies and relative intensities of the v(C-0) infrared bands for the family of isotopomers M(12C0),(13CO),-, (x = 0-n) and so determine which model gives the best account of the observed spectrum. Figure 10 shows the results of applying this procedure (i) to Fe(C0I4,the primary product of ultraviolet photolysis of Fe(CO), (56), and (ii) to O,MO(CO)~,a prominent intermediate in the train of reactions induced by ultraviolet photolysis of Mo(CO16 in the presence of 0, (57). Hence, there is little doubt that Fe(C0I4 has the CZustructure V or that 02Mo(CO)4 has the tram configuration VI with a square planar Mo(CO)~ fragment.
V
VI
6. The greatly enhanced sophistication of modern quantum chemical calculations means that we are now able to predict with considerable confidence the equilibrium structure of a molecule, the harmonic frequencies of its normal modes, and the intensities of the relevant features in infrared absorption. Accordingly, it has become quite commonplace with matrix as with other studies to test the inferences drawn from experiments against the results of appropriate ab initio or DFT analyses. How closely theory matches experiment is shown
127
COMING TO GRIPS WITH REACTIVE INTERMEDIATES r
1
I
(i i)
A a
?
I
I
I
a 2110
2080 cm-1
2675
2050 2000
19j5
1950
cm-1
FIG.10. (i) IR Spectra showing how the structure of matrix-isolated Fe(CO), was established: (a) observed spectrum of Fe(CO)l with partial W O enrichment isolated in a solid SF,, matrix a t 20 K [bands shown in dotted lines are due to residual Fe(CO),, precursorl; tb) spectrum calculated for the C2,,geometry, V, which gives the best agreement with the observed spectrum; (c) spectrum calculated for a C:+,geometry which provides a poor match to the observed spectrum 156). tii) IR Spectra showing how the structure of trans-0,Mo(CO)4was established: ( a ) observed spectrum of trans02Mo(C0)4with partial W O enrichment isolated in a solid CH1 matrix (bands shown in dotted lines are due to photoejected '.'CO); (b) spectrum calculated for a planar Mo (CO), fragment with Dlh symmetry, VI (57).
for the case of matrix-isolated MezAINHzformed in situ by the elimination of CH, from the adduct Me& * NH3 through irradiation with light a t wavelengths near 210 nm (see Fig. 11)(58). In relation to mechanistic enquiries, the principal strength of matrix isolation lies in the identification and specification of real or potential reaction intermediates. From such studies has come, for example, our first sighting of unsaturated molecules such as Fe(C01, (561, Fe2(CO), (591,Cr(CO)5(601, (q5-C5H&W(54),and Ru(dmpeh (dmpe = MezPCHzCH2PMe2)(61). The equilibrium structures of the ground-
2
128
DOWNS AND GREENE
X
1
3400
,
3000
I
1
1600
I
,
,
1000
,
,
I
400
FIG.11. IR Evidence for the formation of MezAINHaby W irradiation (A = 210 nm) of the complex Me&l.NH3: (a) spectrum calculated for Me2AlNH, a t the RMPZ(fc)/ 6-31G* level of theory; (b) difference IR spectrum after irradiation of Me&. NH3 (decreasing absorptions point downward; increasing absorptions point upward). x indicates CH,, y indicates NH,, and z indicates H2 0 (impurity) [reproduced with permission from (58),p. 63721.
state molecules appear to be independent of the matrix material (provided that there is no direct reaction with the matrix), and the case histories of numerous molecules affirm that the same structures persist in the fluid phases. About the electronic and photochemical properties of such molecules there is much also to be gleaned through matrix isolation, so there can be little doubt, for example, that Fe(CO)*has a triplet electronic ground state or that Cr(COI5as initially generated by photolysis of Cr(CO)6is in a vibrationally or possibly electronically excited state. Some clear sign of the reactivity of the molecule may be found. In this regard, few experiments are more striking or more significant than those in which the visible absorption spectrum of Cr(CO)5is found to vary dramatically according to the nature of the matrix in which the molecule is trapped (60).Hence we
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
129
learn just what an extraordinarily powerful Lewis acid Cr(CO&is, for the unsaturated molecule, which has a square pyramidal (C4")structure, evidently interacts to a n appreciable extent through the vacant, sixth coordination site even with a matrix molecule such as CH4,Xe, or Ar that has few claims to be regarded as a normal ligand. Experiments with mixed matrices have shown that the token ligand is interchanged by irradiating selectively into the characteristic electronic absorption band a t lowest energy. Because the molecules are not free to rotate, it is possible to select those in certain orientations by the use of plane-polarized photolyzing radiation and so generate products in orientations that can be analyzed spectroscopically, also with the aid of polarized light. Hence, the interchange reactions are found to occur by expulsion of one token ligand, followed by a Berry pseudorotation, and uptake of another token ligand (Fig. 12). There is still much that we do not know about Cr(CO)5and its photochemistry, but these matrix studies have surely opened our eyes to the nature of this particular intermediate and the part it plays in the supposedly simple substitution reactions of Cr(CO), . However, matrix studies are limited ultimately by the low mobility of the caged species, allied to the narrow temperature range over which they can be held. The very immobilization that helps to preserve a highly labile molecule in a solid matrix becomes a n Achilles' heel when we wish to explore its reactions. The restraining effect of the matrix cage is much in evidence in the way photodissociation or photoisomerization reactions are immediately arrested and in the way bimolecular reactions induced by photolysis are normally restricted to neighboring pairs of molecules. Reaction channels that are important in the mobile fluid phases may thus be debarred, whereas other channels of minor importance in the fluid phases may well be favored. For example, photolysis of the recently identified chlorine oxide ClC10, (formally a dimer of C10) results in isomerization to ClOClO and ClOOCl when the molecules are confined to a matrix, but yields Clz and O2 as the only detectable products in the gas phase (62).We have already considered the scope that exists for recognizing and characterizing reactive intermediates in the gas phase. What options are there for extending our investigations to the liquid phase and to solutions?
D. SOLUTION STUDIES AT Low TEMPERATURES Various problems are inherent in the solution state when it comes to tracking reactive intermediates. First, the solvent may be a far
130
DOWNS AND GREENE
Xe
FIG.12. Schematic representation of the photochemical behavior of the Cr(CO)Sfragment in a mixed Ne/Xe matrix a t low temperatures. * represents the Cq. fragment in the excited 'E state [reproduced with permission from (60),p. 1361.
from innocent spectator; second, as with matrix samples, rotational transitions are not a feasible source of information; third, the spectroscopic features due to a solute molecule are appreciably broader than those of the same molecule in a solid, inert matrix, so that isotopic effects are much harder to evaluate; and last, most solvents absorb strongly in many regions of the electromagnetic spectrum, thereby obscuring our vision of the solute species in these regions. Nevertheless, the simple expedient of cooling a solution may be sufficient to extend the lifetime of an intermediate to the point at which it can be detected by conventional spectroscopic means. A good example is the q1-C2H3)(q2-C2H,)H (VII). This is vinyl hydride complex ( q5-C5H5)Ir(
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
13 1
VII
one of the intermediates formed by photolysis of the corresponding bidethenel complex ($-C5H5)Ir($-C,H,), with light having wavelengths >290 nm (63).The photoinsertion into a C-H bond of a coordinated ethene molecule can be engineered simply by photolyzing a frozen [2H81tolueneglass at 77 K, and the vinyl hydride can be recognized definitively by thawing the glass and measuring the 'H and I3C NMR spectra of the resulting solution at ca. 230 K, raising the temperature to 273 K results in slow decomposition, with reversion to the starting material. There is much to be gained from the use of a very weakly coordinating solvent such as a liquefied or even supercritical noble gas (64,651. The noble gases may each have a very narrow liquid range a t low pressure, but by working at pressures of 20 bars or more it is possible to span the entire range from 77 K to ambient temperatures with liquid argon, krypton, and xenon. I11 adapted though these liquids may be for the dissolution of polar compounds, they are capable of dissolving more or less freely many organic and organometallic materials as well as gases such as H2, N,, GO, and CHI, and analysis of the resulting solutions is hugely facilitated by the absence of solvent absorption in the infrared, visible, and normal ultraviolet regions, with the result that much longer pathlengths are admissible than for conventional solution studies. Accordingly, we can draw on a range of conventional spectroscopic methods to examine the solutions. Bandwidths may be appreciably greater than for matrix spectra, but problems of matrix site effects are eliminated and the bands are regular and symmetrical in shape. Moreover, the solutions give much flatter spectroscopic baselines than do matrices; allied to the advantage of the long pathlength, this increases considerably the capacity to detect weak bands. More fundamentally still, and in contrast with matrix studies, solution-based measurements provide the opportunity of exploring the kinetics and positions of equilibrium of reactions under thermal control. At the same time, we lose sight of some intermediates revealed by matrix isolation or other means that do not survive
132
DOWNS AND GREENE
in solution, at least long enough to be detected by normal means. Solution experiments may thus succeed in recording the fate of a labile, coordinatively saturated organometallic molecule such as CI-(CO)~(N~) but not of a coordinatively, unsaturated one such as CI$CO)~. Figure 13 shows the design of one such cell as used by Poliakoff and Turner and their group at Nottingham for experiments of this sort (64). The cell is cooled by liquid nitrogen, the flow of which is pulsed to vary the temperature. An important feature not illustrated in the figure is the presence within the cell of a microscale magnetic stirrer “flea”; hence, even solutions that are strongly absorbing in the ultraviolet or visible regions can be made to react photochemically because the portion of solution immediately adjacent to the irradiation window is constantly being renewed. For photochemically generated compounds that have lifetimes in the order of seconds, it may be
FIG.13. Schematic view of a variable temperature cold cell used for infrared studies of liquid noble-gas solutions. Cooling is achieved with a pulsed flow of liquid N2 (LN2) controlled by the output from one of the two thermocouples T so as to stabilize the temperature. The whole cell fits into a vacuum jacket (not illustrated), which is pumped through the tube V in the top flange of the cell. The solution under study can be passed through the cell from a room temperature reservoir via the two tubes marked In and Out [reproduced with permission from (811, p. 5551.
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
133
necessary to record the spectrum while the solution is being irradiated. Some idea of the compass of such studies may be gained from the case of the dihydrogen complex Cr(CO)S(q2-H,),which for all its 18electron configuration, is too labile to be isolated or spotted by ordinary methods of attack. The complex can be detected following photolysis of Cr(CO)6 or Cr(COI5(NH3)in an H2-doped matrix, but the matrix testimony has to rely principally on v ( C - 0 ) bands in the infrared spectrum; the broad v(H-H) feature near 3000 cm-' is some 20 to 30 times weaker than the least intense of the v(C-0) bands and, not surprisingly, is barely visible above irregularities in the baseline. These findings have been largely eclipsed, though, by the results of experiments involving photolysis of Cr(CO&dissolved in liquid xenon under a pressure of H2. Using these conditions, the Nottingham group has succeeded in detecting a weak, broad infrared absorption fulfilling all the requirements for assignment to the v(H-H) mode of the coordinated H2 moiety (66).The absorption comes a t 3030 cm-' when Hz is the ligand, shifting to 2241 cm-' when Hz gives way to Dz. On the other hand, similar experiments starting from the bis(ethene) complex ( q5-C5H5)Rh(q2-CzH,)z reveal the formation not of a dihydrogen complex but of the unstable dihydrido derivative (q5-C5H5)Rh($-CzHJ(H)z (67). Ultraviolet irradiation of a liquid xenon solution of Cr(CO)6without any other agent generates a new species that, with a half-life of about 2 s at 175 K, is sufficiently long-lived to be detected by conventional infrared measurements. The species has d C - 0 ) absorptions very close in frequency and intensity pattern to those assigned to the matrix-isolated species Cr(CO)5----Xe (64).Its identity can be checked, first by deliberately doping the solution with a range of likely impurities in order to rule out the possibility of its being some other compound of the type Cr(CO)5L (L = Nz, H20, etc.), and second, by carrying out experiments with Cr(CO)6 dissolved in pure liquid krypton and in liquid krypton doped with 5% xenon. Hence, the realization is dawning that the coordination of Cr(COI5 by xenon involves not a generalized solvation but a specific interaction with the formation of a Cr-Xe bond of appreciable strength. Similarly, the transient Ni(CO),(N,) has been identified following its generation by photolysis of Ni(CO), in liquid krypton doped with Nz a t 114 K. Ultraviolet irradiation establishes a steady-state concentration of Ni(CO),(N,); as soon as the photolysis source is switched off, the dinitrogen complex begins to decay, reacting thermally with CO to regenerate Ni(CO), . The kinetics of the decay reaction,
134
DOWNS AND GREENE
have been investigated in some detail over the temperature range 112-127 K (68). The rate shows a first-order dependence on the concentration of Ni(CO),(Nz)and a more complex dependence on the concentrations of CO and Nz. Hence, it appears that the reaction proceeds via two simultaneous paths, one dissociative and the other probably associative. An estimate of about 42 t- 4 k J mol-' for AH$ for the dissociative path gives a measure of the Ni-N2 bond dissociation enthalpy in Ni(CO),(NZ). In an extension of the technique we can now trade on the remarkable properties of supercritical fluids (64, 69). These combine a number of properties of gases and liquids; like gases they are highly compressible and are completely miscible with Hz or Nz, but like liquids they are relatively dense and dissolve a wide range of nonionic solids. In fact, most materials are considerably more soluble in supercritical xenon than in liquid xenon, despite the fact that the density of the supercritical material at the critical point is only about one-third that of liquid xenon. Moreover, the solubility of any compound in a supercritical fluid increases with increasing pressure at a constant temperature, with the result that the solvating power of the medium can be tuned by varying the pressure. It is the superior solubility of compounds that gives supercritical solvents a major advantage over their normal liquid counterparts and particularly in reactions involving gases such as Hz or Nz. Thus, for a given pressure of gas, the effective concentrations of Hz in a supercritical solution can be nearly an order of magnitude higher than in a liquid solution. As a result, it is altogether easier to detect, preserve, and even synthesize known or potential reaction intermediates. Previously unknown dihydrogen and dinitrogen complexes, which would normally decay rapidly at ambient temperatures, have thus been formed in supercritical xenon, where they are sufficiently robust at room temperature to be characterized quite normally by their infrared spectra (69, 70). Kinetic measurements between 283 and 353 K show that (r16-C6HSMe)Cr(CO)z(772-Hz) reacts with CO in accordance with Eq. (141, with an activation enthalpy of 70 t- 5 k J . mol-', which may correspond to the Cr-H2 bond dissociation enthalpy :
+
+
( 776-CGHsMe)Cr(CO)z(~-H2) CO -+ (q6-C6HsMe)Cr(CO), H2.
(14)
This conclusion finds support from more recent studies, also at high pressure but in n-heptane solution and using photoacoustic calorime-
135
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
try to measure the enthalpy change of Reaction (15) (711, M(CO)G + L + M(CO),L + CO
(15)
(M = Cr or Mo; L = n-heptane, H2or N2); the Cr-H2 and Cr-N2 bond enthalpies are estimated in this case to be 78 and 81 kJ . mol-', respectively. Figure 14 illustrates the region of the infrared spectrum associated with z4N-N) modes for a supercritical xenon solution of the rhenium carbonyl ( T ~ - C ~ H ~ ) R ~under ( C O )a~pressure of NZ (ca. 90 bars) and shows how this changes with time on photolysis with ultraviolet light (70). The spectrum bears clear witness to the sequential formation of C~H ) ,, ) R ~ ( C O )and ( N ~(q5-C5H5))~, the species ( T ~ - C ~ H ~ ) R ~ ( C O( T) ~ (- N Re(N2I3a t room temperature, and the last two turn out to be surprisingly stable under these conditions. In suitable circumstances, solutions in supercritical solvents are amenable even to NMR measurements (72). Supercritical fluids are thus affording a new dimension and are giving direct access to new chemistry involving reactive intermediates, as well as a new approach to problems of analysis, synthesis, and materials processing.
0.30
0.15
0 2160
2130
2100
2070 v/cm-'
FIG.14. IR spectra showing the sequential formation of ( v ~ - C ~ H ~ ) R ~ ( C O (dark )~(N~) peaks), (7)"-CSH5)Re(CO)(Nz)z (medium peaks), and ( T ~ - C ~ H ~ ) R (light ~ ( Npeaks ~ ) ~ ) from ($-C,H5)Re(CO)3 and Nz in supercritical xenon solution [reproduced with permission from Poliakoff, M.; Howdle, S. Chem. Br. 1995, 31, 1201.
136
DOWNS AND GREENE
IV. Experimental Characterization of Reaction Intermediates: Time-Resolved Methods
To retard or suppress chemical reactions it has been necessary to work under rather unusual conditions. It is only natural to want to know about the behavior of intermediates under more normal conditions. However, if the species are highly reactive, detection and monitoring of their fate demand very rapid spectroscopic measurements. From the more or less conventional type of spectroscopic analysis that has been the mainstay of our investigation so far we pass to methods that can be adapted to real-time measurements and in which spectral evolution or decay can be monitored at intervals sometimes as short as femtoseconds. It will soon be apparent?though, that the knowledge about the kinetic correlation between intermediates, reagents, and products thus gained can only be at the expense of detailed knowledge about the physical properties of the molecules-sometimes even their very identities. The function of the gas-phase, supersonic-jet, matrixisolation, and low-temperature solution techniques described in the preceding sections is in the provision of a “spectral library” of more or less well-authenticated reaction intermediates, whether real or model ones; this library is going to prove essential to the interpretation of many time-resolved experiments. The most popular procedure involves flash photolysis or, in the terms of laser technology, pump-and-probe experiments (73-76). Flash photolysis was developed by Norrish and Porter in Cambridge in the 1940s and earned them the Nobel Prize in 1967. The basis of the technique is simplicity itself. A pulse of light is used to generate the transient, typically in a fluid sample at or near ambient temperatures, and a spectroscopic record of the sample in absorption or emission is then kept as a function oftime. The time-scales of the reactions that can be studied are limited by the duration of the initiating light pulse. The flash lamps used by Norrish and Porter, which give pulse durations of several milliseconds, have now been largely superseded by high-powered lasers giving pulse durations of nanoseconds or less. By means of a process called mode-locking it is possible even to generate laser pulses lasting only a few picoseconds or less; the world record is currently reported to be 4 femtoseconds. Such lasers offer a means of tracking very fast kinetic phenomena, such as the isomerization of an electronically excited alkene, the rotational motion of a large molecule in solution, or some of the fast processes occurring in photosynthesis. Although these ultrafast processes (77) are not “reactions” in the sense that we have discussed so far, they are none-
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
137
theless important in the contributions they may make to overall reactions. An alternative technique for the rapid generation of transient species is that of pulse radiolysis triggered not by a light flash but by a short pulse (10-9-10-6 s) of high-energy electrons (1-5 MeV); this approach has been employed primarily for studies of liquid phase kinetics and especially of the solvated electron, but it is a very flexible method with applications also for gas phase studies (78). Representative of this method of attack is the formation of the peroxynitrate anion 02NOO- less than 2 ms after pulse irradiation of aerated aqueous solutions containing nitrate and formate a t relatively low concentrations (79). The anion, which can be detected and monitored by its UV absorption near 290 nm, is a strong oxidizing agent notable for its potential mediation in some of the reactions that make living systems vulnerable to nitrogen dioxide, one of the most toxic components of polluted atmospheres. It is of some consequence therefore that studies of several oxidation reactions should reveal two distinct pathways by which the peroxynitrate decays, one proceeding through the parent acid OzNOOH and the other probably evolving through secondary reactive intermediates derived from the decomposition of the peroxynitrate. Once the transient species has been formed, it has to be monitored by some form of kinetic spectroscopy, typically with ultraviolet-visible absorption or emission, infrared (time-resolved infrared or TRIR) (74), or resonance Raman (time-resolved resonance Raman or TR3) (80) methods of detection. The transient is usually tracked by a probe beam a t a single characteristic frequency, thereby giving direct access to the kinetic dimension. Spectra can then be built up point by point, if necessary, with a n appropriate change of probe frequency for each point, although improvements in the sensitivity of multichannel detectors may be expected to lead increasingly to the replacement of the laborious point-by-point method by full two-dimensional methods of spectroscopic assay (that is, with both spectral and kinetic dimensions). Ultraviolet-visible absorption has traditionally been the basis of detection in flash photolysis experiments. It offers a number of advantages in sensitivity and efficiency and has certainly delivered much vital information about the reactivity of transient species. On the other hand, the ultraviolet-visible absorption bands characteristic of any but the simplest molecules tend to be broad and relatively uninformative as regards identity and structure, and so we may run into problems not only with the overlap of absorptions due to different
138
DOWNS AND GREENE
species, but also with deciding on the nature of the carrier of a particular absorption spectrum. Altogether more discriminating, albeit less sensitive, are the TRIR methods that have been successfully developed in the past 15 years or so (74, 81). With the sort of apparatus depicted schematically in Fig. 15, it has been possible to draw on the spectral library built up from matrix isolation and other experiments in the “slow” regime to identify transients formed in the gas phase or in solution at ambient temperatures, as well as investigating their kinetic behavior. Just what can be achieved in this way will be demonstrated by some examples representative of recent studies. As a start, we return to that remarkable Lewis acid Cr(COI5.Flash photolysis experiments have been carried out with Cr(CO)6either in solution or in the gas phase and with either infrared or visible probes (9,601. For example, photolysis of gaseous Cr(CO)6with the output of an XeF laser yields a predominant transient that can be tracked by a IR spectrometer
n
HeNe alignment laser
Sample cell
Oscilloscope Pulse generator
\
0.-
Micrmmputer
Detector
Transient digitiser
FIG.15. Schematic representation of the time-resolved infrared (TRIR) flash photolysis apparatus used a t Nottingham. The UV pulse laser generates transient species; the continuous IR laser monitors the change in transmission a t a particular IR frequency, producing a trace showing the IR absorbance as a function of time. The experiment is repeated a t different IR frequencies so that a complete IR spectrum of the transient can he built up [reproduced with permission from (97),p. 1031.
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
139
line-tunable, liquid-nitrogen-cooled CO laser, changes in intensity of the transmitted infrared light being recorded by a high-speed InSb detector. The v(C-0) bands clearly imply that the species is “naked” Cr(CO), (freed a t last from any token ligand) with the same square pyramidal geometry as the molecule identified previously by matrix isolation (see Section 111,C).The identification is confirmed by the decay of the transient a t a rate matching the rate of reappearance of Cr(CO)6.The rate constant for the reaction of Cr(CO), with CO a t 293 K comes out to be 1.5 t 0.3 X 10’”dm3-mol I . 5 - l . Similar studies of the reaction of Fe(C0I4with CO give a rate constant about two orders of magnitude smaller (56).This difference can be rationalized, a t least in part, with the realization that Fe(CO), has a triplet ground state whereas Cr(CO), almost certainly has a singlet ground state. Closer examination shows that Cr(CO), is generated within 1 picosecond of light absorption and that it is formed in vibrationally excited states that relax over tens if not hundreds of picoseconds (60).These details emerge not only from time-resolved ultraviolet-visible studies but also from analogous TR3 experiments with cyclohexane solutions in which two 5-ps pulses a t A = 266 nm are employed; the first dissociates Cr(CO)6,and the second probes the resonance Raman spectrum of the photoproduct Cr(CO),. Hence, it appears that the naked intermediate, albeit vibrationally excited, can exist for a few picoseconds before taking up a solvent molecule. Extension of these studies to supercritical fluids provides a way of making TRIR measurements on intermediates incorporating a bond between a metal center and a weakly coordinating ligand such as a noble gas or CO,. Hence, for example, a series of organometallic noble-gas compounds, W(CO),(Ar), M(CO),(Kr), and M(CO),(Xe) (M = Cr, Mo, or W), have been detected transiently in fluid solution a t room temperature (82).The second-order rate constants for the reaction of M(CO15Lwith CO show that the reactivity for a given metal M varies in the order L = Kr > Xe = C 0 2 , whereas the reactivity for a given ligand L varies in the order M = Cr = Mo > W. The reaction of W(CO),(Xe) with CO has an enthalpy of activation of 34 k J . mol-’, representing a lower limit for the W-Xe bond enthalpy. The bond is therefore twice as strong as a conventional hydrogen bond, and its mechanistic impact cannot be ignored. The photochemistry of the binuclear iron carbonyl [(q5-C5H5)Fe( C O ) z ] 2VIII, , provides another outstanding example of the interplay between time-resolved and alternative strategies (76). According to the infrared spectrum of this compound in a suitable matrix a t low temperatures (83), ultraviolet photolysis results in the elimination of free CO, but the pronounced cage effect of the medium entirely sup-
140
DOWNS AND GREENE
presses the separation of (q5-C5H5)Fe( C0lz radicals, X. Experiments with a l3C0-labeled sample demonstrate that the iron-containing product formed by CO loss has a binuclear framework incorporating three CO bridges, M: 0
IX
VIII
X
Pulsed laser photolysis (at A = 308 nm) of VIII in an alkane solution at room temperature can be shown by TRIR measurements to generate not one but two metal carbonyl species with quite different kinetic properties (84).One transient displaying two infrared bands characteristic of the u(C-0) modes of terminal CO groups decays rapidly by second-order kinetics (with kz = 5 X lo9 dm3.mo1-l. s-') to regenerate the precursor within about 50 ps. The infrared spectrum of this species, allied to its other properties, makes clear that it is the radical transient is the triply CO-bridged ( Q ~ - C ~ H ~ ) F ~ X. ( C The O ) ~ second , complex M,easily recognizable by the close resemblance its infrared spectrum bears to that of the matrix-isolated species. In the absence of added CO, this has a half-life of about 0.6 s-that is, some four orders of magnitude longer than that of the radical. IX reacts with CO, and its decay as well as the regeneration of VIII can also be monitored by TRIR measurements. Similar experiments with another ~ ( CLO=) CH3CN ~L substrate L to form the product ( Q ~ - C ~ H ~ ) ~ F ~[e.g., or P(OMe)3, see Fig. 161 reveal the participation of both intermediates M and X through parallel reaction paths. The reactivity of the radical X may thus be gauged by rate constants close to the diffusioncontrolled limit. For another string to our spectroscopic bow we may also turn to Raman scattering, using a single visible laser to fulfil the dual functions of pump and probe, and detecting the emitted light by a suitable optical multichannel analyser. Cyclohexane solutions of [( Q ~ - C ~ H ~ ) F ~ and ( C Oits) ~permethylated ]~ analog [ ( Q ~ - C ~ M ~ ~ ) F ~ ( C O ) ~ I show resonance Raman behaviors correlating with the lower energy
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
141
U
C
0
n
a
el
1
1950
L
I
1850
I
1750
WO"e""lllber/Clll-'
FIG. 16. Time-resolved infrared spectrum obtained after UV flash photolysis of I($CSH5)Fe(CO)211,VIII (A), and MeCN in cyclohexane solution at 25°C. The bands ar e labeled thus: B, ($-C5H5)Fe(CO)2,X C, ($-C5H5),Fe,(p-CO)n,M;and D (r)"-CSH&FeL(CO).I(MeCN).The first three spectra correspond to the duration of the firing of the UV flash lamp, and subsequent spectra are shown a t intervals of 10 p s . The negative peaks in the first spectrum are due to material destroyed by t h e flash; these have been omitted from the subsequent traces to avoid undue confusion [reproduced with permission from Dixon, A. J.; Healy, M. A,; Poliakoff, M.; Turner, J. J. J. Chem. Soc., Chem.. Commun. 1986, 9941.
electronic absorption bands, thereby shedding light on the molecular orbitals involved in the electronic transitions (85). By the simple device of increasing the incident laser power it is possible to observe additional rich resonance Raman spectra that must originate in the binuclear transients (775-CSRS)zFez(~-CO)~, where R = H (MIor Me.
142
DOWNS AND GREENE
With its different requirements, the Raman effect gives access to vibrational transitions that may well be lost to view in infrared absorption. Time-resolved Raman spectroscopy, both spontaneous and coherent, has developed apace in the past two decades (86). As a means of pursuing short-lived intermediates, spontaneous resonance Raman scattering has proved particularly instructive, partly because of its high selectivity, partly because of its enhanced sensitivity. Since studies of this sort were initiated in 1976, the time resolution has improved from microseconds to subpicoseconds. The method has really come into its own, notably in the hands of Kitagawa and his colleagues at Okazaki (87), in the investigation of heme proteins and of how they bind dioxygen and other ligands. For example, the reaction of dioxygen with the reduced enzyme cytochrome c oxidase (CcO), formed by CO photodissociation in aqueous solution, has been followed by TR3 measurements (87), which reveal within 10 ms the growth and decay of several intermediates identifiable by emissions due to 0-0 or Fe-0 stretching vibrations in the region 300-850 cm-*. The origins of the bands have been traced by experiments with "0enriched samples of 02,and hence it can be shown that the O2 binds initially to the Feassite in CcO in an end-on fashion with an Fe-0-0 bond angle close to 120". Studies with picosecond time resolution witness within 5 ps of photolysis of the CO-bound precursor an intermediate believed t o be the high-spin, five-coordinate heme Fe:; to which a histidine (His) is ligated; this is indicated by the growth of a band at 220 cm-' attributable to the Fe-His vibration (88). Intense interest in how carbon-hydrogen bond activation may be engineered has led to detailed studies of several transition-metal compounds known to compass reactions of this type, which are sometimes initiated thermally, but more often photochemically. Two such compounds are the iron and ruthenium dihydrides M(dmpe)zHz(M = Fe or Ru; dmpe = Me2PCH2CH2PMe2), which react with alkanes RH under irradiation in accordance with Eq. (17) (61, 76):
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
143
The ruthenium compound gives an object lesson in the merits of combining matrix with time-resolved solution studies. Photochemical reductive elimination of Hz may be followed by the decay of the characteristic v(Ru-H) bands in the infrared spectrum or by the growth of three intense visible bands at ca. 740, 540, and 460 nm. Although the product lacks characteristic infrared features, the circumstances of its formation and its response to matrix dopants imply that it is the 16-electron species Ru(dmpe):,, presumably with a square planar skeleton. Flash photolysis of a cyclohexane solution of the dihydride with a pulsed laser yields a transient with a visible spectrum remarkably like that of the matrix photoproduct. The transient signal decays by second-order kinetics over ca. 80 p s as the precursor is regenerated. Addition of H2 even in very low concentration increases the rate of decay, causing it to become pseudo-first order; under these conditions the reaction of the intermediate with H2 is essentially diffusion controlled, being opposed by a minimal activation barrier. The corresponding iron compound Fe(dmpe)2H, cannot be vaporized without decomposition and so does not lend itself to matrix isolation, but flash photolysis of a n alkane solution at room temperature g v e s a transient Fe(dmpeIz differing conspicuously from its ruthenium counterpart. For one thing, its electronic absorption spectrum shows but a single band a t low energy, and that not in the visible but in the nearultraviolet region. The reactivities are quite different too: both intermediates add substrates to form products of the type M(dmpeIzL, where L = CO, C2H4,or PMe3, as well as reacting with Hz, certain hydrocarbons, or EtySiH to give insertion products of the type M(dmpe12(X)(H)(X = H, organic group, or Eta&), with the kinetic results summarized in Fig. 17. The reactions with Hz, EtsSiH and alkenes are slower by at least two orders of magnitude for Fe(dmpeIz than for Ru(dmpeIz,whereas alkanes and arenes find Fe(drnpe&much the more reactive of the two. One plausible explanation of the spectral and kinetic differences is that the intermediates have different structures, with Fe(dmpe12adopting a “butterfly” configuration unlike the square planar form of its ruthenium analog, although we lack definitive evidence on this point. The reduced rate of the back-reaction with H2 to regenerate the dihydride precursor must be one of the key factors making Fe(dmpe)z more reactive than Ru(dmpeIz toward C-H insertion reactions, particularly with alkanes and arenes. Similar experiments involving photolysis of the ruthenium hydride place Ru(PP3)Hz,in which the two bidentate dmpe ligands have given place to the tetradentate phosphine PPy= P(CHzCH2PPh2),, reveal the
144
DOWNS AND GREENE
'0
--..co
..HZ
-co
-Et&H
2cyclopantene 6. 109 k2 4 -
2 -
-benzene. max -cyclopentene -EtjSIH H2
-pentane -cyclohexane
0
FIG. 17. Schematic comparison of the intermediates Ru(dmpe)a and Fe(dmpe):, contrasting their UV-visible spectra and reactivities; the (dmpe = MeaPCH2CHaPMe2) latter are expressed as log k a where k a is the second-order rate constant for reaction with the substrate a t room temperature deduced by laser flash photolysis [reproduced with permission from (611, p. 3641.
transient intermediate Ru(PP3) (89). Here the PP3 ligand obliges the molecule to adopt not a square planar RuP4 skeleton but a pyramidal (XI) or butterfly one (XII), and it is noteworthy that the UV-visible
XI
XI1
spectrum resembles that of Fe(dmpeI2with but a single absorption a t 395 nm. In its reactivity, moreover, Ru(PP3) is quite different from R ~ ( d m p e )being ~, much slower, for example, to react with H2. The
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
145
corresponding osmium intermediate, which has a similar structure on the evidence of its W-visible spectrum, does not even react with H2 but inserts into the C-H bonds of alkanes with rate constants in the order of lo5 dm” mol-’. s at 295 K. Hence, it appears that the enforced nonplanar geometry effectively prepares Ru(PP3),and more especially Os(PP3),for C-H activation (89). In the first application of ultrafast TRIR spectroscopy (that is, with time-scales of to IO-’s), reductive elimination and oxidative addition of H2 have been traced following flash photolysis of the related ruthenium carbonyl dihydride complex Ru(PPh,),(CO)H, in benzene solution (90). This precursor is of particular note because it is known to catalyze insertion of alkenes into C-H bonds a t the unsaturated carbon center of alkenes or arenes in a position p to a carbonyl group. The course of events has been monitored from excitation with an ultrafast UV laser (giving pulses a t A = 304 nm with a n energy of 1-2 pJ, a full width of 4 ps a t half of maximum duration, and a repetition rate of 1.05 kHz) to 2000 ps after the laser pulse using the v(C-0) mode as a reporter, IR absorption measurements being made with a CO laser or with a diode laser via a detection system employing “up-conversion.” Photolysis results in reductive elimination of H2 and the formation of an unsaturated transient, Ru(PPh3)&CO), which can be identified unequivocally on the basis of its rapid buildup, the kinetics of its reaction with H2, the partial regeneration of starting material, and the frequency of its v(C-0) band. Moreover, the photoelimination step-involving H-H bond formation, Ru-H bond cleavage, and any reorganization of the coordination geometry at ruthenium-is found to be complete within 6 ps. Nor has the speed limit yet been reached, for there are even ways of clocking chemical events a t the femtosecond level. For example, Ahmed Zewail and his colleagues a t the California Institute of Technology have developed a novel approach relying on ultrafast lasers giving femtosecond pulses to study in real time the transition-state dynamics of charge-transfer and other reactions (91, 92). The entire molecular system is prepared on a reactive potential energy surface and in a well-defined impact geometry. A femtosecond light pulse induces a charge-transfer transition, and the dynamics of the ensuing events are then followed by interrogating the transition state or the reaction products; this is done with probe femtosecond light pulses causing photoionization and mass spectrometric sensing of the resulting fragments. The method is well illustrated by the dissociative charge-transfer reaction of the benzene-iodine complex (91 ):
146
DOWNS AND GREENE
rcnc tniit
transition state
products
The transition state is directly accessed by excitation into the charge-transfer absorption band of the complex formed initially by expanding a gas mixture containing benzene, iodine, and helium through a pulsed valve. The resulting molecular beam is skimmed and intersected by the femtosecond laser pulses in the acceleration region of a linear time-of-flight mass spectrometer. The pulses of the femtosecond pump-laser (at A = 275 nm) initiate the reaction and define t = 0. The probe pulses (at A = 304.7 nm) then detect the free iodine atoms through a resonantly enhanced multiphoton ionization (REMPI) process. Transient I atom growth is measured by monitoring the It signal as a function of the pump-probe delay. In addition, by using linearly polarized pump pulses and having provision for varying the orientation of the plane of polarization with respect to the axis of the mass spectrometer, it is possible to analyze the velocity distributions of the I atoms. Hence, microscopic elucidation of the reaction dynamics and mechanism becomes a practical proposition. We are able to witness the evolution and decay of the transition state and to show that dissociation with the formation of I atoms occurs via one of two exit channels. One (ionic) follows the ionic potential of the charge-transfer state CsH& to produce c6Hg.1- + I; the other (neutral) involves intermolecular electron transfer through coupling of the transition state to neutral, locally excited, iodine-repulsive states, and in this case the products are C6Hs.I + I. With a timescale of 250 fs, the neutral process is ultrafast, outstripping the ionic process, which proceeds at a leisurely 800 fs or so. The kinetic energy characteristics of the I atoms even go so far as to reveal a near-axial geometry for the transition state. A similar approach has been adopted to show that the two-center elimination reaction (19) is a two-step, nonconcerted process evolving via the intermediate CzFJ (92):
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
147
Here the time-scales for the breaking of the two C-I bonds differ by two orders of magnitude: the primary step takes about 200 fs, whereas the secondary step occupies 25 ps. The speed of the first step reflects the repulsive force in the C-I bond caused by the promotion of a nonbonding electron from the highest occupied molecular orbital (HOMO) to the a * lowest unoccupied molecular orbital (LUMO), whereas the second elimination is governed by the energy redistribution attending the concerted fission of a C-I a-bond and making of a C-C .rr-bond. The high-resolution spectroscopic methods involving, for example, infrared diode lasers that have been successfully applied to the detailing of gaseous transients have also been adapted to the time domain (10). High temporal resolution cannot of course be reconciled with high spectral resolution, but measurements made a t even millisecond or microsecond intervals can still provide invaluable information about many reaction systems, with the bonus of unambiguous identification of intermediate species made possible by the high spectral definition. Reactions have been induced typically by irradiation with a high-power source, such as a n excimer or CO, laser, or on occasion by pulse radiolysis, and the subsequent course of events has been monitored by a suitable high-resolution spectroscopic probe. For example, Curl’s group a t Rice University in Houston, Texas has investigated the reaction between the radicals NH, and NO, which appears to proceed via more than one channel (93):
NH,
+ NO
P
N, t H,O
(a)
H + N,+OH
(b)
HN,(?) + OH
(c).
(20)
This is of considerable interest because of the key role of the reaction in the thermal “de-NO,” process involving the catalyzed reaction of ammonia with nitrogen oxides; hence, in principle, the gaseous emissions from industrial plants and power stations can be purged of the
148
DOWNS AND GREENE
noxious oxides, which are known to be major sources of smog and acid rain. Reaction (20) can be initiated by excimer laser photolysis of ammonia in the presence of NO and monitored by comparing the intensity of the infrared absorption due to NH3 with that due to NH, as functions of time, and also with those due to OH and H,O. Hence, the branching ratio of the OH channel or channels has been determined at different temperatures, with results suggesting that an additional channel may be implicated. Replacing the probe laser of these studies with electron pulses offers a means of imaging structural changes by electron diffraction, but observation of chemical changes in this way is nontrivial, and only recently have the odds against the experiment been significantly shortened (94-96). The essential experimental principles of “stroboscopic” electron diffraction were first devised more than a decade ago, with a pulsed electron beam being generated by photoemission with millisecond to picosecond time resolution. Major advances have come with the application of online data-recording techniques, involving either photodiode array or charge-coupled device (CCD) detection. Quite apart from the low intensity of coherent molecular scattering compared with incoherent background scattering, however, numerous problems remain to be solved. How, for example, is one to determine in situ the zero of time in a chemical change? And how is one to take account of molecular ensembles in nonequilibrium (that is, non-Boltzmann) vibrational distributions? In fact, a temporal resolution of 15 ns has been achieved with a pulsed laser-driven electron source. Hence, the structural and vibrational kinetics of CS have been explored during the first 120 ns following the photodissociation of CS, at A = 193 nm (94):
The observed changes of vibrational population with time have been interpreted on the basis of inelastic collisions of CS with sulfur atoms in the electronically excited ‘D state. Further hazards await any attempts to probe picosecond or subpicosecond changes by electron diffraction: space-charge effects broaden the electron pulse-width, and temporal overlap of the initiating photon pulse and the probe electron pulse must somehow be established. For ultrafast studies it is necessary then (i> to measure the electron pulse-width, (ii) to have the means of accurate clocking of the reaction, and (iii) to be able to detect single electrons (so as to reduce the electron flux and minimize space-
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
149
charge broadening). Zewail and his group have risen to this challenge by developing the new apparatus shown in Fig. 18 (96).This consists of a femtosecond laser, an ultrafast electron gun, a free-jet expansion source, and a newly designed two-dimensional single-electron detec-
Thermoelectric
I Fibre-optic raper
FIG. 18. ( a ) Ultrafast electron diffraction apparatus consisting of a n electron gun chamber, a diffraction chamber, and a detector chamber. Two fs laser pulses are used, one to initiate the chemical change and the second to generate the electron pulse. (b) Detector system: incident electrons either directly bombard a small CCD or strike a phosphor-coated fused fiber-optic window. Light emitted from the phosphor is amplified by an image intensifier and brought to a scientific-grade CCD. Both CCDs are thermop. 1601. electrically cooled [reproduced with permission from (96),
150
DOWNS AND GREENE
I
0
1
3
2
r
4
5
(A)
FIG.19. Radial distribution functions f ( r )and Af(r) derived from the molecular scattering curves for gaseous CHzIza t different delay times between the photodissociating fs laser pulse (at A = 310 nm) and the ps electron pulse (pulse-width 15 ps). The corresponding theoretical f ( r ) curve for CH212is supeorimposedon the -20-ps data set. The changes observed are a t r = ca. 2.0 and 3.5 A corresponding to the C-I and I----I internuclear spacings, respectively [reproduced with permission from (96), p. 1611.
tion system. Femtosecond laser pulses are created from a collidingpulse mode-locked ring dye laser, the output from which is directed through a four-stage pulsed dye amplifier with no pulse compression (620 nm, 2-3 mJ, 30 Hz, 300 fs). To initiate the reaction, 95% of this beam is doubled (310 nm, ca. 250 pJ);the remainder (also doubled) is focused on a back-illuminated, negatively biased photocathode to generate the electron pulses. Critical to the success of the experiment is the detector, a two-dimensional CCD operating in a direct electron bombardment mode. Hence, a temporal resolution in the picosecond range can be realised. The first such studies have monitored the photodissociation of CHJz a t A = 310 nm; the molecular scattering and radial distribution functions vary significantly with time (see Fig. 19, for example) in a manner wholly consistent with Reaction (22): CH&&
CHzI -t I.
(22)
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
151
A more detailed quantitative analysis may be expected to elucidate the structure of the CHJ intermediate. Modulation and difference detection techniques should provide a way of suppressing the background scattering from unreactive species, thereby enhancing the precision with which structure changes can be evaluated. To understand the whole of a photochemical pathway requires a knowledge of the excited electronic state or states (75, 97). In addition, some thermally activated processes proceed through the formation of intermediates or of products in excited states (witness, for example, the formation of singlet dioxygen). However, inorganic and organometallic molecules in these states are generally far too shortlived to be characterized directly by conventional spectroscopic observations, and there is no physical device open to us that will significantly extend their lifetimes. Figure 20 illustrates schematically a simple system involving just the ground and one excited state, and
f
Exated sfafe
I
2.
P
c
w
Resonance
Raman
Qk (distortioncoordinate)
FIG.20. Schematic representation of the potential energy curves for the ground and an excited electronic state of a molecule and some of the spectroscopic connections that can be made between them [reproduced with permission from (7.9, p. 1191.
152
DOWNS AND GREENE
some of the spectroscopic connections that can be made between them. In principle, the optical absorption and excitation spectra provide a means of probing the vibrational levels of the excited state, whereas the emission spectrum reports not only on the vibrational levels of the ground state but also on the lifetime of the excited state. The intensities of the vibrational components of these spectra contain the seeds of information about the distortion the molecule experiences with the switch from the ground to the excited state. Unfortunately, though, relatively few molecules, and then at low temperatures, afford the sort of fine structure needed to make this a realistic proposition. The metallocene intermediates (q5-C5H5)zM (M = W or Re) provide two cases in point, well-resolved vibrational progressions in the optical spectra of the matrix-isolated species admitting meaningful estimates of the changes in the metal-to-ring-centroid distance accompanying the electronic transition at lowest energy (61, 76).Resonance Raman spectra also give access to similar structural information. The case of the binuclear iron carbonyls (q5-C5R5)2Fez(C0)4 (R = H or Me) has already been alluded to (85), and similar studies have shown that the W-N and trans C-0 bonds of W(CO)5(pyridine)are extended by 18 and 12 pm, respectively, on excitation to the photoactive state (75). However, none of these methods involves probing of the excited state in real time, with the advantages this would bring to its structural and temporal delineation. One way of observing directly the vibrational spectrum of a molecule in an excited state is to engage TR3 measurements (801, as with W(CO),(diimine) complexes (diimine = 2,2'-bipyridine or related species) in the tungsten-todiimine charge-transfer excited state at lowest energy (98). The totally symmetric cis-carbonyl v ( C - 0 ) mode has a frequency about 50 cm-' higher than it does in the ground state, implying some reduction of metal-carbonyl a-back-bonding in the excited state. For all its merits, however, this approach is somewhat limited in its application, and there are real attractions in being able to record the infrared spectrum of a molecule in an excited electronic state. The new developments in TRIR spectroscopy now make this possible, at least in suitable cases. Thus, fast TRIR measurements are able to detect and monitor the lower metal-to-ligand charge-transfer state (MLCT where L = 4-cyanopyridine) of the tungsten pentacarbonyl complex W(C0)5(4-cyanopyridine),XI11 (99). Excitation causes the v(C-0) infrared bands to shift to higher frequency, confirming that the metal center is oxidized in the excited state. This state can be observed to decay, partly reverting to the ground state and partly with dissocia-
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
153
0
XI11
tion to form solvated W(CO)5and 4-cyanopyridine. The second reaction channel depends on rapid thermal equilibration between the MLCT and a second excited state, which is metal-centered or ligand field (LF) in nature, and it is this state that is dissociative with respect to separation of the 4-cyanopyridine ligand. confirmation of this comes from the finding that the yield of W(CO), varies with temperature, giving an estimate of some 4000 cm-' for the energy gap separating the lowest MLCT and LF states (see Fig. 21). This example illustrates the sort of subtle details that time-resolved studies are now able to chart with regard to excited states and the chemical processes they beget. About reactions occurring in the solid state there is not generally the same urgency that distinguishes so much of the chemistry of the more mobile phases. There may not be the same call for ultrafast LF
FIG.21. Scheme showing the low-energy states of the photosystem W(C0),(4-CNpyr) (4-CNpyr = 4-cyanopyridine).k , and k2 represent the sums of the radiative and nonradiative rate constants for the decay of the levels 1 and 2, respectively; kd is the rate constant for photodissociation to W(CO), and 4-CNpyr [reproduced with permission from (99), p. 35461.
154
DOWNS AND CREENE
methods of detection and analysis, but the ability to follow the evolution of a solid-state transformation is no less desirable for a proper kinetic and mechanistic understanding of the change. One such change currently attracting much interest is intercalation whereby mobile guest species enter a crystalline host lattice that contains an interconnected system of empty lattice sites (100). The staging of intercalation in Ago,,,TiS2and Hg,TiS2 has been effectively witnessed, for example, as a function of time by high-resolution transmission electron microscopy. There has been developed, in addition, a reaction cell that enables intercalation reactions to be monitored in situ using real-time X-ray diffraction techniques. Exploiting the high flux, white X-ray radiation from a suitable synchrotron radiation source and energy-dispersive diffraction techniques permits a large energy window of the powder X-ray diffraction spectrum to be recorded simultaneously at a single fixed detector angle, with acquisition times in the order of seconds. Figure 22 illustrates strikingly (100) the progressive intercalation of the electron-rich metallocene (T~’-C,H,),CO into a tin disulfide host. Still more dramatic is the very recent tour de force in protein science in which a group led by Michael Wulff and Keith Moffat at Chicago has virtually “watched” a protein function (101). Working at the European Synchrotron Radiation Facility, this team has devised a way of collecting pulsed Laue X-ray diffraction patterns with nanosecond time resolution. Hence, it has been possible to trace the structural changes that accompany the process of heme
FIG.22. A stack plot showing the evolution of the energy dispersive X-ray diffraction spectrum of SnS2 following the injection of (q5-C,H,),Co. Each spectrum took 10 s to record. The signals a t 25 and 29 keV arise from resonances from the tin K., and & core electrons and have been used to normalize the other signals [reproduced with permission from (IOO),p. 1831.
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
155
and protein relaxation brought about first by photodissociation and then by rebinding of the carbon monoxide complex of myoglobin.
V. Experimental Characterization of Reaction Intermediates: Flow and Other Methods
The basis of flow methods has been outlined earlier. We have to find a way of generating our intermediate continuously in a flowing system-sometimes by mixing gases or solutions, sometimes by the action of a discharge-in a way that can be matched to its kinetic properties, the flow rate, and the time-scale of the spectroscopic method of detection. Even here it may be necessary to keep to a minimum the time taken to record the spectrum. In the past much of the work with flowing gases relied on photographic recording of electronic emission or absorption spectra, but only for the smallest molecules are these likely to give detailed information about the vibrational and other properties of any intermediates. By contrast, current laser techniques allow much more subtle interrogation of transient molecules, both small and large, in flow systems, leading sometimes to the identification and precise characterization of these molecules ( l o ) ,as well as reporting on the kinetics of the reactions that feed upon them. The sort of apparatus that has been used for the study of a gaseous reaction initiated by a discharge is shown schematically in Fig. 23 (73).The time between initiation and reactant or product detection at some point downstream in the flow tube can be calculated if the velocity of the gas mixture is known. Detection and monitoring of the relevant species can be effected by various spectroscopic techniques, usually mass spectrometry, resonance fluorescence, laser-induced fluorescence, ESR, or laser magnetic resonance; high-resolution techniques such as microwave and infrared diode laser spectroscopy are also compatible with the analysis of reactive gas flows, although they lack the sensitivity of the more widely used methods which are better suited to kinetic studies. Resonance fluorescence is a well-tried method for tracking the concentrations of atomic transients (e.g., H, N, 0, F, C1, or Br), whereas laser-induced fluorescence (LIF) offers greater sensitivity and applicability, especially for molecular transients (e.g., OH, CN, or CH90), at the cost of greater experimental complexity. Laser magnetic resonance (LMR) employs a CO, laser to produce a variety of sharp rovibrational lines to one Qf which a rovibrational line of a radical may often be tuned by the application of
156
DOWNS AND GREENE
To
r able ctor
Excimer laser I
Trigger pulse
FIG.23. Discharge flow apparatus used to study reactions of OH radicals with detection by laser-induced fluorescence. The OH radicals are generated in situ by the reaction between H atoms (themselves formed by the action of a microwave discharge on a n H2/Hemixture) and NOz. The distance between the points of initiation and detection is vaned by moving the inlet tube [reproduced with permission from (73),p. 301.
a magnetic field. The biradical methylene is generated in its triplet electronic ground state by the reaction of oxygen atoms with ketene:
0 + CHz=C=O + 3CH, + O=C=O.
(23)
At slightly higher energies there exists a singlet state. Whereas singlet methylene can be detected by LIF, only the triplet state responds to LMR. Hence, it will be evident that the detection techniques are often complementary, and that no one technique is ideal for all transients. The reaction of 3CHzwith ethene,
3CHz+ CHz=CHz+ CH3CH=CHz,
(24)
157
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
has been studied by monitoring the LMR signal as a function of the distance between the mixing point and the laser detector with a range of different ethene concentrations; on this basis the relevant kinetic parameters have been evaluated (731. Flow methods have found frequent application in the direct study of C10 radicals and related species noteworthy as sources and reservoirs of the halogens in the atmosphere (17).In a typical experiment an inert carrier gas is pumped at pressures near 1-10 torr down a glass or metal flow tube; the first reactant may be added at a point upstream to the carrier flow, and the second reactant is added subsequently. The evolution of concentration with distance after mixing is then followed. Hence, for example, the reactions of C10 generated in a discharge-flow system have been explored, and mass spectrometric measurements have been used to determine the overall rate constant for its decay by the second-order process [Eq. (25a)l (102);resonance fluorescence associated with C1 atoms indicates that Reaction (25b) is the major pathway by which C10 decays. Mass spectrometric detection has also been used to determine the relative importance of the two channels [Eq. (25c)l followed by the reaction between C10 and the NO3 radical (103), whereas ESR detection of the C10 provides a means of exploring the kinetics of the reaction with HN03, which probably proceeds according to Eq. (25d), and so allows this to be discounted as having a significant role in the stratosphere (104).
+ C10 + C10 C10 + NO, C10 + HNO,
C10 C10
+=---
all products c1 + ClOO ClOO + NO2 OClO + NO2 HCl + NO2 + 0,. (d)
<
(25)
J
Flow methods are not without their limitations (73). In practice they are usually limited to reactions with time-scales no shorter than the millisecond range. Reactions involving more than one reagent pose particular problems through the time taken for efficient mixing to occur-in the order of a fraction of a millisecond at a gas pressure of about 1 torr. A further restriction on kinetic studies is the need to retain a uniform flow velocity along the entire cross section of the tube. In fact, these conditions of “plug flow’’ can be attained only at low pressures, so flow experiments are generally restricted to gas pressures below 10 torr. There are also the seeds of potential compli-
158
DOWNS AND GREENE
cations in unwanted heterogenous reactions occurring at the walls of the flow tube that may catalyze the reaction under investigation, promote secondary reactions, or scavenge the reaction intermediates. Most flow studies have been directed toward the kinetic elaboration of the reactions of intermediates that have been characterized previously by independent means. However, the flow method has also been exploited rather differently to help elicit detailed structural and spectroscopic properties of numerous molecules and molecular ions having but a fleeting existence under normal conditions. Hirota and his group at the Institute for Molecular Science at Okazaki have worked wonders with a flow reactor, using discharge, thermal, or photolysis methods to generate the transients and typically surveying the rotational or rovibrational spectrum of the flowing vapor at high resolution by a microwave or tunable infrared diode laser probe (10).The methods of production and sampling depend on the lifetime of the transient-that is, whether it is generated inside or outside the cell and how fast it must be pumped through the cell to create a workable effective pathlength for absorption spectroscopy. Instead of probing the flowing gas laterally for information about changes of composition with distance, the practice now is usually to probe the gas column longitudinally so as to intercept as many molecules as possible, and often with multipass facilities to give optical pathlengths of tens of meters. For example, the NO, radical-a prominent player on the atmospheric scene-has been generated in a flow system by admixing NOz with an excess of 0, and characterized by its antisymmetric N-0 stretching mode (v,) by means of a tunable infrared diode laser spectrometer (105). Here Zeeman modulation turns out t o be invaluable in picking out the absorption lines due to paramagnetic NO, from among the many lines due to diamagnetic species. The results indicate that NO, is planar and conforms to DBhsymmetry in the 2Ai ground electronic state with an N-0 bond length of 124.0 pm. On the other hand, close scrutiny reveals some intriguing anomalies, possibly arising from interaction between the ground state and a neighboring 2E' excited state. The transient SiClZ, an intermediate in various chemical processes, has also been generated chemically (by passing Sic& vapor over heated silicon powder) and characterized in a flow reactor by both LIF and microwave studies. The microwave measurements fix the equilibrium structure of the molecule with remarkable precision, giving r&Si-Cl) = 206.531 pm and LCl-Si-C1 = 101.324", together with harmonic, cubic, and third-order anharmonic potential constants (106).
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
159
Flow systems are not limited to the study of gaseous reactions, and with certain alterations can be applied to the study of reactions in the liquid phase ( 3 , 73). Because mixing is significantly slower in the liquid phase, attention has to be paid to the design of the mixing chamber, which may include, for example, an elaborate set of baffles to promote turbulent flow and mixing of two reactant solutions. A familiar and more economical variation on the flow theme is the stoppedflow technique. As with continuous flow, the reactant solutions are injected from syringes into a mixing cell, whence the mixture flows into an observation tube. After a few milliseconds the flow is abruptly arrested. The real-time evolution of the species present in solution is then monitored; in effect, the method is exactly analogous to conventional batch-mixing kinetic studies and the stopped flow is simply a device for the rapid mixing of two solutions. The most common detection technique is electronic absorption spectrophotometry, a relatively blunt instrument when it comes to the identification and characterization of reaction intermediates in the condensed phases. No better examples of the use of stopped-flow methods can be found than in the work of Dale Margerum and his colleagues a t Purdue University, who have developed a pulsed-accelerated-flow spectrophotometer with UVvisible detection that permits the measurement of pseudo-first-order rate constants as large as 5 X lo5 s-’ (tllz = 1.4 ps) (107). Typical of the kinetic studies carried out by this group on a wide range of “classical” inorganic reactions in aqueous solution are those implying (i) that hydroxylamine is oxidized by iodine by way of the adduct I2 * NH20H, which undergoes general-base-assisted deprotonation to give the intermediate INHOH (108), and (ii) that bromide ions are oxidized by nitrogen trichloride through the intermediate agency of NBrClz (109).Here we are relying on kinetic measurements mainly, if not exclusively, for whatever inferences may be drawn about reaction intermediates. In contrast to the gas phase, solutions labor under the disadvantage that kinetic measurements are confined to a relatively narrow temperature range, normally 40°C or less, thereby impairing the precision of any estimates of activation parameters. Indeed, recent years have seen a perceptible trend away from attempting any such estimates. From our present perspective, though, no level of confidence in reaction rates can entirely redeem the uncertainties that are apt to cloud the identities of any transients stopped-flow and related studies may disclose, whether directly or indirectly. For example, stopped-flow experiments have been carried out t o study the reaction occurring in
160
DOWNS AND GREENE
aqueous solution between cerium(IV) and an excess of peroxotitanium(IV), Ti(02)2+(110). This yields a transient, which, on the evidence of its ultraviolet absorption spectrum, conditions of formation, and decay properties, was believed to be a superoxo-titanium(1V) species existing apparently in two forms that were tentatively formulated as Ti(OO)3' and Ti(OO)OH2+.A separate investigation of the decomposition kinetics of this species suggests that it is more likely to be TiO(H02.)2+, that is, a complex of the aquated Ti02+ion with the perhydroxyl radical (110). ESR measurements may shed a more distinctive light on paramagnetic intermediates. Thus, the oxidation of chromium(II1) to chromium(VI) by hydrogen peroxide in basic media occurs via a branching mechanism: from the initial formation of a chromium(II1)-peroxide intermediate, one pathway leads to a chromium(IV) intermediate and a second pathway leads to a chromium(V) intermediate, which can be identified empirically by its ESR spectrum (111).Nevertheless, doubts assailing us about the true nature of intermediates like these contrast starkly with the conviction high-resolution studies bring to the identification and characterization of a gaseous intermediate like Sic&. For all its immense power and versatility, NMR spectroscopy has kept a relatively low profile in the story so far. There is no doubting the potential of real-time NMR measurements to report directly on a reaction and on all the components of the reaction mixture. There are certainly ways of allying rapid mixing, stopped-flow, temperaturejump or other methods with NMR detection (3, 112),but no technical stratagem can remedy the want of sensitivity that is inherent in the NMR experiment. Nevertheless, the ability of the technique specifically to monitor a free ligand can be turned to good account, for example to resolve the three stages in the displacement of three dmso molecules (dmso = Me2SO)in the cation [Al(dmso),13+by one terdentate ligand 2,2': 6'2"-terpyridine (terpy), XIV, in nitromethane solution (113).The implication is that there are two intermediates involving
XIV
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
161
mono- or bidentate coordination of the terpy ligand to the metal center. Rates for the three stages have been determined using a special stopped-flow apparatus, with ‘H NMR measurements reporting on the release of coordinated dmso molecules. Hence, activation parameters have been determined for the first bond formation and the two slower chelate-ring closures that ensue. More recent experiments have turned to the much more elaborate puzzle of protein folding, where NMR studies are privy to unique structural insights (114). The resulting spectra relay information not only about kinetic folding events but also about the structures of folding intermediates, partly folded states, peptide fragments, and unfolded or denatured proteins. Recent developments in mass spectrometry, with “electrospray” ionization, provide another relatively fast device for monitoring solution processes in real time, and one that has been applied to the teasing problems of protein folding (115, 116). Such folding is assisted by a host of helper proteins including the molecular chaperones, which are thought to safeguard newly synthesized proteins from unproductive interactions that may lead to aggregation. This raises the problem of how to determine the conformational properties of a relatively small substrate protein in the presence of a large chaperonin oligomer like GroEL. One new perspective shown schematically in Fig. 24 (115) seeks to elicit the information from hydrogen exchange protection (a technique that has revolutionized studies of protein folding in vitro) specifically relating to complexes relevant to folding in the cell. For example, a partially folded protein bound within the GroEL central cavity can be directly studied by electrospray ionization mass spectrometry (ESI-MS), allowing the visualization of both the GroEL subunits and the ligand in the same spectrum. When combined with hydrogen exchange labeling, this approach admits the study of individual components of the complex that retain the history of the complexed state by virtue of their deuterium content; it thus obviates the need to quench exchange and dissociate the complex before measurement. Hence, we are able in effect to measure selectively the protection of individual components in a mixture of species and in proteins that are well beyond the molecular mass range open to detailed NMR analysis, and also to observe states too short-lived to be characterized, say, by X-ray crystallography. In addition, differences in the cooperativity of folding events can be detected directly by the ESI-MS method (116). Yet another option is being elaborated by Fraser Armstrong and his group a t Oxford through fast-scan electrochemical techniques, which have now advanced to the point where they are capable of probing
162
DOWNS AND GREENE DETECTED IN MASS SPECTROMnER
+
GroEL
1-
Compkx Is diluted 1 &fold into H p solution.
Compkx dbscciates in maw spectrometer. &and retains bbelling. Extent of labelli k
according to iu stmctum
FIG.24. Schematic diagram outlining the experiment designed to monitor hydrogen exchange in the GroEL: [3SS] BLA complex (BLA = bovine a-lactalbumin) by electrospray ionization mass spectrometry (ESI-MS). Before formation of the complex, all exchangeable sites in a-lactalbumin were deuterated by incubation of the apoprotein in DzO. Hydrogen exchange is initiated by a 10-fold dilution of the complex in H,O, pH 5.0, a t 4°C. Dissociation of the GroEL into monomers in the gas phase releases the bound polypeptide ligand while maintaining the protection in the substrate protein. In this way, the rate of hydrogen exchange of [3SS1 a-lactalbumin (and GroEL) in the complex can be measured directly as a change in their masses a s a function of the incubation time in HzO by ESI-MS [reproduced with permission from (115),p. 6471.
transient intermediates formed during redox reactions (117).In the technique of “protein-film voltammetry,” the redox protein molecules under investigation are adsorbed on the rotating electrode surface of a voltammetric assembly and interrogated there electrochemically. Cyclic voltammetry at scan rates up to 1000 V s-l may then reveal pairs of oxidation-reduction peaks, which may serve as markers by which the status of redox centers within the protein can be identified and quantified. In effect, we have an interactive “spectrum,” albeit one providing no structural information; for that we must turn to true spectroscopic methods, such as ESR, to examine the species generated in solution at the appropriate potentials. Protein-film voltammetry has already been successfully exploited to track active-site redox transformations over a range of conditions and so reveal the intramolecular electron relays operating in the multicentered enzyme. Large systems like these hold out little prospect of furnishing the sort of structurally explicit information that is now available for simple molecules like SizHzor Cr(CO),. On the other hand, what we lose in struc-
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
163
tural definition may well be compensated by a clearer perception of the thermodynamic and kinetic properties that determine the reactivity of a given intermediate. Problems of reliable identification are no strangers to studies that concentrate on kinetics, whether they involve the stopped-flow technique, relaxation methods, or other strategies. We may know very well the speed and order of a particular reaction; we may even be well versed in the dependence of the kinetic parameters on various conditions; however, we cannot always be sure of the exact nature of the reagents and the products, still less the intermediates formed on the way. To that extent, any mechanistic interpretation is bound to be impaired. Indeed, the situation has something of the surreal about it, reminiscent of the conversation between Alice and the Cheshire Cat: “Would you tell me, please, which way I ought to go from here?” “That depends a good deal on where you want to get to,” said the Cat. “I don’t much care where . . . ,” said Alice. “Then it doesn’t matter which way you go,” said the Cat. “. . . So long as I get somewhere,” Alice added as an explanation. “Oh, you’re sure to do that,” said the Cat, “if you only walk long enough.” Lewis Carroll, “Alice’s Adventures in Wonderland”
Vi. Conclusions
In this account we have sought to highlight reaction intermediates and the crucial bearing they have on the way chemical processes actually take place. We have concentrated on how such intermediates can be directly observed, identified, characterized, and tracked by their spectroscopic properties. Frustration of the reactions t o which the intermediates are natural prey can be achieved by control of temperature, pressure, and environment, and in these circumstances it may be perfectly possible to form a clear and detailed picture of a given intermediate on the basis of the same sort of spectroscopic methods that are used to study normal, long-lived species. To chart the chemistry of such an intermediate requires a closer acquaintance with the reality of normal conditions, and a t least some concession not only to mobility, concentration, and temperature, but also to time as a variable; this commonly involves some form of time-resolved spectroscopic
164
DOWNS AND GREENE
probe, as in flash photolysis, or the induction of reaction not in a static but in a uniformly flowing gas or liquid. A recurring example has been chromium pentacarbonyl, Cr(CO), , an intermediate most easily generated photochemically en route t o substitution products of chromium hexacarbonyl (60, 611. Matrix-isolation experiments were the first to fix the structure of this molecule and to alert us to its extraordinary reactivity. Experiments with fluid samples were then needed t o elaborate on the ability of Cr(CO)5to bind supposedly inert molecules such as Xe, H2, and CH, , and timeresolved spectroscopy has opened our eyes to the energetics of these interactions as well as the photochemistry and kinetics of the processes involving Cr(CO)5. So strongly is CI-(CO)~solvated that it is hard to conceive of the molecule in solution without its “token” ligand. Naked Cr(CO)5plays a role, it is true, but certainly does not live long enough t o “wait” for a collision in dilute solution; instead, in the space of picoseconds, it picks up a token ligand from within its solvent cage via the pseudorotation illustrated in Fig. 13. Hence, the orthodox notion of a simple dissociative mechanism for a substitution reaction starting from saturated 18-electron Cr(CO), and involving the intermediate formation of an unsaturated 16-electron intermediate must give place to an interchange model that is altogether more intricate and subtle and in which the solvent is no mere spectator. Even now, for all our knowledge of the chemistry of CI-(CO)~at time-scales of a nanosecond and longer, controversy and doubts continue to surround the very fast processes implicating the fragment. No less striking have been the advances made in our understanding of the substitution reactions peculiar t o the much studied binuclear iron carbonyl, VIII (83-85). Here, too, a variety of experimental techniques has been brought to bear on the intermediate photoproductsone a binuclear species IX and the other a mononuclear radical Xwhich differ hugely in their reactivity. The ability of a reaction intermediate such as CI-(CO)~ to bind methane may be counterintuitive but commands attention for its significance in relation to C-H bond activation, a teasing but vital issue in the context of the chemical industry (118).Numerous complexes of alkanes with unsaturated transition-metal fragments (including atoms and ions) have now been detected in both the gas and condensed phases. It is not surprising that all of them are unstable at room temperature. Experiments in which alkanes undergo oxidative addition to, or reductive elimination from, transition-metal complexes have revealed, nonetheless, the intermediacy of alkane complexes.
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
165
For example, TRIR studies of liquid xenon solutions of the rhodium complex (v5-CSMe5)Rh(CO), witness on photolysis the expulsion of CO and formation of a transient characterized by its v(C-0) mode; a similar band is observed a t much the same frequency when liquid krypton is used as the solvent but belongs to a considerably more reactive carrier (119).The transients are most likely to be (q5-CSMe5)Rh(CO)(Xe)and (v5-C5Me5)Rh(CO)(Kr), respectively. In the presence of a small amount of cyclohexane, however, the intermediates transform rapidly into the insertion product (v5-C5Me,)Rh(CO)(C,H,,)(H).The kinetics of formation of this product in liquid krypton are consistent not with a dissociative mechanism but with a reaction scheme incorporating a n equilibrium between krypton and cyclohexane complexes of (v5CSMeS)Rh(CO)prior to oxidative addition. The failure to observe a distinct v(C-0) mode for the alkane complex ( vS-C5Me5)Rh(CO)(C6Hl2) is probably due to overlap with the corresponding band of the krypton complex. Unambiguous though the case for transition-metal alkane complexes may be on the evidence of these and many other studies (1181, the mode of coordination of the alkane has yet to be fathomed properly. Plainly these are but the early stages of a mechanistic expedition that is likely to take many years to complete. In very different territory we have noted how a molecule such as C10 that is normally short-lived in the laboratory is still capable of exercising a major influence on the fate of stratospheric ozone (17). There is particular interest, for example, in the dimerization of C10 and subsequent photolysis of the products [see Eqs. 25(a) and 25(b)l because these processes may have a marked impact on the ozone depletion of the atmosphere above the polar regions. Of especial note in this context has therefore been the first preparation of another labile molecule with the composition Cl2OZby halogen exchange between gaseous FC10, and AICIB(62, 120); trapping in a solid noblegas matrix enables this species to be identified as chloryl chloride, C1C102,XV, on the basis of its infrared and ultraviolet spectra. From the vibrational frequencies of four independent isotopomers, the known properties of related compounds, and the results of ab initio calculations, the following geometric parameters have been estimated: r(C1-C1) = 222 pm, r(Cl=O) = 144.0 pm, LCl-Cl=O = 103.5", and ~ 0 = C 1 = 0= 116.0'. What is intriguing about the photochemistry of this new isomer of Cl2OZis that it decays under appropriate conditions not only t o the more familiar peroxo form ClOOCl, XVI, but also to a third, photolabile isomer, chlorine chlorite, ClOClO, XVII. At room temperature and partial and total pressures of 1 and
166
DOWNS AND GREENE
4 torr, respectively, CICIOz has a half-life of 1 min. Matrix studies involving selective irradiation adduce the first evidence for a second
xv
XVI
XVII
photodissociation channel disposing of ClOOCl (and giving 2C10), and also make it clear that dimerization of C10 can lead to both ClOClO and ClOOCl. These results must surely be heeded by gas-phase kineticists and others seeking to model the atmospheric fates of the chlorine oxides. “Everything’s got a moral, if you can only find it,” and to this tale there are not one but two morals. The first is that reaction intermediates, however ephemeral, can and should be explored directly. The second is that no one experimental or theoretical technique has a monopoly of the truth we seek; to arrive at a comprehensive and realistic picture the best policy must be to draw on as many sources as possible. It is true that the Promised Land of a proper understanding of the mechanisms of major chemical reactions is still a great way off and that the intermediates by which the reactions proceed are not going to be quick to yield up their secrets. However, it is equally true that without these secrets even the most inspired perceptions of mechanism are either incomplete or insecure. In Robin Perutz’s trenchant words (611, “If you don’t understand the intermediates, you don’t understand the mechanisms.” Not only is there but one way of doing things rightly, but there is only one way of seeing them, and that is, seeing the whole of them. J. Ruskin, “The Two Paths”
REFERENCES 1. Tobe, M. L. “Inorganic Reaction Mechanisms”; Nelson: London, 1972. 2. Jordan, R. B. “Reaction Mechanisms of Inorganic and Organometallic Systems”; Oxford University Press: New York and Oxford, 1991. 3. Wilkins, R. G. “Kinetics and Mechanism of Reactions of Transition Metal Complexes”; 2nd ed., VCH: Weinheim and New York, 1991. 4. Darensbourg, D. J. Adu. Organomet. Chem. 1982,21, 113. 5. Howell, J.A. S.; Burkinshaw, P. M. Chem. Rev. 1983,83, 557.
COMING TO GRIPS WITH REACTrVE INTERMEDIATES
167
6. Woodward, S. In “Comprehensive Organometallic Chemistry 11,” Vol. 5 (J. A. Labinger and M. J . Winter, Eds.), Chap. 4, pp. 215-280, Pergamon: Oxford, 1995. 7. Richter-Addo, G. B.; Legzdins, P. “Metal Nitrosyls”; Oxford University Press: New York and Oxford, 1992. 8. Niu, S.; Hall, M. B. J. Am. Chem. Soc. 1997,119, 3077 and references cited therein. 9. (a) Seder, T. A.; Church, S. P.; Ouderkirk, A. J.; Weitz, E. J. Am. Chem. Soc. 1985, 107, 1432; (b)Fletcher, T. R.; Rosenfeld, R. N. J. Am. Chem. Soc. 1985, 107, 2203. 10. (a) Hirota, E. “High-Resolution Spectroscopy of Transient Molecules”; Springer: Berlin, 1985; (b) Hirota, E. Irrt. Reu. Ph.ys. Chem. 1989,8, 171; (c) Hirota, E. Annu. Reu. Phys. Chem. 1991,42, 1; (d) Hirota, E. Chem. Rev. 1992, 92, 141; ( e )Hirota, E. Ann. Rep. Prog. Chem., Sect. C, Phys. Chem. 1994, 91, 3. 11. Cox, P. A. “The Elements: Their Origin, Abundance, and Distribution”, pp. 93126, Oxford University Press: Oxford, 1989. 12. Manson, E. L., Jr.; Clark, W. W.; De Lucia, F. C.; Gordy, W. Phys. Rev. A 1977, 15, 223. 13. Huber, K. P.; Herzberg, G . “Molecular Spectra and Molecular Structure. IV.Constants of Diatomic Molecules”; Van Nostrand Reinhold: New York, 1979. 14. Ram, R. S.; Bernath, P. F. J . Mol. Spectrosc. 1996, 180, 414. 15. Lorenz, M.; Agreiter, J.; Smith, A. M.; Bondybey, V. E. J. Chem. Phys. 1996, 104, 3143. 16. Bredohl, H.; Dubois, I.; Houbrechts, Y.; Nzohabonayo, P. J . Mol. Spectrosc. 1985, 122, 430. 17. ( a ) Wayne, R. P. “Chemistry of Atmospheres,” 2nd ed., Clarendon Press: Oxford, 1991; (b) Wayne, R. P.; Poulet, G.; Biggs, P.; Burrows, J. P.; Cox, R. A,; Crutzen, P. J.; Hayman, G. D.; Jenkin, M. E.; Le Bras, G.; Moortgat, G. K.; Platt, U.; Schindler, R. N. Atmos. Enuiron. 1995,29, 2677. 18. Anderson, J . G.; Toohey, D. W.; Brune, W. H. Science 1991,251, 39. 19. (a) Cohen, E. A.; Pickett, H. M.; Geller, M. J. Mol. Spectrosc. 1984, 106, 430; (b) Burkholder, J. B.; Hammer, P. D.; Howard, C. J.; Maki, A. G.; Thompson, G.; Chackerian, C., Jr. J. Mol. Spectrosc. 1987, 124, 139; (c) McLoughlin, P. W.; Park, C. R.; Wiesenfeld, J. R. J . Mol. Spectrosc. 1993, 162, 307. 20. Herzberg, G.; Johns, J. W. C. Proc. Roy. Soc. l967,298A, 142. 21. Kawaguchi, K. J . Chem. Phys. 1992,96, 3411; Can. J . Ph,ys. 1994, 72, 925. 22. Kawashima, Y.; Kawaguchi, K.; Hirota, E. J. Chem. Phys. 1987,87, 6331. 23. Ebsworth, E. A. V.; Rankin, D. W. H.; Cradock, S. “Structural Methods in Inorganic Chemistry,” 2nd ed., pp. 334-343, Blackwell Scientific Publications: Oxford, 1991. 24. See, for example, Blake, A. J.; Brain, P. T.; McNab, H.; Miller, J.; Morrison, C. A,; Parsons, S.; Rankin, D. W. H.; Robertson, H. E.; Smart, B. A. J. Phys. Chern. 1996, 100, 12280. 25. Fujiwara, H.; Egawa, T.; Konaka, S. J . Mol. Struct. 1995, 344, 217. 26. Fujiwara, H.; Egawa, T.; Konaka, S. J . Am. Chem. Soc. 1997,119, 1346. 27. Molnar, J.; Marsden, C. J.; Hargittai, M. J . Phys. Chern. 1995, 99, 9062. 28. Hedberg, K.; Hedberg, L.; Buhl, M.; Bethune, D. S.; Brown, C. A,; Johnson, R. D. J. Am. Chem. Soc. 1997,119, 5314. 29. Carlowitz, M. V.; Oberhammer, H.; Willner, H.; Boggs, J. E. J. Mol. Struct. 1983, 100, 161. 30. Cleaver, W. M.; Spath, M.; Hnyk, D.; McMurdo, G.; Power, M. B.; Stuke, M.; Rankin, D. W. H.; Barron, A. R. Orgunometullzcs 1995, 14, 690. 31. Legon, A. C. Chem. Br. 1990,26, 562.
168
DOWNS AND GREENE
32. Legon, A. C. In “Atomic and Molecular Beam Methods,” Vol. 2 (G. Scoles, Ed.), pp. 289-308, Oxford University Press: New York, 1992. 33. Legon, A. C. J. Chem. SOC.,Chem. Commun. 1996, 109. 34. Bloemink, H. I.; Hinds, K.; Legon, A. C.; Thorn, J . C. Chem. Eur. J. 1995, I, 17. 35. Bloemink, H. I.; Hinds, K.; Holloway, J . H.; Legon, A. C. Chem. Phys. Lett. 1995, 245, 598. 36. Campbell, E. J.; Kukolich, S. G. Chem. Phys. 1983, 76, 225. 37. Campbell, E. J.; Read, W. G. J. Chem. Phys. 1983,78, 6490. 38. Bloemink, H. I.; Evans, C. M.; Holloway, J. H.; Legon, A. C. Chem. Phys. Lett. 1996,248, 260. 39. Barnes, M.; Hajigeorgiou, P. G.; Kasrai, R.; Merer, A. J.; Metha, G. F. J. Am. Chem. SOC.1995, 1 17, 2096. 40. Davies, P. B. Chem. SOC.Rev. 1995,24, 151. 41. Hirahara, Y.; Ohshima, Y.; Endo, Y. J . Chem. Phys. 1994,101, 7342. 42. Yamada, C.; Hirota, E. Phys. Rev. Lett. 1986, 56, 923. 43. Bogey, M.; Bolvin, H.; Demuynck, C.; Destombes, J. L. Phys. Reu. Lett. 1991, 66, 413. 44. Colegrove, B. T.; Schaefer, H. F., 111. J. Phys. Chem. 1990,94, 5593. 45. Cordonnier, M.; Bogey, M.; Demuynck, C.; Destombes, J.-L. J. Chem. Phys. 1992, 97, 7984. 46. Harper, W. W.; Ferrall, E. A.; Hilliard, R. K.; Stogner, S. M.; Grev, R. S.; Clouthier, D. J. J . Am. Chem. SOC.1997,119, 8361. 47. See, for example, (a) Davies, P. B.; Martineau, P. M. J . Appl. Phys. 1992,71, 6125; (b) Naito, S.; Ito, N.; Hattori, T.; Goto, T. Jpn. J. Appl. Phys., Part 1 1994, 33, 5967; (c) Fukuzawa, T.; Obata, K.; Kawasaki, H.; Shiratani, M.; Watanabe, Y. J. Appl. Phys. 1996,80, 3202. 48. Perutz, R. N. Chem. Rev. 1985,85, 77. 49. Downs, A. J.; Pulham, C. R. Adu. Inorg. Chem. 1994,41, 171. 50. Sawyer, D. T. “Oxygen Chemistry,” pp. 156-158, Oxford University Press: New York, 1991. 51. (a) Almond, M. J.; Downs, A. J . Adv. Spectrosc. 1989, 17, 1-511; (b) Andrews, L., and Moskovits, M., Eds., “Chemistry and Physics of Matrix-Isolated Species,” North Holland: Amsterdam, 1989. 52. Downs, A. J . In “Low Temperature Molecular Spectroscopy” (R. Fausto, Ed.), pp. 1-93, NATO AS1 Series, Series C: Vol. 483, Kluwer: Dordrecht, 1996. 53. Bondybey, V. E.; Smith, A. M.; Agreiter, J . Chem. Rev. 1996, 96, 2113. 54. Chetwynd-Talbot, J.; Grebenik, P.; Perutz, R. N. Inorg. Chem. 1982,21, 3647. 55. Tague, T. J., Jr.; Andrews, L. J . Am. Chem. SOC.1994, 116, 4970. 56. Poliakoff, M.; Weitz, E. Acc. Chem. Res. 1987,20, 408. 57. Crayston, J. A.; Almond, M. J.; Downs, A. J.; Poliakoff, M.; Turner, J . J. Znorg. Chem. 1984,23, 3051. 58. Muller, J. J . Am. Chem. SOC.1996, 118, 6370. 59. Fletcher, S. C . ; Poliakoff, M.; Turner, J. J. Zmorg. Chem. 1986,25, 3597. 60. Turner, J. J. In “Photoprocesses in Transition Metal Complexes, Biosystems and Other Molecules. Experiment and Theory” (E. Kochanski, Ed.), pp. 125-140, NATO AS1 Series, Series C: Vol. 376, Kluwer: Dordrecht, 1992. 61. Perutz, R. N. Chem. SOC.Reu. 1993,22, 361. 62. Jacobs, J.; Kronberg, M.; Muller, H. S. P.; Willner, H. J. Am. Chem. SOC.1994, 116, 1106.
COMING TO GRIPS WITH REACTrVE INTERMEDIATES
169
63. Bell, T. W.; Haddleton, D. M.; McCamley, A.; Partridge, M. G.; Perutz, R. N.; Willner, H. J . Am. Chem. SOC.1990, 112, 9212. 64. Poliakoff, M.; Turner, J. J. Adu. Spectrosc. 1995, 23, 275. 65. Van der Veken, B. J. In “Low Temperature Molecular Spectroscopy” (R. Fausto, Ed.), pp. 371-420, NATO AS1 Series, Series C: Vol. 483, Kluwer: Dordrecht, 1996. 66. Upmacis, R. K.; Poliakoff, M.; Turner, J. J. J. Am. Chem. SOC. 1986, 108, 3645. 67. Duckett, S. B.; Haddleton, D. M.; Jackson, S. A,; Perutz, R. N.; Poliakoff, M.; Upmacis, R. K. Organometallics 1988, 7, 1526. 68. Turner, J. J.; Simpson, M. B.; Poliakoff, M.; Maier, W. B., 11. J. Am. Chem. SOC. 1983,105, 3898. 69. Howdle, S. M.; Healy, M. A,; Poliakoff, M. J. Am. Chem. SOC. 1990, 112, 4804. 70. (a) McHugh, M. A,; Krukonis, V. J . “Supercritical Fluid Extraction: Principles and Practice,” 2nd ed., Butterworth-Heinemann: Boston, 1994; (b) Poliakoff, M.; Howdle, S. Chem. Br. 1995, 31, 118. 71. Walsh, E. F.; Popov, V. K.; George, M. W.; Poliakoff, M. J. Phys. Chem. 1995, 99, 12016. 72. Rathke, J. W.; Klingler, R. J.; Krause, T. R. Organometallics 1991, 10, 1350. 73. Pilling, M. J.; Seakins, P. W. “Reaction Kinetics”; Oxford University Press: Oxford, 1995. 74. Poliakoff, M.; Weitz, E. Adu. Organomet. Chem. 1986, 25, 277. 75. Turner, J. J . In “Photoprocesses in Transition Metal Complexes, Biosystems and Other Molecules. Experiment and Theory” (E. Kochanski, Ed.), pp. 113-123, NATO AS1 Series, Series C: Vol. 376, Kluwer: Dordrecht, 1992. 76. Perutz, R. N. In “Low Temperature Molecular Spectroscopy” (R. Fausto, Ed.), pp. 95-124, NATO AS1 Series, Series C: Val. 483, Kluwer: Dordrecht, 1996. 77. El-Sayed, M. A,, Tanaka, I., and Molin, Y., Eds., “Ultrafast Processes in Chemistry and Photobiology”; Blackwell Science: Oxford, 1995. 78. Farhataziz; Rodgers, M. A. J. “Radiation Chemistry-Principles and Applications”; VCH: New York, 1987. 79. Goldstein, S.; Czapski, G. Znorg. Chem. 1997, 36, 4156. 80. Tripathi, G. N. R. Adu. Spectrosc. 1989, 18, 157. 81. George, M. W.; Poliakoff, M.; Turner, J. J. Anal-yst 1994, 119, 551. 82. Sun, X.-Z.; George, M. W.; Kazarian, S. G.; Nikiforov, S. M.; Poliakoff, M. J . Am. Chem. Sac. 1996,118, 10525. 83. (a) Hooker, R. H.; Mahmoud, K. A,; Rest, A. J. J . Chern. SOC.,Chem. Commun. 1983, 1022; (b) Hepp, A. F.; Blaha, J. P.; Lewis, C.; Wrighton, M. S. Organometallics 1984, 3, 174. 84. (a) Moore, B. D.; Simpson, M. B.; Poliakoff, M.; Turner, J. J. J . Chem. SOC.,Chem. Commun. 1984, 972; (b) Dixon, A. J.; George, M. W.; Hughes, C.; Poliakoff, M.; Turner, J . J . J. Am. Chem. Sac. 1992,114, 1719. 85. Vitale, M.; Lee, K. K.; Hemann, C. F.; Hille, R.; Gustafson, T. L.; Bursten, B. E. J. Am. Chem. SOC.1995,117, 2286. 86. Hamaguchi, H.-o.; Gustafson, T. L. Annu. Reu. Phys. Chen. 1994,45, 593. 87. Kitagawa, T.; Ogura, T. Prog. Inorg. Chem. 1997,45, 431. 88. Schelvis, J. P. M.; Deinum, G.; Varotsis, C. A,; Ferguson-Miller, S.; Babcock, G. T. J . Am. Chem. SOC.1997, 119, 8409. 89. Osman, R.; Pattison, D. I.; Perutz, R. N.; Bianchini, C.; Casares, J . A.; Peruzzini, 1997, 119, 8459. M. J. Am. Chem. SOC. 90. Colombo, M.; George, M. W.; Moore, J. N.; Pattison, D. 1.; Perutz, R. N.; Virrels, I. G.; Ye, T.-Q. J. Chem. SOC.,Dalton Trans., 1997, 2857.
170
DOWNS AND GREENE
91. Cheng, P. Y.; Zhong, D.; Zewail, A. H. J. Chem. Phys. 1996, 105, 6216. 92. Zhong, D.; Ahmad, S.; Zewail, A. H. J . Am. Chem. SOC.1997, 119, 5978. 93. Stephens, J. W.; Morter, C. L.; Farhat, S. K.; Glass, G. P.; Curl, R. F. J. Phys. Chem. 1993,97, 8944. 94. (a) Ischenko, A. A,; Schafer, L.; Luo, J . Y.; Ewbank, J. D. J. Phys. Chem. 1994, 98, 8673; (b) Ischenko, A. A,; Ewbank, J. D.; Schafer, L. J. Phys. Chem. 1995, 99, 15790. 95. (a) Williamson, J. C.; Zewail, A. H. J. Phys. Chem. 1994,98, 2766; (b) Dantus, M.; Kim, S. B.; Williamson, J . C.; Zewail, A. H. J . Phys. Chem. 1994, 98, 2782. 96. Williamson, J . C.; Cao, J.; Ihee, H.; Frey, H.; Zewail, A. H. Nature 1997,386, 159. 97. Turner, J. J.; George, M. W.; Johnson, F. P. A.; Westwell, J. R. Coord. Chem. Reu. 1993, 125, 101. 98. Perng, J.-H.; Zink, J . I. Znorg, Chem. 1990,29, 1158. 99. Glyn, P.; Johnson, F. P. A,; George, M. W.; Lees, A. J.; Turner, J. J. Znorg. Chem. 1991,30, 3543. 100. O’Hare, D. In “Inorganic Materials,” 2nd ed. (D. W. Bruce and D. O’Hare, Eds.), Chap. 4, pp. 171-254, Wiley: Chichester, 1996. 101. (a) Srajer, V.; Teng, T.-y.; Ursby, T.; Pradervand, C.; Ren, Z.; Adachi, S.-i.; Schildkamp, W.; Bourgeois, D.; Wulff, M.; Moffat, K. Science 1996,274, 1726; (b) Eaton, W. A,; Henry, E. R.; Hofrichter, J. Science 1996,274, 1631. 102. Clyne, M. A. A,; McKenney, D. J.; Watson, R. T. J. Chem. Soc., Faraday Trans. I 1975, 71, 322. 103. Kukui, A.; Jungkamp, T. P. W.; Schindler, R. N. Ber. Bunsenges. Phys. Chem. 1994,98, 1619. 104. Zagogianni, H.; Mellouki, A.; Poulet, G. C. R. Acad. Sci. Paris, Ser. IZ 1987, 304, 573. 105. (a) Ishiwata, T.; Tanaka, I.; Kawaguchi, K.; Hirota, E. J. Chem. Phys. 1985, 82, 2196; (b) Kawaguchi, K.; Hirota, E.; Ishiwata, T.; Tanaka, I. J. Chem. Phys. 1990, 93, 951. 106. Fujitake, M.; Hirota, E. Spectrochim. Acta 1994, 50A, 1345. 107. Bowers, C. P.; Fogelman, K. D.; Nagy, J. C.; Ridley, T. Y.; Wang, Y. L.; Evetts, S. W.; Margerum, D. W. Anal. Chem. 1997,69, 431. 108. Liu, R. M.; McDonald, M. R.; Margerum, D. W. Znorg. Chem. 1995,34, 6093. 109. Gazda, M.; Kumar, K.; Margerum, D. W. Inorg. Chem. 1995,34, 3536. 120. Bourke, G. C. M.; Thompson, R. C. Inorg. Chem. 1987,26, 903. Rotzinger, F. P.; Griitzel, M. Inorg. Chem. 1987,26, 3704. 111. Knoblowitz, M.; Morrow, J. I. Znorg. Chem. 1976, 15, 1674. 112. Moore, P. Pure Appl. Chem. 1985,57, 347. 113. Brown, A. J.; Howarth, 0. W.; Moore, P.; Parr, W. J. E. J. Chem. SOC.,Dalton Trans. 1978, 1776. 114. Dyson, H. J.; Wright, P. E. Annu. Reu. Phys. Chem. 1996,47, 369. 115. Robinson, C. V.; Gross, M.; Eyles, S. J.; Ewbank, J . J.; Mayhew, M.; Hartl, F. U.; Dobson, C. M.; Radford, S. E. Nature 1994,372, 646. 116. Hooke, S. D.; Eyles, S. J.; Miranker, A.; Radford, S. E.; Robinson, C. V.; Dobson, C. M. J. Am. Chem. SOC.1995,117, 7548. 117. (a) Armstrong, F. A.; Heering, H. A,; Hirst, J. Chem. Soc. Reu. 1997, 26, 169; (b) Heering, H. A.; Weiner, J . H.; Armstrong, F. A. J. Am. Chem. SOC.,1997, 119, 11628. 118. (a) Crabtree, R. H. Chem. Reu. 1996, 95, 987; (b) Hall, C.; Perutz, R. N. Chem. Reu. 1996,96, 3125.
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
171
119. (a) Bengali, A. A,; Arndtsen, B. A.; Burger, P. M.; Schultz, R. H.; Weiller, B. H.; Kyle, K. R.; Moore, C. B.; Bergman, R. G. Pure Appl. Chem. 1995, 67, 281; (b) Arndtsen, B. A,; Bergman, R. G.; Mobley, T. A.; Peterson, T. H. Acc. Chem. Res. 1995,28, 154. 120. Miiller, H. S. P.; Willner, H. Znorg. Chem. 1992,31, 2527.
This Page Intention ally Left Blank
ADVANCES IN INOKGANI(’ CHEMISTKY VOL
46
TOWARD THE CONSTRUCTION OF FUNCTIONAL SOLID-STATE SUPRAMOLECULAR METAL COMPLEXES CONTAINING COPPER(1) AND SILVER(1) MEGUMU MUNAKATA, LIANG PING WU, and TAKAYOSHI KURODA-SOWA Department of Chemistry, Kinki University, Kowakae 3-4-1, Higashi-Osaka 577, Japan
I. Introduction 11. Helical Frameworks A. Infinite Single-Helical Complexes B. Double-Helical Complexes C. Triple-Helical Complexes 111. S----SContact-Assembled Frameworks A. dmit and the Related Ligands B. BEDT-TTF C. TTC,,-TTF Systems D. C,H& IV. Hexagonal Frameworks and Graphite-like Structures A. Metal Cyanide-Regulated by the Metal Ion B. Pyrazine Systems-Regulated by Substituents C. Phenazine and Benzothiadiazole Systems-Regulated by Counteranions V. Hydrogen-Bond-Assembled Frameworks A. Three-Dimensional Supramolecular Cu(1) Complexes with Channels B. Hydrogen-Bonding- and a-a-Stacking-Assembled Cu(I) Complexes VI. a-a-Interaction-Assembled Frameworks A. Intermolecular a--TIInteraction in Discrete Coordination Compounds B. Inter- and Intrapolymer a-a Interaction VII. Diamondoid Frameworks A. Bridged by Pyridine or Pyrazine Derivatives B. Bridged by Bisnitrile Ligands C. Other Bridging Ligands D. T-a Interaction in Diamondoid Frameworks VIII. Other Frameworks Based on Covalent Bonds A. Infinite-Chain Structures B. Two-Dimensional Structures C. Three-Dimensional Structures IX. Concluding Remarks References 173 Copyright 11 1999 hy Academic Press N I rights of reproduction i n any form reservcd.
O ~ ~ B - B H W Y Y $za.nn
174
MUNAKATA, WU, AND KURODA-SOWA
1. Introduction
Recent years have witnessed considerable interest in the development of rational synthetic routes to supramolecular architecture from self-assembly of component metal complexes. These solid materials with well-defined, discrete network topologies are attractive to chemists not only for aesthetic reasons but also for their potential applications in many areas (1-13). A major difficulty in controlling such a self-assembling process is the fact that it involves several stages, such as recognition between the components, correct orientation so as to allow growth, and termination of the process leading to the predesigned species ( 1 4 ) .In addition, the intermolecular forces that control these stages are directional and weak enough to enable closely related molecules to give widely different aggregates under slightly modified conditions. Usually, the synthetic strategy involves selecting the coordination geometry of metal ions, the chemical structure of organic ligands, and the favorable reaction conditions such as solvents and counterions. In this context, a fundamental principle of the designed chemistry of the intermolecular bond is the utilization of the intermolecular forces, which may consist of hydrogen bonding, S----Scontacts, aromatic stackings, host-guest interactions, van der Waals forces, and other electrostatic attractions. Unfortunately, these forces are much less well understood than classical chemical bonds in terms of their energetic and geometric properties. Therefore, the control of molecular assembly using supramolecular interactions is probably a new challenge to chemists. The copper(1) and silver(1) ions are regarded as extremely soft acids favoring coordination to soft bases, such as ligands containing S and unsaturated N (15, 16). Copper(1) and silver(1) complexes with these soft ligands give rise to an interesting array of stereochemistries and geometric configurations, with the coordination numbers of two to six all occurring. The most common stereochemistries for both ions are the linear two-coordinate and the tetrahedral four-coordinate geometries with some distortions of the environment, particularly in the presence of chelating type ligands, attributable to the spherical dl0 configuration. Under suitable conditions, these simple coordination compounds with the presence of two rodlike or four sticky sites can be used as tectons to form the self-assembly of predictable supramolecular aggregates. On the other hand, account should be taken of careful selection of the multifunctional organic ligands and controlling the assembly and orientation of the individual building blocks-in other words, a combination of coordination bonds and non-
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
175
covalent intermolecular interactions in a mutually compatible manner. Much study has centered on the use of multifunctional ligands. One of the simplest and most widely employed methods is to use a bifunctional rodlike diatomic CN- anion and bidentate N,N'-donor linking groups such as pyridine-, pyrazine-, and imidazol-based ligands with a preference for binding metals at each end in a linear fashion, together with a metal center with a preference for a polyhedral arrangement of ligands. The recent efforts in this field have even been extended to the novel oligopyridines, calixarenes, crown ethers, cryptands, and tetrathiafulvalene derivatives. Such an approach opens the possibility of rational synthesis of functional solid-state supramolecular metal complexes containing copper(1) and silver(1) with multidimensional helical, honeycomb channel, interwoven diamondoid, and graphite frameworks among other novel structures, Scheme 1. rotaxane
cale na ne
honeycomb
k
linear chain
graphite
SCHEME 1. The self-assembly pathway for formation of copper(1) and silver(1) coordination supramolecules.
176
MUNAKATA, WU, AND KURODA-SOWA
Polynuclear Cu(1) and Ag(1) compounds belong to very diverse stoichiometries. In general, they can be prepared either directly by a reaction between reactants or by electrochemistry. It is the aim of the present review to give a cursory examination of recent developments in construction of the supramolecular frameworks, which combine the covalent bond-forming capability of the metal ion with the ligand surface capable of forming noncovalent interactions. Because of the great volume of literature on compounds, it has been necessary to be highly selective in the choice of material for inclusion. Emphasis throughout is on compounds isolated in the solid state. In a review some 130 pages long one cannot expect to do justice to the depth and extent of investigation into these systems. Some of the topics were reviewed previously (17, 181, but the articles soon become outdated because of the rapid growth of the fields.
II. Helical Frameworks
In this section helical complexes of copper(1) and silver(1) are reviewed as examples of self-assembly in metallosupramolecular systems. One of the most intriguing dissymmetric shapes in the natural system is the helix. Organic and inorganic polymers existing in helical structures are of special interest because of their structural similarities to nucleic acids. They are also intrinsically interesting for their potential applications in the fields of supramolecular chemistry, asymmetric catalysis, and nonlinear optical materials. Since the early pioneering work of Lehn on double-helical copper(1) complexes with oligopyridines ( I , 19), there has been an enormous worldwide interest in helical complexes over the past decade. Most of the work has concentrated on the use of oligopyridines and oligophenanthrolines to control the assembly of helical supramolecular systems. An enormous number of infinite single-helical complexes and double- and triple-stranded helicates have appeared in the literature (see Table I). Many comprehensive reviews have appeared in recent years, usually covering helical structures in general without special reference to copper(1) or silver(1) complexes (8, 20-22). In this section, the complexes with single-, double-, and triple-stranded structures are dealt with separately. The ligands involved in this section are listed in Fig. 1.
COPPER(1, AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
177
TABLE I COPPER(1)AND SILVER(1)COMPLEXES WITH HELICAL FIWEWORKS Complex
Stereochemistry of M
Structure
Ref.
tetrahedral tetrahedral linear linear 3-coordinate tetrahedral linear, 3-coordinate trigonal pyramidal tetrahedral
infinite single strand infinite single strand infinite single strang infinite single strand infinite single strand tetrahelix chain infinite single helical chain 3-D single helix dinuclear double helicate
2.3 24 25 25 26 27 28 29 30, 31
tetrahedral tetrahedral tetrahedral tetrahedral tetrahedral tetrahedral tetrahedral tetrahedral octahedral, tetrahedral octahedral, tetrahedral trigonal bipyramidal tetrahedral 5-coordinate linear, tetrahedral linear, tetrahedral tetrahedral tetrahedral linear linear 2-coordinate
dinuclear double helicate dinuclear double helicate dinuclear double helicate dinuclear double helicate dinuclear double helicate dinuclear double helicate trefoil knot dinuclear double helicate dinuclear double helicate
32 33 34 35 36 37 38, 39 40 41
heterodinuclear double helicate dinuclear double helicate
42
dinuclear double helicate dinuclear double helicate dinuclear double helicate dinuclear double helicate trinuclear double helicate trinuclear double helicate infinite double helicate infinite double helicate triple helicate
43 44 45 46 47 48 50 51 61
43
A. INFINITE SINGLE-HELICAL COMPLEXES The number of one-dimensional infinite single-helical copper( I) and silver(1) complexes is rather limited. They are characterized by a single strand of ligands twisting around the helical axis defined by metal ions [Fig. 2(a)l. It may be left-handed or right-handed. When ligand L, reacted with two equivalents of [Cu(MeCN),IBF.,in a mixture of
178
MUNAKATA, WU, AND KURODA-SOWA
PhzPHzC
CHzPPhz bMe
L4
L5
L6
LlO
Ll1
L12
FIG.1. List of organic ligands in copper(1) and silver(1) helical complexes.
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
L20
L19
L22
179
LP1
L23
FIG.1. (continued)
CHC& and MeCN, bright orange crystals of [Cu(LI)lBF4.CHClB were obtained ( 2 3 ) . In the cation each Cu atom is four-coordinated, with two pyridylethylimine bidentate units from two different ligand molecules in a distorted tetrahedral geometry, and each L, exhibits a bis (bidentate) fashion, bridging two metal centers forming a linear helical strain (Fig. 3). The recently synthesized ligand bisl'342-pyridyl) pyrazol-1-yllphosphinate (L,) contains two bidentate compartments linked by a flexible phosphinate bridge ( 2 4 ) .Its flexibility means that it can adapt to the specific preferences of the metal ions. Its complex
180
f
MUNAKATA, WU, AND KURODA-SOWA
f
'j
k
FIG.2. Schematic views of copper and silver helical arrangements: (a) infinite single strand; (b) 3-D single helical copper(1) complex with mixed ligands; (c) dinuclear double helicate; (d) trinuclear double helicate; (el infinite double helicate; (f) infinite chiral double helicate of copper(I1) with arginine and rn-phthalate; (g) triple helicate of silver(1).
with Ag', [Ag(L2)I* 2H20 contains infinite helical chains of the cation in which each ligand donates one N,N'-bidentate arm to each of two metals and each metal ion is four coordinated by two arms from different ligand!. There are interligand aromatic stacking interactions (3.2-3.6 A) both within each helical strand and between strands. The strands are further held together via a hydrogen-bonding network involving the phosphinate groups and lattice water molecules (Fig. 4). Tetrahedral coordination of the metal ion is not an essential requirement for formation of this type of helical structure. Twocoordinate silver(1) ion plays an important role in the formation of a helical framework, that is, the stereoconformation of the ligand itself is maintained on coordination to Ag(1) and arranged in a onedimensional helical chain. Reaction of the bidentate optically active ligands (4R,5R)-and (4~,5S)-4,5-bis(2-(2-pyridyl)ethyl)-1,3-dioxolane (R,R-L3 or S,S-L3)with silver(1) trifluoromethanesulfonate in metha-
COPPER(1) AND SILVER(I1 SUPRAMOLECULAR METAL COMPLEXES
181
FIG. 3. Structure of [Cu(LI)IBF4.CHC1,. (From Fig. 1 in Gelling, 0. J.; van Bolhuis, F.; Feringa, B. L. J . Chern. SOC.,Chern. Comrnun. 1991, 917.)
no1 gave two isomeric complexes, [Ag(R,R-L,)IO,SCF, and [Ag(S,SL,)]OsSCF3(25).Each complex cation has an extended structure consisting of Ag' with a slightly distorted linear geometry and the bridging ligands. Projection array along the screw axis for each isomer exhibits the left-handed helicity for the former (Fig. 5 ) , and the righthanded for the latter. Helical complexes with nonpolypyridine ligands can give diverse structures and functions. The three-coordinate silver(1) complex with helical structure was observed in [Ag(LJ(NO,)I, where L4 is the chiral ligand (R,R)-(4R,5R)-truns-4,5-bis[(diphenylphosphino)methyll-2,2-dimethyl-1,3-dioxalanel (26). The structure contains an infinite right-handed helical strand consisting of silver atoms, each coordinated by two phosphorus atoms of two adjacent L4 ligands and an oxygen atom of the nitrate ion. The unusual iodine tetrahelix was observed in [(C,H,),P],'[Cu,,I,I, in which tetraphenylphosphonium ions are accompanied by a helical chain of face-sharing
182
MUNAKATA, WU, AND KURODA-SOWA
FIG.4. Structure of [Ag(L2)1.2Hz0.(From Fig. 3 in Psillakis, E.; Jeffery, J. C.; McCleverty, J. A.; Ward, M. D. J. Chern. SOC.,Dalton Trans. 1997, 1645.)
tetrahedra formed by iodocuprate(1) anions (27). Another infinite single-helical structure was reported in a polymeric silver(1) cryptate (28).The polymeric cation [Ag{Ag2(L5)}l consists of dinuclear [Ag2(L,)I units in which each of the two silver atoms located inside the cavity of the ligand is coordinated to two imino nitrogens and one bridgehead nitrogen in a distorted trigonal environment. The third Ag ion links two dinuclear units by coordinating to one of the imino nitrogens of two different L6 ligands. Therefore, the cation can be regarded as the conjugate bis(iminobenzene1 moieties linked together by linear N-Ag-N bridges in an alternate in- and out-conformations relative to the benzene rings, resulting in the formation of an infinite singlehelical chain (Fig. 6). The structure determination reveals that L6 is a unique polydentate ligand with all eight N atoms involved in coordination to the metal. To satisfy the steric requirement of the assembling process, both bond distances and bond angles of the ligand show appreciable differences compared with those in the corresponding cryptate compounds. At the same time, the coordination environment around the silver ion is significantly distorted from ideal linear or trigonal geometry as a stereochemical compromise for formation of the helix.
COPPERII)AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
183
b
FIG. 5 . Structure of [Ag(R&-LdlO,SCF,. (From Fig. 1 in Suzuki, T.; Kotsuki, H.; Isobe, K.; Moriya, N.; Nakagawa, Y.;Ochi, M. Znorg. Chem. 1995,34, 530.)
Recently an unusual type of three-dimensional single-helical framework, [Fig. 2(b)l, has been reported in the copper(1) complex [Cu(Lg) (Me,CO), 5JBF4,where L, is a derivative of pyrazine, 2-pyrazinecarboxamide (29). In the extended structure of the cation each metal atom is linked to two L6 ligands, forming a distorted trigonal planar structure, and axially bridged to another metal center by one acetone molecule with a Cu-0 distance of 2.423(9) giving rise to an infinite helicaI structure (Fig. 7). The most remarkable feature of the complex is that the infinite helices generate a three-dimensionally extended hexagonal array of Cu atoms with a large cavity in which the counteranions are placed. In this structure the primary coordination involves L6 tridentately bridging two copper atoms, and the helical
A
184
MUNAKATA, WU, AND KURODA-SOWA
P
FIG.6 . Structure of the polymeric cation [Ag{Ag2(L6)}l.(From Fig. 2 in Yu, S.-Y.; Luo, Q.-H.; Wu, B.; Huang, X.-Y.; Sheng, T.-L.; Wu, X.-T.; Wu, D.-X. Polyhedron 1997, 16, 453.)
structure is further stabilized by secondary interactions between Cu atoms and the bridging acetone as the continuous strand. It demonstrates that assembled helical structures of copper(1) complexes with tridentate oligopyrazine ligands can be achieved by introducing a suitable spacer groups such as acetone between the metal-binding sites to match the metal ion stereochemical preference.
COPPERU) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
185
n
FIG. 7. The hexagonal framework in [Cu(L6)(Me2CO)os]BF, ( a ) is generated by the helical cyclic array (b), where only copper atoms are shown. (c) Space-filling model of the complex, where BF; are omitted. (From Fig. 4 in Munakata, M.; Wu, L. P.; KurodaSowa, T.; Maekawa, M.; Moriwaki, K.; Kitagawa, S. Znorg. Chern. 1997,36,5416.)
B. DOUBLE-HELICAL COMPLEXES Double helicates constitute the most abundant species among the copper(1) and silver(1) helical complexes. In the presence of the tetrahedral Cu(1) and AgU) ions, two oligopyridines can wrap around each
186
MUNAKATA, WU, AND KURODA-SOWA
other in a double-helical fashion with the metal ions holding them together [Fig. 2(c)]. Rational design of double-helical structures containing up to five metal ions and those with infinite frameworks has been developed by different groups (8,20-22). Among double-helical complexes, dinuclear helicates with two ligand molecular threads twisting around two metal ions predominate. Based on the description of these complexes made by Constable (81, they can be defined as [4 + 41, [5 + 51, and [6 + 41 dinuclear helicates, the two numbers in the square bracket indicating the number of nitrogens to which the two metal ions are bounded. The most common are the dinuclear [4 + 41 helicates with the tetradentate quaterpyridine derivatives. They include complexes of quaterpyridine in (30, 31 ), the sym[CU&)ZI(PF~Z, [AgdL7)~I(PF6),, and [Agz(L7)21(BF4)z metrically methylthio- and phenyl-substituted quaterpyridine deriva(32) ~ 7and } 2 1[CU&)ZI(PF~)Z ~PF6~z (331. To tives in ~ ~ u 2 { 4 ’ , 4 - ~ ~ e ~ ~ 2 probe the steric control in the self-assembly of directional helicates, Constable has carried out systematic studies on asymmetrically alkyl( P F).~It) ~ substituted quaterpyridine derivatives such as [ C U ~ ( L ~ ) ~ I (34 is found that substituents do not control the self-assembly of quaterpyridines but are responsible for the detailed structure. All these dinuclear helicates with quaterpyridines, L7-LB,have general structure (Fig. S), though there are some features as exemplified by [CU~(L,)~I+ slight differences in structural parameters as result of the introduction of the substituents and changes of the metal cations and counteranions. The cation is dinuclear, with each metal ion in a distorted four-coordinate environment, and the two quaterpyridine strands are
FIG.8. Structure of dinuclear double helicate with quaterpyridine [Cu2(L&]+.(From Fig. 6 in Constable, E. C.; Elder, S . M.; Hannon, M. J.; Martin, A,; Raithby, P. R.; Tocher, D. A. J . Chem. SOC.,Dalton Trans. 1996,2423.)
COPPERU) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
187
wrapped around the binuclear core in a double-helical array. It is also observed that n-stacking interactions between aromatic rings play a n important role in stabilizing the double-helical geometry. The fact that the oligopyridine disfavors binding to only a single strand in the system suggests that it contains sufficient molecular information for the recognition of the metal-binding domain in self-assembling processes. Conductivity measurements and UV-visible spectra show that the dinucleating bisbidentate ligand Ll0also forms similar double helices with copper(1) in [ C U ~ ( L ~ O ) ~ I (HC2~0O ~)~ (35). The copper(1) complex with CHzCH2-bridgedbipyridine ligand LI1is ( C ~again O ~ ) the ~ , two equivalent ligands formulated as [ C U ~ ( L ~ ~ ) ~ I and are arranged in a helical manner by complexation with two copper ions ( 3 6 ) .Reaction of L12with Ag(1) resulted in the formation of another [4 + 41 dinuclear double-helical complex [Ag2(L12)z1(PFs)z (37). In this complex, the distorted four-coordinate environment of each silver is maintained, but with weak interactions with a fifth ligand. The helical structure is stabilized by n-stacking interactions between parallel pyridyl rings. Another example worthy of comment is [ C U ~ ( L ~ ~ ) ~ ] utilized ( B F ~ ) ~as, a precursor of the trefoil knot system in which the two coordinating molecular threads, L13 , are well interlaced on two copper(1) centers, forming a dinuclear double-stranded helical complex (38).The synthesis of the first trefoil knot was reported by the same group five years earlier, with the chelate unit, LI3,bridged by (CH2)4 rather than a phenylene group (39). Modification of the linker between the two 1,lO-phenanthroline chelates certainly increased the yield, but in both cases, the organic precursors are found nicely wound and well adapted to the formation of a knot by connecting the appropriate ends of the strands (Fig. 9). The metal-ion stereochemical preference and the compatible ligands are the first elements of helical structural control. Interestingly, L14with pyridine bridging two imidazole apartments formed a dinuclear double-helical complex, IC U ~ ( L ~ ~ ) ~ I (similar C ~ O ~to) ~those , described earlier (40).By contrast, the ligand obtained through the replacement of the pyridine in LI4 by a benzene group gave a dinuclear nonhelical complex with Cu(I), suggesting the fundamental role played by the spacer for the helical twist. Fully understanding the elements that guide the preferential combination of ligands and metal ions can surely allow the controlled assembly of multicomponent systems. Copper(I) and silver(1) complexes derived from odd-numbered oligopyridines present a different situation. Quinquepyridine (LIE)in dinuclear helicates usually involves a rotation about a n interannual
-
188
MUNAKATA, WU, AND KURODA-SOWA
FIG. 9. The dinuclear double-stranded helical complex [CU~(L,,)~I(BF,)~ is utilized as the precursor of the trefoil knot system. (From Scheme 1 in Dietrich-Buchecher, C. D.; Sauvage, J.-P.; Cian, A. D.; Fischer, J. J. Chem. SOC.,Chem. Commun. 1994, 2231.)
C-C bond to separate the ligand into a tridentate 2,2‘ :6’,2-terpyridine and a 2,2’-bidentate pyridine part. A double-helical array of two L15 ligands presents a total of ten nitrogen donor atoms, which can be arranged to create one six-coordinate [tpy + tpyl and one fourcoordinate [bpy + bpyl, or two five-coordinate (tpy + bpy) dinuclear units (8, 21). So far copper(1) dinuclear [5 + 51 helicates are not known because pentacoordination is not the favorable stereochemistry for Cu(1);instead, it forms a [6 + 41 dinuclear helicate with Cu(I1). Reaction of quinquepyridine and [Cu(MeCNWPFsin air gave a brown mixed-valence complex, [Cuz(L16)zI(PF~)3, in which the divalent copper is in an octahedral [tpy + tpy] and the monovalent copper is in a tetrahedral [bpy + bpyl environment (411. The heterodinuclear double-helical complex containing Ag(1) and COW)has been reported. The reaction of the two mononuclear cations, [ C O ( L , ~ ) ( M ~ ~ Hand ),I~+ [Ag(L16)I+, in a 1: 1 ratio gave (42). The helical structure is reminiscent of that observed for the mixed-valence copper complex [Cuz(L,6)zI(PF6)3, in which the Cu(I1) is replaced by Co(I1). In contrast to these observations, a dinuclear [5 + 51 helical structure has been observed in the silver(1) complex [Ag2~MezLl~~zl~C104>z twists around in which the ligand, 6,6””-dimethyl-quinquepyridine, the dinuclear core, with two nitrogens coordinated to one Ag and the other three nitrogens to the second Ag, adopting the usual [2 + 31
COPPERW AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
189
mode ( 4 3 ) .As expected, both silver(1) ions are five-coordinated with a flattened and distorted trigonal bipyramidal geometry. Interestingly, another quinquepyridine derivative with two further phenyl groups a t the 4‘ and 4”’positions forms a [4 + 41 helicate in [Ag{(MePh)zL15}zl (C10&, in which (MePh)zL15actually acts as a tetradentate ligand and the central pyridyl ring plays the role of a rigid spacer rather than a donor ( 4 3 ) . Because spacers linking the oligopyridine units have a notable impact on the versatility of the compounds in accommodating metal ions with different sizes and different geometries, one of the approaches for rational design of helical structures is to use aliphatic ethylene or aromatic 1,3-phenylene groups as spacers (3 6 ). For example, Lls with 173-phenylenea s a spacer linking two terpyridine units exhibits essentially a pentadentate coordination mode in [Ag2(L16)21(PF,)z ( 4 4 ) .The crystal structure reveals a [5 + 51 doublehelical bimetallic silver complex in which each Ag(1) adopts a n irregular five coordination. Terpyridine is the simplest oligopyridine capable of forming a double-stranded helicate. Copper(1) complexes with terpyridine derivatives have been reported. In [Cuz(L17)z1(PF6)2 ( 4 5 ) and [CUZ&~ZI(PF& (461, the ligands have essentially distributed themselves to present bidentate domains to one metal center and a single pyridine donor to the other, to give a [4 + 21 double helicate. The two copper atoms involve different coordination environments, one with a distorted tetrahedral geometry and the second in a approximately linear twocoordinate environment (Fig. 10). Dinuclear double helicates are too numerous to be fully discussed here. In contrast, tri, tetra, penta-, and polynuclear double-helical complexes of copper(1) and silver(1) appeared only in a limited num-
FIG.10. Schematic representation of the two copper atoms involving different coordination environments in a [4 + 21 double helicate.
190
MUNAKATA, WU, AND KURODA-SOWA
ber. A ligand with three bipyridiyl binding sites such as Llo can bind up to three cations, giving a trinuclear helical complex. Trinuclear double helicates [Cu3(Llo)zl(PF,)3 (47)and [Ag3(Lzo)21(CF3S03)3 (48) were prepared by treatment of sexipyridine and the relevant metal salts. Structure analysis and molecular modeling studies indicate that the two ligand groups fit three metal ions inside the doublehelical array by twisting around the helical axis [Fig. 2(d)]. Each metal center is coordinated to a pair of adjacent pyridine residues from each ligand in a distorted tetrahedral environment. The functionalized oligopyridines of Lz0with n = 4 and 5 formed tetra- and pentanuclear double helicates, respectively, with Ag(1) (48) and Cu(1) (49).Their overall structural features are analogous to those of the trinuclear double helicates, but their total length is estimated to be as long as around 22 A and 27 A, respectively. These self-organized nanostructures promise potential applications in the field of functional nanoscale species and molecular devices. Discrete infinite double helices are quite rare in inorganic and coordination chemistry [Fig. 2(e)l. Ciani has reported that reaction of Lzl with Ag' salts in the ratio 2 : 1gave completely different and noteworthy products, one of which is the infinite double-helical coordination polymer [Ag(Lz1)ICF3SO3 (50). The structure shows the balanced packing of left-handed and right-handed double helices of cationic -Ag-L-Ag-Lchains (Fig. 11). The period of the helices is and the two strands are bridged by weak Ag-Ag aurophlic 21.1 ( C P ~ - ~interactions. '~) The copper(1) complex [ C U ~ ( L ~ ~ ) ~ Ihas (C~O~)~ been prepared and structurally characterized (51). The basic dinuclear unit has a structure similar to that observed in [Cuz~L14~zlz+ (40). The ligands bind in a bis-monodentate coordination mode and form the strands of the helix that twist around the helical axis on
A,
FIG.11. Space-filling model of the infinite double helix of [Ag(L,,)ICF,SO,. (From Fig. 2 in Carlucci, L.; Ciani, G.; W. v. Gudenberg, D.; Proserpio, D. M. Inorg. Chern. 1997,36, 3812.)
COPPER(1) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
19 1
which the copper ions lie. Each Cu(1) center is nominally linear coordinated by two imidazole units, and the interactions with the two bridging pyridines are weak. The double-helical subunits stack together in parallel columns of constant but different helicity, leading to infinite double-helical columns. Chiral helical complexes are of interest in connection with the manifestation of functions such as optical activity, molecular recognition, and enantioselective catalysis (52-56). With spirobisindanol and dimethylbiphenic acid as chiral templates, Siege1 has successfully obtained bipyridine based double helicates containing up to three copper(1) ions by enanthoselective synthesis (57). The spectroscopic results confirmed the complex as a single enantiomer of the head-tohead isomer whose stereochemistry is controlled by the single stereoelement of the template throughout a span of 20 A, demonstrating the ability to transmit stereochemical information over nanometer distances in the compounds. Chiral double-helical structures of copper(I1)-L- and -D-arginine complexes with aromatic dicarboxylates have been reported, although the corresponding copper(1) species are not known (58). These crystals reveal that the complex with L-Arg forms a right-handed helix, whereas the complex involving D - h g gives a left-handed helix [Fig. 2(r?l. This indicates that the handedness of the double helices in the solid state is governed by ligand chirality. These and other synthetic routes open new ways for development of enantioselective reagents and may contribute to the understanding of spontaneous aggregation of conjugated helical molecules in biological systems (59, 60).
C. TRIPLE-HELICAL COMPLEXES There is only one example of a triple helix containing Ag(I), reported recently by Williams and his co-workers (61). The colorless crystals [Ag3(L23)51(BF4)3 were obtained by diffusion of benzene into a solution of the compound in acetonitrile. The structure consists of an equilateral triangle of silver ions with the ligands bridging the sides of the triangle [Fig. 2(g)l. The silver ion is not in a strictly linear environment, with an N-Ag-N angle of 153.3(6)”. Each ligand binds to one metal from below the plane of the silver atoms and to a second metal from above the plane. The structure may be considered as a triple helix in which the ligands wrap around the threefold axis and are held in place by coordination to the Ag, triangle.
192
MUNAKATA, WU, AND KURODA-SOWA
Ill. S - - 4Contact-Assembled Frameworks
Some sulfur-rich dithiolates and the tetrathiafulvalene (TTFI-based species involved in the following discussion are listed in Fig. 12. Following the observation that TTF (L2&forms a stable radical when treated with chlorine (62) and the discovery that a partial charge transfer between separately stacked donor and acceptor molecules leads to the first organic metal, TTF * TCNQ (tetracyanoquinodimethane) (63), the use of sulfur-containing molecules as precursors for synthesis of conductive or superconductive materials has been of unabated interest for over two decades. Because the large sulfur atomic orbitals are capable of promoting effective intermolecular overlap in the organic metals, partial oxidation of TTF and its substituted derivatives, when coupled with suitable electron acceptors, can give rise to low-dimensional arrays of these donors through sulfur-sulfur molecular interactions, providing an effective pathway for electronic conduction. A tremendous number of studies have been carried out on the
L27
M e 0 2 C f ~ s Me02C
s
LZ8
s-
'Xs F3c1sNcxs S
S
F&
L29
S'
NC
S
L30
L31
.=(Ise-
0
Se'
S
0 L32
L33
L36
L34
L35
lTC,-lTF
FIG. 12. Sulfur-rich dithiolates and TTF derivatives.
L38
COPPER(I) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
193
solid-state conductive properties of the TTF . TCNQ-like organic and inorganic metals by modifications of TTF (64-66). By contrast, the study of TTF-based species as building blocks in supramolecular chemistry has attracted attention only recently (67). With a preference for a soft donor such as the sulfur atom, both silver(1) and copper(1) readily form metal complexes with the sulfur-rich TTF-based organic molecule, with a high possibility of generating one- two- and three-dimensional supramolecular and macromolecular system by the covalent bonds as well as by S----Scontacts. In general, the S----Scontacts considered to be effective in sulfur-rich conductive compoundso refer to those having interplanar S----S separations a t 3.35-3.75 A, close to the sum of two sulfur-atom van der Waals radii (3.60A). This section focuses on construction of the supramolecular frameworks through S----Scontacts based on dmit, BEDT-TTF, TTC,TTF, C5H4S,and the related species.
A. dmit
AND THE
RELATED LIGANDS
The sulfur-rich 1,3-dithiole-2-thione-4,5-dithiolate (dmit) (LZ5)has received considerable attention. Using square planar coordinating metal ions such as Ni, Pd, and Pt, many electrically conducting materials derived from planar metal complexes [M(dmit)J- have been synthesized, and some of them have been found to undergo a superconducting transition at low temperature by applying high pressure. The nonplanar bulky metal complexes [M(drnit),l"- (M = Fe, Re, Mo, or W) have received equal attention. Numerous papers and reviews have focused on the study of the role of the intermolecular and intramolecular interactions in determining the electrical properties in the system (68-83). As a n example, Fig. 13 shows a uniform segregated stacking arrangement of the constituent Ni(dmitI2 units with the open-shell 7r-acceptor TTF and the closed-shell cation (CH3I4Nin the superconductors (TTF)[Ni(drnit)& (84) and [(CH,),N][Ni(dmit)zlz(851, respectively, where close intermolecular S----Scontacts occur between units in adjacent stacks leading to a n infinite network. Furthermore, extensive studies have also been carried out by using other ligands resembling dmit, such as L26-L33 (86-92). Along the same line, the selenium analogs (L34and L35)are also of much interest, because even more effective molecular interactions are expected to occur owing to Se having more extended orbitals than sulfur (93,941.
Although a number of Cu(I1) complexes with the dmit-like ligands have been reported in which the metal ion involves a significant dis-
194
MUNAKATA, WU, AND KURODA-SOWA
0
b
(b) FIG. 13. Molecular packing of superconductors (TTF)[Ni(Lp)& (a) and [(CHAN] [Ni(La&lz(b). (From Figs. 5 and 6 in Bousseau, M.; Valade, L.; Legros, J.-P.; Cassoux, P.; Garbauskas, M.; Interrante, L. V. J. Am. Chem. SOC.1986, 108, 1908, and Figs. 1 and 4 in Kim, H.; Kobayashi, A.; Sasaki, Y.; Kato, R.; Kobayashi, H. Chem. Lett. 1987, 1799.)
COPPERW AND SILVERU) SUPRAMOLECULARMETAL COMPLEXES
195
tortion from planar to tetrahedral (73, 83, 89, 931, few nonplanar Cu(1) or Ag(1) metal complexes of dmit are known. The most interesting of these is [mpy]z[Cu4(dmit)31, prepared, like its Cu(I1) analog, by a route of direct reaction of coppert11 salt and Nazdmit in the presence of excess of [mpylI (95). The structure contains dimerized anion units, each consisting of a tetranuclear Cuds6cluster. Each Cu(1) ion involves a distorted tetrahedral coordination comprised of four dmitsulfur atoms. The dimeic units further interact with each other through close sulfur-sulfur contacts to form a two-dimensional molecular interaction sheet. The cations are located between the anion sheets. B. BEDT-TTF BEDT-TTF (L36) is another electron donor molecule and in combination with inorganic anions has provided several air-stable superconductors at low temperature (96-99). In fact, both dmit and L36 superconductors have the characters of (i) good orbital overlap between extended r-electron systems (S----Scontacts) and (ii) partial filling of the conduction band through either partial oxidation or partial reduction. Here again the radical cation salts of square planar transition metal complexes dominate, and they feature S----S interactions between planar L36'+ molecules leading to segregated columns of donors. On the other hand, the reported copper(1) and silver(1) coordination complexes of L36 are scarce. They include (LSt)Cu2Br3,prepared by a redox reaction between L36 and [CuBr4I2-(100).The most unique feature of the complex is that the L36 molecules are not stacked in a column but are coordinated to the tetrahedral copper(1) centers in a -Cu-Br-Cu-(p-Br)z-Cuchain (Fig. 14). Due to lack of close S----Scontacts between radical cations, the complex shows low conductivity .
FIG.14. Structure of (L,+)[Cu2Br31.
196
MUNAKATA, WU, AND KURODA-SOWA
C. TTC,-TTF SYSTEMS Several copper(1) and silver(1) complexes with tetrakis(alky1thio) tetrathiafulvalene (TTC,-TTF) (L3J have appeared in the literature, and various S----Scontact-assembled frameworks have been observed. L37 contains a central TTF (L2J r-system, which keeps the electrondonor property. As single-component organic semiconductors with low electrical conductivity, their physical properties have been widely investigated. They can be easily oxidized to the stable radical cation; accordingly, their chemistry is dominated by electrotransfer processes. They react with organic electron acceptors such as TCNQ and HCBD (hexacyanobutadiene), and inorganic oxidants to form partially or completely oxidized materials containing L37’+, where p ranges from about 0.6 to 2.0 (101-104).In most of these chargetransfer compounds the donors and acceptors are stacked in segregated columns. Construction of two- and three-dimensional coordination polymers using the diversity of copper halide frameworks is an area of considerable importance. Besides its electron-donor character, Lg7possesses alkylthio groups, which have coordination ability to metal ions through the sulfur atoms. This gives Lg7species a unique ability in linking metal ions to form coordination polymer structures. Six Lg7 complexes with copper(1) halides, [(CUX)~(L~,)] (where n = 1 or 2 and X = C1-, Br-, or I-), have been structurally characterized (105-107). The syntheses of these compounds are straightforward, usually carried out by direct reaction of copper(1)halide and Lg7in a nonaqueous solution. The structural determination of these complexes reveals three different copper halide frames, namely rhomb, helix, and zigzag chain (Table 11).Although the metal ions in all cases involve a tetrahedral coordination comprising two halogen atoms and two sulfur atoms of the Lg7moiety, with a certain degree of distortion, which in turn acts as a bidbidentatel linker between the two metal centers through alkyl thioether groups, the resulting compounds often exhibit for exnovel structures dramatically changed by the halogen. In LgTa, ample (1051,the chloride complex consists of novel two-dimensional sheets of L37a molecules arranged between zigzag frames of {CuCl}, . Within the two-dimensional sheet no significant S----Scontacts are observed because the long Cu----Cuseparation of 5.78 A on the same side of the CuCl zigzag frame gives a long interplanar spacing (4.90 A) of Lg7,,. However, between the parallel sheets there are close intergiving a three-dimensional netmolecular S----Scontacts of 3.53 work (Fig. 15). By contrast, a novel framework of {CuBr},, helixes
A,
COPPER(I) AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
197
TABLE I1 BONDANGLESAND CuX FRAMEWOKKS I N TTC,,-TTF(Ls,) COMPLEXES OF COPPEKU) HALIDE 01
P
Y
6
113.4 116.0" 109.1" 112.6"
113.1" 111.8 113.6 106.5
88.8 89.9 91.1"
116.1 116.1
116.8 112.5
111.8
113.2
CuX framework
Ref.
rhomb rhomb rhomb rhomb
105 107 106
87.1 86.9
helix helix
105 106
90.9
zigzag chain
105
108
Average value.
< =c,X
-c u
X
rhomb
infinite helix
zigzag chain
which are further linked to each other by coordination of L37ato the metal was observed in the bromide complex. The L37amolecules are stacked together, with the dihedral angle between the neighboring molecules being essentially zero on the stacking axis. The unique feature of the structure is that the {CuBr}, helixes orientate alternately in the reverse direction and the L37amolecular stacks are mutually perpendicular to each other along the c-axis. The S----S contacts of 3.68 A between the stacked L37amolecules lead to a two-dimensional framework. The iodide complex contains an infinite chain of the bent LS7, molecules perpendicularly bridged on the methyl thio2ther groups by rhomboid CuzIzmoieties. The S----Sdistances of 3.75 A between the chains indicate weak interaction present, associating the adjacent chains of -L-Cfi-Ltogether, giving a two-dimensional network. Similar results have been observed in the corresponding LS7,, complexes (106, 1071, in which the chloride complex involves helical
198
MUNAKATA, WU, AND KURODA-SOWA
3.53.i
GI
!
12Dsheet
I
-
I uc’ --uc
II
CI I
C-ICU-
I
I
I
FIG.15. Schematic views of different frames of CuX in [(CuX),(L,,,)I complexes.
{CuCl}, frames whereas the bromide and the iodide both contain CuzXzrhombs. Bond angles of copper(1) halide complexes with LSmare summarized in Table 11. It can be seen that the main factor determining the framework of the CuX unit seems to be connected with the Cu-X-Cu angle (a).As a increases smoothly from 66.6 to 1 1 1 3 , the framework changes from rhomb to zigzag chain through infinite helix, whereas other angles do not show such correlation. The L37a+ cation coordination complex (L37,)[CuBrz1zalso falls nicely into this category (108). As a comparison, it is instructive to briefly mention the corresponding LS7b+complexes of copper(I1) halide (L3&[Cu2&1 (where X = C1or Br-) prepared by oxidation of L3%with CuC12 or CuBrz (109). Unlike the copper(1) halide complexes with LS7bridging between two metal atoms, both structures consist of two segregated stacks of L3%+ donors and CU&~- acceptors associated via short S----Scontacts (Fig.
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
199
FIG. 16. Packing diagram of (L,,)[Cu,CI,I showing S----Sand S----Cl.(From Fig. 1 and Fig. 2 in Wu, L. P.; Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y. Mol. Cryst. Liq. Cryst. 1996,285, 75.)
16). The donors face to each other to assume a dimeric structure, with close intradimeric S - - 4conducts of 3.43-3.52 A. Consistent with the absence of significant interdimeric S----S interactions, the two compounds exhibit low electric conductivity. The diversity of copper halide frameworks, affected by several elements such as the metal ions, coordination of the halogen atoms, and the stereofactors of the ligand, makes it possible to design two- and three-dimensional coordination polymers. Copper(1) perchlorate and tetrafluoroborate salts have also been found to form coordination compounds with the neutral ligands L3, (n = 2 or 3), but the structures vary little due to nonparticipation of the counteranions C10,- and BF,- for their weak coordinating ability (110, 111). In fact, they are 1:1 metal-ligand compounds in which
200
MUNAKATA, WU, AND KURODA-SOWA
each Cu(1) ion is tetrahedrally coordinated by four sulfur atoms from the bridging Lg7molecules to form infinite polymeric chains separated from each other by the anions. The shortest S----Sdistances between suggesting negligible the adjacent chains in all cases are over 3.8 S----Sinteractions in the system. Same conclusion can be drawn from the Ls7, complex [Ag(L3,,)(03SCF3)I(1121, in which each Ag(1) ion is six-coordinated by four sulfur atoms from the bridging Lg7*and two oxygen atoms from the bidentate 03SCF3anion, resulting in a structure of two interwoven polymericochains. The closest S----Sdistance between the two chains is 4.00 A, precluding any significant S----S contacts present. Apart from the varieties of interesting structures observed in the Lg7system, it is worth noting the preparation of these complexes and the conductivity of the Iz-doped species. As mentioned previously, these complexes are easily prepared without elaborate synthetic procedures or particular precautions. However, successful preparation of the single crystals of these complexes largely depends on the reaction temperature. Again take Lg7,, for example. The preparation was carried out by allowing the mixture solution of CuX salt and L37ato stand in a water bath at a constant temperature of 70, 55, and 23°C for the chloride, bromide, and iodide, respectively (105).The same reactions performed at other nondesignated temperatures may lead to a mixture of precipitate and tiny single crystals unsuitable for X-ray analysis. This is also true for the preparation of the corresponding L,,, complexes (106,107).Iodine oxidation is a viable synthetic route to highly conducting mixed-valence materials. Although copper(1) and < silver(1) complexes with neutral Lg7 ligands are insulators S . cm-') at room temperature, their iodine-doped products behave as semiconductors, with conductivities increasing in the order [(CUC~)Z(L~~)I < [(CuBr),(L,,)I < [ ( C U I ) ~ ( L(Fig. ~ ~ ) I17). A partial oxidation can be assumed to occur in the 12-dopedspecies and the mixedvalence ligand-ligand interactions might cause electrical conduction pathways.
A,
4,5-Ethylenedithio-l,3-dithiole-2-thione (C5H,S5)(Lss)is well known as an electron donor (113, 114).It is also a derivative of dmit and has a structure similar to half of LSs. The structure determination of its metal complexes has demonstrated the unique coordination versatility of Lg8(Table 111). It can act as a monodentate, bidentate, or even
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
201
-1.5
-1.7 -2.1
E
-
%? D
-2.5
-rn 0
lrdoping
.-c>i .-
12-doping
c)
0
3
'0 C
8
-3.6 -
-3.5
orange crystal
= CIg
Bry
I
orange crystal
orange crystal
L37a = tetrakis(methy1thio)tetrathiafulvalene
FIG.17. Preparation scheme and conductivity of [(CuX)2(Ls~,)l.
tridentate ligand, binding one and bridging two or three metal ions through the thiocarbonyl group alone or by both thiocarbonyl and thioether groups. Up to now, three silver(1) and one copper(1) complexes of L38have been synthesized and crystallographically characterized. They display a variety of one- to three-dimensional frameworks based on coordination bonds as well as intermolecular S----S contacts. Reaction of AgC104.H20 and L,, in acetonitrile gave red brick crystals of {[Ag(L38)31C104~CH3CN}2 (115).The compound has a dimeric structure
202
MUNAKATA, WU, AND KURODA-SOWA
TABLE 111 COORDINATION MODESOF C5H4S5(LS), THE METALCOMPLEX STRUCTURE AND THE CONDUCTIVITY OF THE I 2 - D 0 pPRODUCTS ~~ Mode"
Structure of the complex
Conductivity S . cm-'
Ref.
[Agz(Lss)J(C104)~ I, (11) 111 [Ag(Lsd,NO,I [ A ~ ( L s ~ ) C F ~ S O ~ IV I I, I11 [CuaI4(Lsa)al
dimer infinite dinuclear chain 2-D sheet infinite tetranuclear chain
6.9 X 3.0 x 10-5 1.5 x 10-4 2.2 x
115 116 115 117
Complex
_______
~
Coordination modes of C5H4S5.
n s
g
sKs S
IM
mode I
A
kS-" sHs-M A
A
sKs
sKs
s)-+s
sKs JSI
M M mode I1
S
IM mode 111
JSI
M M mode IV
with two nonequivalent intercoordinated quasi-trigonal [Ag(L3&1+ units, which are further assembled by very close S--------S contacts of 3.284(4)-3.569(2) to form a one-dimensional chain structure (Fig. 18).Replacement of the anion by triflate in preparation gave a poly-
A
FIG.18. Schematic representation of S----Scontacts in {[Ag(LS)31C104. CH,CN}, .
COPPER(I) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
203
nuclear complex with a different stoichiometry, [Ag(L,,)CF,SO,I ( I 15). The structure contains two crystallographically independent five-coordinate silver(1) ions, very rare stereochemistry for Ag(1). Both the L3, group and the CF,SO, ion bridge between two Ag atoms, forming a one-dimensional chain along the c-axis. All the ligand groups are approximately perpendicular to the chain, and between the c h a i p there are several close S----Scontacts (the shortest one being 3.262 A), resulting in a two-dimensional polymer structure (Fig. 19). The third silver(1) complex belonging to this family is [Ag(L,,),NO31 (116).Each L,, displays a bidentate fashion through both thiocarbonyl sulfur and thioether sulfur atoms bridging two octahedral silver atoms, giving a n infinite chain. The strong S----Sinteractions with a distance of 3.41 A between two nearby chains along the b-axis generate a twodimensional network. The corresponding copper(1) complex [ C U ~ I ~ ( L ,was ~ ) ~prepared I by reaction of L3, and CuI in acetonitrile (117). The structure consists of stepped Cu414clusters bridged by L3, groups, and the resultant polymeric chains are further assembled by short interchain S----Scontacts (3.257 A)to form a three-dimensional network (Fig. 20). Unusually short S----S contacts have been noticed among theo L38 complexes; for example, 3.23 A in (TTF)[Pt(L3&I3 and 3.26 A in (TTF)[Pd(L3d212,[Ag(L3dCF3S031, and [ C U ~ I ~ ( L(115). ~ ~ ) ~ The I extremely short S----Scontacts are attributable to metal-metal bonding in the first two cases, and to the coordination bond linkage as well as the effect of molecular packing in the rest. The copper(1) and silver(1) complexes of L3, were partially oxidized by iodine doping, and the electrical resistivity of the compressed pellet of the compounds as measured by the conventional two-probe technique. Like those of L3,, they are nonconducting a t room temperature but display semiconduc-
FIG. 19. Schematic representation of S----Scontact assembled network in IAg(Ls8) (CF8OdI.
204
MUNAKATA, WU, AND KURODA-SOWA
54
D: S9...S9' (3.62A)
FIG.20. Structure of rCu4I4(L8&1.
tivity after Iz doping, with conductivities of 10-4-10-5 Sacm- (Table 111).
IV. Hexagonal Frameworks and Graphite-like Structures
The scaffolding and related microporous materials are receiving widespread interdisciplinary attention, and the borders between chemistry, physics, and materials science are therefore vague and
COPPER^ AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
205
subjective. Several general reviews and textbooks have provided a comprehensive overview of the chemistry and structural properties of these materials (118-120). The literature relating to assembly of copper(1) and silver(1) complexes into network structures with a graphite-related lattice has reflected a general interest and advancement in our understanding of the fundamental chemistry and processing characteristics of the system and its potential applications as molecular sieves, catalysts, and optical materials. This section is concerned with construction of hexagonal and graphite-like lattices regulated by the metal-ion stereochemistry, the ligand, and the counteranions, placing emphasis on the nitrogen-containing systems such as pyrazine, phenazine, 2,1,3-benzothiadiazoles,and quinoxaline (quin) (Fig. 21). Metal cyanide will be treated in a cursory manner, leaving the readers to consult recent literature and reviews.
A. METALCYANIDE-REGULATED BY THE METAL ION Metal cyanide, which has received little attention for years in many texts concerned with inorganic structures, is now found to possess the most outstanding feature as a building block for infinite polymeric frameworks. Some rather elegant metal cyanide structures with extended frameworks have recently been discovered (121-128). The synthetic strategy is based on the combination of a diatomic bifunctional rodlike CN- ion with a considerable preference for binding metals at each end in a linear fashion together with modification of
L39
(Pz)
L40
L43
L 41
L45
L44
L46 (phz)
L42
L47
(bW
L48
FIG.21. Some nitrogen-containing ligands in copper(1) and silver(1) complexes having hexagonal frameworks.
206
MUNAKATA, WU, AND KURODA-SOWA
the frameworks by metal ions favoring different stereochemistries. The crystal structures of [N(CH,),I[Cu'Zn"(CN)~]and two isomorphous compounds, Zn(CNI2 and Cd(CNI2,have shown that the tetrahedral metal ions favor formation of interpenetrated diamondlike frameworks (127). A three-dimensional honeycomb structure containing hexagonal channels has been observed in complexes consisting of simple cyano-cadmium frameworks in which the metal centers are of two types: octahedral and distorted five-coordinate or octahedral and tetrahedral (122,128).By replacing CN bridges in multidimensional CNlinked octahedral cadmium complexes with linear NC-Ag-CN or NC-Ag(CN)-Ag-CN units, Iwamoto has successfully obtained two cadmium-silver complexes with double and triple interpenetrating frameworks (123). An infinite three-dimensional framework related to PtS has been observed in [NMe,l[CuPt(CN),], in which tetrahedral Cu(1) and the square planar Pt(I1) centers are linked together by CN rods, generating two mutually perpendicular and equivalent sets of hexagonal channels of large cross section together with large empty square channels perpendicular to the hexagonal channels as shown in Fig. 22 (121).
B. PYRAZINE SYSTEMS-REGULATED BY SUBSTITUENTS Pyrazine (LS9)and substituted pyrazines have long been known to act as em-didentate ligands to linearly bridge metal ions, generating oligomeric and polymeric metal complexes with infinite chain and
FIG. 22. Extended framework containing hexagonal and square channels observed i n [CuPt(CN)J:-. (From Fig. 2 in Gable, R. W.; Hoskins, B. F.; Robson, R. J. Chem. SOC. Chem. Commun. 1990, 762.)
207
COPPER(1) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
pleated sheet structures, double and triple interpenetrating frameworks, and interwoven honeycomb architecture (123, 129-132 ). The essential features of some structurally characterized copper(1)and silver(1) complexes with pyrazine and its derivatives including quinoxaline and phenazine are summarized in Table IV. Among them those containing two-dimensional six-membered rings of metal ions linked by LBSbridges are rather common and thus have received considerable attention recently. From the point of view of rational synthesis, it would be possible to obtain polymeric graphite-related cationic lattices in which the three-coordinate metal ions are bonded only to bridging LBSligands by utilizing noncoordinating anions as the counterions in the synthesis of LSs complexes. However, owing to the effects of substituents, the six-membered rings obtained differ in detail from complex to complex in terms of distortion of planarity, which can be grouped under four heads as shown in Fig. 23. It has been suggested that more than two substituents on a pyrazine ring would hinder four coordination of the metal ion, giving infinite zigzag chains or sheets of the cross-linked chains as observed in disubstituted pyrazine systems (135).Therefore, controlling spatial factors plays a key role in building infinite polymeric structures, and the modification of linking ligands by substituents on pyrazine is a practical method for the synthesis of designed copper(1) polymers. Slightly distorted hexagonal frameworks defined by six LBS-bridged copper(1)ions are observed in [CU~(L,,,)~(PF,)~I (L40= 2,5-dimethylpyrTABLE IV Cu(1)AND Ac(1) COMPLEXESWITH PYRAZINE AND ITS DERIVATIVES Compound
CN"
Frameworks
Ref. ~
[Cuz(L,o)dPFs)zl four three [cuz(L41)31(c1o4)2 [Cu2(~~31(c104~, three [Cuz(Lsd31SiFs three ~ C ~ Z ( L ~ , ~ ~ I ( C I O , ~ ~ ( M four, ~ ~ C five O)~ [C~,(LS~)~(CH~CN)ZI(PF~)Z four four [CUZ(L,)4 sl(ClO4)z four [cu(L,o)zlPF~ [Cu,(4,)31(C104)z three, two ~cuz~L4~~31~c1o~~1CIo4 four [Agz(La)(NO3)zI three [Agz(LddBFdz three
2-D hexagonal 2-D hexagonal 2-D hexagonal 2-D hexagonal 2-D hexagonal 2-D hexagonal 2-D hexagonal, square 3-D adamantane linear 3-D adamantane 2-D hexagonal 2-D hexagonal, 3-D interpenetrating nets
' CN, coordination numbers of the metal ions.
133 134 135 136 29 137 135 133 137 132 138 130
208
MUNAKATA, WU, AND KURODA-SOWA
FIG.23. Cu(1) and Ag(1) complexes of pyrazine and the derivatives having hexagonal frameworks: (a) slightly distorted hexagon; (b) distorted hexagon and square; (c) zigzag 2-D sheet; (d) ternary complex. The molecule in the box represents the ligand involved.
azine) (133), [ C U ~ ( L ~ ~ ) ~ ] (L41 ) ( C ~=O2,3-dimethylpyrazine) ~)~ (1341, [ C U ~ ( L ~ ~ ) ; J ((L42 C ~ O=~2,6-dimethylpyrazine) )~ (135), [Cu2(L3&lSiF6 (136),and [ C U ~ ( L ~ , ) , ] ( M ~ (L43 ~ C O=) ~pyrazineamide) (29).Structural studies have established trigonal environments for the metal. These complexes show an infinite sheet structure of macrocations with uninteracting or weakly interacting counterions. If the interaction between the metal and the anion is ignored, the Cu(1) ion in these com-
COPPER(1) AND SILVER(I1 SUPRAMOLECULAR METAL COMPLEXES
209
plexes is essentially in a three-coordinate trigonal planar geometry and the pyrazine ligands bridge adjacent trigonal metal centers to form a graphite-like lattice, Fig. 23(a). Thompson has successfully isolated two Cu(1) complexes from the same reaction mixture of an ethanolic solution of Cu(PF& and L 4 0 , [Cu2(L40)31(PF,)2 and [ C U ( L ~ ~ ) , I P F ~ (133).The structure of the former is made up of parallel puckered layers of the cation, and within each layer the copper(1) ions are linked by L40,defining hexagons. PF, anions are located between the layers and interact very weakly with the copper(1) ions through one F atom, leaving the Cu(1)ions essentially three coordinate. Except for the puckering of the layers, the structure is reminiscent of that of the intercalated graphite. By contrast, each metal ion in the latter case is bonded to four rather than three L40 molecules, with an angularly distorted CuN4chromophore. As a result, the structure exhibits a diamondoid framework of copper(1)ions linked by the rodlike L 4 0 ligands, and the PF; anions are located in cavities within the lattice. Likewise, the reaction of Cu(C104)2with L41 in an aqueous solution yields an air-stable complex, [ C U ~ ( L ~ ~ ) ~ (whose C ~ O ~structure ) ~ ] , is also found to contain a similar two-dimensional graphite lattice of tricoordinate copper(1) atoms bridged by L1lligands (134).Although the exact course for reduction of copper(I1) complexes with nonchelating ligands is not clear, the most plausible explanation lies in tetrahedral distortion of the square planar copper(I1) species by crowding effects of the methyl groups, and consequently, such reduction process is dependent on the substituents, temperature, and solvents (233, 134 ). On the other hand, copper(1)complexes of L39can also be prepared by direct reaction of copper(1) salts, [CU(CH,CN)~]X, and L3, ligands in nonaqueous solutions, such as [CU,(L~~),]S~F, (136).Here again the L3, ligands bridge adjacent trigonal Cu' centers to form hexagonal frameworks with bridged copper cations 6.936 and 6.685 A apart. The SiFi- anions are located in channels and are hydrogen bonded to L3, hydrogen atoms. The most significant feature of the complex is that the hexagonal frameworks are interwoven as shown in Fig. 24. The three-coordinate copper(1)ions can also be linked by the disubstituted Ls9on the 2- and 6-positions as observed in [ C U ~ ( L ~ ~ ) ~ , I ( C ~ O ~ ) , (135).The graphite-like lattice is constructed by the alternating arrangement of the two crystallographically independent three-coordinate copper atoms. The hexagons generated by the Cufi units are slightly distorted due to compression, and the mutually confronted Lea molecules on the sides of the hexagon are sitting out of the plane and parallel to each other, providing a large cavity. It is surprising that neither the anion nor the solvent molecule is included in the cavity,
2 10
MUNAKATA, WU, AND KURODA-SOWA
.......
I................._
i cuqubcu = 1 1
0
LStl C(
:i
/CU-dcu,
u
f c-cu*k
...........................
FIG.24. Schematic view of the interwoven honeycomb grids in [ C U ~ ( L ~ ~ ) (From ~IS~F~. Fig. 4 in MacGillvray, L. R.; Subramanian, S.; Zaworotko, M. J. J. Chem. SOC.Chem. Commun. 1994, 1325.)
they rather form acetone-perchlorate layers intercalated between the copper sheets. Formation of trigonal copper(1) polymers with dimethyl-substituted pyrazines is presumably due to the two substituents on the L3,ring, which hinder the formation of tetrahedral coordination of the metal. No examples of hexagonal frameworks are found to be built up by tetrahedral copper(1) ions. This is understandable because formation of graphite-like sheets would require Cu6motifs on a plane with each metal center linked to three neighbors. Therefore, trigonal, trigonal pyramidal, and trigonal bipyramidal stereochemistries are necessary. The recently characterized copper(1)complex [ C U Z ( L ~ ~ ) ~ ~ ( C ~ ~ ~ ) ~ ( provides the first example of mixed four- and five-coordinate copper(1) ions composing such a hexagon (29). As illustrated in Fig. 25, the cation contains two crystallographically independent copper(1) ions; one involves a distorted trigonal pyramidal CuN30 core composed of the terminal pyrazine nitrogens of three different L43molecules and an axial caboxamide oxygen, and the second involves a distorted trigonal bipyramidal structure with two, rather than one, carboxamide molecule exhiboxygen atoms occupying the apical positions. Each L43
COPPER(1) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
2 11
cu FIG.25. Molecular structure of [ C U ~ ( L ~ ~ ) ~ I ( C I O ~ )in~ which ( M ~ ~the C Ofour) ~ an d fivecoordinate Cu(1) ions are bridged by the tridentate La groups (a) giving a hexagonal channel structure with pores open to accommodate (2104 ions and acetone molecules (b). (From Figs. 1 and 2 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Moriwaki, K.; Kitagawa, S. Znorg. Chem. 1997,36, 5416.)
its a tridentate coordination mode, bridging two metal centers, giving apparently hexagonal channels with pores open to accommodate the anions and acetone molecules as guests. The complexes falling into other groups as shown in Fig. 23(b)-(d), are significantly less numerous. The chlorine-substituted pyrazine is
212
MUNAKATA, WU, AND KURODA-SOWA
reported to form a coordination polymer containing both distorted hexagonal and square frameworks. The reaction of 2-chloropyrazine (La) with [Cu(CzH4)C1O4] in acetone yielded orange brick crystals, . Structural studies have shown both copper(1) ions [CU,(L~~),,I(C~O~), involved tetrahedral environments (135).As L3g and the dimethylsubstituted derivatives do, each L, molecule exhibits a bidentate coordination bridging two metal centers forming a square planar Cu4 framework. These square planes are linked to each other by sharing the edges to afford one dimensional ribbons. The adjacent ribbons are further bridged by L44,leading to two-dimensional pleated hexagonal Cue sheets (Fig. 26). The results further demonstrate that in metal complexes of LSgderivatives, the substituents on pyrazine can greatly
L
Cu(.
FIG. 26. Schematic drawing of infinite sheets in [CU~(L&](C~O~)~. Copper atoms and bridging Ld4molecules are denoted by open circles and solid lines, respectively. Each pyrazine ring is distinguished by alphabetical label. (From Fig. 2 in Kitagawa, S.; Kawata, S.; Kondo, M.; Nozaka, Y.; Munakata, M. Bull. Chem. SOC.Jpn. 1993, 66, 3387.1
COPPERU) AND SILVER(J) SUPRAMOLECULAR METAL COMPLEXES
2 13
affect the stereochemistry of the metal, which in turn plays a key role in determining the framework of the structure. The copper(1) complexes with pyrazine and tetramethylpyrazine (LQ5) may be the best examples to elucidate the steric effects imposed on the substituted pyrazine toward rational synthesis of copper pyrazine polymeric complexes in the single-crystal phase (137).The single crystals of both [ C U ~ ( L ~ ~ ) ~ ( C H S C Nand ) ~ ~[( C P FU~~)(~L ~ ~ ) ~ I ( C ~ O ~ ) were prepared by reaction of the corresponding ligand and the copper(1) salt in acetone. In contrast to the disubstituted pyrazine complexes, in which the metal ions in most cases are found to be in a trigonal environments, the geometry around the copper atom in [CU~(L,~),(CH~CN)~I(PF,), is a distorted tetrahedron comprising three L30 nitrogens and one bent bonding of CH,CN. Consequently, the hexanuclear unit Cu6 present in the infinite cationic sheet is in a chair-type cyclohexane-like framework [Fig. 23(c)l. On the other hand, the L,,-bridged copper(1) complex [ C U ~ ( L ~ ~ ) ~ Icontains ( C ~ O ~a) ~ linear chain polymeric framework. The two copper atoms in the unit cell have different coordination environments, one with trigonal planar and the other with perfect linear geometry. The two-coordinate Cu(1) ions form a linear chain framework, and the trigonal ions are attached to the chain just like a pendant (Fig. 27). The fact that a hexagonal or cyclohexane-like framework of copper atoms is not present in the cation is due to the bulky Lq5ligands. The four methyl groups on all the substitutable positions of pyrazine give rise to steric constraints and, as a result, prevent formation of a hexagonal ring unit and give a linear link as a less steric form.
y-N\ 112.1"
cu\
cu
I
I
112.10
C"
I
C"
\cu'cu\cu~c"\cu~
I
I
I
5" 7" 1"
FIG. 27. Schematic views of the chairlike hexagonal framework in [Cu2(L,,), (CH,CN),I(PF& (a) and a n infinite chain structure in [Cu2(LQ)31(C104)2 (b). (From Fig. lb in Kitagawa, S.; Munakata, M.; Tanimura, T. Znorg. Chem. 1992,3I,1714.)
214
MUNAKATA, WU, AND KURODA-SOWA
C. PHENAZINE AND BENZOTHIADIAZOLE SYSTEMS-REGULATED BY COUNTERANIONS
Like pyrazine, phenazine (La), 2,1,3-benzothiadiazoles (L4,), and quinoxaline (L4&are of interest in the construction of the molecular assembly of the metal complexes because they can also act as bridging ligands linking two metal centers (138, 139). Phenazine formed 1: 1 donor-acceptor complexes with PMDA (pyromellitic dianhydride) and TCNQ, whose structures show alternate stackings of donor and acceptor molecules (140, 141 ). Several copper(1) and silver(1) complexes of La have been prepared and their structures determined, and among these silver(1) nitrate, [Ag2(L4,)(NO,),],is of particular interest (138).The compound was prepared by reaction of AgNO, and L46 in methanol at ambient temperature. In this complex each silver atom is coordinated to two nitrate anions and one L46 molecule in a distorted trigonal geometry, and each L46 molecule in turn links two trigonal silver ions as expected, Fig. 23(d). The most remarkable feature of the ternary complex is that the counteranion NO, acts as a spacer linking two L16-bridged Ag atoms, leading to a framework of sixmembered rings of silver atoms extending in the direction of the band e-axes (Fig. 28). The Ag-0 bond distances of 2.432(6) and 2.569(7) A are in the range of 2.367 to 2.689 A for Ag(1) complexes with nitrate, suggesting fairly strong interaction between the metal and the nitrate ion. The six Ag atoms on a two-dimensional sheet lie
-.
0
I3----jJ
c
FIG. 28. Hexagonal framework in [Ag,(L,,)(NO,),I is constructed by six metal ions interconnected by L, and NO;. (From Fig. 10 in Munakata, M.; Kitagawa, S.; Ujimaru, N.; Nakamura, M.; Maekawa, M.; Matsuda, H. Inorg. Chem. 1993,32,826.)
COPPER(I) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
2 15
approximately on the same plane, and each sheet is weakly linked to its neighbors by the interaction of Ag and the third oxygen atom of the nitrate anion with an Ag-0 distance of 2.669(7) A, giving a threedimensional structure composed of the two-dimensional sheets. By contrast, in the corresponding silver perchlorate complex of L46, [Ag(L4&1O4, the perchlorate ions are not coordinated, and subsequently the complex is composed of infinite -Ag - -Ag - linear chains along the a-axis and a a-a interaction of the L6 aromatic rings between the chains. It is informative enough that in the mononuclear copper(1) nitrate complex of L 4 6 , [ C U ( L ~ ) ~ ( N Othe ~)], Cu(1) ion is coordinated to two terminal L46 molecules and chelated by two oxygen atoms of the nitrate in a distorted tetrahedral environment. The structural differences between the copper(1) nitrate and silver(1) nitrate complexes of L46 are obviously due to the pronounced tendency of copper(1) ion for a tetrahedral coordination. Therefore, in the metal-phenazine system the counteranion as well as metal ion stereochemistry play an important role in regulation of the molecular assembly into different frameworks and stoichiometries. Three copper(1) complexes of L4, have been reported and it reveals that the frameworks in the compounds are regulated by the counteranions as shown in Fig. 29 (139).Because perchlorate ion has weak
1
X=CI04’
FIG.29. Regulation of the six-membered ring structures by anions in a n L4, system.
2 16
MUNAKATA, WU, AND KURODA-SOWA
coordination ability, in [ C U ~ ( L ~ ~ ) ~ ( C ~one O ~ C10; ) ] C ~isO noncoordi~ nating and the other binds weakly with only one copper ion rather than as a interconnecting ligand. This leaves the two copper atoms in different coordination environments, tetrahedral and trigonal geometry, respectively. As a result, the complex is composed of a [cu&L47)G]6+framework containing a six-membered ring of copper atoms interconnected only by L47molecules (Fig. 30). Copper atoms in the two-dimensional sheet are arrang$d in a staggered conformation with a maximum deviation of 0.50 A from the mean plane. Replacement of copper(1) perchlorate by copper(1) hexafluorophosphate in reaction with L47isolated another polymeric complex with the for-
FIG.30. The six-membered ring of copper atoms in [ C U ~ ( L ~ , ) ~ ( C ~ O ~is) composed IC~O~ of [CU,(L~,)~]~+ unit. The coordination of perchlorate is not shown.
COPPERU) AND SILVERO) SUPRAMOLECULAR METAL COMPLEXES
217
mula [CU(L~~)(HPO~F)]. In this complex each Cu(1) ion is coordinated to two nitrogen atoms of different L47ligands and two oxygen atoms of different HP03F- ions in a distorted geometry. An important feature of the structure is that the HPO& anion acts as a n interconnecting ligand between two Cu atoms and takes part along with L47 in formation of the significantly distorted hexagonal framework composed of [C~(L47)4(HP03F)412+ units (Fig. 31). The copper atoms in the two-dimensional sheet are arranged in a ladder pattern in the acplane and a shallow roof pattern in the ab-plane. As discussed earlier, nitrate ion is expected to change the framework of the six-membered rings of copper atoms because it often functions as a bridging ligand. The copper(1) nitrate complex of L47[Cu(L4,)(N03)Iwas prepared by reduction of copper(I1) nitrate trihydrate under an ethylene atmosphere followed by reaction with L47in THF. In the complex the tetrahedral coordination of each copper ion is achieved by two L47ligands
'1-
\
,
FIG.31. The anion HPO$ acts as an interconnecting ligand between two Cu atoms (only the coordinating oxygen atom is shown) and takes part in formation of the hexagonal framework in ICu(L,,)(HPO,F)J.
218
MUNAKATA, WU, AND KURODA-SOWA
FIG.32. Top (a) and side (b) views of the stacks of L4, molecules within and between 2-D sheets in rCu(L4,)(N03)1.(From Fig. 6 in Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Nakamura, M.; Akiyama, S.; Kitagawa, S. Inorg. Chem. 1994,33, 1284.)
and two nitrate anions (Fig. 32). The six-membered rings of metal ions interconnected by L4, and NO, are extremely distorted, and are in fact in a chair form, with the ma5imum deviation of the Cu atom from the mean plane being 2.1-2.4 A. Examined along the c- and baxis, the copper atoms in the sheet are arranged in a ladder and a roof pattern, respectively. In addition, the shortest interplanar spacing distanFes of L4, moeties within and between the sheets are 3.30 and 3.39 A, respectively. Consistent with the observed V-v interactions in the system, the complex in powder form shows semiconductivity, with electric conductivity cr being 10-6.3S . cm-'.
COPPERU) AND SILVERW SUPRAMOLECULAR METAL COMPLEXES
2 19
It is proposed that a combination of three- and four-coordinate metal ions and bridging pyrazine and pyrazole molecules is highly preferred for construction of a two-dimensional sheet structure composed of six-membered rings of metal ions interconnected by aromatic ligands. The participation of suitable counteranions in coordination environments of the metal ions can regulate the frame of the ring, which governs the orientation of the aromatic ligands and the stack of the aromatic ligand within and between the sheets (139). V. Hydrogen-Bond-Assembled Frameworks
Self-associated intermolecular and intramolecular hydrogen-bonding assemblies have captured the attention of many research groups involved in supramolecular chemistry, molecular recognition, and crystal engineering because suitable matching of hydrogen bond donors and acceptors in number and orientation frequently leads to formation of new molecular aggregates (142-167). Hydrogen bonds are of paramount importance in biochemistry for determining the secondary structure of proteins and binding substrates to enzymes, receptors, and carriers. They are weaker than conventional covalent bonds but stronger than van der Waals interactions. Although there is no universal agreement on the best description of the nature of the forces in the hydrogen bond, it is generally accepted that hydrogen bonding, X-H----Y, occurs only between a hydrogen atom bound to an electronegative atom X and another atom Y that is also highly electronegative and has one or more lone pairs, enabling it to act as a base (142). The concept of hydrogen bonding was originally used to explain physical properties of simple organic and inorganic compounds such as abnormally high boiling points and heats of vaporization, but in recent years it has been noted that the introduction of a hydrogen-bonding interaction between ligands in transition-metal complexes is an indispensable tool for creating a variety of molecular architectures in a predictable fashion via self-assembly and molecular recognition. Several important reviews related to this field now are available ( 142-1 4 7 1. With the combination of the covalent bond forming capability of the metal ion and the inherent capability of the ligand for formation of complementary hydrogen bonds, a diversity of H-bonded structures has been obtained, including intramolecular hydrogen-bonded monomers (148), intermolecular hydrogen-bonded dimers, tetramers (149-151 ) and polymers having one-dimensional chain (152), two-
220
MUNAKATA, WU, AND KURODA-SOWA
dimensional sheet (153-158) and three-dimensional networks (159163). The most commonly encountered hydrogen-bonding interactions in H-bonded supramolecular frameworks are 0 -H----0 and N -H----0.Additionally, some weak hydrogen bond interactions such as C-H--0, C-H----Cl and even X-H----arene are also reported (146).The interaction may involve only one hydrogen bond donor and one acceptor, may be bifurcated or trifurcated, and may occur between atoms, molecules, or ions (142).In this respect, the bifunctional ligands should contain the simple functional groups such as carboxyl, amide, amino, cyano, and pyridone. Recent reports have shown that hydrogen bonds can occur in the system between ligand functional groups and some inorganic anions such as BF,, C104, SO$-, and halides, and even water and ammonia molecules (146, 148, 158). The ligands involved in the following discussion of the hydrogen-bonded complexes of Cu and Ag are listed in Fig. 33. Mingos has shown in his recent review how the application of molecular recognition principles based on triple hydrogen bonding resulted in the crystal engineering of aggregates based on ligands that can simultaneously form stable and inert metal-ligand bonds and have recognition sites for complementary arrangements of hydrogenbond donors and acceptors (144),which has been excellently demonstrated by self-assembly of the copper(I1) complex [Cu(L,,), . 2 melamine] (HL4, = 5-(2-pyridylmethylene)-hydantoin)in which L4, involves simultaneous coordination with Cu and an ADA=DAD (A = hydrogen-bond acceptor, D = hydrogen-bond donor) triple hydrogenbonding arrangement with melamine molecules (Fig. 34)(156). Among extended polymeric structures, two-dimensional hydrogenbonded frameworks are enormously popular, and the resulting assemblies resemble practically crinkled tapes (153,154)and infinite sheets (155-158). Smith and co-workers have structurally characterized a
b 2
HL53
L54
FIG.33. List of ligands in copper(1) and silver(1) complexes having hydrogen-bonding networks.
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
221
FIG. 34. Hydrogen-bonding networks in [Cu(L4&.2 melamine]. (From Fig. 2 in Chowdhry, M. M.; Mingos, D. M. P.; White, A. J. P.; Williams, D. J. J. Chem. SOC., Chem. Commun. 1996,899.)
number of silver(1) carboxylates, and among these phthalate and trimesate (LEO) have been found to form hydrogen-bonded polymers involving the ammonia molecule (155). The silver(1) complex [Ag2(phthalate)(NH3)21 contains a phthalate-bridged dimeric unit in which each Ag atom is coordinated to one ammonia molecule and a single phthalate carboxylate oxygen in an essentially linear fashion. All three amine hydrogens are involved in intermolecular hydrogenbonding interactions, giving a two-dimensional hydrogen-bonded sheet structure extending across the bc-direction of the cell (Fig. 35). The adjacent sheets are further linked by hydrogen bonds via the uncoordinated carboxyl oxygens. In the corresponding trimesate complex, [NH,l[Ags(L5,),(NH,)z(H20)21 . HzO, a two-dimensional sheet framework is based on a pseudo-centrosymmetric S-type trimer unit linked by the carboxylate groups of two independent trimesate residues, and the hydrogen bonds between the uncoordinated carboxylate oxygens and amine and water molecules stabilize the structure. Likewise, in the copper(I1) complex of 4,4’-bipyridine ( L d , the coordinating water molecule bridges between the metal center and uncoordinated LS1,affording two-dimensional rectangular grid sheets (158). The structural determination of the Cu(1) complex of 2-cyanoguanidine (LS2),[{Cu~L51)2},~L51)][BF~l,~ MeCN, reveals that the structure contains a dinuclear cation in which two T-shaped copper(1) centers are bridged by L51and terminally coordinated by two L5%molecules
222
MUNAKATA, WU, AND KURODA-SOWA
FIG. 35. Hydrogen-bonding networks in [Ag,(phthalate)(NH&I. (From Fig. 2 in Smith, G.; Reddy, A. N.; Byriel, K. A.; Kennard, C. H. L. J. Chem. Soc., Dalton Trans. 1995, 3565.)
(157).An extensive hydrogen-bonding network between BF, ions and LS2molecules gave a two-dimensional sheet structure as shown in Fig. 36. It is intresting to note that all four F atoms for each anion are involved in hydrogen bonding, three of which are sitting within the two-dimensional sheets and the fourth bridging the sheets. The principles of molecular recognition can also be used to assemble three-dimensional structures, but additional requirements on the number and orientation of the complementary hydrogen bonding motifs must be met (159-163). Nakasuiji has selected biimidazole, pterine, lumazine, and glycoxime as the target ligands because their chelating ability to a metal element and multi-H-bonding sites, and has constructed a number of three-dimensional hydrogen-bonded networks of metal ions such as Ni(I1) and Cu(I1) (160, 161). Molecules in which the motifs are related by Td or even S4symmetry are strong potential candidates for the engineering of a three-dimensional network exhibiting microchannels or diamondoid structures (146). The system we have worked with are literally double hydrogen bonding of the ligands combined with tetrahedral copper(1) ion in an attempt to design unprecedented two-dimensional and three-dimensional molecular architectures.
A. THREE-DIMENSIONAL SUPRAMOLECULAR Cu(1) COMPLEXES WITH CHANNELS (HL53) The bifunctional ligand 3-cyano-6-methyl-2(1H)-pyridinone as shown in Fig. 33, possesses both a coordination group (CN) and a
COPPER(1) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
223
FIG.36. Two-dimensional sheet structure of [(Cu(Lsz)2}~(L~,)11BF112 MeCN showing hydrogen-bonding associations. (From Fig. 2 in Batsanov, A. S.;Begley, M. J.; Hubberstey, P.; Stroud, J. J . Chem. Soc., Dalton Trans. 1996, 1947.)
hydrogen-bonding site (pyridone), and it would be a suitable candidate for the formation of H-bonded metal complex supramolecules. The second reason for choosing it in the study is its relatively small size, which would reinforce the unidentate coordination of each ligand to the tetrahedral copper(1) ion without steric hindrance. The reactions of copper(1) salts with HLssin acetone have isolated four coordination polymers, [Cu(HL,,),IX, where X = ClO,, BF; , PF;, and CFs SO, (164). Although each structure contains a three-dimensional framework of tetrahedral CuNj centers linked by intermolecular hydrogen bonds through pyridone N and 0 atoms in a head-to-tail mode, different patterns of hydrogen bonding give rise to two types of different frameworks-namely, square channel and superadamantane networks-depending on the kinds of counteranions. In the complexes with relatively small anions, perchlorate and tetrafluroborate, each HL53molecule is hydrogen bonded to two adjacent others through pyridone N and 0 atoms; that is, each I C U ( H L ~ ~ ) entity ~ I X is connected to eight neighboring counterparts as shown in Fig. 37. The dihedral
FIG.37. Part of the hydrogen-bonded structure (a), packing diagram (b), and spacefilling model (c) of [Cu(HL,),]CIO,. (From Figs. 2 and 3 in Munakata, M.; Wu,L. P.; Yamarnoto, M.; Kuroda-Sowa, T.; Maekawa, M. J.Am. Chem. SOC.1998, 118, 3117.)
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
225
angle between the hydrogen-bonded HLS3groups is 83.44" and 84.98" for the perchlorate and the tetrafluroborate, respectively. This results in four [Cu(HL,,),IX monomers interconnected through hydrogen bonds, forming a layer with an open cavity occupied by the counteranion. The overall structure is made of such layers, each consisting of an infinite square array of copper atoms coordinated with HLs3groups and bridged by hydrogen bonds, leading to the cross-linking of a given layer with those immediately above and below it. A square channel framework is generated with a relatively small size of cavities (ca. 12.1 A in diameter) filled with guest anions. Let us name such hydrogen bonding as A type, illustrated schematically in Fig. 38. The hydrogen bonding network in [CU(L,~)(HL,,),Iis also found to belong to this type (165). When the small anion is changed to the drastically larger one, PF; or CF,SOi, the cavities in the square channel lattice for the perchlorate and tetrafluroborate cannot be effectively adapted for large guest ions only by elongating the hydrogen bond distances. If the stoichiometry is to remain the same, one way for the system to respond and avoid this destablizing repulsion is to rearrange the hydrogenbonding mode as B type, that is, each HLS3moiety in the monomeric unit of [Cu(HL,,),]X is head-to-tail hydrogen bonded to only one adjacent HL,, molecule, rather than two, with a rather smaller dihedral angle (0"-11.43"). Thus, one [CU(HL,,)~IXentity is connected to four adjacent others, and these repeating units form a three-dimensional adamantane architecture of metal atoms (Fig. 39). The diamondoid frameworks are stacked with each other in such way that all the copper centers are found on lines parallel to the c-axis, which flefines enormous linear chambers with larger open cavities (ca. 13.3 A in diameter), that just fit the large anions PF; and CF3SOi. To show how the assembling process would respond in the system
I
0-1 1 '
FIG.38. Schematic presentation of two types of' hydrogen bonds in HL, complexes.
226
MUNAKATA, WU, AND KURODA-SOWA
t
t
CU
FIG.39. Part of the hydrogen-bonded structure (a), packing diagram (b), and perspective view of diamondoid framework (c) in [Cu(HL,)JPF,.
COPPER(1) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
227
when H-bonding interaction is absent, the structures of two copper(1) complexes of L53were determined (165, 166). It reveals a completely different stoichiometry and framework in both [c~g(L63)4]c104and [Cu,o(L,3),](BF,)2, in which each deprotonated LS3monoanion employs all its three functional groups bridging three rather than two copper atoms, forming a supramolecular channel architecture containing a Cud cluster (Fig. 40).These findings suggest that the bifunctional ligand group m e 3 possesses unique ability for the molecular recognition, which has decisive influence on crystal engineering of coordination polymers. Modification of the hydrogen-bonding mode and distances as well as the direction of H-bonding sites at molecular level can effectively facilitate transition of one framework to the other.
FIG.40. Part of the molecular structure (a) and extended framework (b) in [Cu,(L,,),I
c10,.
228
MUNAKATA, WU, AND KURODA-SOWA
B. HYDROGEN-BONDINGAND ~-~-STACKING-~SEMBLED Cu(1) COMPLEXES The second bifunctional ligand selected to construct the metal architecture is 2-hydroxyquinoxaline (Lu), because it also possesses both N-coordination sites and potential double H-bonding sites (CN=COH or CNHC=O if it undergoes tautomerization). The reaction of copper(1) perchlorate and Ls4under an atmosphere of ethylene gave the complex [ C U ( L ~ ~ ) ~ ( C ~ H ~ whose ) ] C ~structure O~, was determined by X-ray analysis (167). Its unique feature is the cooperative effect of hydrogen-bonding and aromatic-stacking interactions simultaneously present in the system. The metal center is in a trigonal planar geometry achieved by coordination to one ethylene and two Ls4 molecules. The cation maintains the fundamental feature of the free ligand in the sense that two LE4molecules are linked to each other by head-to-tail double hydrogen bonds, resulting in formation of an infinite zigzag chain (Fig. 41). The IR spectrum of the complex shows low v(NH) stretching frequency in the region 2950-3050 cm-', consistent with the short NH----0hydrogen bonds observed. Between the adjacent chains the H-bond assembled La4moleculeoplanes are stacked to each other, with the average distance of 3.30 A. This gives a unique two-dimensional cooperating structure stabilized by hydrogen bonding and r-r interactions as well as covalent bonds, reminiscent of a proton-electron transition (PET) system. It is worth mentioning that ethylene also plays an important role in the construction of the resultant architecture as a spacer because of its relatively small size, which reduces the packing volume, ensuring the effective stacking of the aromatic planes.
VI. a-a-lnteraction-Assembled Frameworks
A. INTERMOLECULAR n-n INTERACTION IN DISCRETE COORDINATION COMPOUNDS Planar coordination compounds with aromatic ligands (Fig. 43,especially those having extended 7~ systems, show r-r interaction in solid states. In alkene or alkyne n-bonded Cu(1) complexes, in-plane coordination of a C =C or C = C bond to trigonal planar Cu(1) centers often leads to planar molecular conformations (167-1 72). The infinite n-r stacking columns are confirmed in the 2,2'-bipyridine (Lss) com-
COPPER(I) AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
229
FIG. 41. Packing diagram of LCu(L,)e(C2H,)IC104; side view showing H-bonded networks (a) and top view showing aromatic stacking interactions. (From Fig. 2 in Dai, J.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata, M. Inorg. Chern. 1997,36,2688.)
plex [CU(L~~)(C~H,)]C~O, (169)and the 1,lO-phenanthroline (LW)complex [ C U ( L ~ ~ ) ( C ~ H C O ~ E(1 ~ )71 ]C O ~ nearest carbon-to-carbon ) ~ with distances of 3.31 and 3.37 (172,1731,respectively, and dimer forma(171) tion through n-rr interaction can be seen in [Cu(LW)(CZHz)IC1O4 [the nearest carbon-to-carbon distances are 3.42 ( 172)l. Flat conformations are also observed in cationic parts of dimeric Cu(I) and Ag(1) compounds bridged by two l&naphthyridine ligands, (1741. The former comLS7, [ C U ( L ~ ~ ) ] ~ ( C and ~ O[Ag(L67)12(C10,)2 ~)~ pound forms a dimer through n-n interaction of the neighboring L67
A
A
230
MUNAKATA, WU, AND KURODA-SOWA
L56
L55
ae L59
L61
L62
Me.
Me
L63
L65
*@
\ /
L67 L68
L69
L66
ip /
L70
Me
@ \ /
L71
L72
L73
L74
FIG.42. Ligands that appear in Sections VI and VII.
molecules; the latter shows one-dimensional infinite :tacking column of the cationic part with interplane distance of 3.40 A (172) as shown in Fig. 43. Another flat dimer, [Cu(L5,)l2(175), also shows a onedimensional a-a stacking column, with interplanar separation of 3.41
A (172).
Not only planar compounds but also nonplanar ones show a-a stacking interaction. In the crystal structure of CUI(L& (176), two
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
23 1
3 L -3 -c 3.40 A
Co=-o; 3 - J - 3
Ag
napy
FIG.43. A crystal-packing view of [Ag(L5,)1~(C104)2 showing r-n interaction between planar cationic parts.
2,6-dimethylpyridines Lse and a n iodine coordinate to a Cu(1) ion in T-shaped trigonal geometry, and the dihedral angles between two ligand planes and the CuINp plane are 87.6 and 71.2", indicating nonplanar conformation. However, one-dimensional n-n stacking can be easily recognized in Fig. 44, with a nearest carbon-to-carbon distance of 3.447 A. Changing the counteranion from an iodide to a perchlorate ion gives two-coordinated copper(1) or silver(I) compounds. Whereas planar cationic parts of [M(Lsg),1X(M = Cu, Ag; X = ClO, , NO,) (177) with two-coordinated M(1) centers show only weak n-n interytion, with the nearest carbon-to-carbon distances being 3.59 to 3.66 A, one of the polymorphs of [ C U ( L ~ ~ ) ~ I(178), C ~ Owith ~ a dihedral angle between two LS9ligands of 5 6 2 , shows a close n-n contact of 3.33 A (172) between two Ls9ligands residing in adjacent molecules. We can see not a one-dimensional n-n stacking column but a n-n-interactionassembled one-dimensional structure like ----Ls9 - Cu -Lse---L59- Cu-L59----. Although a trigonal planar geometry is maintained around a Cu(1) center in the ethylene-coordinated complex [Cu(L54)z(CzH4)IC104 (1671, the two Ls4 molecules coordinated to the Cu(1) ion have a dihedral angle of 102.5", giving a butterfly structure. Repetitions of double hydrogen bonds between two Ls4ligands in neighboring complexes form an infinite zigzag chain (Fig. 41). Additionally, a n-n interaction between two LMligands residing in neighboring chains (the interplanar separation of 3.30 A) forms a two-dimensional structure as shown in Fig. 41. This gives a unique two-dimensional cooperating structure, which is a fundamental characteristic of proton-electron transfer (PET) systems (179, 180).
232
MUNAKATA, WU, AND KURODA-SOWA
I
3.447 A
c sinp
FIG.44. A crystal-packing view of CuI(L6& showing r-a interaction between L60 ligands. (From Fig. 5 in Healy, P. C.; Pakawatchai, C.; White, A. H. J. Chem. SOC., Dalton Trans. 1983, 1917.)
The copper(1) ion in [ C U ( L ~ ) ~ ] CMeOH ~ O ~ (181) has a distorted trigonal geometry with three nitrogen atoms of three 1-aminopyrenes (Lm).Two of them are parallel to each other, whereas the rest lie approximately perpendicular to the other two as shown in the molecularpacking view (Fig. 45). Intra- and intermolecular n-n interactions of the former two form a one-dimtnsional stacking column with interplane separations of 3.40-3.43 A. Moreover, intercolumn n-n interactions between LBOligands with an interplane separation of 3.52 A result in the formation of a two-dimensional structure. The copper(1) ion in [Cu(L&H,0)IC1O4 (138)has a distorted trigonal geometry with coordination of two nitrogen atoms of two terminal 9
COPPERU) AND SILVERW SUPRAMOLECULAR METAL COMPLEXES
233
FIG. 45. Molecular packing view of [Cu(L,)JClO,. MeOH. (From Fig. 2 in Munakata, M.; Dia, J.;Maekawa, M.; Kuroda-Sowa, T.; Fukui, J. J . Chem. SOC.,Chem. Cornmun. 1994, 2331.)
L4 molecules and a water oxygen. The coordinated L4 molecules form an alternating r-n- st9cking column, with the nearest intermolecular separation being 3.46 A. Because basicity of acridine is stronger than that of La (1821, adding a solution of L61 in methanol to a solution containing copper(1) and L4 gives a discrete complex, [ C U ( L ~ ~ ) , N O , IH(,L O)~,,~ ~(183).The copper(1) ion is trigonally coordinated by the two nitrogen atoms of L, and one of the oxygen atoms of the NO; anion. Two coordinated L61 molecules are not parallel to each other at a dihedral angle of 63.6", and each has a r-n- interaction with an adjacent Ll molecule coordinated to another copper atom (Fig. 46). The interplane separations of these intermolecular r-r interactions are 3.41 and 3.49 A. A metalfree L4 is sandwiched between L61 ligands, forming an infinite columnar stacking along the a-axis with a L46-L61-]LB, repeating unit. The closest C----Cdistance of 3.39(1)A indicates a significant n--r interaction between L4 and 4 1 . An L46-bridgedCu(1) dimer complex, [CU,(L~,)~(M~OH),I(L~~)(PF,), (138) shows an alternate r-r stacking of coordinated and uncoordinated La. Two copper(1) ions are crystallographically the same: a T-shaped three-coordination of a methanol and terminal and bridging La with a remarkably large N(termina1 L&Cu-N(bridging La) angle of 156.1". A metal-free L4 molecule is located near the bridging L4, with a dihedral angle of 7.6"and with a nearest carbon-to-carbon sepa-
234
MUNAKATA, WU, AND KURODA-SOWA
FIG. 46. Molecular views of [Cu(L6&N03](La.H 2 0 ) , ,indicating r-r interaction between Lp6and La1 (a) and between two LBLligands (b). (From Fig. 4 in Kuroda-Sowa, T.; Munakata, M.; Matsuda, H.; Akiyama, S.; Maekawa, M. J. Chern. Soc., Dalton Trans. 1995, 2201.)
ration of 3.42A, indicating a significant ?T-T interaction between them. Molecular packing along the c-axis reveals one-dimensional T-?T stacking of bridging and metal-free La alternately.
B. INTERAND INTRAPOLYMER r-rr INTERACTION One-dimensional polymeric structures are observed in copper(1) complexes with 2,9-dimethyl-l,lO-phenanthroline, [CU(L~~)(CN)] and [CU(L62)(NCS)](184). Cn- or NCS- acts as a bridging ligand, and L62 chelates to a copper(1) ion to block two of four coordination sites of the tetrahedral center, resulting in the infinite zigzag chain. In both compounds, zigzag chains are connected through T-?T interaction between L62 molecules residing in neighboring chains, which form two-
COPPERU) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
235
dimensional polymeric structures. Similar two-dimensional network formations are also observed in the trinuclear copper(1) complex with 2,2’-biquinoline (La) [CU,(L,)~(CN),] (1851, in which tetrahedral copper(1) ions coordinated by chelating L63 ligands are bridged alternately by Cu(CN)i and CN-. When La is used as a bridging ligand to construct coordination polymer compounds, interpolymer T-T interactions are often observed. A one-dimensional zigzag chain composed of bridging L,, and a tetrahedral copper(1) ion can be seen in [CU(L,~)(NO~)] (183) (Fig. 47). The copper(1) ion is coordinated by the two nitrogen atoms of the two L46 molecules and by the two oxygen atoms of the NO, anion.
FIG.47. Top (a) and side views (b) of the crystal packing of [Cu(L,,)(N03)]showing the La columnar stacks. (From Fig. 2 in Kuroda-Sowa, T.; Munakata, M.; Matsuda, H.; Akiyama, S.; Maekawa, M. J. Chem. SOC.,Dalton Trans. 1995, 2201.)
236
MUNAKATA, WU, AND KURODA-SOWA
Two L46 molecules coordinated to the same copper atom are inclined to each other at a dihedral angle of 82.2'. As can be seen in Fig. 47, the L46 molecules are stackFd in the b-axis direction, with an interplane separation of 3.47 A. Thus, the compound consists of onedimensional zigzag chains of copper atoms and L46 molecules, interconnected through L4&n-n interactions, resulting in a two-dimensional interaction. A similar zigzag chain composed of bridging 4 6 and a trigonal planar copper(1) ion is observed in [cu(L46) (MeCN)],(L,)(PF,), (La : pyrene) (183).The L46 molecules are stacked in the c-axis direction with an interplane separation of 3.47 A. Thus, the compound consists of one-dimensional zigzag chains of copper atoms and L46 molecules interconnected through L, n-n interactions, resulting in a two-dimensional interaction. A series of halogen-bridged ~ o p p e r ( I ) - L compounds, ~~ [CU~(~-X)~ (L46)l (X = I, Br, or C1) (1861, show one-dimensional (X = I) or twodimensional (X = Br or C1) polymer frameworks. CuzIzrhomboids in the iodide compound are connected by bridging L, ligands through trigonal planar Cu(I) centers, forming an infinite straight chain. The dihedral angle of the rhomboid and L46 plane is 75.28'. The n-n interaction between L46 molecules of adjacent chains (interplanar distances of 3.46 A) gives the complex two-dimensional structure. On the other hand, both the bromide and the chloride compounds have almost same structure: CuX infinite stairs bridged by L , through coordination to distorted tetrahedral (Cu(1) ions, forming a two-dimensional network structure as shown in Fig. 48 for the bromide compound. Intrasheet r-n interactions betwen LlS molecules are also observed in both compounds, with interplanar distances of 3.40 and 3.36 A for the bromide and the chloride compounds, respectively. In the solid-state 13C NMR spectra of these compounds, the increase of the upfield shifts of the resonances assigned to the quaternary carbon atoms of L46 upon coordination (-2.4, -2.9, and -3.9 ppm for the iodide, the bromide, and the chloride compounds, respectively) are well correlated to the decrea!e in the interplanar distances of L 4 6 molecules (3.46, 3.40, and 3.36 A, respectively). Not zigzag but completely straight one-dimensional chains can be seen in [Ag(L46)I(c104) (138),in which the Ag(1) ion has a linear two coordination of two nitrogen atoms of bridging L46 ligands. All the L 4 6 molecules are parallel, as shown in Fig. 49, which results in the formation of a one-dimensional chain structure like a flat ribbon. The shortest intermolecular distance is 3.36 indicating significant T-n interaction, though the overlap between them is not so large. Because one chain interacts with four neighboring chains through n-n interac-
A,
COPPER(1)AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
237
FIG.48. Top (a)and side views (b) of the packing arrangement of ICu2(pBr)2(p-L4B)I. (From Fig. 4 in Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Honda, A.; Kitagawa, S. J. Chem. SOC.,Dalton Trans. 1994, 2771.)
tions, this system has a three-dimensional network assembled by TI-n interaction. A two-dimensional sheetlike structure in [Ag2(L,,J(N03),I (138) is composed of bridging L46 and bridging nitrate anions coordinated to trigonal Ag(1) ions, forming six-membered rings of Ag(1) ions. The significant n-n interaction between two L46 molecules in the neighboring sheets (the shortest intermolecular distance of 3.34 A) indicates the formation of a a-n-interaction-assembled three-dimensional network structure. Perylene molecules &4), having an extended n-system, easily form a-n interaction. Four peripheral C=C moieties of Lfi9coordinate to
238
MUNAKATA, WU, AND KURODA-SOWA
FIG.49. Molecular structure (a) and a perspective view of the packing arrangement (b) of [Ag(Le)l(CIOa).(From Figs. 7 and 8 in Munakata, M.; Kitagawa, S.;Ujimaru, N.; Nakamura, M.; Maekawa, M.; Matsuda, H. Znorg. Chern. 1993,32,826.)
four Ag(1) ions in $-fashion in [Ag2(L6s)(C104)2] (187).The silver(1) ion has a distorted tetrahedral geometry with two C=C groups of two L69 molecules and two oxygen atoms of two perchlorate anions. The resulting two-dimensional sheet shows a W-type wavy conformation 5s shown in Fig. 50, which enables effective n-n interaction (3.31 A) between L@ molecules in adjacent sheets. Thus, the n-n interaction promotes the network dimensionality from two to three dimensions. Chair-formed six-membered rings of Cu(1) ions in [Cu(L4,)(NO,)I (139)form a two-dimensional sheet with a certain thickness, in which each Cu(1) ion is bridged by two L4, and two nitrate anions. Because L4, molecules are almost parallel to the two-dimensional sheet, both intra- and intersheet n-n interaction can be seen, with inter-L4, separations of 3.30 and 3.39 A respectively. The latter interaction enables the three-dimensional network structure assembled through n-n interaction.
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
239
(4
FIG. 50. W-type wavy conformations of two-dimensional sheets in IAg,(LBB)(C104)ZI. (From Fig. 2 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Sugimoto, K. Znorg. Chern. 1997,36,4903.)
240
MUNAKATA, WU, AND KURODA-SOWA
VII. Diamondoid Frameworks
We can imagine a formation of a cubic or a hexagonal diamondrelated framework (Fig. 51) from a combination of a tetrahedral metal ion and a rodlike bridging ligand. Although both types of frameworks are found in minerals or polymorphs of ice (1881, so far as we know, only the former type of framework is known in coordination compounds. Hereafter, we use the terms “diamond” or “diamondoid framework” to mean cubic diamond-related framework. As Hoskins and Robson have proposed (1271, a compound having a diamondoid framework, if it can have a large cavities or channels inside, can offer a number of features of potential interest such as molecular sieve properties, heterogeneous catalytic properties, and mechanically strong materials with an unusually low density. Additionally, T-T interaction between aromatic bridging ligands often controls the degree of interpenetration of diamondoid frameworks and sometimes gives electronic conducting material. In this section, we will focus on coordination compounds with diamondoid frameworks, especially containing copper(1) and silver(1) ions. These ions having d10 electronic configuration are suitable for a tetrahedral metal center in a diamondoid framework. Table V lists the coordination polymers having diamondoid frameworks reported so far, together with other coordination polymers having related frameworks.
A. BRIDGED BY PYRIDINEOR
PYRAZINE
DERIVATIVES
Pyrazine, LS8,and its derivatives are the shortest bridging ligands (133)has a tetrahenext to cyanide. The copper(1)ion in [CU(L~~),IPF, dral geometry, with coordination of four N atoms of Ld0.The other end of the N atoms of each L40 coordinate to other copper(1) ions re-
%%
FIG.51. Cubic (left) and hexagonal (right) diamond-related frameworks.
COPPER^ AND SILVER^ SUPRAMOLECULAR METAL, COMPLEXES
241
TABLE V COORDINATION POLYMERS HAVING DIAMONDOID AND RELATED FRAMEWORKS Compound
Bridged M~-;-M Degree of distance (A) interpenetration
Remarks
5.03 5.11 5.46 6.99 8.856" 8.9 5.04, 9.54 9.93 ca. 10 10.90 11.16 11.6 11.76 11.9 12.09 12.76 13.5 13.55 16.4 17.0
Ref. 127 127,204,205 127,204,205
I33 127 71-71
71-71
8-8
8-71
71-77 71-8
193 206 190 208 207 136 190 4 195, 196
194 197, 798 164
8-71
192
71-71
201,202 203
71-8
Zinc-blende structure.
', Cu----Cudistance.
sulting in the formation of a diamondoid framework (Fig. 52). The shortness of the bridged Cu----Cuseparation of 6.99 A together with the presence of the PF; counteranion and methyl groups in L4,,prevent the interpenetration of any other framework. Although coordination chemistry of Ag(1) and L39(130, 289) revealed several compounds having interesting three-dimensional observed such as cw-ThSi2 or ReOs types, no diamondoid framework has been observed in this system. When 4,4'-bipyridine, Leo,is used instead of L4,,fourfold interpenetrated diamondoid frameworks are obtained for both Cu(1) and Ag(1) ions. The Cu(1) compound [Cu(L&lPF6 1136) exists as four independent concatenated diamondoid frameworks with Cu----Cuseparations of 11.16 A. The Cu(1) centers occupy crysta!lographic 4 positions and hence all Cu-N bonds are identical, 2.034 A. The PF; counteranions
242
MUNAKATA, WU, AND KURODA-SOWA
FIG.52. A single diamondoid framework in [Cu(L&lPFs. For clarity, the PF, anions are omitted.
occupy channels that are parallel to the c-axis and they sit on fourfold (190) also crystallographic axes. The Ag(1) analog [Ag(LBo)~CF3So3 contains four interpenetrating diamondoid frameworks (Ag----Agintraframe separations of ca. 11.6 A). The Ag(1) cations display a distorted tetrahedral geometry, with Ag-N contacts different for the two Interconnection of independent LWligands (mean 2.270 vs. 2.380 trigonal Cu(I) centers through LBO was achieved by hydrothermal synthesis (191). The resultant crystalline compound [cU(L60)1.5]* NOs(H20),,26 shows sixfold interpenetrated a-ThSi2-typeframeworks. By lengthening the linking bipyridyl ligand via insertion of a trans C=C double bond between the pyridyl units, the length of the resultant briding ligand, 1,2-truns-(Cpyridyl)ethene (L70), is extended by about 2.4 A and the interpenetration in diamondoid frameworks is also changed. The Locompound with Cu(I), [ C U ( L ~ ~ ) ~ ~ B F ~ ( O . ~ C H ~ C ~ (192),has fivefold interpenetrated diamondoid frameworks. The Cu----Cu distances bridged by L70 range from 13.33 to 13.82 A to create large cavities within the diamondoid framework as shown in Fig. 53. All five independent frameworks are polycatenated with channels throughout the structure: these channels accommodate BF; counteranions and CH2C12solvent molecules. Among the bipyridine derivatives, L64 can also act as a linear bridging ligand if the transoid conformation is maintained. [Cu(L,&]X (X = BF4, PF,) (1931, having a twofold interpenetrated diamondoid framework, can be obtained as yellow-green triangular crystals. The
A).
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
243
FIG. 53. View of the structure of [Cu(L,,),JBF,. (From Fig. 2 in Blake, A. J.;Champness, N. R.; Chung, s. s. M.; Li, W.-s.; Schroder, M. J. Chem. Soc., Chem. Commun. 1997,1005.)
LM has surely a transoid conformation, but the rings twist out of plane relative to each other by about 30". In addition, the Cu-N bond is bent out of the plane of the aromatic ring by 19,leading to a closing of the Cu-L-Cu angle to 134". Bridged by this large ligand, the resulting framework contains copper atoms separated by 8.9 A and significant void volume. The extra space is occupied by an identical network forming a twofold interpenetrated structure (Fig. 54).
B. BRIDGED BY BISNITRILE LIGANDS Nitrile compounds are generally accepted to be weaker donors than pyridine derivatives; still, they can coordinate to d'O metal ions. A bisnitrile compound, if the two CN groups have opposite directions, can act as a rodlike bridging ligand in the construction of diamondoid frameworks. Although directions of two CN groups in alkyldicarbonitrile, NC(CH,),CH, are not fixed a priori due to free rotation around methylene carbons, a Cu(1)-adiponitrile ( n = 4) complex, [CU(NC(CH~)~
244
MUNAKATA, WU, AND KURODA-SOWA
FIG.54. A schematic illustration of twofold interpenetrated diamondoid frameworks of [Cu(L&lX (X = BF,, PFB).The spheres represent the copper atoms, and the twisted bipyridine ligands are shown as the bet cylinders linking the copper centers together. (From Fig. 2 in Lopez, S.; Kahraman, M.; Harmata, M.; Keller, S. W. Znorg. Chem. 1997,36, 6138.)
CN),INO, (1941, shows a diamondoid framework. In the crystal, four methylene groups of the adiponitrile form a planar zigzag chain and the terminal CN groups align almost antiparallel, which allows the adiponitrile to behave as a linear bridging ligand. The bridged Cu----Cudistance of 12.09 A together with the small size of the counteranion (NOi3)enable it to show sixfold interpenetration of independent diamondoid frameworks. Terephthalonitrile, L66, having two CN groups at the 1- and 4-positions in a benzene ring, is a good candidate for a rodlike bridging ligand for d'" metal coordination polymers. When the aceonitrile liare substituted by L65, the crystals formed gands of [CU(CH~CN)~]BF~ have the composition [Cu(L6,),]BF4(4). Diamondoid framgworks are indeed formed (the bridged Cu----Cudistance of 11.76 A), but the structure contains fivefold interpenetrated independent frameworks. Channels of rhombic cross section are generated in which the BF, ions are located. Methyl substitution of La6 causes a drastic change in the interpenetration of the diamondoid frameworks. Crystals of [Cu(L,,)21X(L66)(THF)(X = BF4, C104) (195, 196) were obtained by solution into the corresponding copper(1)soluslow diffusion of the LBB tion. X-ray crystal structure analysis revealed that both compounds contain threefold interpenetrated diamondoid frameworks as shown in Fig. 55. Because the bridged Cu----Cudistance in these compounds (11.9 A) is almost same as that bridged by L65 in Cu(L&BF4 (41, the lesser interpenetration in LBBcomplexes should be caused by the bulkiness of methyl groups in L66. This is also strongly correlated to the incorporation of guest molecules, a metal-free LBB,and THF. The metal-free dmtpn incorporated in the cu(L& lattice participates in the formation of a T-T stacking column (discussed later).
COPPER(I) AND SILVERW SUPRAMOLECULAR METAL COMPLEXES
245
FIG. 55. A threefold interpenetrated diamondoid framework in [ C U ( L ~ ) ~ I X ( L ~ ) (THF). For clarity, the ligands are represented by lines and the PF, molecules are omitted.
Among the compounds having diamondoid frameworks listed in Table V, the copper complex with 2,5-dimethyl-NJV’-dicyanoquinonediimine [Cu(L7,),I (197, 1981, has a quite interesting conducting property. Recent developments on this and related compounds show other interesting features including metal-insulator-metal transition (reentrant behavior) (199),three-dimensional Fermi surface character (2001, and so on. The structure of [Cu(L,J21shows sevenfold interpenetrated diamondoid frameworks with strong n-n interaction (the interplanar distances between the neighboring L71 molecules are 3.13.2 A). It should be noted that not even-number-fold but sevenfold interpretation in this class of complexes probably disturbs the dimer formation between adjacent LT1molecules, which contribute to the aforementioned conducting behavior. A ninefold interpenetration of diamondoid frameworks, the highest degree thus far reported, is observed in a series of Ag(1) complexes with 4,4’-biphenyldicarbonitrile,L72 (201, 202). Crystallization of L72 with AgX (X = PF6,AsF6, SbF,) (1: 1molar ratio) by heating and slow cooling from ethanol produces yellow crystals of [Ag(L72)21X (X = PF,, AsF6, SbF,). All of these structures show ninefold interpenetrated diamondoid frameworks. This highest degree of interpenetration is a result of the rather large ligand used (the bridged M----Mdistance of 16.4A). The large amount of void volume created by a single network is filled by eight identical nets (Fig. 56).These nets consist of parallel ligands offset along the long axjs and display n-n stacking at a planeto-plane distance of 3.4to 3.6A. Columns of counterions are revealed
246
MUNAKATA, WU, AND KURODA-SOWA
FIG. 56. A ninefold interpenetation of diamondoid frameworks of [Ag(L&IX (X = PF,, AsFs, SbF,). For clarity, the L,, are represented by lines and the PF, are omitted. (From Fig. 1 in Hirsch, K. A,; Wilson, S. R.; Moore, J. S. Chem. Eur. J. 1997, 3, 765.)
down the fourfold c-axis. A packing model for this system proposed by Moore et ul. (202) explains a relationship between the degree of the interpretation and the bridged M----Mdistance. Their model also explains the effect of the anion size on the height of adamantoid cages in a series of compounds [Ag(L&IX (X = PF6,AsF6, SbF6). The longest M----Mdistance in a diamondoid framework is observed in [Ag(L,3)21(C104)H20 (203).The trunsoid conformation of 3,3'-dicyanodiphenylacetylene, L73,bridges two Ag(1) ions separated by 17.0 A. The obtained structure shows eightfold interpenetrated diamondoid frameworks (Fig. 57). A majority of the large amount of void space created in a single diamondoid framework is filled through interpenetration, resulting in an eightfold diamondoid network. The remaining space within the lattice is filled by perchlorate ions and water. Interpenetration in this structure is mediated by T-T stacking of LT3(discussed later).
C. OTHERBRIDGING LIGANDS One of the simplest bridging ligands in constructing diamondoid frameworks is a cyanide, CN-. It is well known that Zn(I1) and Cd(I1)
COPPER(1jAND SILVER(1)SUPRAMOLECULAR METAL COMPLEXES
247
FIG. 57. Diamondoid framework in [Ag(L~:3~dI(C101)H20 representing 7-n stacking along the maxis. (From Fig. 8 in Hirsch, K. A.; Wilson, S. R.; Moore, J. S. Inorg. Chem. 1997, 36, 2960.)
ions bridged by cyanide form doubly interpenetrated diamondoid frameworks of M(CN12(127, 204, 205). A partial replacement of the metal site by a Cu(U ion needs incorporation of countercations due to the charge neutrality, as exemplified in the case of NMe,[CuZn(CN),I (127). The incorporated NMe4 cation prevents the interpenetration of another network, resulting in the formation of a single framework, although the framework is no longer a diamondoid but a zinc blende structure. A detailed structural analysis concluded that carbon atoms in cyano groups coordinate to Cu(I) centers to form Cu-CN-Zn rods. A similar zinc blende structure is also seen in [Cu(C(CGH4CN),IBF4 .xC6H5N02 (x = 7.7) (125, 1271, in which C6H4CN can be regarded as a bridging ligand between Cu(1) and C" centers. The large void space created by the bridging ligand is filled by a counteranion and solvent molecules. A bridging ligand having mixed functional groups also affords a diamond-related framework, 4-Cyanopyridine, LG7,having both a cyano group in one end and a pyridine nitrogen in the other, is such a typical asymmetric bridging ligand. Cuboidal crystals of [Ag(L6,)21BF, (190)were obtained by slow diffusion of an ethanolic solution of AgBF, placed over LG7dissolved in THF (molar ratio 1:2). The fourfold interpenetrated diamoadoid frameworks have a Ag----Agintraframe separation of 9.93 A. The Ag(1) ions, lying on mm special positions, exhibit markedly flattened tetrahedral geometry, with
248
MUNAKATA, WU, AND KURODA-SOWA
N(py)-Ag-N(py), N(cn)-Ag-N(cn), and N(py)-Ag-N(cn) angles of 127.7(5), 126.5(5), and 101.45(12Y, respectively. These local distortions result in a global compression of the diamondoi! framework along c. The Ag-N bond lengths have values of 2.270(6) A [Ag-N(py)l and 2.350(10) A [Ag-N(cn)]. The presence of iso-oriented anisobidentate (with donor ends differing both in steric need and in basic character) ligands gives rise for [Ag(b7)JBF4to a polar axis (c). A T-T stacking column of aromatic ring! could be observed along the c-axis (interplanar distance of 4.0 A). Le7 also forms a diamond-related framework with CuCN. A Cu(1) ion in [CuCN(L,,)I (206) is coordinated by one carbon atom and three different types of nitrogen atoms. The structure was first interpreted as triply interpenetrated threedimensional polymers with one-dimensional zigzag chains of CuCN bridged by Le7 and has recently been reappraised to have diamondrelated networks (190). 2,7-Diazapyrene (LT4),having an extended T-system, also produces a Cu(1) coordination polymer, [Cu(L7,),1PF6(207), which shows threefold interpenetrated diamondoid frameworks. The spaces within each diamondoid framework in [Cu(L,,),IPF, are filled by two other diamondoid frameworks, and the three interpenetrating arrays are related to each other by a 90"rotation (Fig. 58). This form of interpenetration contrasts with that observed for other diamondoid frameworks appearing in Table V, which are related t o each other by simple translation. Adjacent frameworks interact with each other via face-to-face T-T interactions between L7, molecules, which are arranged so that the N----Naxes of the ligand are at 90" to one another. Because of the greater lateral steric bulk of L,,, the number of interpenetrating frameworks of three is smaller than other bridging ligands, giving similar M----Mdistances (Table V). This represents the identification of another factor controlling the degree of interpenetration in diamondoid frameworks. Although HL63 has both a CN group and a pyridone group, a dimerization of two HL53molecules through head-to-tail hydrogen bonds between the pyridone groups brings a symmetric rodlike bridging li~IX = PFs, CF,SOJ (164), each gand. In the crystals of [ C U ( H L ~ ~ )(X HL63 moiety in the monomeric unit of [ C U ( H L ~ ~ )is~ head-to-tail IX hydrogen-bonded to one adjacent HL53 molecule. Thus, one [Cu(HL,,),IX entity is connected to four adjacent others, and these repeating units form a three-dimensional diamondoid framework. The structure exists as fourfold interpenetrated diamondoid frameworks with intraframe Cu----Cuseparations of 13.5 A in both cases. The diamondoid frameworks in both complexes are stacked with each other in such way that all the copper centers are found on lines parallel to the
COPPERU) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
249
FIG. 58. Threefold interpenetrated diamondoid frameworks observed in [Cu(L,,),] PF,. (From Fig. 1 in Blake, A. J.; Champness, N. R.; Khlobystov, A. N.; Lemenovskii, D. A,; Li, W.-S.; Schroder, M. J. Chern. Soc., Chern. Cornrnun. 1997, 1339.)
c-axis, which defines enormous linear chambers. The counteranions PF6and CF3S03occupy channels that run along the fourfold crystallographic axes. Not a single metal ion but a tetranuclear silver cluster plays a role of a tetrahedral center in a diamondoid framework in [Ag2(02C(CH2)2C02)I (208). The compound consists of planar tetrasilver clusters bridged by succinate anions as shown in Fig. 59. The Ag, cluster exhibits pseudo-S, symmetry and is linked though its four ligands to four neighboring clusters, resulting in the formation of triply interpenetrated diamondoid frameworks. This is the first example of a three-dimensional coordination polymer containing metal clusters.
D.
T-T
INTERACTION IN DIAMONDOID FMEWORKS
Interpenetration in diamondoid frameworks is often mediated by interaction between aromatic bridging ligands (Table V). A pack-
T-T
250
MUNAKATA, WU, AND KURODA-SOWA
FIG. 59. A view of the Ag, cluster and the surrounding ligands in (CHz)zCOz)I. (From Fig. 1in Michaelides, A.; Kiritsis, V.; Skoulika, S.; Aubry, Chem., Znt. Ed. Engl. 1993,32, 1495.)
ing model for such systems proposed by Moore et al. (202) explains a relationship between the degree of the interpenetration and the bridged M----Mdistance, and also explains why two polymorphs can be obtained in [Ag(L7,),1AsFs(202). Face-to-face n--n--stacking interactions between Fets of L70 ligands in [ C U ( L ~ ~ ) ~ ~ B F ~ ( O(192) .~CH (3.659 ~ C ~and ~ ) 3.855 A) control the separations between adjacent diamondoid frameworks. The a-n--stacking interaction is clearly important in the overall control of this extended structure, and the presence of this interaction is along the direction in which the long-range structure is formed during crystallization. In case of [Cu(L&IX(L,)(THF) (X = BF4, C10,) (195, 1961,metalfree and coordinated LBB ligands stack alternately; with a nearest carbon-to-carbon separation of 3.37(2) and 3.51(2) A, to form a one-dimensional n--n-stacking column as shown in Fig. 55.
COPPERU) AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
25 1
r-a-Stacking columns in [Ag(L7,),I(C104)H,0(203)occur along both the a- and b-directions, with a plane-to-plane distance of 3.51 A. Along a , adjacent ligands in the stack are flipped 180" relative to one another such that nitrile groups are not overlaid (Fig. 57). The stacking along b is unusual in that a large offset angle causes a LT3molecule to stack between two fragments each consisting of two halves of L7, coordinated to silver(1).To allow stacking in two directions, mutually perpendicular stacks organize in a!ternating layers. A close interplane separation of 3.38 A between adjacent HL,, molecules in [Cu(HLC3),]X(X = PFs, CF,SO,) (164) suggests the presence of strong r-r interactions betwten pyridone rings. This separation is slightly shorter than the 3.40 A observed in the copper(I)-LGOcomplex (186).
VIII. Other Frameworks Based on Covalent Bonds
Apart from several of the classes of intermolecular forces based in supramolecules described earlier, coordination polymers with extended frameworks based mainly on covalent bond forces represent another well characterized class of transition metal supramolecules. One might anticipate that application of the strategies that have worked so well for the preparation of copper(1) and silver(1) coordination polymers with a diversity of novel frameworks based on intermolecular forces would allow a similar easy access to formation of the covalent-force-based frameworks by the selection of suitable ligands and spacers and the introduction of flexibilizing groups, either in the polymer backbone or in the side-group structure (Fig. 60). Recent publications did reveal several such notable examples of infinite chains, two-dimensional sheets, and three-dimensional networks formed by metal cations linked through multidentate organic ligands.
A. INFINITE-CHAIN STRUCTURES 1. Light-Induced Crystal Oscillation In addition to their novel frameworks, low-dimensional metalcontaining coordination polymers are of interest with respect to their electrical, magnetic, and optical properties, and possibly their catalytic behavior. For example, the study of photochromic compounds, which undergo thermal irreversible and fatigue-resistant photochro-
252
MUNAKATA, WU, AND KURODA-SOWA
L75
E
Me
1 b
Me-A-C-C-
-Me HZ HZ Me
c) L7Q
L78
181
180
L82
L86
N-
N L89
L90
191
192
FIG.60. List of ligands in copper(1) and silver(1) complexes with linear chain, 2-D and 3-D networks.
COPPERU) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
L99
L98
A c:v3
253
Lg7
H3
LlOl
mic reactions, is one of the key points in the current revival of interest in designing light-triggered molecular and supramolecular devices (209).Recent work by Munakata and co-workers has yielded a range of interesting and well-characterized materials with intriguing prop-
254
MUNAKATA, WU, AND KURODA-SOWA
erties (210, 211). Reaction of copper(1) perchlorate solution with cis1,2-dicyano-1,2-bis(2,4,5-trimethyl-3-thienyl)ethene (L75) yielded single crystals of [ C U ( L , ~ ) ~ ] C the ~ O ~structure , of which is composed of noninteracting anions and macrocations in which each metal center is coordinated with one CN group of the four dithienylethene molecules and each L,s in turn bridges two copper(1) ions with two cyano groups, leading to an infinite network of metal cations as illustrated in Fig. 61(a). The most striking feature of the complex is that it shows reversible ring-closed and ring-open transformation under different wavelengths of light when probed by both electrical and NMR spectroscopy. This cycle can be repeated many times, indicating that the reversible cyclization reaction takes place in the crystalline state. The light-induced crystal oscillation is suggested to be one possible pathway to account for the approach and rotation of the two thiophene rings as shown in Fig. 61(b). Such a system may be used as an optical memory. 2. Organometallic Polymers
Several organometallic polymers containing copper(1) and silver(1) ions have been reported that, by definition, involve unanimously direct interaction between the carbon atoms of the ligand and the metal ions. The organic ligands studied include alkynyls, isonitriles, and a-conjugated systems. Addition of AgC104 to truns-[Pt(C= CH)2 (PMezPh)J gave a 1 : 1adduct of [PtAg(C104)(C=CH)2(PMe2Ph)2] (212) in which the trans coordination of the Pt is maintained and the Ag atom is coordinated on the same face of each Pt coordination plane, forming a zigzag chain with a perchlorate ion in each cavity along the chain as shown in Fig. 62. Recently Bertrand reported the first bis(carbene1-silver polymer [Ag(L&F3S03 obtained at -30°C from an acetonitrilelether solution of 1,2,4-triazole-3,5-diylidene (L76) and AgCF3S03 (213). In the cation each ligand links two linear silver atoms with C-Ag-C bond angles of 175-180",and due to the alternation in the orientation of the five-membered rings, the complex is regarded as a one-dimensional polymer in which all the rings are coplanar. Despite the fact that silver(1)-aromatic complexes of benzene, cyclophane, indene, acenaphthene, naphthalene, and anthracene have been reported, the corresponding organometallic polymers of pyrene (Led and perylene (L6g) have been crystallographically characterized only recently (187).X-ray structure determination of the complex with pyrene [Ag2(Lm)(ClOJ21 reveals that it exists in the solid state as an
255
COPPERW AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
*
405nm
550nm
twisted ring-open form
coplanar ring-closed form
Light-induced crystal oscillation
FIG.61. Schematic presentation of chain structure (a) and the light-induced crystal oscillation model (b) in [ C U ( L , ~ ) ~ ~ C ~ O ~ .
arene-linked polymer of dimers as shown in Fig. 63. Within the dimer, there are two independent Ag(1) ions coupled by one perchlorate-oxygen bridging with Ag(l)----Ag(2) separation of 4.39 Each pyrene moiety exhibits a tetra-q2-coordination fashion sequentially bridging four metal centers, resulting in a polymeric W-type sandwich
A.
256
MUNAKATA, WU, AND KURODA-SOWA
L L
PMezPh
FIG. 62. Structure of the linear [PtAg(C104)(C-CH)2(PMe2Ph),3.(From Fig. 3 in Yamazaki, S.; Deeming, A,; Speel, D. M.; Hibbs, D. E.; Hursthouse, M. B.; Malik, K.M. A. J . Chem. SOC.,Chern. Commun. 1997, 177.)
of alternating aromatic and silver(1)perchlorate groups running along the b-axis as shown in Fig. 63(b). Several research groups have used isonitriles to prepare organometallic polymers (214,215).End-on coordination at the terminal carbons and linear geometry at the nitrogens impart to these molecular structures the characteristic of association of metal complexes into di-, tri-, and even polymeric species. Harvey has recently reported a series of new organosilver polymers of 1,8-diisocyano-p-methane (LT7), [Ag(L,,),]X (X = BF;, NO, or ClO;), with tubular structures in the solid state (214). Each ligand group bridges two tetrahedral silver atoms, forming zigzag chains that crystallize side by side separated by two rows of the counteranions (Fig. 64). Before this study two series of organosilver polymers with 2,5-dimethyl-2’,5‘-diisocyanohexane (L78)had been reported (215). Complexes of the first series contain one diisonitrile per metal with the formula [Ag(L78)IX(X = BF;, PF;, or NO;). The structures contain infinite chains of silver atoms alternating with bridging ligands. The second series crystallized as [Ag,(L78)31X2, and again, the extended trans-conformer of L7*alternating with Ag atoms generates infinite chains. However, the unique structural difference lies in a adjacent pairs of chains cross-linked by extra ligand groups, leading to a ladderlike pattern.
COPPERU) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
257
FIG.63. The W-type sandwich arrangement of an arene-linked polymer of dimers in [Agl(L,)(C104)2].(From Fig. 1 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Sugimoto, K. horg. Chem. 1997,36,4903.1
258
MUNAKATA, WU, AND KURODA-SOWA
0
J
FIG. 64. Structure of the polymeric chain in [Ag(L&]NO,. (From Fig. 1 in Fortin, D.; Drouin, M.; Turcotte, M.; Harvey, P. D. J. Am. Chem. SOC.1997,119, 531.)
3. Heterometallic Cluster Chain Heterometallic polymeric cluster compounds have received considerable attention due to their useful properties, which are derived from low-dimensional structures. Haushalter and co-workers recently initiated a study designed to determine the possibility of introducing unpaired electrons to closed-shell, low-dimensional Zintl-phase materials by suitable structural modifications or elemental substitutions, with the aim of increasing their electrical conductivity (216).Figure 65 shows the structure of one such example, (Et4N)4[A~(Agl-xA~,)2 Sn2Te9],which contains a semiconductive B-A-wide, one-dimensional chain composed of four different elements surrounded by insulating organic material. The magnetic susceptibility of the compound [Fig.
FIG. 65. Structure of the 1-D polymeric chain [ A U ( A ~ , _ ~ ~ ) (From ~ S ~ Fig. ~ T 1 ~ ~ I ~ ~ . in Dhingra, S. S.; Seo, D.-K.; Kowach, G . R.; Kremer, R. K.; Shreeve-Keyer, J. L.; Haushalter, R. C.; Whangbo, M.-H. Angew. Chem., Znt. Ed. Engl. 1997,36,1087.)
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
259
66(a)l indicates that the chain is diamagnetic in nature and is nearly independent of temperature down to 4 K. Both resistivity data obtained by the microwave cavity perturbation technique and optical diffuse reflectance measurements as shown in Fig. 66(b) and ( c ) respectively, support the claim that the compound is a semiconductor and the delocalized one-dimensional band structure undergoes a Peierls distortion above room temperature. Heterometallic polymeric cluster compounds [W4Ag5S161[M(dmf),l (M = Nd or La) derived from tetrathiotunstate and silver(1) have been reported (217). The onedimensional polymeric anion can be regarded as octanuclear cyclic cluster fragments, [WAg4SI6l4-, linked through Ag' ions as shown in Fig. 67. The significantly short W-Ag bond length of 2.928(1) A is observed in the analogous compound [Ag4WS41(NH4), suggesting that Ag- > WS, electron delocalization and stronger continuous metalmetal interactions are present in this type of heterometallic chains. Reaction of (NEt4),MS4(M = Mo or W) with CuCl and KSCN (or NH4SCN)afforded a set of mixed metal-sulf'ur compounds containing infinite anionic chains Cu,(NCS),MS:- and (CuNCS),WSP- or twodimensional polymeric dianions (CuNCS),MSi- (218). Short W----Cu distances ranging from 2.62 to 2.70 A are observed in these compounds. Incorporation of polyoxoanions into one-dimensional coordination polymer matrices, either through direct condensation to form oxo-bridged arrays of clusters or through transition-metal coordination compounds acting as inorganic bridging ligands, provides a route to composite organic/inorganic materials. In this respect, two composite compounds containing Cu(1) complex ions of 4,4'-bipyridine (L51), 8H20, have recently been pre[CU(L51)4MO@26]and [CU(L~,)~MO,,O~,~ pared by intercalation of molybdenum oxide clusters into the void space in the Cu-LSl subunits (219). By comparison, in a new mixedligand C d I ) polymer containing LS1,[(PPh3)2Cu2C12(L51)1, a different lattice appeared (220). This structure consists of (PPh3)2Cu2C12 units that are bridged by L51 (Fig. 68). The significance of these studies is that members of the vast family of polyoxoanion clusters may be employed as structural motifs to fill tunable void volumes created by extended cationic frameworks, and it would appear that this synthetic approach provides a method for structural modification of metal oxide and, consequently, tuning of electronic, magnetic, and optical properties of the oxide phases. 4. Mixed-Valence Copper Complex with Chain Structure
The mixed-valence copper complexes with polymeric chain structures have been described. Reaction of the cyclic thioether ligand tet-
260
MUNAKATA, WU, AND KURODA-SOWA
a
I
R I%
COPPER(I) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
261
FIG. 67. Structure of the l-D polymeric anion [W&g4Sl6l4-.(From Fig. 1 in Huang,
Q.; Wu, X.; Wang, Q.; Sheng, T.; Lu, J. Angew. Chem., Int. Ed. Engl. 1996,35,868.)
rahydrothiophene (LT9)with CuClz.2H20 in acetone yielded the polymeric, mixed-valence complex [ C U ~ ~ C U ~ ~ ( (221 L , ~).) ~The C ~structure ~I contains two distinct types of copper atoms: The divalent copper atoms involve an intemediate geometry between square planar and tetrahedral, comprising four chlorine atoms, and the monovalent copper atoms have distorted tetrahedral SzC12donor sets. The magnetic susceptibility data, ESR, and electronic reflectance spectra for the compound indicate that no intervalence interactions occurred between two Cu(1) and Cu(I1) sites.
5. Diradical Ligand Complex with Chain Structure Polymeric transition metal complexes of organic radical ligands have received recent attention in magnetic systems. The ligands concerned include the diradical 1,1',5,5'-tetramethyl-6,6'-dioxobis(verdazyl) (LBO). Reactions of LBO with copper(1)halides in acetonitrile or copper(I1) halides in methanol gave copper(1) complex [Cu2X2(L8,)I(2221, whose structure is found to be a polymeric chain with the tetrahedral copper atoms alternately bridged by two halide ions and two verdazyl diradicals [Fig. 69(a)l. Magnetic susceptibility measurements indicate that the spins couple in one-dimensional chains with alternating exchange parameters [Fig. 69(b)]. Although the intraligand electronic exchange is still antiferromagnetic in the system, the triple excited state is considerably stabilized compared with the free ligand as a result of an increasingly important superexchange through the copFIG.66. Physical properties of [Au(Ag,-,Au,),Sn~Te,l' . (a) Temperature dependence of the magnetic susceptibility ,y; (b) temperature-dependent resistivity p ; (c) optical diffuse reflectance R versus wavelength A of incident light. (From Fig. 4 in Dhingra, S. S.; Seo, D.-K.; Kowach, G . R.; Kremer, R. K.; Shreeve-Keyer, J. L.; Haushalter, R. C.; Whangbo, M.-H. Angew. Chem., Int. Ed. Engl. 1997,36,1087.)
262
MUNAKATA, WU, AND KURODA-SOWA
FIG.68. View of the [(PPh,)zCu,Clz(L,,)]chain. (From Fig. 2 in Lu, J.;Crisci, G.; Niu, T.; Jacobson, A. J. Znorg. Chem. 1997,36,5140.)
Temperature IK
FIG. 69. View of the chain structure (a) and temperature dependence of magnetic susceptibility (b) in [CuzXz(L,)I. (From Fig. 3 and Fig. 7 in Brook, D. J. R.; Lynch, V. Conklin, B.; Fox, M. A. J. Am. Chem. SOC.1997, 119, 5155.)
COPPERU) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
263
per(1) center. Another example of coordination of diradical with copper metal is reported by Oshio in a copper(I1) complex of imino nitroxide [CuYL8JI(PF& obtained from the reaction of [Cu1(CH3CN),I PF6 with LSl in methanol (223), where HLBl is 2-(1‘-oxy-4’,4’,5’, 5’-tetramethyl -4’,5’ - dihydro - 1’H-imidazole- 2‘- yl)- 6 - ( 1”- oxyl- 4,4”,5”, 5”-tetramethyl-4,5”-dihydro-1”H-imidazole-2”-yl)pyridine. In this complex the square planar copper(I1) ions are bridged by the iminohydroxyamino anions to form a one-dimensional helical structure. Here again, the magnetic susceptibility data revealed antiferromagnetic interactions. From the investigation of the coordination chemistry of these unusual diradicals it can be expected that the bridging ligands LBOand the polypyridyl type of imino nitroxide hold great promise as building components for future design and construction of polyradical magnetic materials.
6. Carboxylate Ligand Complexes Although the silver(1) ion is regarded as a typical soft Lewis acid, it forms a variety of coordination compounds with carboxylate ligands. Mak classified the coordination modes of Ag carboxylates into four types, most of which are constructed from either dimeric units or polymeric networks of dimeric subunits (224). The same author prepared and characterized two polymeric silver(1) complexes of betaine, ~ A g z ~ M e ~ N C H z C O O ~ z ~ H zand O ~ z [Agz(C~H~NCH~COO),(C10~)21 ~NO~~~l (224).The two complexes are structurally similar, both consisting of centrosymmetric carboxylato-bridged Ag,(carboxylato-O,0’)2 dimers, which extend into a stairlike polymer through the coordination of each metal center by a carboxylate oxygen atom from an adjacent dimer. It is worth noting that both polymers are found to involve very short intradimer Ag----Agdistances of 2.898(1) and 2.814(2) respectively, compared with other dinuclear silver(1) carboxylates.
A,
7. Thio-Ligand Complexes Although both copper(1) and silver(1) metal ions are expected to readily form coordination compounds with thio ligands by covalent bonds, few polymeric structures for thioether crown complexes have been reported. Among these few examples, two silver(1) complexes with thioether macrocyclic ligands are particularly interesting (225). Reaction of [24]aneS8(Lag)with 2 molar equivalents of AgCF3S03gave a one-dimensional polymer, [Agz(L82)(CF3S03)z(MeCN)21, in which each Ag atom is coordinated to four S-donors in a distorted tetrahedral geometry. As illustrated in Fig. 70, the four S-donors come from two different ligand molecules to generate an infinite ladder polymer
264
MUNAKATA, WU, AND KURODA-SOWA
0 0
0 0 0 0 FIG.70. View of the polymeric chain in [Ag2(L8,)I2+. (From Fig. 1 in Blake, A. J.; Li, W.-S.; Lippolis, V.; Schroder, M. J . Chem. Sac., Dalton Trans. 1997,1943.)
along the b-axis. A similar reaction with the crown thioether containing four S-donors, [16]aneS4(LES),gave a three-dimensional polymer, [Ag(LEs)(BF4)1, in which each Ag(1) ion is coordinated by four symmetry-equivalent S-donors in a tetrahedral geometry, and each ligand group in turn bridges four different Ag atoms. The work illustrates the potential of thioether crowns as building blocks for the synthesis of inorganic architectures using their ezo-orientated S-donors. The rodlike ligand 2,l l-dithia[3,3lparacyclophane (Lm)is found to be a particularly versatile organic ligand for the formation of polymeric chains in the presence of bridging anions (226, 227). The arrangement of such chains in the lattice is controlled and modified by the metal ions and coordination of a suitable anion. Two copper(1)and one silver(1) complex of LE4have been reported (226), in one of which, [ C U ~ B ~ ~ ( L ~ ) ( M the ~CN two ) ~Br ] , atoms bridge pairs of Cu(1) ions to form a rhombic CuBrCuBr ring as shown in Fig. 71 and each ligand molecule links two separate metal cations on each side through two sulfur atoms, resulting in a one-dimensional polymeric chain running parallel to a diagonal axis of the triclinic cell. Silver(1) complexes of 1,4-thioxane (L,) have been reported (228). Both [Ag(L,)(CF,SO,)I and [Ag2(L86)(CF3S03)21 were isolated from the mixture of Lg5 and silver triflate at ambient temperature in a 1: 1 mixture of dichloromethane and acetonitrile. The former involves silver(1) cations tetrahedrally coordinated by two sulfur atoms of two Lgg molecules and the oxygen atoms of two bridging triflate anions, giving in the crystal an infinite one-dimensional chain with bridging only through the triflate moieties. Although the immediate coordination environment around the silver(1) ion in the latter complex is similar t o that in the former, bridging between Ag' cations occurs via the participation of both 1,4-thioxane sulfur and triflate oxygen atoms, leading to a two-dimensional lattice.
COPPERU) AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
265
FIG. 71. View of the infinite chains in [CuzBrz(Llcl)(MeCN),l. (From Fig. 1 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Nakagawa, S. J. Chem. Soc., Dalton Trans. 1996,1525.)
Silver(1) complexes of the nonchelating dithioether ligand have been described (229). Crystallographic studies of [Ag(MeSCH&H, CH2SMe)lBF4revealed a chain structure involving trigonal (p,-S),,S ligation with one S donor bridging two adjacent Ag cations while another S donor of the dithioether ligand is nonbridging. In contrast, revealed monodenX-ray analysis of [Ag(PhSCH2CH2CH2SPH)21BF4 tate S ligation of four distinct dithiother ligands to the Ag atom, giving an approximately tetrahedral geometry at the metal ion, with the second S donor of each ligand linking to four other Ag' ions, hence generating a three-dimensional infinite lattice. Due to the biological relevance of copper-sulfur coordination compounds to metalloproteins, the copper(1) complex of thiourea, ICU{SC(NHCH~)~)~INO~, has been prepared from an aqueous solution of copper(I1) nitrate and N,"-dimethylthiourea (230).The structure contains substantially distorted CuS, tetrahedra, each being linked by common edges with two neighbors, resulting in infinite isolated chains parallel to the c-axis of the cell. The nitrate anions are not coordinated to the silver(1) ions, but associate the individual chains of edge-sharing CuS, tetrahedra together. 8. Other Nitrogen-Containing Ligand and Mixed-Ligand Complexes
Mingos and co-workers have recently reported several novel silver(1) complexes with simple polydentate acyclic nitrogen-donor ligands such as diethylenetriamine (dien) and tris(2-aminoethy1)amine
266
MUNAKATA, WU, AND KURODA-SOWA
(LE6)(231). The dien complex [Ag(dien)lPF, is slightly light-sensitive, but may be stored indefinitely under nitrogen at -30°C in the dark. The complex consists of two crystallographically independent silver atoms and two nonequivalent dien ligands, which link adjacent T-shaped silver centers to form a catena structure. The tren complex [Ag(LBs)]PF6 also has a zigzag chain structure and possesses similar structural features except that the two silver(1) ions involve distorted tetrahedral geometries and tren acts essentially as a tetradentate ligand. Luminescent coordination compounds continue t o attract considerable attention. Zink recently reported a new mixed-ligand copper(1) polymer that shows interesting photoluminescence (232). The complex [CuCl(L,,)Ph,P] consists of a one-dimensional chain lattice of metal ions bridged by both C1- ions and pyrazine molecules. The compound shows conductivity of less than lo-* S cm-'. The absorption spectrum of the complex shows a band at 495 nm, which could be interpreted as the promotion of an electron from the valence band to the conduction band. On the basis of resonance Raman spectra, the lowest excited state in the polymer is assigned to the Cu(1)-topyrazine metal-to-ligand charge-transfer excited state. In the previous section, l-cyanoguanidine (L5,) was quoted as an example of the multifunctional ligands for the formation of H-bonded frameworks (157). Because this molecule can act either as a monodentate ligand (nitrile N donor) or as a bidentate bridging ligand (nitrile and amino N donor) a number of 1 : l and 1 : 2 l-cyanoguanidinecopper(1) halide adducts have been reported (233). The complex [CuX(L5,)I,where X = C1- or Br-, is characterized by two mutually perpendicular chains of [CuX(L,,)], and (CuXi), joined at a common halogen. Within these two individual chains the copper atoms are found to involve different distorted tetrahedral and trigonal planar stereochemistries, respectively. Among the mixed-ligand-complexed copper(1) complexes, the one formed by pyrazine (Lso) and 3-methylpyridazine (L8,) deserves to be mentioned (234). The compound [cuZ(L8,)Z(LS~>~1(c104)2 was prepared by the stepwise reaction of copper(1)with two ligands. The two copper centers involve different coordination environments, one in a threecoordinate Y-shaped form and the other in a typical four-coordinate tetrahedral form. The structure consists of a tetracopper unit bridged alternately by Lg, and LDBgroups (Fig. 72). In addition, the cyclic voltammogram of the complex reveals two reversible redox couples with Ellz = +0.22 and +0.53, which are ascribed to the Cu'Cu'/Cu'Cu" a
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
0
4-coordinate copper
0 3-coordinate copper
+
267
inUaunit prz
nmpz
FIG.72. Structure of an infinite cationic chain in [Cu2(L,,),(L,,),I(C10,),. (From Fig. 1 and Fig. 2 in Kitagawa, S.; Munakata, M.; Tanimura, T. Chern. Lett. 1991, 623.)
and CurCu"/CurrCu" reactions of the closely contacting dicopper moiety, respectively. In a n attempt to use dicopper(1) complex unit as building block for synthesis of metal-containing polymers, Kitagawa reported his interesting results in synthesis and structural characterization of a n infinite stair-type chain compound (235).The complex of thiochrome (L88) [CU~(L,&I(C~O,)~ was prepared by reduction of copper(I1) perchlorate followed by reaction with the ligand. The structure contains a dinuclear cationic unit in which the two T-shaped metal ions are doubly bridged by two thiochrome molecules with extremely short Cu----Cu
268
MUNAKATA, WU, AND KURODA-SOWA
A.
contact of 2.476(3) The dimeric units are mutually linked by the 2-hydroxyl group of the coordinated ligand group, giving a novel stairtype chain. The cyclic voltammogram in the solid state gives a single quasi-reversible wave, indicating that the dicopper site undergoes redox reactions. Coordination compounds of biheteroaromatic ligands have been extensively studied (236). The nitrogen-rich ligand, 3,5-bis(2-pyridyl) pyrazole (Lee),in which the pyrazole ring is attached to bulk nitrogen heterocycles a t the 3-and 5-positions, belongs to this class. Polymeric silver(I) complexes containing the uninegative pyrazolate ion had been known for over a century when the formation of an insoluble silver pyrazolate salt Ag(pz) was reported; however, its structure has never been crystallographically established (237). Fortunately, the structure of the polymeric silver(1) complex of the pyrazole derivative ligand Leuhas been determined recently (238).Along with it, Munakata has also reported a copper(I1) and another silver(1) complex with Lsa. The reaction of Leuand AgC10, in different solvents at various molar ratios surprisingly leads to the isolation of two totally different - 2MezC0, and one silver(1) complexes, one dimer, [Ag2(L8u)41(C104)z polymer, [Ag(L,,)]ClO,. In the polymeric structure the coordination around the metal ion involves a distorted trigonal planar core, AgN3, and each LSl molecule bridges two metal centers, giving an extended zigzag chain as shown in Fig. 73. Between the chains there is no effective interaction present due t o the presence of anion column separating the cationic chains. Like Lee,hexamethylenetetraamine (LW)is also a potentially tetradentate ligand. Two silver(1) coordination polymers with this polydentate N-donor base have been reported, and one involves a one-dimensional lattice (239). Reaction of AgSbF, and Luoin a 1: 1 molar ratio in ethanol-CHuClz isolated the compound [Ag(Luo)lSbF,.HzO. The structure contains a novel type of ribbon formed by hexagonal meshes of alternate Ag' ions and Lw,molecules. The water molecules coordinated to the silver ions form O-H----N hydrogen bonds with the bo groups of adjacent ribbons, which generate a three-dimensional network of unprecedented topology (Fig. 74). In terms of Pearson's hard-soft acid-base principle, the soft Lewis acid copper(1)ion is not compatible with the hard base HzO molecule, but a recent article provided the first example of a copper(1)-water bond (240).The novel copper(1) complex with 2,3-diphenylquinoxaline (Ld, [Cu(LUJH~O)IC10,,is diamagnetic and indefinitely stable in air in the solid state. Its structure consists of an infinite polymeric cationic chain in which adjacent metal centers are bridged by the aro-
COPPERU) AND SILVERII) SUPRAMOLECULAR METAL COMPLEXES
269
FIG.73. View of the chain structure in [Ag(L,)]-. The dashed line shows the contact between the nearest silver atoms, Ag----Ag5.67 A. (From Fig. 3 in Munakata, M.; Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Kawata, S.; Kitagawa, S. J . Chem. SOC.,Dalton Trans. 1995, 4099.)
matic nitroeens of the bidentate ligand of bl.The Cu-0 bond length of 2.154(6) A is longer than the calculated distance of 1.99 A based on the bond-valence-sum models. Therefore, the coordination of the copper is best described as distorted trigonal planar CuNIO with a weak interaction with the water molecule.
Q
Q
FIG. 74. One-dimensional ribbon in [Ag(L,,)lSbFs. H,O. (From Fig. 2 in Bertelli, M.; Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J. Muter. Chem. 1997, 7, 1271.)
270
MUNAKATA, WU, AND KURODA-SOWA
Several copper(1) and silver(1) complexes with trans-1,2-bis(2-pyridy1)ethylene (Lo,) have been reported (241). In [M(L,dIX, where M = Cu or Ag and X = PF, or ClO;, a distorted linear coordination geometry completed by the nitrogen atoms of two different ligands is observed for both Cu' and Ag' ions. Both complexes have a similar chain structure, and they only differ in the mode of polymerization, being of rectangular and triangular wave chain types, respectively (Fig. 75). Additionally, the pyridine ring contact between the adjacent chains in the copper complex is 3.46 indicative of r-r interactions present. In another copper(1) complex, [Cu(Lo,)(CO)(CH,CN)]PF,,an infinite
A,
b
i (b)
FIG. 75. ORTEP views of the cationic chains in [Cu(L,)IPF6 (a) and tAg(L)1C1O4 (b). (From Fig. 1 in Powell, J.; Horvath, M. J.; Lough, A. J. Chern. SOC.,Dalton Trans. 1996,1669.)
COPPER(I) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
271
one-dimensional chain based on bridging of Lg2between two metal ions is also observed. However, with additional coordination of carbony1 and acetonitrile, each metal center involves a tetrahedral core CUN~N'C. Coordination compounds containing simple alkyl and aryl halides have been structurally characterized (242, 243). Novel examples include silver(1)-iodocarbon complexes obtained from the reaction of AgPFs and AgN03 with diiodomethane, 1,3-diiodopropane, and simple aryl iodides (243).The compound [Ag(NO,)(l,2-12C6H4)lis composed of spiral chains of [Ag(p-N03)I, as shown in Fig. 76, in which the nitrate uses only one oxygen atom to bridge two silver ions. In [Ag{I(CH2)31}al PF6,[Ag(ICH21)21PF6, [Ag(l,2-I2C6H4),lPF6, and [Ag(l,2-BrIC6H4)41PF~, the structure consists of tetrahedrally coordinated Ag' ions and bridging iodocarbon ligands, giving rise to a chain polymer array.
STRUCTURES B. TWO-DIMENSIONAL 1 . Polycatenane and Polyrotaxane Complexes
Catenanes are a group of interlocked or catenated molecules, and rotaxanes are a class of molecules in which a dumbbell-shaped compo-
FIG.76. Structure of the spiral chains in IAg(N03)(1,2-I,C,H,)].(From Fig. 8 in Powell, J.; Horvath, M. J.; Lough, A. J. Chem. SOC.,Dalton Trans. 1996, 1669.)
272
MUNAKATA, WU, AND KURODA-SOWA
nent is encircled by a macrocycle. The efficient synthesis of such systems fascinated and inspired many scientists in the 1960s and 1970s (244). Due to the introduction of metal templating and supramolecular approaches for molecular threading and interlocking, this area has experienced a renaissance in recent years (245,246).Multicomponent polycatenane and polyrotaxane molecular systems incorporating metal-ion templates represent a particularly interesting class of supramolecular species that might exhibit intriguing chemical topology and display some fascinating electronic, optical, and magnetic properties (247-251 ). Schroder and co-workers have successfully isolated a unique polycatenated undulating molecular ladder that forms interwoven twodimensional sheets (252). The reaction of [CU(M~CN)~]PF~ with 1,4bis(4-pyridy1)butadiyne (Lg3)in MeCN-CH2C12 yielded the complex [ C U ~ ( M ~ C N ) ~ ( L ~ ~ )The ~ ] (compound P F ~ ) ~ . exists as a network of molecular ladders in which the two independent Cu' centers are each coordinated in a tetrahedral geometry t o three LBsgroups and one MeCN molecule. The lattices are polycatenated to give a remarkable twodimensional layer structure (Fig. 77). The sheets of interwoven molecules are separated by PF, counteranions and solvent molecules. The structure is further stabilized by the 7r-v interactions between adjacent, symmetry-related ladders. Robson has constructed a two-dimensional polyrotaxane sheet from the rodlike ligand 1,4-bis(imidazol-l-yl-methyl)benzene(I&,) (253). The compound [Ag2(L,)3(N03)21 was obtained by reaction of LB4and silver nitrate in aqueous methanol. The coordination polymer consists of one-dimensional chains in each of which the metal center is located
& -l
L93
FIG. 77. View of the polycatenated sheet structure in [Cu2(MeCN),(L,),I(PF,), . (From Fig. 3 in Blake, A. J.; Champness, N. R.; Khlobystov, A.; Lemenovskii, D. A.; Li, W.-S.; Schroder, M. J . Chem. SOC.,Chem. Cornmun. 1997, 2027.)
COPPER(1) AND SILVER(I1 SUPRAMOLECULAR METAL COMPLEXES
273
within the plane of the N3 donor set. The individual chains are associated t o produce the polyrotaxane sheets as shown in Fig. 78, in which the LDpmoiety provides both the rodlike segments and two half-loops for each of the rings. K m and Whang have constructed a polyrotaxane containing cyclic beads threaded on two-dimensional coordination polymer networks (254). The compound [Agz(L9s)a~cucurbituril)31(N0,),~ 40H20, where LDs = N,N’-bis(4-pyridylmethyl)-1,4-diaminobutane dihydronitrate and cucurbituril = C36H36N24012, was prepared by the route shown in Fig. 79. The compound is a polyrotaxane in which the cyclic beads of cucurbituril are threaded on a two-dimensional coordination polymer network. The two-dimensional polyrotaxane network forms layers stacked on each other, which in turn fully interlock with themselves. 2. Two-Dimensional Network Comprising Copper Clusters Copper(1) thiolate complexes show remarkable diversity of structure, and the most common structural unit in the known thiolate is a three-coordinate copper atom bonded to p2-bridging thiolates (15). Parish has recently reported a novel compound in which the thiolate sulfur atom involves unusual four-way bridging (2551. The compound with composition [ C U , ~ C ~ ~ ~HzO ( S R(R ) ~=I CHzCHzNH3) ~ was obtained by reaction of CuCl and cysteamine hydrochloride in an aqueous solu-
FIG.78. A 2-D polyrotaxane network in [Ag,(L94),3(N03)21. (From Insertion Fig. V and Fig. 2 in Hoskins, B. F.; Robson, R.; Slizys, D. A. J.Am. Chem. SOC.1997, 119, 2952.)
274
MUNAKATA, WU, AND KURODA-SOWA
FIG.79. Construction scheme for the 2-D polyrotaxane network of [Ag,(L&(cucurbi40Hz0. t~rilhI(N0~ ) ~ . (From Scheme 1in Whang, D.; Kim, K. J. Am. Chern. SOC.1997, 119, 451.)
tion. X-ray analysis revealed a polymeric structure containing a centrosymmetric Culz cubo-octahedron unit. A regular octahedron of sulfur atoms, bearing organic chains pointing outward, has each edge bridged by a CuCl unit. There are thus three mutually perpendicular, planar, eight-membered [(CuCl)(p2-SR)],rings intersecting at the sulfur atoms, giving a gimba-like “Atlas-sphere” configuration (Fig. 80). The Cu12C112(SR)6 clusters are further linked to form two-dimensional sheets with different linkages in the two directions. 3. Two-Dimensional Network Containing Hexagonal Meshes and a Zeolite-like Framework
Ciani and co-workers have reported the self-assembly of two remarkable polymeric networks based on an anionic acetonyl derivative of tetracyanoethylene (256).The attempts to obtain novel extended frameworks of Ag ions bridged by the neutral tetracyanoethylene
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
275
FIG.80. Structure of the basic [Cu13C113(SR)61 unit and the layers of clusters forming sheets. (From Fig. 1 in Parish, R. V.; Salehi, Z.; Pritchard, R. G. Angeru. Chern., Int. Ed. Engl. 1997, 36, 251.)
have led to the unexpected formation of [Ag(L,)I, where L, = 1,1,2,2tetracyanopentan-4-one-1-ide. The structure consists of twodimensional undulated nets of hexagonal meshes formed by alternate three-coordinate silver ions and tridentate anion LBs(Fig. 81). Two independent nets of this type interpenetrate to give a layer structure. The targeted synthesis of [ C U ( M ~ C N ) ~ ]with P F ~the preformed conju-
276
MUNAKATA, WU, AND KURODA-SOWA
FIG.81. View of the twofold interpenetrated layer of [Ag(L,)l. (From Fig. 1 in Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Angew. Chem., Znt. Ed. Engl. 1996, 35, 1088.)
gated acid of LBBin acetone gave the corresponding copper(1) polymer [Cu(LBB)I. Me2C0.Although the compound holds the same stoichiometry as the silver complex, its structure consists of a three-dimensional neutral network of four-coordinate copper(1)ions linked by anions LW. The framework resembles a honeycomb with large channels that include guest acetone molecules. The large difference between the two structures can be related t o the higher tendency of Cu' ions over Ag' ions to achieve a tetrahedral coordination. Therefore, by using this methodology it should be possible to generate other topologies with different metal ions and/or with similar polycyano anions. The assembly of a long-range solid-state superstructure with a zeolite-like framework has been exploited by a number of research groups (128,227).Recently, it has proven possible to synthesize coordination networks with microporous structures that maintain crystal integrity upon loss of guest species. Such zeolite-like behavior of a coordination network represents a new challenge for chemists to look for new class of microporous substances. Reaction of silver triflate and the tritopic ligand 1,3,5-tris(3-ethynylbenzonitrile)benzene(Lu,) yielded solid [Ag(L,,)(CF,SO,)]. 2C6H6 (257).The compound contains a two-dimensional network in which the coordination geometry around silver is trigonal pyramidal, with three nitriles of the network in the basal plane and a trilate counterion in the apical position as shown in Fig. 82. The sheets are stacked, creating a channel structure in which solvent benzene molecules reside in the micropores. The com-
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES A
277
t
FIG. 82. Structure and coordination network in [Ag(L,,)(CF,SO,)I.(From Fig. 1 in Venkataraman, D.; Gardner, G. B.; Lee, S.; Moore, J. S. J.Am. Chem. SOC.1996,117, 11600.)
pound shows stability toward partial loss of inclusion molecules. The TGA trace of the compound reveals two discrete mass losses at 110 and 145°C (Fig. €431, corresponding to the removal of distorted benzene molecules. At 145”C, a solid-to-solid phase transition occurs concomitant with the loss of the remaining benzene molecules. DSC and optical microscopy indicate no evidence of a phase change associated with the first mass loss. Further observations demonstrate that the sample can reversibly remove and reabsorb benzene molecules without collapse of the channel network analogous to that of zeolite-type materials. 4. Two-Dimensional Network Containing an Undulated Wave
Layer Structure In the Section VIILA, “Infinite-Chain Structures,” we noted that Mak et al. characterized a number of carboxylato-bridged silver(1) complexes with chain structures (224). The same research group has
278
MUNAKATA, WU, AND KURODA-SOWA
temperature ("C)
FIG.83. Overlay of TGA and DSC traces for [Ag(L,,)(CF,SO,)]. (From Fig. 2 in Venkataraman, D.; Gardner, G. B.; Lee, S.; Moore, J. S. J.Am. Chem. SOC.1996,117, 11600.)
also employed dicarboxylate-like ligands rneso-2,5-bis(trimethylammonio)adipate (Log)and rneso-2,5-bis(pyridinio)adipate(L,) to synthesize silver(1) complexes with two-dimensional networks (258).Structure analysis revealed that in each of the four complexes [Agp(LQe)l(C104)z, [Agz(LQ9~I(C104)z, [Agz(L~g)l(NO,),,and [Ag~(Lw)l (NO,), , a dimeric unit is involved in which each pair of adjacent metal atoms is doubly bridged by coplanar syn-syn p-carboxylate-0,O' groups. The dimeric subunits are extended into a step polymer through the linkage of each metal center to a carboxylate group of an adjacent dimer. The basic coordination environment for the metal is completed by three carboxylato oxygen atoms in a T-shaped geometry, if the interaction between the anion and the metal center is ignored, as occurred in two perchlorate complexes. Because the nitrate anion functions as a unidentate and 0,O'-bridging mode in the third and the last complex, respectively, it leads correspondingly to a distorted tetrahedral and unusual square pyramidal five-coordination at Ag' ion. Figure 84 shows the structure and packing drawing of the layer structure of the five-coordinate silver complex. All compounds exhibit short Ag----Agseparation in the dimeric unit, ranging from 2.794(1) A to 2.878(2) A, substantially smaller than that in metallic silver (2.89A). It is debatable whether these distances are a result of the bridging ligand or suggestive of strong metal-metal interactions present. Crown thioether can not only form a one-dimensional chain structure, it can also form two-dimensional sheet frameworks (225).The crown thioether ligand (OH),[lGlaneS, (Llm)reacted respectively with silver nitrate and silver acetate, giving two polymeric silver(1) com-
COPPER(I) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
279
FIG.84. Structure skeleton of the asymmetric unit (a) and packing drawing of the layer structure (b) in [Ag2(L,)(NOa)pl.(From Fig. 5 in Wu, D.-D.; Mak, T. C. W. J. C h e n . Soc., Dalton Trans. 1995, 2671.)
plexes, [Ag(Lloo)lOzCMe and [Ag(Lloo)lN03 (259).The two compounds are not isostructural, but their structures and dimensionality seem to be controlled by change of the anions. In the complex with silver acetate, each metal center is bound to four thioether sulfur atoms in a rather distorted tetrahedral fashion. The silver atoms are in a twodimensional sheet arrangement interconnected by the macrocycle in which two crystallographically independent Ag' ions are stacked alternately along a diagonal to the a- and c-axes of the unit cell (Fig. 85). By contrast, the complex with silver nitrate consists of a three-
280
MUNAKATA, WU, AND KURODA-SOWA
-C
b
FIG.85. Packing diagram for [Ag(LlOo)lO2CMe with dashed lines showing the hydrogen bonds. (From Fig. 4 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M. J. Chern. SOC.,Dalton Trans. 1995,3215.)
dimensional tetrahedral network linked by an exodentate coordination of the thioether.
5. Conductive r-Complex Although the silver(1)-pyrene complex [Agz(Lee)(CIO&]is found to display a chain structure as already mentioned, a similar complex with perylene, [Agz(L6s)(C104)z], consists of a two-dimensional framework of the metal ions bridged by the bidentate counterions and the tetra-$-arene groups shown in Fig. 50 (187). Such two-dimensional sheets are further connected by the superpose! intersheet aromatic rr-r stackings at an average distance of 3.31 A, generating a threedimensional network structure. The physicochemical property measurements of two complexes show that a t room temperature no strong ESR spectrum was observed for each case, but the light-irradiated samples show a characteristic g value attributable to the aromatic radicals, suggesting that upon irradiation electron transfer partially takes place from the aromatic donor to the silver (I) ion, giving an organic radical cation and silver(0). In addition, both compounds are electrically nonconducting, whereas their Iz-doped samples display semiconducting behavior at ambient temperature, which is presumed
COPPER^ AND SILVER^ SUPFLAMOLECULAR METAL COMPLEXES
281
to connect mainly with the nonintegral formal oxidation state of the crystallized organic radicals. These conducting radical cation complexes might constitute new functional materials among lowdimensional molecular solids similar to polymetallocenes and multidecker metallacarboranes. A further example of a silver(1) n-complex with a two-dimensional sheet framework is found in a silver(& (243). The solid-state iodocarbon complex, [Ag2(02PF2),(p-IC6H4Me)l structure of the compound is composed of an infinite two-dimensional sheet array of [Ag(O,PFz)l, in which each p-iodotoluene bonds to two Ag' ions via one I-Ag bond and an $-arene bond involving two of the carbons not bonded to I or Me. 6. Inorganic Polymers
The development of inorganic polymer chemistry based on incorporation of transition metals into a polymer main chain has resulted in diverse arrays of interesting structures in recent years. These polymers are expected to constitute a new type of inorganic functional material. It is not the aim in this section to undertake a comprehensive treatment of this field; for this, the reader is directed to the recent reviews and the references herein (260,261).Here we just list a few examples of inorganic polymers that incorporate copper(1) or silver(1) ions with two-dimensional sheet frameworks. Pfitzner has carried out a systematic investigation on new adducts of copper(1) halides with neutral chalcogen fragments or polychalcogenide anions (262).One of them, (CuI)&uzTeSy,was found to contain layers of the complex thioanion [TeSJ- embedded between layers of iodide ions. Copper atoms are distributed in the arrangement of iodine and sulfur atoms. Kanatzidis and Zhang have utilized sulfur-rich mixed polysulfide/ telluride fluxes in combination with transition metals to synthesize novel solid-state materials (263). Thus, six new quaternary copper(1) and silver(1) compounds with composition AMTeS3,where A = K, Rb, or Cs and M = Cu or Ag, were prepared. The structure consists of anionic [MTeSJt- layers and charge-compensating alkali ions between the layers. Each layer is composed of tetrahedrally coordinated Cu' or Ag' centers and trigonal pyramidal TeSi- units, joined together via bridging S atoms. Similar low-dimensional quaternary compounds, KCu,AsS3 and KCu4AsS4,have also been reported, and a unique layered framework was observed in both compounds in which the Cu(1) ions are linked in a complex manner by a series of trigonal ASS:- groups or ASS:- groups as well as S2-ions (264).
282
MUNAKATA, WU, AND KURODA-SOWA
7. Inorganic Grid and Heterobimetallic Aggregates Inspired by the idea that an n-topic ligand would form an n X n inorganic grid based on n2metal ions and 2n ligand components, Lehn and co-workers have reported the self-assembly of the inorganic 3 X 3 grid consisting of an array of nine silver(1) metal ions and six tritopic ligands (265). The complex [Ag,(Llol)sl(CFBSO~)9, where Llol = 6,6’-bis[2-(6-methylpyridyl)]-3,3’-bipyridazine, contains an arrangement of a 3 X 3 grid of nine Ag’ ions chelated by six bidentate ligand molecules as shown in Fig. 86. Such an architecture opens the way to a whole family of polynuclear inorganic m X n grids, which represent a new facet in the controlled arrangement of metal ions into specific arrays and patterns. Among the previously discussed heterobimetallic aggregates of copper(1) with thiotungstates and thiomolybdates, those containing the two-dimensional polymeric dianions (CuNCS),MSq- are noteworthy (218). X-ray diffraction results show that four edges of the tetrahedral MSq- (M = Mo or W) core are coordinated by copper atoms forming WS4Cu4 aggregates linked by eight-membered rings of - CU(NCS)~CU-(Fig. 87). 8. Two-Dimensional Sheets Containing Cu6Hexagons
The rodlike ligand La was also found to form a polymeric complex with copper(1) (2261, but the complex obtained by reaction of La with 1 molar equivalent of CuI in thf, [Cu2I2(La),1.thf, is surprisingly not isostructural of the corresponding complex of the bromide. Instead, it
FIG. 86. Crystal structure of the self-assembled 3 X 3 inorganic grid [Ags(L,,,),] (CF,SO,),. (From Fig. 3 in Baxter, P. N. W.; Lehn, J.-M.; Fischer, J.: Youinou, M.-T. Angew. Chem., Znt. Ed. Engl. 1994,33,2284.)
COPPER(I) AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
283
6
FIG. 87. Perspective view of zig-zag anionic chains in (CuNCS),MS; . (From Fig. 2 in Manoli, J. M.; Potvin, C.; Secheresse, F.; Marzak, S. Inorg. Chem. Actu 1988, 150, 257.)
is composed of a two-dimensional framework in which a large ring containing six copper atoms includes a thf molecule as guest inside the open cavity (Fig. 88). As in the bromide, the rhombic CuICuI ring formed by bridging the two iodide ioas between pairs of copper atoms leads to a Cu----Cudistance of 3.18 A and an I----Idistance of 4.30 A. Each L, fragment bridges two separate metal cations through two sulfur donors, resulting in a two-dimensional sheet arrangement of copper(1) ions. The framework contains six-membered rings of Cu6 hexagons incorporating the thf molecule in the open cavity. In contrast to the formation of the rhombic ring by copper(1)halide ions, the one-nitrate oxygen bridging two metal ions in the silver(1) complex [Ag(L,,)(NO,)l yielded a unique three-dimensional channel framework of silver(1) ions (226). For these reasons, both metal ions and counteranions are considered to be the fundamental factors controlling crystallization and variation of frameworks.
STRUCTURES C. THREE-DIMENSIONAL 1. Construction of Channeled Frameworks
The rational design of multidimensional coordination compounds incorporating large cavities or channels capable of hosting small mol-
2 84
MUNAKATA, WU, AND KURODA-SOWA
(b)
FIG.88. Molecular packing diagram of complex [Cu212(L&I. thf (a) and a view of the 2-D sheet consisting of Cue hexagons (b) where only the metal centers are presented as open circles. (From Fig. 3 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Nakagawa, S. J. Chen. Soc., Dalton Trans. 1996, 1525.)
ecules is currently recognized as one of the most important issues in the synthesis of functional materials for molecular sieves, shapeselective catalysis, ion exchangers, and many other applications ( 4 , 266-270). The approach generally employs rigid and highly symmetric organic units as building blocks via intermolecular forces such as
COPPER^ AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
285
hydrogen bonds (159-163) or coordination to metals. Such zeolite-like behavior of the coordination network in copper(1) and silver(1) complexes is not limited to two-dimensional structures (257). In fact, the majority of examples are found in three-dimensional copper and silver complexes of nitrogen ligands derived from organocyanides (121-128, 271 1, 4,4’-bipyridyl, and hexamethylenetetraamine. The simple ligand 4,4’-bipyridyl has been chosen by many groups as a rod linking together metal centers to give extended solids with diverse topologies (272-275). Reaction of AgNO, and L51 in MeCN gave the silver(1) polymer [Ag(L,,)INO, (273,274j. The crystal structure reveals a triple interpenetrated cationic network schematically shown in Fig. 89, each consisting of linear [Ag(L51)lru chains cross-linked by Ag-Ag bonds of 2.970(2)A. The resultant open framework contains large interconnected cavities that form stairlike microchannels of approximate cross section 6 X 23 A. Rather small channels are observed in the corresponding complex of copper(1) chloride, [Cu(LS1)C1], obtained by diffusing solutions of CuCl and L5,into ethylene glycol (275). The structure contains chloro-bridged Cu’ dimers, which form a sheet framework with 4,4’-bipyridyl molecules. The overall structure thus formed is a neutral three-dimensional framework, in which the interpenetrating two-dimensional sheets fill yp most of the pore space with channels of only small diameter (2 X 4 A) remaining open, too small to accomodate acetonitrile or DMSO solvent molecules as guests in the crystal. Paralleling the approach used in supramolecular organic chemistry of adamantyl templates with different synthons (266-268), Ciani and
FIG.89. Schematic representation of the 3-D framework in [Ag(Lsl)lNO:,,with circles representing Ag cations, filled bars representing L51 and empty bars representing Ag-Ag bonds. (From Scheme 1-D in Robinson, F.; Zaworotko, M. J . Chem. Soc. Cheni. Commun. 1995, 2413.)
286
MUNAKATA, WU, AND KURODA-SOWA
co-workers have developed a scheme using the potentially tetradentate ligand hexamethylenetetraamine (L,) for the preparation of supertetrahedral networks with metallic synthons (239,276-278). They employed reaction of Lm with different silver(1) salts of noncoordinating anions and isolated a number of polymeric silver complexes with three-dimensional frameworks. The first of this series, [Ag(L,)I PF6.HzO,was reported in 1995 (276). The compound was obtained by slow evaporation of an ethanolic solution of AgPF6 layered on a solution of L, in CHzClz.The structure contains a molecular-based framework topologically related t o the three-dimensional, three-connected cubic net of highest symmetry as schematically shown in Fig. 90. The PF; anions and the guest water molecules occupy the large octagonal channels and form an extended hydrogen-bonding network. Similar reactions under different conditions afforded three other compounds. The structure of [Agll(L,)6](PF6)l,.14Hz0 consists of an open threedimensional cationic frame, with all the Lm molecules acting as tetradentate ligands while nine silver cations are biconnected and two are triconnected to the ligands per formula unit (239). The (3,4)-connected network schematically shown in Fig. 91 is composed of triconnected (silver ions) and tetraconnected (L,) centers in the ratio 1: 3. Such connection generates two types of parallel channels to host the anions and the water molecules. The compound [Ag5(LW))6](PF6)5. 3CHzC12is - 3EtOH gives a threean oligomer, whereas [Ag,(Lso)(HzO)l(PF6)4 dimensional network with large cavities and channels of the hexagonal sections, including anions and solvent molecules as guest species (277). in a different molar ratio yielded two Reaction of AgC104 and LBO polymeric complexes (278). The crystal structure of [Ag(L,)IClO, consists of two-dimensional infinite layers of hexagonal meshes formed by alternate triconnected silver ions and L, molecules. The structure
FIG.90. A schematic view of the 3-D network in [Ag(L,,)IPF,. HzO. (From Chart 1 in Carlucci, L.; Ciani, G.; Proserpio,D. M.; Sironi, A. J. Am. Chem. Soc., 1995,117, 12861.)
COPPERU) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
287
FIG. 91. A schematic view of the 3-D network in [Agll(Leo)sl(PFs)ll~ 14H20. (From Fig. 7 in Bertelli, M.; Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J. Muter. Chem. 1997, 7, 1271.)
of [Ag3(Lso)21(C104)3. 2H20 contains an open three-dimensional cationic network formed by [Ag(Lso)]hexagonal layers, which are joined by biconnected silver ions (Fig. 92). Thus, the three-dimensional framework is a (3,4)-connected net comprising triconnected and tetraconnected centers in the ratio 1: 1. Guest water molecules occupy the large hexagonal channels and interact with the silver ions of the layers. Thermal analysis has shown that these water molecules can be reversibly removed from the crystals by thermal activation. Another highly symmetrical coordination polymer was formed in the reaction of [CU(CH,CN)~]C~O, and 2,4,6-tri(4-pyridyl)-1,3,5-triazine (Lloz).The compound formulated as has a cu-
FIG.92. A schematic view of the 3-D network in [Ag3(Ld21(C10,)32Hz0. (From Fig. 2 in Carlucci, L.; Ciani, G.; v. Gudenberg, D. W.; Proserpio, D. M.; Sironi, A. J . Chem. SOC.,Chem. Commun. 1997, 631.)
288
MUNAKATA, WU, AND KURODA-SOWA
bic (3,4)-connectednet with large cavities (279).All the ligand groups are equivalent and are attached to three copper centers at the corners of an equilateral triangle, and each copper ion is coordinated by four pyridine donors in a distorted tetrahedral arrangement. This results in an infinite (3,4)-connected three-dimensional network containing large octahedral chambers of diameter 18.241(4) A. Each chamber is connected by its six copper vertices to six others whose centers are arranged octahedrally around, generating an infinite cubic collection of chambers (Fig. 93). Reid has extended the investigation of channeled frameworks to the methylene-bridged bidentate ligands MeECHzEMe (E = S, Se, or Te) on the expectation that such ligands would disfavor chelation and might therefore promote formation of ordered extended networks (280).X-ray diffraction studies show that in [ C U ( M ~ S ~ C H ~ S ~ M ~ ) ~ ] P F each copper ion is tetrahedrally coordinated to one Se donor of four diselenoether ligands and the second Se donor of each ligand is ligated to a different adjacent Cu' ion to give an infinite threedimensional cationic network (Fig. 94). The network contains cylindrical channels of diameter ca. 12.5 A incorporating the PF, counteranions. The silver(1) complex [Ag,(MeSeCHpSeMe)*l(BF,),contains a similar three-dimensional lattice with rectangular-shaped channels of cross section ca. 12.6 X 4.1 A to fill in the anions. A large number of host-guest structures are known, but a reversible guest exchange or chemical transformation of the guest inside a coordination polymer framework has been described for only a few (257, 281-283). In most cases the integrity of the inclusion lattice cannot be effectively maintained in the absence of guest molecules
L102
FIG. 93. Schematic representation of two interpenetrating, identical frameworks in [ C U ~ ( L ~ ~ ~ ) ~ I(From ( C ~ OFig. ~ ) 3~ .in Abrahams, B. F.; Batten, S. R.; Hamit, H.; Hoskins, B. F.; Robson, R. Angew. Chern., Znt. Ed. Engl. 1996,35, 1690.)
COPPERU) AND SILVERU) SUPRAMOLECULARMETAL COMPLEXES
289
FIG. 94. View of the 3-D network in [ C U ( M ~ S ~ C H ~ S ~ M(From ~)~]P Fig. F ~1. in Black,
J. R.: Champness, N. R.: Levason, W.; Reid, G. Inorg. Chem. 1996,35,4432.)
(166).By using the trigonal ligands, 1,3,5-tricyanobenzene (L103) and 1,3,5-tris(4-ethynylbenzonitrile)benzene (LlO4),to make networks joined at the vertices by metal ions, Moore and Lee have obtained two silver(1) complexes [Ag(Llo3)(CF,S03)I and [Ag(Llo4)(CF3S03)I by mixing the constituent molecules in benzene (281, 282). The crystal structure of the compound with L103 consists of honeycomb sheets based on alternating LIo3and the trigonalaAg’ units. The honeycomb sheets create cavities of diameter 10.03 A to accommodate CF3SOi ions which are weakly bound to the silver at the axial position of a trigonal pyramid. By contrast, the complex with tbe larger ligand L104 generates larger channels of diameter 15 X 22 A and consequently involves interpenetration of lattices (Fig. 95). The three-dimensional (3,3)-connected net is based on the end-on coordination of L104 to the trigonal pyramidal silver cations. The interpretation of the networks generates hexagonal channels in which the benzene molecules from the crystallization process are located. Investigation of exchange properties within a host crystal were carried out in solution and in vapor with a variety of guest molecules. The process was monitored by powder X-ray diffraction and thermogravimetric analyse (TGAYdifferential scanning calorimetry (DSC) studies, and the results show that the host adduct can be repeatedly voided of and filled with guest molecules without the destruction of the channel framework. 2. Feldspar Structure Without Void Space The term “mineralomimetic chemistry” was proposed to denote the chemistry of the build-up of mineral-like structures using materials
290
MUNAKATA, WU, AND KURODA-SOWA
FIG.95. View of [Ag(L,,)(CF,S03)1down the a-axis with plane waves corresponding to the 002 reflection highlighted. Triflate anions are omitted. (From Fig. lc in Gardner, G. B.; Kiang, Y.-H.; Lee, S.;Asgaonkar, A.; Venkataraman, D. J. Am. Chem. SOC.1996, 118, 6946.)
that never give stable minerals in nature (270). The topologic similarities observed between the well-defined coordination geometries and in natural minerals prompted many chemists to mimic the structures of simple minerals (121). The copper(1) complex of pyrimidine (LlOs), [Cu(LlO,),IBF5,was found to have a three-dimensional structure related to that of the natural mineral feldspar (284). The complex, obtained by reaction of [Cu(CH3CN)JBF4and pyrimidine, contains two independent copper centers in the unit cell, each coordinated in a slightly distorted tetrahedral geometry by four nitrogen atoms from four different ligand molecules. Each Llosin turn bridges two copper ions, leading to a three-dimensional framework. The structure is related to that of feldspar in such that the corner-shared network consists of eight-rings of tetrahedral connected together with four-rings to form elliptical channels (Fig. 96). Due to interpenetration of neighboring layers, such channels are almost completely constricted, leaving no void space in the structure. This phenomenon has been previously described in [Cu(Lsl)C1],where the interpenetrating of the two-dimensional sheets resulted in small channels (275). Thus, it would be possible to increase the volume of the channels by lengthening the spacer ligands that connect the metal centers.
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
291
b
FIG.96. Polyhedral packing diagram of [Cu(L,,,),IBF, (a) and schematic view of the feldspar structure viewed down the a-axis (b). (From Fig. 2 in Keller, S. W. Angew. Chem., Znt. Ed. Engl. 1997,36,247.)
3. Channel Structure Through Siluer(l)-Promoted in Situ Ligand Synthesis Successful construction of multidimensional frameworks largely relies on ligand design to suit different geometries and coordination numbers of the metal ions. A recent report on isolation of an unexpected three-dimensional metal complex by in situ cyclization of the ligand aroused tremendous interest among ligand designers (285). Reaction of 1,2-trans-(4-pyridyl)ethene(L7J and AgBF, in MeCN and CHzClzin the presence of light yielded [Ag(Llw)IBF4,where Llw = 1,2,3,4-tetrakis(4-pyridyl)cyclobutane.The complex exists as a threedimensional cationic polymer in which each Ag' ion is coordinated to one pyridyl group of four different Lloagroups. Each Llw bridges four tetrahedral metal centers leading t o a three-dimensional array of silver atoms containing channels to fill in the BF; anions and MeCN solvent molecules (Fig. 97). N
AgBF,
2
MeCN/CH,CI,
70
L 106
292
MUNAKATA, WU, AND KURODA-SOWA
FIG.97. View of the polymeric array formed by [Ag(Llos)l+ illustrating the channels formed. (From Fig. 3 in Blake, A. J.; Champness, N. R.: Chung, S.S.M.; Li, W.-S.; SchrGder, M. J. Chem. SOC.Chem. Commun. 1997, 1675.)
X-ray structural determination results revealed that the ligand L,,, had undergone a [2 + 21 cyclization to form a tetrapyridyl-substituted cyclobutane ring. The light-induced dimerization is found to be promoted by the presence of Ag' ions. Although the detailed mechanism of the cyclodimerization reaction is not clear, such in situ formation of the ligand nevertheless represents a new approach to future inorganic crystal engineering. IX. Concluding Remarks
The design and synthesis of new functional polymeric coordination complexes through intermolecular interactions or coordination bonds is a continuing challenge. Our aim in this review has been to show what this development has achieved recently in copper(1)and silver(1) chemistry. Throughout we have employed relatively straightforward ideas covering the latest references and reviews. The concept of supramolecular chemistry is not new, but in order to keep this review within bounds, we have had to severely restrict the number of references. Most of these can be traced via a recent comprehensive review (286).In the present review we have tried to emphasize three aspects toward the construction of functional solid-state supramolecular metal complexes, illustrating them by reference to a variety of topological systems. The first point is that the coordination number and stereochemistry of metal ions play an important role in construction of multidimensional networks. Most polymeric frameworks encoun-
COPPER^ AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
293
tered, such as diamondoid, are based upon linear, trigonal, and tetrahedral metal templates. Therefore, the 4-coordinate tetrahedral copper(1) ion and the 2-4-coordination of silver(1) ion are likely to be the best candidates to produce diverse architectures. Second, designing ligands is an important step in making new functional solids. We must realize that the study we are undertaking today is not just limited to appreciation of novel frameworks for the aesthetics that the system can offer. We have already explored a number of functional ligands such as the conductive molecule TTF and derivatives, the photochromic compound cisl,2-dicyano-1,2-bis(2,4,5-trimethyl-3-thienyl) ethane (LT5),and the linear exodentate spacer 4,4’-bipyridine. We also have noted that in the metallic conductive compound [Cu(L,,),], matching of metal d-orbitals with n-orbitals of ligand may be responsible for the observed extremely high electrical conductivity (197200). Thus, in striving for functional coordination materials with potential applications as inorganic devices, the search for functional ligands constitutes a new challenge. Finally, X-ray diffraction in combination with other physical methods has proven to be the dominant means of structure determination of inorganic polymers. To obtain single crystals in the polymeric phase, the art of synthesis must undergo a constant modification. On the other hand, application of newer experimental techniques for structural analysis of the powder sample may provide further advances in the study of polymeric structures. The structural chemistry of coordination compounds continues to pose challenging problems, but a combination of physical methods and revolutionized synthetic techniques should clarify the picture and provide a real opportunity for those involved in this research. It should be encouraging. Such a prospect provides a major impetus to discover other coordination polymers with unexpected frameworks.
REFERENCES Lehn, J.-M. Angew. Chem., Int. Ed. Engl. 1988,27, 89. Lehn, J.-M. Angew. Chem., Int. Ed. Engl. 1990,29, 1304. Vogtle, F. “Supramolecular Chemistry”; Wiley: Chichester, 1993. Robson, R.; Abrahams, B. F.; Batten, S. R.; Gable, R. W.; Hoskins, B. F.; Liu, J. In “Supramolecular Architecture”; Bein, T., Ed.; American Chemical Society: Washington, DC, 1992; Chapter 19; p. 256. 5. “Comprehensive Supramolecular Chemistry”; Lehn, J.-M., Ed.; Pergamon Press: Oxford, 1995; Vol. 9.
1. 2. 3. 4.
294
MUNAKATA, WU, AND KURODA-SOWA
6. “Transitional Metals in Supramolecular Chemistry”; Fabbrizzi, L., and Poggi, A., Eds.; Kluwer: Holland, 1994. 7. Miiller, A,; Reuter, H.; Dillinger, S. Angew. Chem., Int. Ed. Engl. 1995,34, 2328. 8. Constable, E. C. In “Progress in Inorganic Chemistry”; Karlin, K. D., Ed.; Wiley: New York, 1994; Vol. 42; p. 92. 9. Balzani, V. Tetrahedron 1992,48, 10443. 10. “Supramolecular Chemistry”; Balzani, V., and De Cola, L., Eds.; Kluwer: Holland, 1992. 11. Constable, E. C. Adu. Inorg. Chem. 1989,34, 1. 12. “Supramolecular Photochemistry”; Balzani, V., Ed.; Reidel: Holland, 1987. 13. Lehn, L.-M. In; Williams, A. F., Floriani, C., and Merbach, A. E., Eds.; VHCA Basel, 1992. 14. Baxter, P.; Lehn, J.-M.; DeCian, A,; Fischer, J. Angew. Chem., Int. Ed. Engl. 1993, 32, 69. 15. Hathaway, B. J . In; Wilkinson, G., Gillard, R. D., and McCleverty, J . A,, Eds.; Pergamon: Oxford, 1987; Vol. 5; p. 533. 16. Lancashire, R. J. In “Comprehensive Coordination Chemistry”; Wilkinson, G., Gillard, R. D., and McCleverty, J. A., Eds.; Pergamon: Oxford, 1987; Vol. 5; p. 775. 17. Kitagawa, S.; Munakata, M. Trends Inorg. Chem. 1993, 3. 18. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T. Bull. Chem. SOC.Jpn. 1997, 70, 1727. 19. Lehn, J.-M.; Rigault, A.; Siegel, J.; Harrowfield, J.; Chevirier, B.; Moras, D. Proc. Natl. Acad. Sci. USA 1987, 84, 2565. 20. Lehn, J.-M. “Supramolecular Chemistry”; VCH: Weinheim, 1995. 21. Constable, E. C. Tetrahedron 1992, 48, 10013. 22. Williams, A. Chem. Eur. J . 1997, 3, 15. 23. Gelling, 0. J.; van Bolhuis, F.; Feringa, B. L. J. Chem. SOC.,Chem. Commun. 1991,917. 24. Psillakis, E.; Kotsuki, H.; Isobe, K.; Moriya, N.; Nakagawa, Y.; Ochi, M. J . Chem. SOC.,Dalton Trans. 1997, 1645. 25. Suzuki, T.; Kotsuki, H.; Isobe, K.; Moriya, N.; Nakagawa, Y.; Ochi, M. Inorg. Chem. 1995,34, 530. 26. Wu, B.; Zhang, W.-J.; Yu, S.-Y.; Wu, X.-T. J. Chem. Soc., Dalton Trans. 1997,1795. 27. Hartl, H.; M.-H.-Abadi, F. Angew. Chem., Int. Ed. Engl. 1994,33, 1841. 28. Yu, S.-Y.; Luo, Q.-H.; Wu, B.; Huang, X.-Y.; Sheng, T.-L.; Wu, X.-T.; Wu, D.-X. Polyhedron 1997, 16, 453. 29. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Moriwaki, K.; Kitagawa, S. Inorg. Chem. 1997,36, 5416. 30. Constable, E. C.; Elder, S. M.; Hannon, M. J.; Martin, A.; Raithby, P. R.; Tocher, D. A. J . Chem. SOC.,Dalton Trans. 1996,2423. 31. Constable, E. C.; Hannon, M. J.; Martin, A,; Raithby, P. R.; Tocher, D. A. Polyhedron 1992,11, 2967. 32. Potts, K. T.; Keshavarz-K, M.; Tham, F. S.; Abrufia, H. D.; k a n a , C. R. Inorg. Chem. 1993,32, 4422. 33. Baxter, P. N. W.; Lehn, L.-M.; Rissanen, K. J. Chem. Soc., Chem. Commun. 1997, 1323. 34. Constable, E. C.; Heirtzler, F.; Neuburger, M.; Zehnder, M. J. Am. Chem. SOC. 1997,119, 5606. 35. Piguet, C.; Bernardinelli, G.; Bocquet, B.; Quattropani, A.; Williams, A. F. J. Am. Chem. SOC.1992,114, 7440. 36. Youinou, M.-T.; Ziessel, R.; Lehn, J.-M. Inorg. Chem. 1991,30, 2144.
COPPERU) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
295
37. Constable, E. C.; Holmes, J . M.; Raithby, P. R. Polyhedron 1991, 10, 127. 38. Dietrich-Buchecher, C. D.; Sauvage, J.-P.; Cian, A. D.; Fischer, J. J. Chem. Soc., Chem. Commun. 1994, 2231. 39. Dietrich-Buchecher, C. D.; Sauvage, J.-P. Angew. Chem., Int. Ed. Engl. 1989,28, 189. 40. Ruttimann, S.; Piguet, C.; Bernardinelli, G.; Bocquet, B.; Williams, A. F. J . Am. Chem. SOC.1992,114, 4230. 41. Barley, M.; Constable, E. C.; Corr, S. A,; McQueen, R. C. S.; Nutkins, J . C.; Ward, M. D.; Drew, M. G. B. J. Chem. Soc., Dalton Trans. 1988, 2655. 42. Constable, E. C.; Edwards. A. J . ; Raithby, P. R.; Walker, J. V. Angew. Chem., Znt. Ed. Engl. 1993, 32, 1465. 43. Fu, Y.; Sun, J.; Li, Q.; Chen, Y.; Dai, W.; Wang, D.; Mak, T. C. W.; Tang, W.; Hu, H. J . Chem. Soc., Dalton Trans. 1996, 2309. 44. Ho, P. K.-K.; Peng, S. M.; Wong, K.-Y.; Che, C.-M. J . Chem. Soc., Dalton Trans. 1996, 1829. 45. Constable, E. C.; Edwards, A. J.; Hannon, M. J.; Raithby, P. R. J. Chem. Soc., Dalton Trans. 1994, 1991. 46. Potts, K. T.; Keshavarz-K, M.; Tham, F. S.; Abruria, H. D.; Arana, C. R. Inorg. Chem. 1993,32, 4450. 47. Constable, E. C.; Ward, M. D.; Tocher, D. A. J . Chem. Soc., Dalton Trans. 1991, 1675. 48. Garrett, T. M.; Koert, U.; Lehn, J.-M.; Rigault, A,; Meyer, D.; Fischer, J. J . Chem. Soc., Chem. Commun. 1990,557. 49. Lehn, J.-M.; Rigault, A. Angew. Chem., Int. Ed. Engl. 1988,27, 1095. 50. Carlucci, L.; Ciani, G.; W. v. Gudenberg, D.; Proserpio, D. M. Inorg. Chem. 1997, 36, 3812. 51. Carina, R. F.; Bernardinelli, G.;Williams, A. F.Angew. Chem., Int. Ed. Engl. 1993, 32, 1463. 52. Zelikovich, L.; Libman, J.; Sbanzer, A. Nature 1995,374, 790. 53. Townsend, J . M.; Blount, J . F.; Sun, R. C.; Zawoiski, S.; Valentine, J. D. J . Org. Chem. 1980,45, 2995. 54. Piguet, C.; Bernardinelli, G.; Biinzli, J.-C. G.; Petoud, S.; Hopfgartner, G. J. Chem. SOC., Chem. Commun. 1995,2575. 55. Soghomonian, V.; Chen, Q.; Haushalter, R. C.; Zubieta, J.; OConnor, C. J. Science 1993,259, 1596. 56. Maruoka, K.; Murase, N.; Yamamoto, H. J . Org. Chem. 1993, 58, 2938. 57. Woods, C. R.; Benaglia, M.; Cozzi, F.; Siegel, J . S. Angew. Chem., Int. Ed. Engl. 1996,35, 1830. 58. Ohata, N.; Masuda, H.; Yamauchi, 0. Angew. Chem., Int. Ed. Engl. 1996,35, 531. 59. Batten, S. R.; Hoskins, B. F.; Robson, R. Angew. Chem., Int. Ed. Engl. 1997,36,636. 60. Ramos, E.; Bosch, J.; Serrano, J . L.; Sierra, T.; Veciana, J . J. Am. Chem. SOC. 1996,118, 4703. 61. Provent, C.; Hewage, S.; Brand, G.; Bernardinelli, G.; Charbonniere, L. J.; Williams, A. F. Angew. Chem., Int. Ed. Engl. 1997, 36, 1287. 62. Wudl, F.; Smith, G. M.; Hufnagel, E. J . J . Chem. SOC.,Chem. Commun. 1970,1453. 63. Ferraris, J. P.; Cowan, D. 0.;Walatka, V.; Perlstein, J . H. J . Am. Chem. SOC.1973, 95, 948. 64. Wudl, F. Acc. Chem. Res. 1984, 17, 227. 65. Williams, J . M.; Beno, M. A,; Wang, H. H.; Leung, P. C. W.; Emge, T. J.; Geiser, U.; Carlson, K. D. Acc. Chem. Res. 1985, 18, 261.
296
MUNAKATA, WU, AND KURODA-SOWA
66. Farraro, J. R.; Williams, J . M. “Introduction to Synthetic Electrical Conductors”; Academic Press: New York, 1987. 67. J~rgensen,T.; Hansen, T. K.; Becher, J. Chem. SOC.Rev. 1994, 41. 68. Cassoux, P.; Valade, L.; Kobayashi, H.; Kobayashi, A.; Clark, R. A.; Underhill, A. E. Coord. Chem. Rev. 1991,110, 115. 69. Olk, R.-M.; Olk, B.; Dietzsch, W.; Kirmse, R.; Hoyer, E. Coord. Chem. Reu. 1992, 117, 99. 70. Cassoux, P.; Valade, L. In “Inorganic Materials”; Bruce, D. W., Ed.; Wiley: ChiChester, 1992; p. 2. 71. Matsubayashi, G.; Douki, K.; Tamura, H.; Nakano, M. Inorg. Chem. 1993, 32, 5990. 72. Matsubayashi, G.; Maikawa, T.; Nakano, M. J. Chem. SOC.,Dalton Trans. 1993, 2995. 73. Matsubayashi, G.; Takahashi, K.; Tanaka, T. J . Chem. SOC.,Dalton Trans. 1988, 967. 74. Tanaka, S.; Matsubayashi, G. J. Chem. Sac., Dalton Trans. 1992,2837. 75. Nakamura, T.; Underhill, A. E.; Coomber, A. T.; Friend, R. H.; Tajima, H.; Kobayashi, A.; Kobayashi, H. Inorg. Chem. 1995,34, 870. 76. Kobayashi, A,; Kim, H.; Sasaki, Y.; Kato, R.; Kobayashi, H.; Moriyama, S.; Nishio, Y. Chem. Lett. 1987, 1819. 77. Kobayashi, H.; Bun, K.; Naito, T.; Kato, R.; Kobayashi, A. Chem. Lett. 1992, 1909. 78. Tajima, H.; Inokuchi, M.; Kobayashi, A,; Ohta, T.; Kato, R.; Kobayashi, H.; Kuroda, H. Chem. Lett. 1993, 1235. 79. Kobayashi, A,; Kobayashi, H.; Miyamoto, A.; Kato, R.; Clark, R. A.; Underhill, A. E. Chem. Lett. 1991, 2163. 80. Cornelissen, J. P.; L o n , R. L.; Jansen, J.; Haasnoot, J . G.; Reedijk, J.; Horn, E.; Spek, A. L.; Pomarede, B.; Legros, J.-P.; Reefman, D. J. Chem. SOC.,Dalton Trans. 1992,2911. 81. Pullen, A. E.; Olk, R.-M.; Zeltner, S.; Hoyer, E.; Abboud, K. A.; Reynolds, J. R. Inorg. Chem. 1997,36, 958. 82. Faulmann, C.; Delpech, F.; Malfant, I.; Cassoux, P. J . Chem. SOC.,Dalton Trans. 1996,2261. 83. Piotraschke, J.; Pullen, A. E.; Abboud, K. A,; Reynolds, R. Inorg. Chem. 1995, 34, 4011. 84. Bousseau, M.; Valade, L.; Legros, J.-P.; Cassoux, P.; Garbauskas, M.; Interrante, L. V. J. Am. Chem. SOC.1986,108, 1908. 85. Kim, H.; Kobayashi, A.; Sasaki, Y.; Kato, R.; Kobayashi, H. Chem. Lett. 1987,1799. 86. Faulmann, C.; Errami, A.; Donnadieu, B.; Malfant, I.; Legros, J.-P.; Cassoux, P.; Rovira, C.; Canadell, E. Inorg. Chem. 1996,35, 3856. 87. Sun, S. Q.; Zhang, B.; Wu, P.; Zhu, D. B. J. Chem. SOC.,Dalton Trans. 1997, 277. 88. Yang, X.; Doxsee, D. D.; Rauchfuss, T. B.; Wilson, S. R. J. Chem. SOC.,Chem. Commun. 1994,821. 89. Pullen, A. E.; Piotraschke, J.; Abbound, K. A.; Reynolds, J. R. Inorg. Chem. 1996, 35, 793. 90. Narvor, N.; Robertson, N.; Weyland, T.; Kilburn, J . D.; Underhill, A. E.; Webster, M.; Svenstrup, N.; Becher, J. J. Chem. SOC.,Chem. Commun. 1996, 1363. 91. Narvor, N.; Robertson, N.; Wallace, E.; Kilburn, J. D.; Underhill, A. E.; Bartlett, P.; Webster, M. J. J. Chem. SOC.,Dalton Trans. 1996, 1363. 92. Schultz, A. J.; Wang, H. H.; Soderholm, L. C.; Sifter, T. L.; Williams, J. M.; Bechgaard, K.; Whangbo, M.-H. Inorg. Chem. 1987,2626, 3757.
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
297
93. Pullen, A. E.; Zeltner, S.; Olk, R.-M.; Hoyer, E.; Abboud, K. A,; Reynolds, J. R. Znorg. Chem. 1996,35, 4420. 94. Pullen, A. E.; Zeltner, S.; Olk, R.-M.; Hoyer, E.; Abboud, K. A.; Reynolds, J. R. Inorg. Chem. 1997,36, 4163. 95. Matsubayashi, G.; Yokozawa, A. J. Chem. SOC.,Chem. Commun. 1991,68. 96. Bellitto, C.; Bonamico, M.; Fares, V.; Serino, P. Znorg. Chem. 1996,35, 4070. 97. Geiser, U.; Beno, M. A.; Kini, A. M.; Wang, H. H.; Schultz, A. J.; Gates, B. D.; Cariss, C . S.; Carlson, K. D.; Williams, J. M. Synth. Met. 1988, 27, A235. 98. Mori, H.; Mori, T.; Kato, K.; Maruyama, Y.; Inokuchi, H.; Kirabayashi, I.; Tanaka, S. Solid State Commun. 1992, 82, 177. 99. Horiuchi, S.; Yamochi, H.; Saito, G.; Sakaguchi, K.; Kusunoki, M. J. Am. Chem. SOC.1996, 118, 8604. 100. Inoue, M. B.; Inoue, M.; Bruck, M. A.; Fernando, Q. J. Chern. SOC.,Chem. Commun. 1992,515. 101. Mori, T.; Wu, P.; Imaeda, K.; Enoki, T.; Inokuchi, H.; Saito, G. Synth. Met. 1987, 19, 545. 102. Katayama, C.; Honda, M.; Kumagai, H.; Tanaka, J.; Saito, G.; Inokuchi, H. Bull. Chem. SOC.Jpn. 1985,58, 2272. 103. Imaeda, K.; Enoki, T.; Shi, Z.; Wu, P.; Okada, N.;Yamochi, H.; Saito, G.; Inokuchi, H. Bull. Chem. SOC.Jpn. 1987, 60, 3163. 104. Honda, K.; Goto, M.; Kurahashi, M.; Anzai, H.; Tokumoto, M.; Ishiguro, T. Bull. Chem. SOC.Jpn. 1988,61, 588. 105. Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Hirota, A,; Kitagawa, S. Znorg. Chem. 1995,34, 2705. 106. Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Misaki, Y. Polyhedron 1995, 14, 1343. 107. Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M. Bull. Chem. SOC.Jpn. 1994,67, 3009. 108. Kuroda-Sowa, T.; Hirota, A,; Munakata, M.; Maekawa, M. Mol. Cryst. Liq. Cr.yst. 1996, 285, 69. 109. Wu, L. P.; Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y. Mol. Cryst. Liq. Cryst. 1996, 285, 75. 110. Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Yamamoto, M. Polyhedron 1995,14, 1647. 111. Yamamoto, M.; Gan, X.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata, M. Inorg. Chim. Acta 1997,261, 169. 112. Munakata, M.; Wu, L. P.; Gan, X.; Kuroda-Sowa, T.; Suenaga, Y . Mol. Cryst. Liq. Cryst. 1996,284, 319. 113. Svenstrup, N.; Becher, J. Synthesis 1995, 215. 114. Dai, J. Ph.D. Thesis, Kinki University, 1997. 115. Dai, J.; Kuroda-Sowa, T.; Munakata, M.; Maekawa, M.; Suenaga, Y.; Ohno, Y. J . Chem. SOC.,Dalton Trans. 1997, 2363. 116. Dai, J.; Munakata, M.; Kuroda-Sowa, T.; Suenaga, Y.; Wu, L. P.; Yamamoto, M. Inorg. Chim. Acta 1997,255, 163. 117. Dai, J.;Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Suenaga, Y. Inorg. Chim. Acta 1997, 258, 65. 118. Iwamoto, T. In “Inclusion Compounds”; Atwood, J. L., Davies, J. E. D., MacNicol, D. D., and Eds.; Oxford University Press: Oxford, 1991; p. 177. 119. Iwamoto, T. In “Chemistry of Microporous Crystals”; Inui, T., Namba, S., Tatsumi, T., and Eds.; Kodansha/Elsevier: Tokyo, 1991; p. 1.
298 120. 121. 122, 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153.
MUNAKATA, WU, AND KURODA-SOWA
Haushalter, R. C.; Mundi, L. A. Chem. Muter. 1992,4, 31. Gable, R. W.; Hoskins, B. F.; Robson, R. J . Chem. SOC.,Chem. Commun. 1990,762. Kim, J.; Whang, D.; Koh, Y.-S.; Kim, K. J. Chem. SOC.,Chem. Commun. 1994,637. Soma, T.; Yuge, H.; Iwamoto, T. Angew. Chem., Int. Ed. Engl. 1994,33, 1665. Abrahams, B. F.; Hoskins, B. F.; Michail, D. M.; Robson, R. Nature 1994,369, 727. Hoskins, B. F.; Robson, R. J. Am. Chem. SOC.1989,111, 5962. Abrahams, B. F.; Hoskins, B. F.; Robson, R. J. Chem. SOC., Chem. Commun. 1990, 60. Hoskins, B. F.; Robson, R. J. Am. Chem. SOC.1990, 112, 1546. Abrahams, B. F.; Hoskins, B. F.; Liu, J.; Robson, R. J . Am. Chem. SOC.1191, 113, 3045. Jung, 0.-S.;Pierpont, C. G. J . Am. Chem. SOC.1994, 116, 2229. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J. Am. Chem. SOC.1995, 117, 4562. Darriet, J.; Haddad, M. S.; Duesler, E. N.; Hendrickson, D. N. Inorg. Chem. 1979, 18, 2679. Lumme, P.; Lindroos, S.; Lidell, E. Acta Crystallogr., Sect. C 1987,43, 2053. Otieno, T.; Rettig, S. J.; Thompson, R. C.; Trotter J. Inorg. Chem. 1993,32, 1607. Turnbull, M. M.; Pon, G.; Willett, R. D. Polyhedron 1991, 10, 1835. Kitagawa, S.; Kawata, S.; Kondo, M.; Nozaka, Y.; Munakata, M. Bull. Chem. SOC. Jpn. 1993,66, 3387. MacGillivray, L. R.; Subramanian, S.; Zaworotko, M. J. J. Chem. SOC.,Chem. Commun. 1994, 1325. Kitagawa, S.; Munakata, M.; Tanimura, T. Inorg. Chem. 1992,31, 1714. Munakata, M.; Kitagawa, S.; Ujimaru, N.; Nakamura, M.; Maekawa, M.; Matsuda, H. Inorg. Chem. 1993,32, 826. Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Nakamura, M.; Akiyama, S.-I.; Kitagawa, S. Inorg. Chem. 1994,33, 1284. Karl, N.; Ketterer, W.; Stezowski, J. J. Acta Crystallogr. 1982, B38, 2917. Goldberg, I.; Shmueli, U. Acta Crystallogr. 1973, B29, 440. Aakeroy, C. B.; Seddon, K. R. Chem. SOC.Rev. 1993,397. Etter, M. C. Acc. Chem. Res. 1990,23, 120. Burrows, A. D.; Chan, C.-W.; Chowdhry, M. M.; McGrady, J. E.; Mingos, D. M. P. Chem. SOC.Rev. 1995,329. MacDonald, J . C.; Whitesides, G. M. Chem. Rev. 1994, 94, 2383. Subramanian, S.; Zaworotko, M. J . Coord. Chem. Rev. 1994, 137, 357. Bernstein, J.; Davis, R. E.; Shimoni, L.; Chang, N.-L. Angew. Chem., Int. Ed. Engl. 1996,34, 1555. Batsanov, A. S.; Hubberstey, P.; Russell, C. E. J . Chem. SOC.,Dalton Trans. 1994, 3189. Begley, M. J.; Hubberstey, P.; Stroud, J. J. Chem. SOC.,Dalton Trans. 1996, 2323. Burrows, A. D.; Mingos, D. M. P.; White, A. J. P.; Williams, D. J. J . Chem. SOC., Dalton Trans. 1996, 3805. Tadokoro, M.; Toyoda, J.; Isobe, K.; Itoh, T.; Miyazaki, A.; Enokie, T.; Nakasuji, K. Chem. Lett. 1996, 613. Yamanari, K.; Kida, M.; Yamamoto, M.; Fujihara, T.; Fuyuhiro, A,; Kaizaki, S. J . Chem. SOC.,Dalton Trans. 1995, 2627. Burrows, A. D.; Mingos, D. M. P.; White, A. J . P.; Williams, D. J . J. Chem. SOC., Dalton Trans. 1996, 149.
COPPERU) AND SILVERO) SUPRAMOLECULARMETAL COMPLEXES
299
154. Itoh, T.; Toyoda, J.; Tadokoro, M.; Kitagawa, H.; Mitani, T.; Nakasuji, K. Chem. Lett. 1995, 41. 155. Smith, G.; Reddy, A. N.; Byriel, K. A.; Kennard, C. H. L. J. Chem. Soc., Dalton Trans. 1995, 3565. 156. Chowdhry, M. M.; Mingos, D. M. P.; White, A. J. P.; Williams, D. J. J. Chem. Soc., Chem. Commun. 1996,899. 157. Batsanov, A. S.; Begley, M. J.; Hubberstey, P.; Stroud, J. J. Chem. Soc., Dalton Trans. 1996, 1947. 158. Blake, A. J.; Hill, S. J.; Hubberstey, P.; Li, W . 3 . J . Chem. Soc., Dalton Trans. 1997, 913. 159. Blake, A. J.; Fallis, I. A,; Heppeler, A,; Parsons, S.; Ross, S. A,; Schriider, M. J . Chem. Soc., Dalton Trans. 1996, 31. 160. Nakasuji, K.; Tadokoro, M.; Toyoda, J.;Mitsumi, M.; Itoh, T.; Iijima, K. Mol. Cryst. Liq. Cryst. 1996, 285, 241. 161. Mitumi, M.; Toyoda, J.; Nakasuji, K. Znorg. Chem. 1995, 34, 3367. 162. Wu, L. P.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y. Inorg. Chim. Acta 1996,249, 183. 163. Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, T.; Fukui, J.; Munakata, M. In.org. Chim. Acta 1996, 239, 165. 164. Munakata, M.; Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M. J . Am. Chem. SOC.1996, 118, 3117. 165. Munakata, M.; Yamamoto, M. unpublished results, 1997. 166. Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata, M. J . Chem. Soc., Dalton Trans. 1996, 2031. 167. Dai, J.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata, M. Inorg. Chem. 1997,36, 2688. 168. Thompson, J. S.; Whitney, J. F. Inorg. Ch,em. 1984,23, 2813. 169. Masuda, H.; Yamamoto, N.; Taga, T.; Machida, K.; Kitagawa, S.; Munakata, M. J . Organomet. Chem. 1987,322, 121. 170. Masuda, H.; Machida, K.; Munakata, M.; Kitagawa, S.; Shimono, H. J. Chem. Soc., Dalton Trans. 1988, 1907. 171. Munakata, M.; Kitagawa, S.; Kawada, I.; Maekawa, M.; Shimono, H. J . Chem. Soc., Dalton Trans. 1992, 2225. 172. Not in the original paper, but we reexamined from the crystallographic data of ref. 173. 173. Hubberstey, P., private communication. 174. Munakata, M.; Maekawa, M.; Kitagawa, S.; Adachi, M.; Masuda, H. Inorg. Chim. Acta 1990, 167, 181. 175. Lee, S. W.; Trogler, W. C. Inorg. Chem. 1990,29, 1659. 176. Healy, P. C.; Pakawatchai, C.; White, A. H. J . Chem. Soc., Dalton Trans. 1983, 1917. 177. Engelhardt, L. M.; Pakawatchai, C.; White, A. H.; Healy, P. C. J . Chem. Soc., Dalton Trans. 1985, 117. 178. Munakata, M.; Kitagawa, S.; Shimono, H.; Masuda, H. Inorg. Chim. Acta 1989, 158, 217. 179. Rumpel, H.; Limbach, H. H. J. Am. Chem. Soc. 1989, 111, 5429. 180. Nakasuji, K.; Sugiura, K.; &tagawa, T.; Toyoda, J.; Okamoto, H.; Okaniwa, K.; Mitani, T.; Yamamoto, H.; Murata, I.; Kawamoto, A,; Tanaka, J. J. Am. Ch,ena. SOC.1991, 113, 1862.
300
MUNAKATA, WU, AND KURODA-SOWA
181. Munakata, M.; Dai, J.; Maekawa, M.; Kuroda-Sowa, T.; Fukui, J. J . Chem. Sac., Chem. Commun. 1994,2331. 182. Albert, A.; Goldacre, R.; Phillips, J. J. Chem. Sac. 1948, 2240. 183. Kuroda-Sowa, T.; Munakata, M.; Matsuda, H.; Akiyama, S.; Maekawa, M. J . Chem. SOC.,Dalton Trans. 1995, 2201. 184. Morpurgo, G. 0.;Fares, V.; Dessy, G. J. Chem. Sac., Dalton Trans. 1984, 785. 185. Dessy, G.; Fares, V.; Imperatori, P.; Morpurgo, G. 0.J. Chem. Sac., Dalton Trans. 1985, 1285. 186. Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Honda, A.; Kitagawa, S. J. Chem. Sac., Dalton Trans. 1994, 2771. 187. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Sugimoto, K. Inorg. Chem. 1997,36, 4903. 188. Wells, A. F. ‘‘Structural Inorganic Chemistry,” 5th ed.; Oxford University Press: Oxford, 1984. 189. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Angew. Chem., Znt. Ed. Engl. 1995,34, 1895. 190. Carlucci, L.; Ciani, G.;Proserpio, D. M.; Sironi, A. J. Chem. Sac., Chem. Commun. 1994,2755. 191. Yaghi, 0. M.; Li, H. J. Am. Chem. Sac. 1995,117, 10404. 192. Blake, A. J.; Champness, N. R.; Chung, S. S. M.; Li, W.-S.; Schroder, M. J. Chem. Sac., Chem. Commun. 1997, 1005. 193. Lopez, S.; Kahraman, M.; Harmata, M.; Keller, S. W. Znorg. Chem. 1997,36, 6138. 194. Kinoshita, Y.; Matsubara, I.; Saito, Y. Bull. Chem. Sac. Jpn. 1959,32, 1221. 195. Kuroda-Sowa, T.; Yamamoto, M.; Munakata, M.; Seto, M.; Maekawa, M. Chem. Lett. 1996, 349. 196. Kuroda-Sowa, T.; Horino, T.; Yamamoto, M.; Ohno, Y.; Maekawa, M.; Munakata, M. Znorg. Chem. 1997,36, 6382. 197. Aumuller, A,; Erk, P.; Klebe, G.; Hunig, S.; Schutz, J. U.; Werner, H.-P. Angew. Chem., Znt. Ed. Engl. 1986,25, 740. 198. Ermer, 0.Adu. Muter. 1991,3, 608. 199. Tomic, S.; Jerome, D.; Aumiiller, A.; Erk, P.; Hunig, S.; V. Schiitz, J . U. Synth. Met. 1988,27, B281. 200. Uji, S.; Terashima, T.; Aoki, H.; Brooks, J. S.; Kato, R.; Sawa, H.; Aonuma, S.; Tamura, M.; Kinoshita, M. Phys. Reu. B 1994,50, 15597. 201. Hirsch, K. A.; Venkataraman, D.; Wilson, S. R.; Moore, J . S.; Lee, S. J. Chem. SOC.,Chem. Commun. 1995, 2199. 202. Hirsch, K. A.; Wilson, S. R.; Moore, J . S. Chem. Eur. J. 1997,3, 765. 203. Hirsch, K. A.; Wilson, S. R.; Moore, J. S. Znorg. Chem. 1997, 36, 2960. 204. Zhdanow, H. C. R. Acad. Sci. URSS 1941,31, 350. 205. Shugam, E.; Zhdanow, H. Acta Physiochim. URSS 1945,20, 247. 206. Cromer, D. T.; Larson, A. C. Acta Crystallogr., Sect. B 1972, B28, 1052. 207. Blake, A. J.; Champness, N. R.; Khlobystov, A. N.; Lemenovskii, D. A,; Li, W.-S.; Schroder, M. J. Chem. Sac., Chem. Commun. 1997, 1939. 208. Michaelides, A.; Kiritsis, V.; Skoulika, S.; Aubry, A. Angew. Chem., Znt. Ed. Engl. 1993,32, 1495. 209. Irie, M. In “Photoreactive Materials for Ultrahigh Density Optical Memory”; Irie, M., Ed.; Elsevier: Amsterdam, 1994; p. 1. 210. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Furuichi, K. J. Am. Chem. Sac. 1996,118, 3305.
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
301
211. Wu, L. P.; Suenaga, Y.; Kuroda-Sowa, T.; Maekawa, M.; Furuichi, K.; Munakata, M. Inorg. Chim. Acta 1996,248, 147. 212. Yamazalu, S.; Deeming, A.; Speel, D. M.; Hibbs, D. E.; Hursthouse, M. B.; Malik, K. M. A. J. Chem. SOC.,Chem. Commun. 1997, 177. 213. Guerret, 0.;Sole, S.; Gornitzka, H.; Teichert, M.; Trinquier, G.; Bertrand, G. J . Am. Chem. SOC.1997,119, 6668. 214. Fortin, D.; Drouin, M.; Turcotte, M.; Harvey, P. D. J . Am. Chem. SOC.1997, 119, 531. 215. Dartiguenave, M.; Dartiguenave, Y.; Mari, A.; Guitard, A,; Olivier, M. J . ; Beauchamp, A. L. Can. J. Chem. 1966, 66, 2386. 216. Dhingra, S. S.; Sea, D.-K.; Kowach, G. R.; Kremer, R. K.; Shreeve-Keyer, J . L.; Haushalter, R. C.; Whangbo, M.-H. Angew. Chem., Int. Ed. Engl. 1997,36, 1087. 217. Huang, Q.; Wu, X.; Wang, Q.; Sheng, T.; Lu, J . Angew. Chem., Int. Ed. Engl. 1996, 35, 868. 218. Manoli, J. M.; Potvin, C.; Secheresse, F.; Marzak, S. Inorg. Chem. Acta 1988, 150, 257. 219. Hagrman, D.; Zubieta, C.; Rose, D. J.; Zubieta, J.; Haushalter, R. C. Angew. Chem., Int. Ed. Engl. 1997, 36, 873. 220. Lu, J.; Crisci, G.; Niu, T.; Jacobson, A. Inorg. Chem. 1997,36, 5140. 221. Ainscough, E. W.; Brodie, A. M.; Husbands, J . M.; Gainsford, G. J.; Gabe, E. J.; Curtis, N. F. J . Chem. SOC.,Dalton Trans. 1965, 151. 222. Brook, D. J . R.; Lynch, V.; Conklin, B.; Fox, M. A. J . Am. Chem. SOC.1997, 119, 5155. 223. Oshio, H.; Watanabe, T.; Ohto, A,; Ito, T. Inorg. Chem. 1997, 36, 1608. 224. Chen, X.-M.; Mak, T. C. W. J. Chem. SOC.,Dalton Trans. 1991, 1219. 225. Blake, A. J.; Li, W.-S.; Lippolis, V.; Schroder, M. J. Chem. SOC.,Dalton Trans. 1997, 1943. 226. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Nakagawa, S. J. Chem. SOC.,Dalton Trans. 1996, 1525. 227. Yamamoto, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata, M. Inorg. Chim. Acta 1997,258, 87. 228. Buchholz, H. A,; Prakash, G. K. S.; Vaughan, J . F. S.; Bau, R.; Olah, G. A. Inorg. Chem. 1996,35, 4076. 229. Blake, J. R.; Champness, N. N.;Levason, W.; Reid, G. J . Chem. SOC.,Chem. Comn u n . 1995, 1277. 230. Dubler, E.; Bensch, W. Inorg. Chim. Acta 1986, 125, 37. 231. Plappert, E. C.; Mingos, D. M.; Lawrence, S. E.; Williams, D. J . J . Chem. Soc., Dalton Trans. 1997, 2119. 232. Henary, M.; Wootton, J . L.; Khan, S. I.; Zink, J. I. Inorg. Chem. 1997, 36, 796. 233. Begley, M. J.; Eisenstein, 0.;Hubberstey, P.; Jackson, S.; Russell, C. E.; Walton, P. H. J . Chem. Soc., Dalton Trans. 1994, 1935. 234. Kitagawa, S.; Munakata, M.; Tanimura, T. Chem. Lett. 1991, 623. 235. Kitagawa, S.; Matsuyama, S.; Munakata, M.; Osawa, N.; Masuda, H. J . Chem. SOC.,Dalton Trans. 1991, 1717. 236. Constable, E. C.; Steel, P. J. Coord. Chem. Reu. 1989,93, 205. 237. Buchner, E. Chem. Ber. 1669,22, 842. 238. Munakata, M.; Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Kawata, S.; Kitagawa, S. J. Chem. Sac., Dalton Trans. 1995, 4099. 239. Bertelli, M.; Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J . Muter. Chem. 1997, 7, 1271.
302
MUNAKATA, WU, AND KURODA-SOWA
240. Naskar, J . P.; Hati, S.; Datta, D.; Tocher, D. A. J. Chem. SOC.,Chem. Commun. 1997, 1319. 241. Kitagawa, S.; Matsuyama, S.; Munakata, M.; Emori, T. J . Chem. SOC.,Dalton Trans. 1991, 2869. 242. Kulawiec, R. J.; Crabtree, R. H. Coord. Chem. Reu. 1990,99, 89. 243. Powell, J.; Horvath, M. J.; tough, A. J . Chem. SOC.,Drzlton Trans. 1996, 1669. 244. Schill, G. “Catenanes, Rotaxanes and Knots”; Academic: New York, 1971. 245. Jager, R.; Vogtle, F. Angew. Chem., Int. Ed. Engl. 1997,36, 930. 246. Dietrich-Buchecker, C. 0.;Sauvage, J.-P. Chem. Reu. 1987,87, 795. 247. Cardenas, D. J.; Gavifia, P.; Sauvage, J.-P. J. Am. Chem. SOC.1997,119, 2656. 248. Solladie, N.; Chambron, J.-C.; Dietrich-Buchecker, C. 0.; Sauvage, J.-P. Angew. Chem., Int. Ed. Engl. 1996, 35, 906. 249. Baxter, P. N. W.; Sleiman, H.; Lehn, J.-M.; Rissanen, K. Angew. Chem., Int. Ed. Engl. 1997,36, 1294. 250. Piguet, C.; Bernardinelli, G.; Williams, A. F.; Bocquet, B. Angew. Chem., Int. Ed. Engl. 1995, 34, 582. 251. Sleiman, H.; Baxter, P.; Lehn, J.-M.; Rissanen, K. J. Chem. SOC.,Chem. Commun. 1995, 715. 252. Blake, A. J.; Champness, N. R.; Khlobystov, A,; Lemenovskii, D. A.; Li, W.-S.; Schroder, M. J. Chem. Soc., Chem. Commun. 1997,2027. 253. Hoskins, B. F.; Robson, R.; Slizys, D. A. J , Am. Chem. SOC.1897, 219, 2952. 254. Whang, D.; Kim, K. J. Am. Chem. SOC.1997,119, 451. 255. Parish, R. V.; Salehi, Z.; Pritchard, R. G. Angew. Chem., Int. Ed. Engl. 1997, 36, 251. 256. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Angew. Chem., Int. Ed. Engl. 1996,35, 1088. 257. Venkataraman, D.; Gardner, G. B.; Lee, S.; Moore, J. S. J. Am. Chem. SOC.1995, 117, 11600. 258. Wu, D.-D.; Mak, T. C. W. J. Chem. SOC.,Dalton Trans. 1995, 2671. 259. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M. J . Chem. SOC.,Dalton Trans. 1995, 3215. 260. Manners, I. Angew. Chem., Int. Ed. Engl. 1996,35, 1602. 261. Schon, J . C.; Jansen, M. Angew. Chem., Int. Ed. Engl. 1996,35, 1286. 262. Pfitzner, A.; Zimmerer, S. Angew. Chem., Int. Ed. Engl. 1997,36, 982. 263. Zhang, X.; Kanatzidis, M. G. J. Am. Chem. SOC.1994, 116, 1890. 264. Jerome, J. E.; Wood, P. T.; Pennington, W. T.; Kolis, J. W . Inorg. Chem. 1994, 33, 1733. 265. Baxter, P. N. W.; Lehn, J.-M.; Fischer, J.; Youinou, M.-T. Angew. Chem., Int. Ed. Engl. 1994, 33, 2284. 266. Desiraju, G. R. Angew. Chem., Int. Ed. Engl. 1995,34, 2311. 267. Janiak, C. Angew. Chem., Int. Ed. Engl. 1997, 36, 1431. 268. Yaghi, 0. M.; Li, H.; Groy, T. L. J . Am. Chem. SOC.1996,118, 9096. 269. Yaghi, 0. M.; Davis, C. E.; Li, G.; Li, H. J. Am. Chem. SOC.1997,119, 2861. 270. Iwamoto, T.; Nishikiori, S.; Kitazawa, T.; Yuge, H. J. Chem. SOC.,Dalton Trans. 1997,4127. 271. Hoskins, B. F.; Robson, R.; Scarlett, N. V. Y. J , Chem. SOC.,Chem. Commun. 1994,2025. 272. Fiijita, M.; Kwon, Y. J.; Washizu, S.; Ogura, K . J . Am. Chem. SOC. 1994,116, 1151. 273. Robinson, F.; Zaworotko, M. J . Chem. SOC.,Chem. Commun. 1995, 2413. 274. Yaghi, 0. M.; Li, H. J . Am. Chem. SOC.1996,118, 295.
COPPER(I) AND SILYERU) SUPRAMOLECULAR METAL COMPLEXES
303
275. Yaghi, 0. M.; Li, G. Angew. Chem., Int. Ed. Engl. 1995,34, 207. 276. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J . Am. Chem. Soc. 1995, 117, 1286 1. 277. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Znorg. Chem. 1997,36, 1736. 278. Carlucci, L.; Ciani, G.; v. Gudenberg, D. W.; Proserpio, D. M.; Sironi, A. J. Chem. Soc., Chem. Commun. 1997,631. 279. Abrahams, B. F.; Batten, S. R.; Hamit, H.; Hoskins, B. F.; Robson, R. Angew. Chem., Int. Ed. Engl. 1996, 35, 1690. 280. Black, J. R.; Champness, N. R.; Levason, W.; Reid, G. Znorg. Chem. 1996,35, 4432. 281. Gardner, G. B.; Venkataraman, D.; Moore, J. S.; Lee, S. Nature 1995, 374, 792. 282. Gardner, G. B.; Kiang, Y.-H; Lee, S.; Asgaonkar, A,; Venkataraman, D. J. Am. Chem. Soc. 1996,118, 6946. 283. Yaghi, 0. M.; Li, G.; Li, H. Nature 1995, 378, 703. 284. Keller, S. W. Angew. Chem., Znt. Ed. Engl. 1997, 36, 247. 285. Blake, A. J.; Champness, N. R.; Chung, S. S. M.; Li, W . 3 ; Schroder, M. J . Chem. Soc., Chem. Commun. 1997, 1675. 286. Philp, D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 1996,35, 1154.
This Page Intention ally Left Blank
ADVANCES IN INORGANIC CHEMISTRY, VOL. 46
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS: PROPERTIES, STRUCTURES, AND REACTIVITY NEIL A. LAW, M. TYLER CAUDLE, and VINCENT L. PECORARO Department of Chemistry, The University of Michigan, Ann Arbor, Michigan 48109
I. Introduction 11. The Enzymes A. Manganese Superoxide Dismutase, MnSOD B. Manganese(I1) Dioxygenase C. Manganese Peroxidase, MnP D. Manganese Ribonucleotide Reductase, MnRR E. Manganese Catalase F. Binuclear Manganese Enzymes: A Comparison of Three Enzymes G . The Oxygen-Evolving Complex, OEC 111. Structural Models A. Mononuclear B. Binuclear C. Trinuclear D. Tetranuclear IV. Physical Properties A. Electronic Spectroscopy B. Magnetism C . EPR Spectroscopy D. X-Ray Absorption Spectroscopy V. Reactivity A. Organic Transformations B. Enzyme Model Systems VI. Conclusion References
I . Introduction
Manganese is one of several first-row transition elements that have been employed by biological systems to assist in varied metabolic and structural roles. Manganese is used to give structural support to pro305 Copyiiglit 1999 by Academic P I F S ~ All ilghla of repiudurtion i n any form reserved ORSR-RR JRi9Y $25 00
306
LAW, CAUDLE, AND PECORARO
teins and is a cofactor in chemical transformations that include hydrolytic and redox reactions. Perhaps the best known function, and the one of greatest importance to aerobic life, is in the oxygen-evolving complex, which oxidizes water to dioxygen during photosynthesis. Some representative reactions catalyzed by manganese-containing enzymes are shown in Scheme 1. Oxygen Evolving Complex 2 HzO
*4Hf+
4e-
+
O2
Manganese Catalase
Manganese Superoxide Dismutase
Arginase arginine
ornithine+ urea
Manganese RibonucleotideReductme
*
RNA
DNA
SCHEME 1.
Several enzymes are known that naturally contain manganese. Most of these employ manganese in nonredox roles (I),for example conducting hydrolytic reactions, a Lewis acid role, or utilizing Mn" to provide structural support to a protein so that it may maintain a specific conformation. For example, pyruvate carboxylase binds Mn" tightly, but the Mn" is not involved in the enzyme's catalytic process, thus playing a structural role (2). Several binuclear metallohydrolases are known that contain a binuclear Mni' site (3-5). Many of these may be categorized as phosphohydrolases, hydrolyzing phosphate groups from a substrate molecule. Hydrolases are often activated by more than one MI1 element, although at least one is specific for manganese. One particularly key metallohydrolase that contains manganese is the enzyme arginase (4-7), an enzyme for which a crystal structure
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
307
at 2.1-A resolution has recently been solved for the arginase enzyme from rat liver (Fig. 1) (8). This enzyme catalyzes the hydrolytic conversion of arginine to urea and ornithine. Arginase is found in a great variety of organisms for the removal of nitrogen-containing wastes and is therefore integral to the biological cycle of nitrogen. Many of the binuclear hydrolases may be active with metals other than Mn at their active sites, but arginase has been shown to specifically require manganese to be fully functional ( 5 ) .The active site contains a binuclear Mnp dimer bridged by a bidentate carboxylate and two p2-"O" donor bridges, a monodentate carboxylate, and a solvent molecule derived hydroxide (or perhaps water a t lower pH, the crystals were grown at pH 8.5). The catalysis of this reaction involves a bridging p-OH unit, which is formed according to a mechanistic proposal by loss of a proton from a water molecule upon binding of substrate. This hydroxide is believed to be ideally situated for forming the intermediates required for arginase hydrolysis (8). Finally, arginase has recently been reported to exhibit some redox activity, catalyzing the disproportionation of hydrogen peroxide and peroxidase activity (discussed later) (9,10). Manganese(I1) ions may also be employed in the hydrolases that compose the ribonuclease H domain of reverse transcriptases ( 3 , 4 ) . Many of these enzymes, which may employ either Mn" or Mg", are found in a variety of organisms where they may or may not be essential (e.g., Escherichia coli) ( 3 , 4 ) . This class of enzymes catalyzes the hydrolysis of DNA-RNA hybrids. However, retroviral reverse transcriptases are critical for the replication of retroviruses, and Mn" may be the required cofactor for these ribonuclease hydrolases to function ( 3 , 4 ) . One example of this family of hydrolase enzymes is the RNase H enzyme from HIV-I. This enzyme has been crystallographically characterized ( 1 1 ) . In the crystal structure at 2.4-A resolution, the two Mn" ions are separated by a distance of about 4 A. They are bound to carboxylate residues that are located near the surface of the enzyme. One of these carboxylates bridges the two manganese
FIG.1. Structure of the active site of the rat liver arginase enzyme based on (8).
308
LAW, CAUDLE, AND PECORARO
ions in a motif that is now recognized as a common structural element in many binuclear metalloenzyme active sites (discussed later). Some controversy with respect to the active sites of ribonuclease enzymes does exist, with some reports suggesting that there may be only one metal ion present (or perhaps required) in the active enzyme (12, 13). Concanavalin A, a lectin, is another interesting nonredox manganese-containing enzyme that is found in various forms in plants and animals (14, 15). Lectins are a class of enzymes known to bind small carbohydrate molecules (16, 17). The active site of crystallographically characterized concanavalin A enzyme shows a six-coordinate Mn" bridged via two carboxylate-bearing protein residues to an approximately seven-coordinate Ca" ion (Fig. 2) (14,015,18).The calcium and manganese ions are separated by about 4.2 A. Concanavalin A is an interesting case of a mixed metal active site in a biological system. If the Call is replaced by another Mn" ion, an EPR spectrum consistent with a weakly coupled MnH system is observed (19, 20). Newer crystal structures of concanavalin A with the substrate methyl-a-Dmannopyranoside have been reported. The more recent structure in 1994 has a resolution of 2.0 A. These structures show that methyl-an-mannopyranoside is bound near the metal binding site, about 8.7 A from the calcium and about 12.8 A from manganese (21). The 1994 structure also indicates that a water bound to the calcium may be involved in hydrogen bonding to the carboxylate-bearing protein residues, which are in turn hydrogen bond t o the methyl-a-D-mannopyranoside (18). Finally, a new crystal structure for concanavalin A at 0.94-A resolution has been solved (22).
H
N FIG.2. Structure of the manganese and calcium dimer in concanavalin A, based on ( 1 4 , 15, 18,221.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
309
Redox chemistry is another facet of the biological chemistry of manganese, and various enzymes employ manganese a t mono-, bi-, and tetranuclear sites. Three mononuclear manganese enzymes are currently known that catalyze redox-based chemical transformations. These are manganese superoxide dismutase, manganese peroxidase, and manganese” dioxygenase. Catalase and the manganese ribonucleotide reductase are two known manganese enzymes that undertake redox transformations that utilize a binuclear manganese site, and a tetranuclear cluster of four manganese is found in the oxygen evolving complex (OEC) of Photosystem I1 (PS 111, the site where water is oxidized during the photosynthetic process to yield dioxygen. The focus of this review will be on mono- and multinuclear redox active manganese systems, and in particular, the relation of bi-, tri-, and tetranuclear manganese complexes as models for the multinuclear manganese-containing enzymes. Aspects of the structurally characterized compounds, spectroscopic features of synthetic compounds and magnetochemistry will also be discussed. The reactivity of complexes will also be addressed, with respect to epoxidation and catalase models, for example. Throughout, discussion will relate our current knowledge of manganese chemistry to the broader area of biological systems. Before addressing the synthetic complexes and their properties in detail, a brief overview of some of the redoxactive manganese enzymes will be presented. Finally, several reviews of manganese coordination complex chemistry have appeared in recent years. Among these are reviews by Wieghardt, Pecoraro and co-workers, Vincent and Christou, Brudvig and Crabtree, and Dismukes and co-workers, which have focused on a wide range of model systems, manganese-dioxygen interactions, mechanistic considerations of reactivity studies, and single systems such as the OEC (5, 23-50). The 1992 book “Manganese Redox Enzymes” (28) also contains much relevant information about biological manganese chemistry by a variety of researchers. A 1996 issue of Chemistry Reviews focusing on bioinorganic chemistry contains several reviews that are pertinent to the biological role of manganese (3, 33, 37). In addition, several reviews aimed solely at the OEC are also available, including reviews by Debus (42) and the recent book “Oxygenic Photosynthesis: The Light Reactions’’ edited by Ort and Yocum (48). Finally, the Proceedings of the Photosynthesis Conference, held periodically, presents short updates from research groups around the world as they work to comprehend this intricate system (51).
310
LAW, CAUDLE, AND PECORARO
II. The Enzymes
A. MANGANESE SUPEROXIDE DISMUTASE, MnSOD Manganese superoxide dismutase (MnSOD) is a redox-active manganese enzyme that employs a mononuclear manganese ion at its active site. Discovered in 1970 (52),this enzyme catalyzes the dismutation of superoxide (HOz), to dioxygen and hydrogen peroxide, as shown in Scheme 2. Superoxide is the radical, one-electron-reduced Mn3+ Mn2+
+ +
-
HOP' HOP' + H+
Mn2+ + O2 + H' Mn3+ + H202
SCHEME 2.
form of dioxygen. MnSOD is found in a broad range of species that run the gamut from bacteria to humans (53).It is a member of one of the two families of superoxide dismutase enzymes, the first of which is the family that contains MnSOD and the analogous iron superoxide dismutase, FeSOD. A high degree of sequence homology exists between the two types of enzymes, and the active sites of both enzymes employ the same types and orientations of protein residues to generate a trigonal bipyramidal geometry about the active site metal (54611. The second family contains the structurally distinct copper/zinc superoxide dismutase, CuZnSOD (62, 63). Recent reports have indicated that a new group of nickel superoxide dismutase enzymes might also exist (64). Superoxide and superoxide dismutases have been the subject of a variety of reviews (5,29, 53, 65). Superoxide, H02, is produced naturally by several enzymes and was first recognized as an intermediate of aerobic respiration in 1968 (62, 66-68). Although it may be utilized beneficially, for example by macrophages t o fight an infection (691,it is generally not a benign byproduct. Therefore, SODS are critical to aerobic life forms, in which they perform the important role of protecting an organism from the deleterious effects of superoxide. Although HOz may not be the direct cause of oxidative damage to an organism due to its relatively short life span (self dismutation occurs at a rate of 2.0-3.2 X lo5 M-' . s - l ) (521, it is sufficiently long-lived to allow it to engender chemical compounds such as hydrogen peroxide or hydroxyl radical, HO., which readily cause oxidative damage to cellular components (70). Another aspect of superoxide chemistry is the growing research into peroxynitrite (71-741, which can be formed from nitrite and superoxide. Thus, superoxide has been implicated as a participant in a vari-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
311
ety of destructive processes affecting DNA, cell membranes, arteriosclerosis and the damage that is caused to cells deprived of oxygen, ischemia, during a stroke or heart attack due to an increased output of superoxide once oxygen is restored to the affected cells and their enzymes (69, 71, 72). MnSOD is often found as a homodimer or homotetramer, with approximately one manganese per monomer subunit (53).A few of these enzymes have been crystallographically characterized, including the enzyme from human mitochondria (75) and the enzyme from the bacterium Thermus thermophilus (76).The active sites of these enzymes have been shown to contain a single manganese bound by three histidines, one aspartate, and either a water or hydroxide in a trigonal bipyramidal geometry as shown in Fig. 3, based on the T. thermophiZus structure (76).In this structure, two of the histidine residues and the aspartate form the equatorial plane, and the third histidine and waterhydroxide occupy the axial positions. Other structures of this enzyme with azide in the active site indicate that the coordination number of the manganese may be expanded to accommodate a sixth ligand, for example, superoxide during turnover (58). MnSOD catalyzes the disrnutation of HOz into dioxygen and hydrogen peroxide. As a redox enzyme, it shuttles between the MnT'and the Mn"' oxidation states (77).This process has been studied, and the enzyme has been shown to catalyze this reaction at a rate of 1-2.2 X 109 M-I.s-I , which is at the diffusion limit (52). A representation of the proposed catalytic cycle is shown in Scheme 3 (78). Several studies on MnSOD kinetics have been published (78-82). Early studies on the reaction of MnSOD from Escherichia coli and Bacillus thermophilus indicated evidence for a four-step process involving two fast and
FIG.3. Structure of the active site of manganese superoxide dismutase, based on (76).
3 12
LAW, CAUDLE, AND PECORARO
two slower steps (79, 80). Later studies utilizing stopped-flow techniques refined this further. Saturation kinetics were observed for the T. thermophilus MnSOD, and the data showed that there were three phases to this reaction: a fast “burst” phase of quick HOz dismutation, a slower second phase, and a final fast phase (78).A “dead-end form of the enzyme was implicated to account for the slow phase (78). This has been formulated as a side-on-bound MnIII-peroxo species based on spectroscopic similarities to manganese model complexes with side-bn-bound-peroxy groups (83). Such phases are not observed for FeSOD (84).
+
H02
t
Mn3+SOD-Oi
Mn2+SOD
“Dead-End”Complex
SCHEME 3.
B. MANGANESE(II) DIOXYGENASE There now exist three reported Mn(I1) dioxygenase enzymes. The most extensively studied of these is the Mn(I1) dioxygenase from Arthrobacter globoformis CM-2 (85).The first Mn(I1) dioxygenase was reported in 1981 from Bacillus breuis (86).Dioxygenases catalyze the incorporation of dioxygen into an organic substrate (87). In this case, the substrate is an aromatic compound that contains a catechol cisdiol structure. These reactions are important in that they catalyze the degradation of aromatic compounds to simpler carbon components for action by other enzymes. In addition, the ability of dioxygenase enzymes to initiate the decomposition of aromatic compounds may be useful in the bioremediation of sites contaminated with aromatic organics of this type (88). Most of the known dioxygenases contain iron, and these enzymes tend to be very specific in their function (87, 89). For example, the
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
313
Fe(II1) dioxygenases have generally been shown to be intradiol dioxygenases, meaning that they cleave the aromatic ring between the cisdiols. The Fe(I1) dioxygenases are classified as extradiol dioxygenases and cleave to the aromatic ring to one side of the diol moiety. Substrate specificity is also exhibited to various degrees, with many dioxygenases preferentially cleaving one substrate over others (89). The Mn(I1) dioxygenases characterized to date belong to the class of extradiol dioxygenases (86, 90). Although none of the Mn(I1) dioxygenases have been structurally characterized, a recent structure has appeared for an Fe(I1) dioxygenase (91).Whereas the Fe"' enzymes have phenolate and histidine ligands, the Fe" was bound to the protein by two histidines and a glutamate, with two waters completing the first coordination sphere. The geometry about the iron atom is square pyramidal, with one of the histidines occupying the axial position (Fig. 4).Finally, some evidence for an Fe"/Mn" dioxygenase correspondence similar to that observed for MnSOD/FeSOD might also exist, as evidenced by the conserved residues that form the metal binding sites of these dioxygenases (92). Thus, this manganese enzyme may have a structure similar to the iron site. The details of the reactions of the iron dioxygenases have been extensively probed for the Fe(II1) enzymes and more recently for the Fe(I1) dioxygenases (87,891. These two groups of enzymes exhibit different catalytic processes. For both of these types of enzymes, though, it has been shown that although the reaction catalyzed is redox in nature, the central metal does not cycle between discrete oxidation states during the reaction. A resonance form of an intermediate, however, has been proposed to exist briefly as Fe"' for the Fe(I1) dioxygenases. A recent proposal for the catalytic cycle of an Fe(I1) dioxygenase is presented in Scheme 4 (89, 93). The catechol substrate is thought to bind as a monoanion to the central metal ion followed by the coordi-
FIG.4. Structure of the active site of an extradiol iron(I1) dioxygenase, based on (91).
314
LAW, CAUDLE, AND PECORARO
SCHEME 4. A proposed mechanism for extradiol Fe(I1)dioxygenases. [Adapted with permission from (89).Copyright 1996 the American Chemical Society.]
nation of a dioxygen molecule to the metal, which might be formally considered to be a superoxide bound t o an Fe”’. It is believed that substrate binding activates the metal center for binding dioxygen, based on the increased affinity for binding nitrosyl to the active site iron in the presence of substrate (89). Attack of the “superoxide group” on the carbon adjacent to the catechol moiety followed by rearrangement to give the product and the product’s release complete the catalytic cycle. This mechanism contrasts with the Fe(II1) dioxygenase mechanism, wherein the iron activates the substrate for direct attack of dioxygen on the aromatic ring of the substrate. The reaction rate of the manganese containing A. globoformis enzyme, with its primary substrate 3,4-dihydroxyphenylacetate, has been measured, yielding values of kcat= 1400 min-l, and K, = 7 pM (85).It is likely
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
315
that a similar mechanism exists between the Fe" and Mn" enzymes (89,90). 1. Modeling Features, MnSOD and Mn(II) Dioxygenase These two enzymes present a unique challenge for the synthesis of small-molecule model complexes. First, the MnSOD has been shown to have an active site with a trigonal bipyramidal geometry about the manganese ion, whereas the Mn(I1) dioxygenase most likely contains a square pyramidal geometry about the manganese ion in accord with the Fe(I1) dioxygenase structure. Few complexes of manganese in a trigonal bipyramidal geometry are known, and these are predominately of Mn", whereas the MnSOD enzyme has been crystallographically characterized with a trigonal bipyramidal geometry about the manganese in both the Mn" and Mn"' oxidation states. In the case of the Mn(I1) dioxygenase, the structure is probably square pyramidal by analogy with the Fe(I1) dioxygenase structure. Square pyramidal complexes are the more common form for five-coordinate manganese complexes, but this geometry is most often observed for the Mnl" oxidation state due to the presence of the pseudo-Jahn-Teller distortion axis, and not Mn". The reactivity of model compounds is also key to gaining a better understanding of an enzyme. The MnSOD presents a number of challenges, the first of which is catalytic dismutation of superoxide. Many Mn compounds are known to interact (discussed later) with superoxide, but few have been shown to be catalytic in nature. The unique characteristics of the catalytic cycle, with the "dead-end" complex, also provides challenges for synthesis. Further, there has been some interest in recent years in the use of manganese complexes as pharmaceutical superoxide dismutase agents to aid in reducing cellular damage by superoxide, for example in the wake of a heart attack, which opens questions of the biological compatibility interactions of the complexes and their solution behavior at physiological pH values.
C . MANGANESE PEROXIDASE, MnP Manganese peroxidase (MnP) is an unique enzyme in many respects. It is an extracellular enzyme that involves a heme protoporphyrin IX for the oxidation of Mn" to Mn"' (94,951. A crystal structure at 2.06-A resolution of the manganese peroxidase from the white rot basidiomycete Phanerocaete chrysosporium, which utilizes this enzyme to degrade lignin, appeared in 1994 (96).The active site (Fig. 5) and the overall structure are quite similar to lignin peroxidase (Lip),
3 16
LAW, CAUDLE, AND PECOFURO
FIG.5. Structure of the active site of MnP. The left half is a side-on view of the active site showing the protoporphyrin M, the bound proximal histidine, and the water and histidine poised on the opposite face. The right half is a “top”-down view showing the heme and the bound Mn”, and the additional protein residues that act as ligands to Mn”. [Adapted with permission from (96).Copyright 1994, The American Society for Biochemistry and Molecular Biology.]
an enzyme with which MnP shares a significant sequence homology (96). The white rot fungi are some of the only species known that are able to degrade lignin (97-99). Furthermore, MnP itself is one of only two, or perhaps three, enzymes that are known to degrade lignin. The related Lips are the other lignin-degrading enzymes that have been known since the early 1980s (98, 99). Meanwhile, a few more recent reports have suggested that some fungi may also utilize a laccase in addition to MnP and Lip to facilitate lignin degradation (100, 101). The MnP and Lip enzymes catalyze the biodegradation of lignin to phenolic residues that can then be further decomposed. Lignin, found predominantly in plant cell walls, is a random three-dimensional polymeric material composed of phenolic residues that is thought to be the second most common biopolymer after cellulose. Lignin comprises a significant portion of woody plants, with lignin content values ranging from 15 to 36%(97-99,102). Lignin is also found in a variety of other vascular plants. The major role of lignin in the plant is to provide support and protection for plant tissue (98, 99). The lignindegrading peroxidases act on lignin by initiating one-electron oxidations that lead to the breakdown of the linkages that connect the building blocks of this material. It has been shown that the decomposition of cellulose is retarded by lignin, probably as a result of lignin
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
317
interfering with access to the cellulose (98, 99). [As has been pointed out, the importance of the degradation of lignin is underscored by the extensive amounts of carbon employed in lignin and cellulose manufacture by plants and the interplay that this may have with respect to the cycling of carbon on the globe (97-991.1 Thus, the Mn-catalyzed degradation of lignin may be designed to allow the fungus access to the cellulose (98, 99). MnP functions by the oxidation of the heme by hydrogen peroxide to form compound I, Ferv=O(P+')(Scheme 5). (95, 103-106). The oxidized heme then oxidizes a Mn" that is bound to at least one of the
oxalate
SCHEME 5.
two propionate arms of the heme group in a site located near the edge of the enzyme, an observation based on the recently solved crystal structure of the P. chrysosporium enzyme (96).This forms a Mn"' ion and Compound 11, Fe(IV)=O. The Mn"' then diffuses away from the enzyme to oxidize lignin. In this way, the enzyme can work on the random structured polymer, where structural specificity may be a problem for an enzyme and where Mn"' may diffuse more easily into the substrate than a larger enzyme would be able to do. The Mn"' initiates a one-electron oxidation, most likely forming an initial phenoxy radical cation, that triggers a cascade of oxidations that eventually break down the linkages between the lignin subunits. Compound I1 is then reduced to the resting Fe"' oxidation state by oxidizing a second equivalent of Mn". This step absolutely requires manganese, preferentially a Mn"-oxalate complex (1051, unlike the first step, which will function with substrates other than manganese (103).A new addition to the chemistry of lignin peroxidases is the observation that Lips are also capable of oxidizing Mn" to Mn"' in the presence of oxalate (107). Thus, some Lips may also act as a manganese peroxidase. Oxalate and similar small molecule chelators play a key role in the
3 18
LAW, CAUDLE, AND PECORARO
function of MnP (104-109). Oxalate and malonate have been observed to be excreted in active MnP-producing cultures (104), and they have been shown to promote enzyme activity (105, 106, 108, 109). One of the roles proposed for the small molecules is to aid in the diffusion process of Mn"' and to stabilize the Mn"' from disproportionating to Mn" and MntVso that it may reach the target lignin and initiate the depolymerization of lignin (95, 103). It is not known whether oxalate binds with manganese to the enzyme during the reduction of Compound I to Compound I1 or whether a small molecule is taken up to assist departure of Mn"' from the enzyme. No evidence for oxalate or other small molecules bound to manganese were observed in the crystal structure other than water. It is known, however, that without the oxalate or a related small chelating molecule, turnover is reduced (104-106). Furthermore, studies suggest that the manganese required for the reduction of Compound I1 must bind with the exogenous ligand oxalate (105, 106). Experiments with a-hydroxy acids such as lactic acid or small organic acids such as malonic acid, showed that such compounds also appeared to promote the activity of MnP (95, 103). However, related compounds that could form stronger complexes with manganese were found to inhibit the enzyme, presumably by out-competing the enzyme system for binding Mn". It was proposed that these compounds might facilitate the diffusion of Mn from the enzyme and in that way lead to enhanced reactivity. Roles for oxalate beyond the reduction of Compound I1 (e.g., assisting in the diffusion of Mn"') remain controversial (104-107, 109). Oxalate is known to be decarboxylated in the presence of Mn"', so it may not be utilized in Mn"' diffusion, although that may be a role for the malonate. One other proposal for the role of oxalate was the reduction of MnIV0 2 which , is known to form near these enzymes, to return the manganese to a functional oxidation state. Some model work has addressed these points, and although unstable, such Mn'II-oxalate complexes may have lifetimes sufficient for them to reach the substrate and initiate oxidation before they would completely decompose. Finally, there is a functional similarity between the MnP and Lip enzymes (110). Lip has been shown to be able to oxidize Mn" (107). Further, it has been shown that a veratryl alcohol is produced in conjunction with Lip, and it has been suggested that this compound may serve Lip in a capacity similar to that served by Mn" in MnP (109).
Functional mimic systems for the MnP enzyme will be discussed in Section V, "Reactivity." An example of monomeric complexes prepared with a-hydroxy acids as functional mimics is presented in Section III.A.3 on monomeric structures.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
3 19
D. MANGANESE RIBONUCLEOTIDE REDUCTASE, MnRR The conversion of ribonucleotides into deoxyribonucleotides for the synthesis of DNA is key to the survival of any DNA-based life form. This process is completed by one of a variety of ribonucleotide reductases that reduce the furanose ring of a ribonucleic diphosphate acid monomer by replacing a hydroxyl functional group at the 2'-position with a hydrogen (Scheme 6) to generate the deoxyribonucleic acid Phosphates.
Phosphates.
* HO
HO
OH
RNA
H
DNA
SCHEME 6
monomers necessary for DNA synthesis ( 111 -1 14). Ribonucleotide reductases have been identified with a wide variety of protein structures and active site compositions. Many are known that contain a diiron core (in the oxidized form of FeRR, the active site contains a n FeB' dimer bridged by a p-0x0 and a p-carboxylato), a stable tyrosine radical, and potentially redox active thiols in the form of cysteines a t the active site ( 1 1 4 , 115). The role of the diiron dimer is to generate the stable tyrosyl radical, which in turn initiates a transient thiyl radical at the active site (114, 115). Two other classes of ribonucleotide reductases have also been identified; one is based on adenosylcobalamin (114, 116) and the other, which is found in strictly anaerobic microorganisms, likely utilizes a glycyl radical in conjunction with S-adenosylmethionine to reduce RNA. These enzymes have been grouped as Class I, Class 11, and Class I11 ribonucleotide reductases, respectively ( 116). A fourth segment of these enzymes is represented by the manganese-based ribonucleotide reductases, which probably function in a manner similar to the iron ribonucleotide reductases (discussed later) and, therefore, may fall under the Class I umbrella (116).The MnRR enzyme is a bacterial enzyme, and is widely distributed among the Coryneform and related bacteria ( 1 1 1 , 1 1 7 , 118). Several members of this group of bacteria have been shown to absolutely require manganese for the transformation of RNA into DNA. The MnRR enzyme from Corynebucterium (formerly Brevibacterium) ammoniugenes has been the most studied member of this group ( 1 1 1 ) . The FeRR system has been extensively studied (112, 114). For this type of ribonucleotide reductase, the enzyme is comprised of two subunits, known as R1 and R2 (or occasionally B1 and B2). The R1 and
320
LAW,CAUDLE, AND PECORARO
R2 subunits are each comprised of a homodimer of the a- and p-polypeptides, respectively (112, 114), and both the R1 and R2 subunits have been crystallographically characterized (119-122). The active site for RNA reduction is located on the R 1 subunit. The oxidized diiron site and the stable tyrosine radical are located in the R2 subunit. The diiron site is comprised of two Fe"' ions,obridged by an 0x0 and a carboxylate, with an Fe-Fe distance of 3.3 A. The reduction of RNA begins with the abstraction of a hydrogen atom from the 3'position of the furanose ring by a transient thiyl radical. This is believed to be generated via an electron-transfer mechanism that allows the tyrosyl radical to form a thiyl radical from one of the cysteines located at the active site. In addition to that cysteine, a pair of redoxactive cysteines is located at the active site. This pair of cysteines is then proposed to be oxidized to a disulfide-linked moiety in conjunction with substrate reduction. In the first step, a hydrogen atom is abstracted by the thiyl radical from the 3'-position on the furanose ring. A proton from one of the redox-active cysteines is transferred to the hydroxyl group of the 2'-position to form a water-molecule-leaving group. Completion of the oxidation to the disulfide and the regeneration of the initial thiyl radical by H-atom transfer to the 3'-position of the furanose yields the DNA monomer product (Scheme 7 ) (114, 115). The disulfide linkage is then believed to be reduced by another enzyme system to regenerate the active site (115).For several of the MnRR systems, a thioredoxin/NADPH-dependent thioredoxin reductase system is thought to complete this phase of the process (123). Although an X-ray crystal structure of MnRR has not been solved to date, a structure with Mn substituted into the iron ribonucleotide reductase has been reported (Fig. 6) (124). In this case, the two manganese(II1) ions were bridged by two carboxylates from protein residues, but without the 0x0 bridge postulated for the active Fe;" form of the FeRR enzyme. This inactive enzyme with manganese was proposed as a model for the inactive Fe;' form of the iron enzyme and is
FIG.6. Structure of the Mn"-substituted FeRR enzyme, based on (124).
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
32 1
Phosphates,
DNA
Phosphates,
HO H20'
/
Phosphates,
O O M S E
H2O Phosphates,
o ) W "
HO
+
HO
SCHEME 7. A partial proposed mechanism for the functioning of iron ribonucleotide reductase. Manganese ribonucleotide reductase may function in a similar manner. [Adapted with permission from (113)and ( 1 1 1 ) . Copyright 1994, Current Biology Publishers, and Copyright 1994 Marcel-Dekker, Inc., respectively.]
quite similar to the structure of the reduced Fekr active site which has since been crystallographically determined. Based on UV-visible spectroscopy of MnRR and the similarity that this spectrum bears to that observed for catalase and to model 0x0-, carboxylato-bridged Mn"' dimer complexes, the active site structure of MnRR is proposed to include the Mn-0-Mn structural motif ( 1 1 1 ) . This would also be similar to the active site structures observed for FeRR. This proposal is
322
LAW, CAUDLE, AND PECORARO
further bolstered by the recent crystal structure of the catalase from the bacterium Thermus thermophilus (125) (discussed later), which generally corroborates the postulated catalase p-carboxylato-, p-oxobridged Mn dimer motif for the catalase active site structure. In addition, the assignment of a Mnf’ oxidation state is supported by the EPR silence of this enzyme with respect to the observation of any signal arising from a manganese dimer (111). MnRR converts all four RNA bases into their corresponding DNA products (111, 123), and the overall enzyme is considered to be structurally similar to FeRR ( 1 1 1 ) . For example, MnRR is comprised of R1 and R2 subunits that are analogous to those of FeRR. A mechanistic similarity between these two systems, however, was not certain. No radical signal equivalent to the FeRR tyrosyl radical had been observed for MnRR prior to 1996 (126).One tantalizing piece of evidence was the hydroxyurea sensitivity of MnRR, a trait also observed for FeRR. Hydroxyurea is known to quench the FeRR tyrosyl radical (118). In the meantime, proposals for this lack of a radical signal had been put forward. One proposal suggested that the coupling of a stable protein radical to the Mn of MnRR might account for the lack of any observed radical by EPR in MnRR (111).Another suggested that the manganese dimer might engage more directly in the RNA-to-DNA conversion process, perhaps by abstracting the initial hydrogen atom itself to initiate the reduction process, in a role distinctly different from that proposed for the diiron site of FeRR (111). Proposals that mirror FeRR, however, may now, in light of the radical signal that has recently been observed in the MnRR R2 subunit, be reasonably considered (126).Unlike the relatively stable FeRR enzyme radical, though, the MnRR radical only has a half-life of 1.5 h. It was only after developing a “fast” isolation technique for this subunit that Auling and co-workers were able to observe this radical signal in the EPR. By analogy, this radical is proposed to be located on tyrosine residue. Finally, the presence and concentration of the radical signal were shown to correlate with the activity of an MnRR sample (126). Otherwise, little is known about the exact identity or proximity of this radical to either the manganese dimer or the substrate binding sites. E. MANGANESE CATALASE Catalase is yet another manganese-containing enzyme that interacts with dioxygen or its reduced forms. The role of this manganese enzyme is to protect the organism from oxidative damage that may be caused or initiated by the presence of hydrogen peroxide, such as in the interaction of hydrogen peroxide with Fe” in Fenton-like chem-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
323
istry to generate H0’63. To date, manganese-containing catalases have only been found in bacteria-for example, the enzymes from Lactobacillus plantarum (127, 1281, Thermoleophilum album (129), and T. thermophilus (130, 131). Several reviews with respect to the catalase itself and the modeling chemistry of this enzyme have appeared in recent years (3,5 , 2 3 , 2 5 , 2 6 , 2 9 ,30, 38). The catalase enzyme that utilizes manganese instead of a heme group to catalyze the disproportionation of hydrogen peroxide t o dioxygen and water was known for several years before the presence of manganese was confirmed in 1983 (128). A low-resolution crystal structure in 1985 indicated that the enzyme from T. thermophilus contained a binuclear manganese site, with a Mn-Mn separation of about 3.6 A (130, 131). This was further supported by evidence from optical spectroscopy that suggested that the active site contained two manganese ions that were bridged by one 0x0, p-02-, and at least one carboxylate ligand based on spectroscopic comparisons to 0x0- and carboxylato-bridged manganese model complexes (30, 132, 133). Extended X-ray absorption fine structure (EXAFS) studies of this enzyme suggested that the first coordination sphere about the manganese of the catalase consisted of O/N donors (134, 135). The presence of the dimer was further bolstered by the observation of EPR spectra consistent with a coupled manganese dimer, for example the Mn”’Mn” EPR spectrum of the “superoxidized” enzyme (136).EXAFS showed that this “superoxidized” catalase has a Mn-Mn vector of 2.7 A, consistent with a Mnz(p2-O2- core. Very recent reports of crystal structures at resolutions of 1.6 A for the reduced, Mnf, and 1.4 A for the oxidized, Mn$, forms of the enzyme from T. thermophilus confirm much of what had been proposed for the structure of the catalase’s active site (125).A representation of these results is shown in Fig. 7. The active site assembly contains two five- or six-coordinate
FIG. 7. Structure of the active site of the 2’. fherrnophilus catalase enzyme, based on (125).
324
LAW, CAUDLE, AND PECORARO
manganese ions bridged by one carboxylate from a protein residue, one water molecule, and a hydroxide ion in the reduced enzyme. The hydroxide is suggested to be converted to an 0x0 bridge in the oxidized form of the enzyme. One additional water molecule, carboxylates, and imidazoles complete the first coordination spheres of the Mn ions. The geometry about both Mn ions has been describtd in terms of square pyramidal (125). The Mn-Mn vectors are 3.18 A in the reduced form and 3.14 A for the oxidized form of the enzyme. Finally, this binuclear manganese site is bound within a bundle of four a-helices. The T. thermophilus enzyme itself is a homohexamer, composed of six identical subunits (125, 131). During its catalytic cycle, the Mn catalases shuttle between the MnP and the Mn$ oxidation states in 2 two-electron reactions (137). Overall this process converts two hydrogen peroxide molecules into one molecule of dioxygen and two water molecules. A proposal for the catalytic cycle of catalase is presented in Scheme 8 (23). When the
, Y o
o r 0
SCHEME 8. A proposed mechanism for the disproportionation of hydrogen peroxide by manganese catalase, based on the mechanism proposed by Penner-Hahn in 1992, and the 2'. thernophilus catalase crystal structure. [Adapted with permission from (24). Copyright 1992 WILEY-VCH Verlag.]
T. thermophilus enzyme is isolated, a mixture of enzyme oxidation states including MnH, MnIIMnIII, MniI', and Mn1"MnWis observed (138, 139). The MnIIMn"' and Mn"'MnTVoxidation states are catalytically inactive but can be reduced to active oxidation states, Mn"Mn" or Mn'IIMnIII, by hydroxylamine (23). The Mnl*'MnTV oxidation state is quite intriguing because it is thought to contain a bis-(p2-02-),core which previously has been implicated in the oxygen-evolving complex
MANGANESE REDUX ENZYMES AND MODEL SYSTEMS
325
(23). The products of the "superoxidized" state may be explained if
one considers the possibility that occasionally during turnover an active site at Mnil' is reduced by only one electron to Mn"Mn"'. The following two-electron oxidation step would then yield the Mn"'Mn'" form of the enzyme, which is only slowly reduced back to an active Mn!" form. Such chemistry was shown to be viable by the introduction of the one-electron reductant hydroxylamine to catalase followed by the addition of hydrogen peroxide (23). These catalases exhibit saturation kinetics and complete the disproportionation of hydrogen peroxide at a rate of 5.75 X lo5 M-l-s-' for L. plantarum (231, 1.7 X lo6 M-. s-l for T. album (129) and 3.2 X lo6 M-'. s-' for T.thermophilus (131, 140). Finally, the catalase enzymes have been shown to be inhibited by azide, chloride, fluoride, and thiocyanate (23). In the case of fluoride inhibition, there is significant controversy both for the number of fluorides that are bound and whether or not the anion acts as a terminal ligand or a bridging p2-F- species. Numerous model complexes that model one or more traits of the catalase enzyme have been prepared over the years (discussed later). Several systems that mimic the function of this enzyme have been reported, and a selection of these will be discussed in Section V, "Reactivity." Other models with respect to structures or physical properties of the catalase enzyme will be presented in the ensuing sections of this review.
F. BINUCLEAR MANGANESE ENZYMES: A COMPARISON OF THREE ENZYMES One of the striking features of the two redox-active binuclear enzymes described earlier, MnRR and catalase, and the hydrolytically active enzyme arginase is the apparent similarity of their active site structures. Yet all three exhibit vastly different reactivities. Both the MnRR and catalase exhibit UV-vis spectra consistent with carboxylato- and 0x0-, hydroxo-, or aquo-bridged cores. This would indicate that both have similar Mn-Mn core structures. An EXAFS study by Stemmler, et al. (9) compared the active sites of these three enzymes. Those studies suggested that arginase and the iron ribonucleotide reductase substituted with Mn had similar structures that tended to be six coordinate. Their data for catalase suggested that the manganese of this enzyme were likely five coordinate. No crystal structure for MnRR has been solved to allow for a specific identification of ligands or a direct comparison of distance parameters and ligand orientations. However, it is possible to compare the arginase (8)and catalase crystal structures in this manner (125). Other comparisons have ap-
326
LAW, CAUDLE, AND PECORARO
peared in the literature (9, 10).These environments are quite similar, as can be seen from Figs. 1 and 7, and the potential correlation between the two sites is strengthened by the low level of catalase activity exhibited by arginase (9, 10). The similarity of the bridging motifs and protein residues as ligands is striking. Both have enzyme-active sites that contain dimers of Mn ions bridged by a bidentate carboxylate and two ,uZ-"O"donor bridges: a monodentate carboxylate and a hydroxide (crystals grown at pH) in arginase, and a water molecule and hydroxide in the reduced catalase. The remaining ligands to the metal ions on each side of the dimer are identical in terms of type, carboxylates, and imidazoles, and in terms of their general positions about the manganese dimer. In the case of arginase, one Mn ion is five coordinate, the other is six coordinate due to a bidentate carboxylate. The catalase contains a water in the area of what would be the corresponding open site on the five-coordinate arginase Mn, and the carboxylate donor on the second manganese is monod$ntate instead of bidentate. The Mn-Mn vector for arginase is 3.3 A compared to 3.14 and 3.18 for the MnH and Mnil' oxidation states of catalase. Presumably, the native MnRR will also have a Mn-Mn vector in the 3.0- to 3.5-A range. The broader structure about the Mn cores of these sites of the arginase and catalase enzymes, however, does vary considerably from one enzyme to the other. Both of these enzymes are located within hydrophobic pockets. However, in the case of the active-site structure of catalase, there is an additional glutamate that may hydrogen bond to two of the imidazoles that are bound to the manganese and a lysine NH3+group is also hydrogen bonded to a manganese ligand carboxylate (125). In arginase, there are only two other polar residues reported to be near the active site beyond those that are bound to manganese (8). One of these is a histidine imidazole, which may function in a proton shuttle capacity. The other is a glutamate, which probably salt-bridges to the guanidinium functional group of the arginine substrate. The protein structures near the active sites are also quite different. In catalase, theobinuclear active site is located within a fourhelix bundle about 18 A from the surface at the bottom of a channel. Overall, the catalase enzyme is composed mostly of helical structural motifs. Meanwhile, the arginase enzyme's active site is not located within a four-helix bundle, and this enzyme contains several helix and P-Fheet structures. The active site of arginase sits at the bottom of 15-A-deep cleft that is located within an area of turns and loops between helices and P-sheets. The reactivity of catalase enzymes and rat liver arginase may also now be directly compared. Arginase has recently been reported to dis-
A
A
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
327
proportionate hydrogen peroxide, albeit much less effectively than the actual manganese catalases (in a buffered solution a t pH 9.5) (9, 10), albeit much less effective than the actual manganese catalases, exhibiting a rate of reaction on the order of 30 s - I versus 200,000 5-l for L. plantarum (10).It was also reported that evidence for the manganesesubstituted iron ribonucleotide reductase exhibiting a low level of catalase activity might also exist (9).No "superoxidized form of arginase could be formed in the presence of hydrogen peroxide and hydroxylamine, but it did show hydroxylamine inhibition. Another difference between these systems was a lowered sensitivity of arginase to the inhibitory effects of halides (10). A final point of comparison is the reported peroxidase activity of arginase. The catalase from T. album has been shown to complete a peroxidase reaction on p-phenylenediamine (129). Arginase was shown to perform peroxidase activity on this compound and o-phenylenediamine (10).Thus, as one might expect from their structural similarities, these enzymes can in some cases catalyze similar chemistry. Thus, despite the initial similarities of the immediate active site dimers, a broader inspection of the enzyme begins to suggest potential explanations for the differences in their reactivity. For example, the existence and type of residues near the active site, with respect to acidhase chemistry, that may attenuate the overall reactivity of the core will play a role. The types of bridging p"0" moieties, monodenate carboxylate and hydroxide in arginase versus hydroxide and water in catalase, probably play a significant role. The hydroxide and the proposed 0x0 (or perhaps two hydroxides) of the MniI' form of the enzyme may be better able to support the higher oxidation state on manganese versus the carboxylatehydroxide bridges of arginase. The two five-coordinate manganese of catalase may play a role, too, if rearrangement were required for the activation of arginase. Overall, these enzymes exemplify a growing theme in bioinorganic chemistry, that of nature utilizing an active site structural motif in more than one enzyme to catalyze quite similar or dissimilar chemistries. Such reapplication of a structural motif has been observed in the recurring themes of binuclear iron ( 1 4 1 ) ,for example. For manganese, it appears that a general binuclear manganese active site design has been reapplied three times in separate settings for at least three different purposes.
G. THEOXYGEN-EVOLVING COMPLEX, OEC Dioxygen is important to all aerobic life forms. It is required for respiration and, therefore, the release of energy required to sustain
328
LAW, CAUDLE, AND PECORARO
biological processes. Dioxygen is a by-product of photosynthesis and is produced by the oxidation of two molecules of water to dioxygen in conjunction with the release of four equivalents of protons and four equivalents of electrons, which will be processed into reducing equivalents suitable for eventual carbohydrate biosynthesis. The catalytic site for water oxidation is known as the oxygen-evolving complex (OEC) of Photosystem I1 (PS 11),located in the chloroplasts of green plants and algae (42, 48). The OEC, sometimes referred to as the WOC, water oxidation complex, or water oxidase, contains a group of four manganese ions at its active site (42).Calcium and chloride are also required for the proper function of the OEC (35).Part of a complex system of several polypeptides and cofactors, which are membrane-bound in the thylakoid membrane of green plants, the OEC’s role is to provide the reducing equivalents necessary to rereduce a specific photoexcited chlorophyll a molecule that initiates the chain of events leading to the production of reducing equivalents. The polypeptides of PS I1 and those surrounding the OEC serve in various roles, many of which have been probed and elucidated ( 4 2 , 4 8 , 5 1 ) .Some of these roles relate directly to the manganese cluster of the OEC and the other inorganic cofactors. Based on the wealth of research in this field related to PS 11, a gross structural similarity of green plant and algal PS 11s is believed to exist with that of the crystallographically characterized (142) bacterial (Rhodopseudomonas) PS 11s (42, 143) that do not employ the OEC. The commonalities between these systems has led to a proposed orientation of the various components of PS I1 (Fig. 8) (24).The OEC is located near the “base” of two polypeptides known as D1 and D2 and is surrounded and stabilized by several additional polypeptide units (42, 144).
+
4
FIG.8. A scheme of the generally proposed layout of the components of Photosystem 11. The polypeptides are labeled, too. The electron flow in PS I1 follows the arrows shown. [Reproduced with permission from (24).Copyright 1998 Springer-Verlag.1
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
329
For more detailed discussions of the biochemical aspects of the PS 11, such as the role of the various polypeptides or differences between the PS IIs, the reader is directed to one of several reviews and compilations. The OEC and PS I1 have been extensively reviewed over the past several years. The review by Debus in 1992 sums up much of the knowledge obtained on PS I1 to that point (42). The Photosynthetic Congress, held every few years, has produced several volumes of material from a broad spectrum of researchers covering the whole of photosynthesis with current results to the time of that Congress (51). A 1996 book edited by Ort and Yocum (48) contains several chapters covering various aspects of photosynthesis, from PS I1 to PS I. Manganese modeling chemistry has also been extensively covered by reviews (24,26-31,34, 37, 45,50).
The OEC is involved in the earliest stages of the overall photosynthetic process, which produces NADPH for carbohydrate biosynthesis. Overall, the photosynthetic system is composed of two halves, PS I1 and PS I (48). PS I1 provides reducing equivalents that are transferred to PS I and are eventually converted into the reducing energy of NADPH. Both PS I and PS I1 absorb light energy to initiate their respective electron flows. The protons released during these steps are likely utilized in the generation of a proton gradient for the synthesis of ATP, which is also required for carbohydrate biosynthesis (48). It has been established that the OEC contains a tetranuclear cluster of manganese ions, and that it requires at least one calcium ion (39, 145, 146) and at least one chloride ion (147, 148) to complete its catalytic cycle. This catalytic cycle involves four successive oxidations of the OEC (42, 48) to eventually yield a highly oxidized system that then completes the oxidation of two molecules of water with the concomitant release of dioxygen and the return of the OEC to an overall oxidation state four levels lower. The sequential oxidation process starts when light energy is harvested by antennae chlorophyll molecules, which transfer this energy to the Psso chlorophyll a molecule of the reaction center. This photoexcited P680 then reduces Pheophytin A to give a charge-separated state of Ps+Boand Pheo-. Pheophytin A then reduces a bound plastoquinone molecule, QA, which then reduces a dissociable plastoquinone molecule, known as QB. Once Q B has been doubly reduced and doubly protonated, it departs to carry the collected reducing equivalents further into the photosynthetic system (Fig. 8). Once oxidized, Psso+ needs to be “reset” so that new reducing equivalents may be initiated, and to prevent charge recombination, or back-reactions, from occurring. This is accomplished by oxidizing the OEC. This oxidation is mediated by a particular tyrosine residue, known as Tyrosine Z, Y;,residue 161 of the D1 polypeptide (149,150).
330
LAW, CAUDLE, AND PECORARO
This tyrosine is initially oxidized by P&, and then it in turn oxidizes the OEC by one electron. In this manner, the OEC is oxidized successively four times, whereupon dioxygen release occurs, the OEC returns to its initial oxidation state, and the successive oxidation cycle begins anew. Four protons are generated in conjunction with the electron withdrawal and dioxygen release, and these are likely utilized in the generation of a proton gradient for ATP synthesis (42, 144). The cycle of four progressive oxidations followed by release of dioxygen with the concomitant reduction of the manganese center of the OEC has been termed the S cycle, or S clock (151-153) (Scheme 9). This 2 H20
/
SCHEME 9.
clock is a five-stage cycle that proceeds from So to S4, with So being the most reduced form of the enzyme whereas S4represents the most oxidized form. Dioxygen release occurs during the S3-S4-So phase of the cycle, presumably once S4 is reached (152, 153). S , is relatively unstable and rapidly returns to the So state (46).The initial evidence for the S cycle was reported by Joliot (151).Those experiments were reproduced and reinterpreted by Kok and co-workers (153), who showed that, for photosynthetic centers given a sequence of single flashes of light, a burst of dioxygen occurred with the third flash of light and a burst of dioxygen then occurred with every fourth flash thereafter (Fig. 9). The initial burst came on the third flash, indicating that PS 11s stored in the dark will equilibrate t o the S1 state, which is one step oxidized from the OEC’s lowest overall oxidation state. Further research has tied at least some S-state progressions to an increase in the overall oxidation of the manganese cluster of the OEC (discussed later). In the absence of crystallographic data, all of the structural infor-
33 1
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
2.4
30 FLASH NUMBER
n
-I
wh
1.6
-
0" Q
w
N
= 0.0 -
J
4
p: 0
0 2 EVOLUTION IN FLASHING LIGHT
z
1
0
6
I
1
12 18 FLASH NUMBER
I
I
24
30
FIG.9. The Four Flash experiment. The peaks in the plot represent the dioxygen bursts that occur with every fourth flash of light administered to a PS I1 sample. This cycle peaks on the third flash due to the resting state of the enzyme being a t S , . The process is eventually dampened by double hits of light leading, for example, to an inhomogenous sample with respect t o the S state. [Reproduced with permission from ( 4 6 ) . Chapter 9. Copyright 1996 Kluwer Academic Publishers.]
mation known about the OEC has been gathered by the use of spectroscopy (24,28,33,42,46). The use of X-ray absorption spectroscopy ( U S ) has provided some particularly useful quantitative structural information. EXAFS (extended X-ray absorption fine structure) is an X-ray absorption spectroscopic technique that provides distance information and some idea about the nature of the first coordination sphere of the element in question. EXAFS studies have shown that the OEC manganese ions are present in a predominantly O/N-donor coordination environment (24, 154). Furthermore, EXAFS data show that there exists at least one Mn-Mn vector, and prob%bly two, of 2.7 A, and a third significant vector is at a distance of 3.3 A (155-159, 477), which may result from a Mn-Mn or Mn-Ca interaction (discussed later). EPR spectroscopy has been another extremely important means for exploring the OEC (24, 28, 42, 46). The S, state is readily prepared
332
LAW, CAUDLE, AND PECORARO
by dark-adapting preparations of PS IIs, from which Sz is achieved by flashing the sample with one pulse of light (42).The Sz state exhibits at least two distinct EPR signals, which have been the focus of extensive research and controversy, too. The first, the multiline, is a multiline signal centered around g = 2 (160, 161), and the second, which depends on sample preparation technique and conditions, appears a t g = 4.1 (162-166). These EPR signals may be interconverted, and are shown in Fig. 10. The multiline signal is indicative of magnetically coupled manganese ions, with at least 18-20 resolved lines being observed. It also indicates that at least two of the manganese are involved in a multinuclear cluster. Based on the expected manganese oxidation states, and more lines than the 12-16 traditionally observed for a MnlIIMnNdimer, a cluster of at least three manganese ions is probable (46).The second EPR signal at first appeared similar to that generated by an isolated MnN, leading to trinuclear/mononuclear proposals for the structure of the OEC. Later EPR evidence showed that this signal did not result from a monomeric MnIVion, but rather most likely arose from a multinuclear manganese structure itself (167, 168). In ammonia-treated, oriented PS I1 samples, a multiline signal was observed in the g = 4.1 EPR spectrum. Further-
'...,.,..,....(,...,...,,....,.... lo00
ZOO0
3000 Maptic Field (G)
....,' 40oo
FIG.10. The g = 4.1 EPR signal from the SS state of the OEC (left) and the g = 2 multiline signal (right). [Reproduced with permission from (46), Chapter 9. Copyright 1996 Kluwer Academic Publishers.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
333
more, another study on this signal suggested that it arose from an S = 512 ground state (1691,which would be inconsistent with a Mn" ion, but not a multinuclear array of manganese. The appearance of the multiline EPR signal upon going from S1 to 9, also supports the existence of metal-centered oxidation (160, 161). For a more detailed discussion on these two signals and the conditions affecting their respective formation, the reader is again directed to one of the various reviews and compilations (42, 46). In 1997, new EPR data on the So state was reported by Messinger and co-workers (170, 171). Until these reports, no EPR spectra had been recorded for this S state. The first set of experiments took darkadapted PS 11s to Soby the addition of reductants (171). This S state is labeled S$ to account for potential alterations to the manganese structure. A multiline signal EPR signal was observed for this S$ state that exhibits 24-26 lines centered near g = 2 [Fig. ll(a)l. This signal is wider than the S , multiline and exhibits more lines, many of which do not line up with transitions observed in the S , multiline. The authors suggest that this added breadth is a result of the lower overall oxidation state of the manganese ions composing Soand that this may be indicative of the presence of a Mn"Mn"' couple based on model complex data. A second set of experiments reported by Messinger, Sauer, Klein, and co-workers (170) took dark-adapted samples through one S-state cycle to poise them at So.This preparation also presented a multiline signal, which did not appear to differ significantly from that observed for S$ [Fig. ll(b)]. For both of these spectra, the presence of methanol (0.5-1.5%) in the sample was required to observe the multiline signals. Without methanol, a broad EPR signal is still observed for the So preparation. The authors theorize that methanol may be binding to the manganese cluster and that this facilitates the observed spectra, perhaps by altering the coupling of the manganese ions. Simulations conducted to ascertain what the possible oxidation states of the manganese ions in So might be have not yet clarified this point, because the overall oxidation states Mn"Mnin, Mn"Mn1I1MniV,or Mni*'MnIVall gave similar spectral widths consistent with the data. However, the simulations did suggest that neither binuclear species nor a tetranuclear cluster arranged as MnI1Mnw/ MnlIIMnIVwould not yield comparable spectral widths. Also in 1997, Styring and co-workers (172) reported a multiline signal for So.Their study showed that by administering light flashes to PS I1 samples, they could produce a multiline signal that they believed to arise from Soafter 3 flashes from S z samples. This spectrum is also wider than that observed for S2, similar to that reported earlier. A second se-
334
LAW, CAUDLE, AND PECORARO
A 3 2.8 2.6 2.4
g value 2
22
1.8
1.6
at
B I
I
.
250
I
I
,
l
.
300
,
.
.
,
.
350
Field (mT)
.
.
.
,
.
400
p=2
L
1900
2700
3500
4300
5100
Magnetic Field (Gauss)
FIG. 11. Sa and SoEPR signals. (a) EPR Spectrum of the Sg state. The top spectrum, A, represents the S-l state that was formed by reducing PS I1 samples with hydrazine. This was illuminated to yield a new multiline signal, C. B is the spectrum of the Sz state multiline from control experiments. Details of sample preparation may be found in the references provided. (b) A and B. Difference spectra of prepared So minus residual S1 signal. B. For comparison, the Sz multiline under the same conditions. Note that So is a broader signal and is shifted with respect to the Sz signal. C. Sospectrum without added methanol. D. Sg*Spectrum produced by reducing S1samples with hydroxylamine. E. Simulated spectrum of So. [(a) and (b) reproduced with permission from (171) and ( 170), respectively. Copyright 1997 the American Chemical Society, respectively.]
quence of light flashes regenerated the So signal. They also utilized methanol in their samples. EPR signals have also been reported for the S, state (173). A very intriguing article (174) has recently appeared in which Britt, Debus, and co-workers showed, via the application of parallel polarization EPR studies, that a multiline EPR signal can be observed for the S1 state (Fig. 12). When illuminated, the amplitude of this signal was “greatly reduced,” and a conventional perpendicular polarization g = 2 multiline signal consistent with the formation of the S, state was
335
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
(a) Parallcl Polarization
?P
I
I
zoo
I
ooo
I
I
1
1MM
1200
1
800 Magnetic Field (G)
600
(b) Perpendicular Polarizatlon
Illuminated minus Dark
9# 1
1
2s00
Moo
1
3500 Magnetic Field (G)
I
I
4000
4500
FIG.12. S, multiline signal. In (a) the spectrum (parallel polarization EPR) labeled Dark represents a dark-adapted PS I1 sample; thus, it is in the S, state. The second trace represents the sample following illumination with light to generate S2. The top, Dark, spectrum minus the middle, Illuminated, spectrum yields the third spectrum in (a). This represents a well-resolved multiline signal arising from a “multinuclear exchange coupled paramagnetic Mn cluster” (174) in the S, state of the OEC. If the subtraction is reversed and the spectra are recorded using perpendicular polarization, then when the Dark spectrum is subtracted from the Illuminated spectrum, an S2 multiline signal is readily observed (b). [Reproduced with permission from (174). Copyright 1998 the American Chemical Society.]
336
LAW,CAUDLE, AND PECORARO
observed. This S, multiline signal is consistent with the interaction of multinuclear clusters of manganese ions, and these authors proposed that it arises from a cluster of manganese ions. This proposal would minimally require that a trimeric cluster of coupled manganese ions exists, versus an arrangement invoking two distinct dimers, one a M$' dimer and the other a Mny, an arrangement based on the assumed overall integer spin of this S state. Furthermore, because Sl is expected to be integer spin, a trimer/monomer motif would suggest that the monomeric manganese should be in the Mn"' oxidation state. If that were true, then a six-line spectrum consistent with a Mn"' should be observable under the conditions employed to observe this new S1multiline signal. The lack of such a signal is suggestive that this new multiline signal actually arises from a tetranuclear cluster of manganese ions. Application of the X-ray Absorption Spectroscopy (XAS) to the observation of Mn K-edge data (24, 32,33, 156, 1751, via X-ray Absorption Near Edge Structure (XANES), has provided insight into the overall oxidation state of various S states. The inflection point of the edge shifts for the various oxidation states of manganese (24, 33). Studies of the OEC show changes in edge energies consistent with an increase in the overall oxidation state of the manganese cluster. Thus, at least some of the oxidations of the OEC occur on the tetranuclear cluster of manganese ions. Numerous XAS studies in conjunction with EPR data have lead to proposals of the oxidation levels of the OEC S states (24, 33, 42, 46). Two of these proposals are presented in Scheme 10. The S-state oxidation levels of the manganese are not s-state
so
s, s2
s, s.4
IIInIInIIIiv IrinIinvnv
riinvnvm rvnvnvnv tvnvnvN
IIIIIIIIIIIIIV IIInIInvm IIIIIVIIVIIV
IIIIIIIIVIIV
nmnvnv IVIIVIIVIIVY I. SCHEME 10.
clear-cut, and some controversy does remain ( 2 4 , 3 3 , 4 2 , 4 6 ) .Much of this controversy revolves around defining oxidation states for the S states from spectroscopic techniques. For example, the transitions from So to Sl and S1 to S2 have strong evidence for manganese-based oxidation state changes. However, the nature of the S2to S, transition is quite controversial because there is no clear-cut alteration to the observed XANES edge energies. Proposals to explain this have invoked oxidized protein residues or structural rearrangements of the
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
337
tetranuclear cluster (176). The imidazole oxidation that had been suggested, however, is unlikely. The signals that were attributed to the oxidized imidazole have now been shown to arise from Y; (177, 178). In the final S , to S4to So phase, little is certain due to the instability of the S, and the transient nature of the Sqstate. Again the oxidation state controversy is centered on whether oxidation occurs on a protein residue, such as Y;, or on a manganese of the OEC. The results of the XAS and EPR data, combined with knowledge from model complexes, (discussed later) allow for the formulation of structural models of the OEC. Many manganese compounds have been proposed as structural models for the OEC, ranging from butterfly arrangements to adamantanes to cubanes (26,27,29-34,50). Most of those proposals, however, are inconsistent with the known structural data for the OEC. Currently the most widely favored model is the “dimer of dimers” model (Scheme 111, first suggested by Klein,
SCHEME 11.
Sauer, and co-workers (32, 154). This model proposes that two manganese dimers that interact with one another exist in the OEC. The dimers are suggested to contain Mnz(pz-02-)2 cores and t o be bridged to one another by 0x0 and carboxylato units. This model is supported by several pieces of current data on the OEC. First, the 2.7-w Mn-Mn vectors of the OEC are most consistent with crystallographically characterized dimeric manganese model complexes that contain Mnz(pz-02-)z cores, and two such vectors are observed for the OEC. Second, such a model would provide the potential for magnetically coupled Mn-Mn interactions to occur. Bridging two Mn2(pz-OZIz cores ) would proto one another by carboxylates and a single ( p 2 - 0 2 -unit vide for an approximately 3.3-w distance, a formulation consistent with known model compounds (discussed later). Furthermore, the
338
LAW, CAUDLE, AND PECORARO
EPR data from So, S , , and S2indicate that there are coupled Mn ions. In particular, the multiline signals are suggestive of coupled manganese structures that require clusters of greater than two manganese ions. The related Electron Nuclear Double Resonance (ESEEM) and Electron Spin Echo Envelope Modulation (ENDOR) spectroscopies have also proven useful in elucidating information about protein residues near the OEC. One of these residues is the important protein radical Tyrosine Z, Y;, which is usually found in an oxidized radical form (42).Study of this protein residue via ENDOR and ESEEM spectroscopies has been enlightening. The presence and protein residue associated with this radical signal was defined several years ago, and Y; was originally assigned the role of an electron-transfer mediator for the electrons moving between the OEC and (150, 179). More recent work has been modifying this view. Tyrosine Z has been shown by Britt and co-workers to be located in close proximity to the manganese cluster, probably within 5 A (180).Probes of the nature of this tyrosyl radical by Babcock’s group have suggested that this tyrosine may play a more active role than solely to act in an electron-transfer capacity (181,182).It does not bear the hallmarks of a protein residue whose sole purpose is to participate in an electron-transfer chain-for example, it does not retain a nearly rigid orientation with respect to its surroundings. Other similarities between this redox-active tyrosine and other metalloenzyme systems that employ radical components suggested that this tyrosine might also act in such as role (47, 183). These observations have lead to recent proposals by Britt and more intimately in the water oxidation mechBabcock that involve Y;, anism (182, 184,185).Instead of merely acting as an electron transfer agent between the OEC and P&, it is now proposed to perform a hydrogen atom abstraction. New models for the function of the OEC have been formulated based on this proposal (discussed later). The role of Yz is still not settled, of course, as evidenced by a recent article arguing against a hydrogen atom abstraction role (186).Finally, spectroscopic studies have also focused on identifying potential protein residues that might be ligands to the OEC. Through ESEEM spectroscopy, the presence of at least one imidazole ligand to the OEC manganese cluster has been established (187). Studies involving calcium and chloride have shown these ions to be critical t o the proper functioning of the OEC (35, 146, 1881, although their exact roles have yet to be elucidated. Without either of these, dioxygen evolution is halted and the OEC cannot complete one full cycle. It has recently been reported, for example, that the OEC can reach Sz without chloride, but cannot be further oxidized to S , o r pro-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
339
ceed from S3 to S, without the addition of chloride (188). There has been speculation that the chloride’s role may be to prevent premature loss of oxidizing equivalents, perhaps by preventing the premature loss of oxidation equivalents in the formation of hydrogen peroxide (189).Another suggestion may relate chloride to substrate binding, an idea that has received some support from small-molecule-binding studies, which indicate that there is a site for a small molecule such as amnonia to bind to the OEC in a process that displaces chloride (35, 190). Calcium may be bound near the manganese and has been proposed to be the source for the Mn-X 3.3-A distance observed by EXAFS (191). One possible role for the calcium may be to act as a water-binding site from which water is transferred at some point to one of the four Mn ions of the OEC (192, 193). Various researchers have studied the effects of substituting for these critical ions. Only strontium has been successfully inserted in place of calcium with retention of any level of function, albeit severely attenuated (35, 194). Substitution for chloride is less sensitive. The addition of bromide does not appreciably alter the ability of the OEC to generate dioxygen (35,421. The OEC is also functional, but at reduced levels in the presence of iodide and nitrate. Fluoride strongly inhibits the OEC and has been suggested to alter the structure of the manganese cluster structure based on EXAFS data (195). Amines also replace the chloride and inhibit the function of the OEC (35, 190). Ammonia is a particularly interesting case in this regard, in that it appears to bind at two different sites within the OEC, suggesting that there may exist a second chloride-inaccessible site, which has been proposed to be a waterbinding site to a manganese (35, 196). Water is of course key to this whole process. Water-binding studies have been attempted to establish when the two water molecules that are oxidized to dioxygen become bound to the OEC and whether they are bound directly to manganese. Early water-binding studies suggested that the water converted to dioxygen is still exchangeable until at least S, . From isotopic labeling studies, it appears that the waters that are oxidized are not bound or highly restricted with respect to exchange until the late S states (197). A recent water-binding study suggested that there were two waters in exchange with the bulk water, but that these waters were exchanging at different rates (198). One proposal put forward to explain this observation was the existence of an Mn=O moiety. A review based on the thermodynamics of water oxidation was published in 1996 (37). Pulling the data together again, a larger picture of some to the structural and reactivity details of the OEC begins to emerge. The new hydrogen atom abstraction role proposed for Yz coupled with the
340
LAW, CAUDLE, AND PECORARO
dimer of dimers model has led to some intriguing new hypotheses for the potential manner in which the OEC functions (46, 47, 177, 178, 181,182,184,185).A synthesis of these new concepts is presented in Scheme 12 (182). If the OEC functions in the newly suggested manner, then it would join a growing class of enzymes that utilize protein a', H
-!-
I
7kl+ Y',
-7Y-',
't.1"
SCHEME 12. A mechanism for dioxygen evolution by the OEC as proposed by Babcock that employs the hydrogen atom abstraction concept. [Reproduced with permission from (182). Copyright 1998, the American Chemical Society.]
residue radicals to complete their catalytic processes (47, 183). The dimer of dimers structural model has been modified t o include hydrogen atom abstraction from a water bound to a manganese by Y ; , which is equivalent to proton loss being coupled to the stepwise oxidation of the OEC. This proposal eventually leads to the generation of a high-valent manganese (IV)or (V) with an 0x0 bound, Mn=O, that then combines with a second water-derived oxygen atom bound to a nearby manganese to form dioxygen during the S4-So step. Another feature of these new models is the retention of charge neutrality throughout the process. Scheme 13 shows the Babcock proposal as modified by our group. For any functional proposal to gain validity, one needs to consider whether or not the energetics of the process involved would be favorable. Recent studies by Pecoraro and co-workers suggest that the energies required for this process to occur are likely to be favorable (discussed later).
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
+'%
2 H,O+ CI-
t
341
\
e- ,H+ YZ'
/e-.H+
s2
SCHEME 13. A mechanism for dioxygen evolution by the OEC involving hydrogen atom abstraction a8 proposed by Pecoraro.
The OEC provides perhaps the most complicated manganese enzyme system yet known and thus is rich with possibilities for the exploration of manganese chemistry. Overall, the OEC and PS I1 are among the more intricate enzymatic systems that are currently being explored. Structural models, spectroscopic models, and reactivity model complexes are all integral to insights that have been attained and those yet to be gained. The following paragraphs detail some of the work in this area. Models related to water oxidation are presented in the Section V, "Reactivity." A selection of the newer multinuclear complexes will be discussed in Sections 1II.B to 1II.D. Again, many reviews have dealt with model complexes for the OEC (29-31, 34, 45, 50). Many bioinorganic chemists have prepared speculative structural models over the past several years to gain a n insight into this system-for example, with respect to possible arrangement of manganese ions-or to provide characterized materials as reference compounds for spectroscopic examinations. The agreement between EXAFS data and the Mn-Mn vector of a Mn&-O), core is one prime example of structural modeling chemistry and spectroscopy combining to elucidate critical data on this system (discussed earlier). Multinuclear clusters are always of interest in this field, and several have been prepared (28-31, 34, 45). The adamantane and cubane structures led to proposals for the possible oxidation of water to dioxygen
342
LAW, CAUDLE, AND PECORARO
that involved rearrangement of a manganese cubane to adamantane upon incorporation of water molecules, and another proposal invoked a butterfly structure that was converted to a cubane during the catalytic cycle (199, 200). These proposals, among others, have been extensively reviewed, and the reader is directed to one of the referenced sources noted earlier. New examples of dimers of alkyl-bridged Mn2(pn-O),cores have been prepared and structurally characterized in recent years (discussed later). A few of these also include waters bound to the manganese. One of the more interesting tetranuclear manganese structures in this class of compounds was reported in 1992 by Armstrong and co-workers, who then further probed its complicated magnetic couplings (201). A variety of other complexes have been prepared that have proven useful in the area of the spectroscopic understanding of the OEC. Many have been used as models for XAS studies to establish what is to be expected from the various oxidation states of Mn. Others have been probed to gain a better understanding of the interactions between manganese nuclei-for example, most Mn"'MnTVcomplexes that are known exhibit 16-line EPR signals (140). However, the Mn"' Mn'" complex prepared with the 2-OHsalpn ligand (202) only exhibits a 12-line spectrum. Another area that model complexes may help to address is the effects of water binding to a manganese cluster. Small molecules such as water can be bound t o the complexes derived from 2-OHsalpn, and these have been probed by ENDOR spectroscopy to ascertain what perturbations might arise from that bound molecule versus the complex without such an additional ligand (203). The energetics of water oxidation is another key point to address. This topic includes the interesting question of how the oxidizing equivalents are stored in the OEC without being released too soon, and how the requisite 3.6 V versus NHE for water oxidation can be built up when Y; is thought to maximally be able t o provide oxidations at a level of 1.1 V per oxidation (41).In addition, the proposal of hydrogen atom abstraction has led t o the exploration of how the homolytic bond dissociation energy of a water bound to a manganese dimer might be altered from the values observed for an individual water molecule in solution. To abstract a hydrogen atom from water would require more energy than is available to Y; (Table I). However, Pecoraro and co-workers have shown that the AG value for removing a proton from a protonated p-0x0 bridge is within a reasonable range for the Y; to be able to abstract a hydrogen atom (Table I) (204,205). These studies employ a concept worked out by Mayer for the oxidation of, for example, toluene by permanganate (206). Studies of water
343
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
TABLE I HYDROGEN BONDDrssOCiArION ENERGIES" Complex
HBDE (kcal/mol)
Reference
86 119 76 77 79 84
457 458 204 204 204 204,459,460
89 94 85 89 82 86
205 205 205 205 205 205
Tyrosine Water IMn'"(salpn)(p-0)I2 LMn'"(3,5-di-C1-salpn)(p-O)12 [MniV(3,5-di-(N02)salpn)(p-0)12 [Mniii'TV(bpy)(p-O)l~+ LMn$"(2-OH(3,5-diClsal )pn),(OHJl [Mn'i'MnTY(2-OH(3,5-d~Clsal)pnJ2(OHL)1' IMn~"(2-OHsalpn),(OH2)] IMn'1iMniV(2-OHsalpn)2(OH,)1 LMnYi(2-OH(3,5-di-t-Bu~alJpn)~(OH~)l
~Mn"iMn"'(2-OH(3,5-di-t-Busal)pn)2(OH2)l
For complete details, including complex pK,'s and electrochemical reduction potentials, the reader is directed to (204)and (205).
bound to complexes derived from the 2-OHsalpn ligand show that hydrogen atom abstraction from these complexes is also in a reasonable energy range, theoretically, for H-atom abstraction by Y; (205). In addition, the normally observed increase in reduction potential of this complex with an hydroxide bound is attenuated. This also suggests that if H-atom abstraction is occurring in the OEC, this might favorably affect the redox potential of the manganese dimer to which that water is bound, further stabilizing the system against early loss of its oxidizing equivalents before water oxidation can occur and thus allowing the potential for oxidation of the manganese cluster to remain in a range accessible to Y;. A report of quantum chemical calculations in 1997 supports these data (207). Other topics of importance to the OEC will be discussed in greater detail later. Water oxidation models will be addressed in Section V, "Reactivity," and models for the catalase-like reaction of the OEC will be addressed in Section V.B.3., "Catalase."
111. Structural Models
Manganese redox enzymes exhibit a range of structural motifs involving mononuclear or bi- and tetranuclear aggregates of manganese ions. There has been a great deal of research in the area of structur-
344
LAW,CAUDLE, AND PECORARO
ally modeling the active sites of manganese enzymes such as the MnSOD, catalase, and the OEC. These small-active-site analogs have been prepared in the hope that structural mimics will also mimic (i) the spectroscopy and (ii) the reactivity of the biological center. Some success has been achieved in modeling the reactivity of catalase and SOD with these models, but there is no structural analog for the OEC that exhibits water-oxidizing activity. The following is an overview of additions in the field of manganese chemistry since the superior review by Wieghardt in 1989 (30).This section will provide a compendium of the structure types that have appeared in the literature over the past eight years but is not meant to be an exhaustive recapitulation of every structure that has been generated during that time.
A. MONONUCLEAR 1. M n (II)
Mononuclear complexes of manganese are quite common and have been prepared in all of the currently recognized biologically relevant oxidation states of manganese, II-V. Mononuclear Mn" complexes exhibit a range of polyhedra about the metal, with coordination numbers of 5-8 dominating. The lack of coordination number specificity is understandable, as this d6 ion does not, in theory, exhibit specific geometric preferences. Numerous complexes have been prepared with predominantly N,O,-donor atom sets, which are often augmented by halide. or pseudohalide ligands and the occasional sulfur donor. In general, this represents the known biologically relevant donor atoms to manganese. Several of these complexes have been generated with the target of mimicking or modeling MnSOD. A variety of five-coordinate complexes of Mn" have been prepared. Crystallographically characterized Mn complexes with a trigonal bipyramidal geometry are still relatively rare, however. Kitajima and co-workers (208) reported two trigonal bipyramidal complexes utilizing hydrotris(3,5-diisopropylpyrazo1-l-yl)borate,HB-i-Pr, a facially binding tridentate ligand (209,210)(Fig. 13).The coordination sphere of the first complex was completed by a bidenate benzoate; the other complex contains a monodentate benzoate and one equivalent of 3,5diisopropylpyrazole. Another variation that has recently appeared is composed of tbp complexes of Mn" with tris(benzimidazoy1-2-methyl)amine bound t o the manganese (211,212) (Fig. 14).The tridbenzimidazoyl-2-methyllamine is a tetradentate tripodal amine ligand that provides four N-donor atoms. The first coordination sphere of these complexes was completed by a chloride. However, these products are
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
345
FIG. 13. [Mn~"~hydrotris(3,5-diisopropylpyrazo1-l-yl)borate)(3,5-diisopropylpyrazole~ (benzoate)] perchlorate. [Reproduced with permission from (208). Copyright 1993 the American Chemical Society.1
not purely five-coordinate compounds. The crystal structure of each bears both a five-coordinate form with one C1- bound to the central Mn and one C1- counterion, and a six-coordinate complex with two chlorides bound to the central Mn (Fig. 14). Another five-coordinate complex appeared in 1990 (213) with a single bis(benzyimidazolyl-1methy1)amine ligand and two acetate molecules (Fig. 15). The amine ligand binds meridionally, with the two acetate ions completing the equatorial plane of the structure. A few Mn" complexes with square pyramidal geometry were also reported. In one case, the basal plane is occupied by two dihydrobis(pyrazo1-l-y1)borate ligands with an axial chloride (Fig. 16) (214).In another, the axial position is occupied by an axial THF molecule with two THF molecules and two phenoxy donor groups in the basal plane (215). The greatest number of complexes characterized are those that are six coordinate. One general structure type involves multidentate li-
FIG.14. Crystal structure (left) of [MnYtris(benzimidazoyl-2-methyl)amino)C1It, the trigonal bipyramidal complex. Crystal structure (right) of the neutral six-coordinate [Mn"(tris(benzimidazoyl-2-methyl)amine)C121. [Reproduced with permission from (211) and (2121, respectively. Copyright 1997 the American Chemical Society and copyright 1995 Elsevier Science, respectively.]
346
FIG. mission
LAW, CAUDLE, AND PECORARO
with per-
gands that fill four of the six available coordination sites, allowing monodentate ligands such as halides or water to occupy the two remaining positions in the first coordination sphere. In complexes that involve bidentate ligands such as phenanthroline, bipyridine, or acetylacetone, which often do not bind opposite to each other in the “equatorial” plane of these compounds, a structural type in which the two remaining coordination sites are cis to one another is promoted. One example of these bis bidentate chelate complexes is the complex (ditriflatolbis(bipyridin0) Mn” (216) and bis (phenanthrolino)(dithiocyanato) Mn” (Fig. 17) (217). One of the more unique examples in this group involves a fused-ring system, which provides a tetradentate N4donor ligand that occupies three coordination sites on one face and one addition site on the other. The two remaining cis positions were filled by bromides and chlorides, ultimately leading to distorted octahedral structures (Fig. 18) (218).This group also includes complexes derived from the tris(benzimidazoy1-2-methy1)amine ligand mentioned earlier (Fig. 14) (211,212). When the chelating ligands are able to occupy the “equatorial plane” of a Mn” complex at the same time, the monodentate ligands
FIG. 16. Square pyramidal [Mnl’(dihydorobis(pyrazol-lyl)borate)C1]’. [Reproduced with permission from (214).Copyright 1990 the Royal Society of Chemistry.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
347
S FIG. 17. IMn"(phenanthr~line),(SCN)~].[Reproduced with permission from (217). Copyright 1993 International Union of Crystallography.]
will bind to the central Mn" ion on the remaining two trans "axial" positions. Variations on this theme include charged chelating ligands with a pair of neutral ligands axial or equatorial neutral chelates, such as a tetraazamacrocyle, for example, with charged axial ligands, such as halide ions (Fig. 19) (219). Yet another variation found for these complexes is one in which all of the ligands about the central Mn" are neutral, for example the complex [bis(bis(2-pyridylmethyl) amino)MnI1l2+(Fig. 19) (220). A less common but rapidly growing family of Mn" complexes are those that are seven coordinate. These complexes may be built up from a variety of ligand systems that range from multiple ligands to a single septadentate ligand. Some of these systems have been tested as MnSOD mimics. Most of the seven-coordinate complexes that were structurally characterized in recent years have adopted a pentagonal bipyramidal geometry. One motif is based on a pentadentate macrocylic ligand with two unidentate ligands in the axial positions. One particular family of pentagonal bipyramidal complexes was prepared by Riley and co-workers from the macrocyclic ligand 1,4,7,10,13-pen-
FIG. 18. Mn"(2,5,8,10,13,16-hexaazapentacyclo~8.6.l,1'~5.09~'H.O'"~'7]o~tadecane)Cl~. [Reproduced with permission from (218). Copyright 1989 Verlag der Zeitschriften fur Naturforschung.]
348
LAW, CAUDLE, AND PECORARO
FIG. 19. [Mn”(l,4,7,10-tetraazadodecane)C12], left, and [Mn”(bis(2-pyridylmethyl) amine)z]2+, right. [Reproduced with permission from (219)and (2201,respectively. Copyright 1996 International Union of Crystallography and Copyright 1992 American Chemical Society, respectively.]
FIG. 20. Two seven-coordinate Mn” complexes that have been tested as MnSOD mimics. [Mn11(1,4,7,10,13-pentaazacyclopentadecane)Cl~] (left) and [Mn”(trans-2,3eyclohexano-1,4,7,10,13-pentaazacyclopentadecane)Clz](right). [Reproduced with per mission from (221) and (222). Copyright 1994 and 1996 the American Chemical Society, respectively.]
0
0
0
FIG.21. [Mn”(tris(ethan~l)arnine)~]. [Reproduced with permission from (217). Copyright 1993 Verlag der Zeitschriften fur Naturforschung.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
349
N
FIG.22. [Mn"(tris(benzimidazoyl-2-methyl)amino)(N0,)Z1. [Reproduced with permission from (211).Copyright 1997 the American Chemical Society.]
taazacyclopentadecane (221, 222) (Fig. 20). The first coordination sphere in these complexes is completed by two axial chlorides. This work was specifically focused on the development of MnSOD mimics, and these workers have generated a wide range of complexes by preparing a series of ligands in which the macrocyclic ligand was modified by additions to the ring's carbon skeleton (222). Another example involves a modified 15-C-5 ligand with axial di-t-butylnaphthylsulfonates (223),and a related complex has been prepared with two axial trifluoromethylsulfonates (224). Pentagonal bipyramidal complexes do not, however, require macrocyles to form. Other pentadentate but open-ended ligands have also proved to be well suited for the generation of seven-coordinate complexes (225-227). Figure 21 shows an example of a complex with an MnLz motif, wherein one tris(ethano1) amine ligand binds as a tetradentate ligand, while the other binds as a tridentate ligand with one unbound ethanol arm. Another example of CN 7 is that of the tetradentate ligand tris(benzimidazoy1-2-methyl) amine and Mn" (Fig. 22) (211), wherein the other three binding sites are occupied by one monodentate nitrite and one bidentate nitrite molecule. A final example is provided by the complex of Mn" with quaterypyridine, one water, and a bidentate acetate (Fig. 23) (226).
FIG. 23. [Mn"(quaterpyridine)(OAc)(H,O)l+.[Reproduced with permission from (226).Copyright 1993 Elsevier Science.]
350
LAW, CAUDLE, AND PECORARO
FIG.24. [Mn~'(1,4,7,10,13,16,19-septaazacyclouneicosane)l~+. [Reproduced with permission from (228). Copyright 1990 the American Chemical Society.]
Two seven-coordinate complexes were attained by utilizing single septadentate ligands. The first involved 1,4,7,10,13,16,19-~eptaazacyclouneicosane, but no specific geometric orientation about the Mn" was reported (Fig. 24) (228). The second structure is of a monocapped trigonal antiprism, prepared by using the septadentate ligand formed by Schiff-base condensation from tris(2-aminoethy1)amine and 2-pyridinecarboxaldehyde t o give tris(2-aminoethy1)amine-pendant pyridyl-2-methyl functional groups (Fig. 25) (229). This complex was tested for MnSOD activity (discussed later). A final structural motif for Mn" is represented by a few eight coordinate structures. One of these was a product of tetraazacyclododecane with pyrazolylyl-l-methyl groups appended to the ring nitrogens
FIG.25. [Mn"(tpaa)12+.[Reproduced with permission from (229). Copyright 1996 the American Chemical Society.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
351
FIG.26. [Mni'~1,4,7,10-(l-pyrazolylmethyl~tetraazacyclododecane)]2+ (left) and [Mn" (tris(2-pyridylmethyl)amine)12+(right). [Reproduced with permission from (230) and (232).Copyright 1992 the Royal Society of Chemistry and copyright 1993 Elsevier Science, respectively.]
(Fig. 26) (230).This led to a Mn" complex having a square prismatic geometry. The second complex is composed of a cryptate ligand that yields a complex with a cubic symmetry (Fig. 27) (231).The third structure is built up from two tetradentate tripodal amine ligands about the Mn" ions (Fig. 26) (232). One final note is of interest. Most of the structurally characterized and otherwise isolated complexes of Mn" tend to be high spin. The structure of at least one low-spin complex was reported during the past few years. This is a six-coordinate complex composed of two methyl-4,5-dihydro-5-{[imino(methylthio)methyllazo}-3,5-dimethyl1H-pyrazole-1-carboximidothioatoligands, which provide an N, first coordination sphere via one aza-nitrogen and two carboximido-nitrogens (Fig. 28) (233).
FIG.27. [Mnii([2.2.2]cryptand)]'+ [2.2.2]cryptand = 4,7,13,16,21,24-hexaoxa-l,lO-diazabicyclo[8.8.8]hexaxosane. [Reproduced with permission from (231).Copyright 1992 Wiley-VCH Verlag.]
352
LAW, CAUDLE, AND PECOFURO
I
FIG.28. Low-spin [Mn"~methyl-4,5-dihydro-5-{~imino~methylt~o~methyllazo}-3,5-dimethyl-lH-pyrazole-l-carboximidothioate)1. [Reproduced with permission from (233). Copyright 1993 the Royal Chemical Society.] 2. Mn(III)
Five- and six-coordinate structures dominate the crystallographically characterized structures reported for Mn"'. Here the pseudoJahn-Teller axis of this d4ion plays a role in the overall structure of the Mn"' coordination complex, so these complexes tend to adopt either a square pyramidal or octahedral geometry. Most of the five-coordinate complexes that have been structurally characterized are found to adopt a square pyramidal geometry. A prime example of such products are the myriad of Mn"' complexes that utilize a tetradentate Schiff-base ligand (234),for example H2-Salpn or H2Salen,with a fifth donor in the axial position t o complete a square pyramidal structure. Such structures are also observed in the Mn"' porphyrin complexes with a single axial ligand (470). This structure ably supports the pseudo-Jahn-Teller axis along the axial, or z, axis. Specific examples of this group of complexes are the complexes utilized for exploring epoxidation chemistry (235).One of these complexes is shown in Fig. 29 (236),with a structure similar to the other members of this type of complex (237).For this type of complex, the Mn'" ion tends to be located out of the equatorial plane toward the axial ligand. Another area of Mn"' complexes are those that utilize a variety of porphyrins as ligands. Five-coordinate complexes were not limited to the square pyramidal geometry. At least two reports of structurally characterized trigonal bipyramidal complexes have appeared in the literature in recent years. The first is of a complex in which the Mn"' is surrounded by an 0, first coordination sphere composed of two bidentate 2,2'-biphenoxide ligands and one monodentate 2,2'-biphenoxide to create a distorted trigonal bipyramidal geometry (Fig. 30) (238).The
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
353
FIG. 29. {Mn1''[3,5-~di-t-buty1~Sa1icylidenimino-~~~)-(1,2-diaminocyclohexane)l}. [Reproduced with permission from (236).Copyright 1996 Wiley-VCH Verlag.]
second is a very intriguing structure with a tegradentate tripodal amine ligand (239).Three of the five donor atoms are nitrogens from amide functional groups (Fig. 31). The unique feature of this complex is the hydroxide bound to the Mn"' ion. In general, structurally characterized complexes of manganese with hydroxide tend to be rare. Six-coordinate complexes of Mn"' are also quite common. Many of these, again, were prepared with tetradentate Schiff-base ligands that occupy the equatorial plane of the complex while two other ligands bind to the central manganese ion on the remaining trans axial positions (234, 240). Structures with Mn"' were not, of course, exclusively prepared with tetradentate Schiff bases or similar ligands, however (241). A unique six-coordinate structure was reported in 1992 for a MnIII-OH complex, another rare example of a structurally characterized Md'I-OH complex (Fig. 31)(242).
FIG. 30. [Mni"(2,2'-bisphenoxide)(2,2'-bisphenoxideH)l.[Reproduced with permission from (238).Copyright 1991 the American Chemical Society.]
354
LAW,CAUDLE, AND PECORARO
od FIG. 31. Rare examples of Mn"'-OH-complexes: [Mn"'(tris(cyclopropylcarbanoylmethyl)amine)OH]- (left) and [Mn'11(bis(2-hydroxy-5-nitrobenzoiminopropyl)methylamine)OH]. [Reproduced with permission from (239) and (242). Copyright 1997 and 1992 the Royal Society of Chemistry, respectively.]
The structure of one rather key six-coordinate Mn"' complex in this area was reported by Kitajima and co-workers in 1994. This complex is composed of the hydrotris(3,5-diisopropylpyrazol-l-yl)borateligand, one monodentate 3,5-diisopropylpyrazole,and a side-on-bound peroxide molecule (Fig. 32) (243).This structure is important with respect to the question of the "dead-end" complex that is proposed to occur during turnover in MnSOD. It is also a rare example of a structurally characterized Mn-peroxide moiety. One of the key features of this system is the two differing structures that were isolated. In the second structure (Fig. 32), the peroxide is presumed to be hydrogen bonding to the extra pyrozole ligand, based on the shortened interaction between the hydrogen on the nitrogen of the pyrazole ring and the oxygens of the peroxide. 3. M n ( W
A less common but growing class of complexes are the mononuclear complexes of MnW.The structurally characterized MnW,d3,complexes tend to adopt six-coordinate structures that are octahedral in nature. Typically, these MnWstructures require charged ligands to support the higher oxidation states. This is also true of the Mn" complexes. In most of these complexes, phenolato, alkoxo, or amido donors have been employed to stabilize these higher oxidation states. Among these are the complexes of MnTVsalpn(acac)(PF6)(244) (Fig. 33), which is octahedral with the salpn2- ligand in a cis-beta conformation, and a MnW(salpn)C12 (245)complex with an equatorial salpn2- and two axial chlorides (Fig. 33). An example of an O6 first coordination sphere is
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
355
FIG. 32. [Mn~"(hydrotris(3,5-diisopropylpyrazol-l-yl)borate)(3,5-diisopropylpyrazole~ peroxide]+ Two different structures were grown at low temperatures. In one structure (bottom) there is a hydrogen-bonding interaction between the 3,5-diisopropylpyrazole ligand and the peroxide moiety. The hydrogen-bonding interactions are indicated in the expanded figures to the right. This is a n important complex with respect to modeling manganese redox enzymes. [Reproduced with permission from (243). Copyright 1994 the American Chemical Society.]
356
LAW,CAUDLE, AND PECORARO
FIG. 33. [MnTY(salpn)(acac)](PF6) (left) and MnN(salpn)C12(right). [Reproduced with permission from (244)and (245).Copyright 1991 the American Chemical Society and copyright 1995 the Royal Society of Chemistry, respectively.]
provided by the Mnw complexes with a-hydroxy acids that were prepared to model manganese peroxidases (Fig. 34) (246).A unique hexadentate ligand, based on octamethyltetraamine with salicylic acid moieties appended to the terminal nitrogens to form amide functionalities, produces the mononuclear Mn'" complex shown in Fig. 35 (247).
The macrocycle 1,4,7-triazacyclononane (2481,tacn, has been a popular ligand in the manganese community in recent years. Two interesting complexes based on this ligand with N303first coordination spheres have been reported. One Mn" example is the MnTVtacn (OCH3I3PF6 complex (249),which has been suggested as able to epoxidize olefins. This complex adopts a structure with a facial array of methoxides due to the nature of the other, facially binding tacn-based ligand. The second is one in which the alkoxides are appended by alkyl arms to the tacn ligand, similar to the structure in Fig. 36. That structure is represented by the complex with an N6 first coordination
FIG. 34. [MnN(2-hydroxyethylbutyric acid)J. [Reproduced with permission from (246).Copyright 1991 the American Chemical Society.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
357
FIG.35. ~Mn'V~l,l0-bis(salicylamido)hexamethylenetetraamine~]. [Reproduced with permission from (247).Copyright 1992 the American Chemical Society.]
sphere and the triply deprotonated ligand 1,4,7-tris(o-aminobenzyl)1,4,7-triazacyclononane (Fig. 36) (250). 4. MnW)
Structurally characterized Mn" complexes are rarer still. The new examples of mononuclear manganese complexes in this oxidation
FIG. 36. [Mn'v(1,4,7-tris(o-aminobenzyl~-1,4,7-triazacyclononane~l+. [Reproduced with permission from (250).Copyright 1995 the American Chemical Society.]
358
LAW, CAUDLE, AND PECORARO
FIG.37. [MnV~N,N'-bis~salicylideneimino~-2,4-dimethyl-2,4-butanediamine~~N)]. [Reproduced with permission from (251).Copyright 1996 the American Chemical Society.]
state have been of nitrodo, N3-, and Mn-oxo derivatives. Most of these adopt a square pyramidal geometry with the organic ligand occupying the axial plane while the 0x0 or nitrido group occupies an axial site (251-253). A manganese-nitrido complex with this orientation appeared with the ligand sale+ and the same ligand with all of the ethylene backbone's protons replaced by methyl groups, saltme+ (Fig. 37) (251). Both the MnV=O group and the MnV=N groups were first characterized in porphyrin systems, but only the MnV=N has been structurally characterized (Fig. 38) (254). More recently, two MnV=O complexes have been crystallographically characterized (255-257), for example the complex shown in Fig. 39. A unique feature of all of these MnVcomplexes is the short Mn-nitrido and Mn-0 bonds. The Mn-N bonds are on the order of 1.50-1.54 A, and the Mn-oxo bonds have been determined to be about 1.54-1.55 A. Most of these complexes are square pyramidal, but some variation is now known. A recent report of a reputed trigonal bipyramidal Mnv-nitrido complex composed of two bidentate ligands has appeared (2531,and a six-coordinate MnVcomplex was produced by Wieghardt and co-workers in 1996 (258) (Fig. 40) that employs a trimethyl tacn ligand, one acetylacetonate ligand, and a terminal nitrido. The Mn-N (nitrido) distance in this complex is 1.518 A. One nitrogen donor atom from the tacn ligand occupies the axial position opposite t o the nitrido at
FIG. 38. [MnV(OEP)(N)].[Reproduced with permission from (254). Copyright 1983 the American Chemical Society.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
359
I
FIG.39. [MnV(LKO)I-.[Reproduced with permission from (256).Copyright 1990 the American Chemical Society.]
an elongated distance of 2.301 8 compared to the other two N donors from this ligand at 2.055 8 and 2.073 8. The structures of both of these types of Mn" complexes, Mn"=O and Mn"-N, are important structural models for epoxidation and aziridination reagents, and for the OEC with respect to the possible formation of a Mn"=O during turnover.
B. BINUCLEAR The majority of model compounds synthesized for studying the biomimetic chemistry of manganese are binuclear complexes, due to the
FIG.40. [MnV(Me3tacn)(acac)N]. [Reproduced with permission from (258). Copyright 1996 the American Chemical Society.]
360
LAW, CAUDLE, AND PECORARO
relative ease and reproducibility of their preparation when compared with complexes of higher nuclearity. There are a number of binuclear manganese enzymes for which binuclear complexes can function as direct structural models (discussed earlier). Furthermore, higher nuclearity clusters in model chemistry and biological chemistry can often be described as assemblies of well-characterized binuclear units. Within the myriad of binuclear manganese complexes synthesized to date, a few core structural motifs appear with regularity. Each core is distinguished by a characteristic Mn-Mn distance that is relatively invariant with changes in the supporting ligands. As such, the Mn-Mn distance, where it is known, is often used as evidence of a particular core structure when a full crystal structure is not available. This is often the case in the biological systems themselves, which have been studied by X-ray absorption spectroscopy and from which Mn-Mn distances are determined even in cases where no crystal structure is available. Table I1 lists the binuclear manganese complexes whose cores have been structurally characterized along with their respective Mn-Mn separations. Although a multitude of dimanganese complexes have been prepared within each core structure, we will restrict the discussion in this section to those that have been structurally characterized by X-ray crystallography or X-ray absorption spectroscopy. TABLE I1 GENERAL TRENDSIN MN-MN DISTANCES OF SOME COMMON BRIDGING MOTIFS Bridging motif
Oxidation state Mn-Mn Distance (8) 2.8 2.6-2.7 2.6-2.7 3.2-3.3 3.1-3.4 3.1-3.2 3.25 3.6 3.3 3.1-3.6 2.6 3.4-3.5 3.5-3.6
References are given in text or are in (280).
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
36 1
The simplest binuclear structure possesses two manganese ions bridged by a single 0x0 or hydroxo ligand. Because dioxo- and dihydroxo-bridged manganese dimers tend to be thermodynamically stable, the mono-oxo and monohydroxo complexes are isolated and characterized only with sterically demanding ligands that do not provide two labile manganese coordination sites in a cis orientation. Only three complexes with biologically relevant ligands have been structurally characterized, these have two manganese bridged by a single hydroxide. The only nonporphyrin manganese dimer bridged by a single hydroxo linkage is the binuclear complex of the macrocyclic ligand tpictn (Fig. 41) (259).The Mn-Mn separation in this highly supported structure is 3.6 8, with a Mn-O-Mn angle of 126.8'. Mn2(tpictn) (p-OH) is a special case, however, in that the Mn,(p-OH) core is supported by a macrocyclic ligand that spans both Mn" ions. The only unsupported complexes having a single hydroxide bridge are the Mn"' porphyrin complexes [Mn"'(OEP)],(p-OH) (260) and [Mn*1*(TPP)12 (p-OH) (261) (Fig. 42). These complexes exhibit Mn-O-Mn angles of 152.7' and 160.4, respectively, consistent with the bridging hydroxo group. The more linear geometry of the TPP derivative is probably due to steric demands of the pendant phenyl groups, resulting in a longer Mn-Mn distance, 3.993 8, for this complex than for the OEP derivative. [Mn"'(5-NOzsaldien)lz(p-O) (262) has a single oxide ligand bridging two Mn"' ions and has a Mn-Mn distance of 3.490 8 (Fig.
Fic. 41. [Mnb'(tpictn)(p-OH)l. [Reproduced with permission from (259). Copyright 1995 the Royal Society of Chemistry.1
362
LAW, CAUDLE, AND PECORARO
Q FIG.42. [Mn"'(OEP)],(p-OH)+ (left) and [Mn"(TPP)N3lZ(p-O)(right). [Reproduced with permission from (261)and (265).Copyright 1996 the American Chemical Society and copyright 1981 the Royal Society of Chemistry, respectively.]
43). The hydrotris[(3,5-diisopropyl)pyrazolyllborate derivative [Mn"' (HB(3,5-iPr2pz)2(3-iPr0-5-iPrpz))12(p-O) (263) (Fig. 43),has a similar Mn-Mn distance, 3.53 A, and a more linear Mn-0-Mn geometry explained by steric demands of the hindered HB(3,5-iPrzpz)2(3-iPrO-5iPrpz) ligand. The MnTV complex with tetraphenyl porphyrin forms an 0x0-bridged complex, [MnW(TPP)N3l2(p-O), with a nearly linear Mn0-Mn angle (264,265) (Fig. 42). Analogs of p-OH-bridged complexes having a single p-alkoxo ligand are known as well. In fact, when analogous p-OH and p-OR bridged complexes are structurally characterized, the structural parameters are often quite similar. At this time the only structurally characterized complexes having a single p-OR ligand involve the highly supported 2-OHsalpn ligand motif (266-268) For example, [Mn1Y2-OH-
FIG. 43. [Mn1'1(5-N02saldien)]2(p-O)and [Mn"'(HB(3,5-iPr2pz)2(3-iPrO-5-iPrpz))lz (p-0).[Reproduced with permission from (262)and (271).Copyright 1990 and 1991 the
American Chemical Society, respectively.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
363
FIG. 44. [Mn'"(2-OH-(5-Clsal)pn)],(CH30H) and [Mn"'Mnw(2-OH-(3,5-(Cl)zsal)pn)2
(THF)]'. [Reproduced with permission from (270).Copyright 1992 the American Chemical Society.]
(5-Clsal)pn)lZ(CH30H)(Fig. 44) has a Mn-Mn distance of 3.808' and a Mn-O-Mn angle of 129" (269), within the range expected for Mn"12(p-OH) complexes (discussed earlier). The analogous mixedvalent [Mn"'Mn" (2-OH-(3,5-(Cl)zsal~pn)2(THF)lt complex also has a
364
LAW, CAUDLE, AND PECORARO
single alkoxide bridge (270) but with a shorter Mn-Mn distance of 3.65 8, and a Mn-O-Mn angle of 126.7"(Fig. 44). Bis-p-oxo dimanganese complexes are the most prevalent structurally characterized dimanganese motif in the literature. A smaller number of bis-p-hydroxo- and bis-p-alkoxo-bridged analogs are also known. Because of the prevalence of similar structures differing only in the supporting ligands, binuclear manganese bis-p-0x0, bis-p-hydroxo, and bis-p-alkoxo complexes are most conveniently considered as discrete manganese cores in which the supporting ligands play a stabilizing role. This view is supported by the remarkable similarity in core structure among complexes having very different ligands. We will begin by discussing lower valent states and continue with increasing oxidation state of the manganese in the binuclear cluster. [Mn11z(HB(3,5-iPrzpz)3)]2(0H)z, with a Mn-Mn distance of 3.314A, is to date the only crystallographically characterized binuclear manganese complex with two hydroxo bridges (271) (Fig. 45). When the bis-p-oxo complex this complex is oxidized by Mn04- or 02, [Mn"'tHB(3,5-iPrzpz>3)l~~O)z is formed (Fig. 45).This oxidized complex has a much shorter Mn-Mn distance of 2.696 8, between two fivecoordinate manganese"' ions. The six-coordinate complex [Mn"'(bispic-
b FIG.45. [Mn~'(HB(3,5-iPr,p~)a)l~(OH)~ (top) and [Mn'1'(HB(3,5-iPrzpz)3)12(0)2 (bottom). [Reproduced with permission from (271).Copyright 1991 the American Chemical Society.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
365
FIG.46. [MnT"(bispicMepen)l~(p-Oj. [Reproduced with permission from (272). Copyright 1994 the American Chemical Society.]
Me2en)12(p-O) (Fig. 461, has nearly the same Mn-Mn distance, 2.699 elongation axis common in Mn"' ions does not include a bridging ligand (272). No bis-p-hydroxo dimanganese(II1) cores have been structurally characterized, but the analogous bis-p-methoxo dimanganese(II1) complex [Mn"'(salpn)], (p-OCH,), (Fig. 47) has been prepared and exhibits a 3.192 A Mn-Mn distance (244). The longer Mn-Mn distance for the alkoxide-bridged complex reflects the Jahn-Teller elongation axis on each of the symmetry-related manganese ions, which includes a bridging alkoxide ligand. The supported [Mn"'(2-OHsalpn)12(Fig. 48) has two alkoxide bridges as well, but with slightly longer Mn-Mn separation due to the highly supported ligand structure (273,274). Mixed-valent dimanganese(II1,IV) complexes with the Mn"'Mn'" (~-0 core )~ have structural parameters very similar to the corresponding MII?(~-O)~ complexes, with Mn-Mn distances in the 2.7-A range. In spite of the similarity to homovalent cores, the prototypical [Mn"' (276) comMn1v(bpy)4(p-0),13+ (275) and [Mn"1Mn'V(phen),(p-O)213+
A, because the Jahn-Teller
FIG.47. IMn"i(3,5-diC1-salpn)(p-OCH,)I,. [Reproduced with permission from (244 j. Copyright 1991 American Chemical Society.]
366
LAW, CAUDLE, AND PECORARO
FIG.48. [Mn"'(2-OH(5-NOz)salpn)]z. [Reproducedwith permission from (273).Copyright 1993 the American Chemical Society.]
plexes are valence localized in the solid state, with one manganese exhibiting the axial elongation associated with isolated Mn"' centers and the other having more octahedral geometry consistent with Mnrv complexes exhibit valence localization (Fig. 49). Other Mn111Mn1V(p-0)2 as well (272,277,278). The only example of a mixed-valent Mn1I1MnW (2791, which bis-p-alkoxo-bridged complex is [Mn111MnTV(2-OHsalpn)zl+
FIG. 49. [Mnr1LMnw(bpy)4(p-O)2]3t. [Reproduced with permission from (275). Copyright 1995 the American Chemical Society.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
367
FIG. 50. [Mn'V(salpn)]z(p-O)z.[Reproduced with permission from (244). Copyright 1991 the American Chemical Society.1
has a Mn-Mn separation of 3.3 A, substantially longer than for bisp-0x0 complexes in the same oxidation state. The higher valent Mnr(p-O), complexes have been prepared with many supportin ligands, but all show the characteristic Mn-Mn distance of 2.7 (34).The prototypical complex [Mn'Ysalpn)lz(p-O)~ is shown in Fig. 50 (280, 281). Each successive protonation of [Mn [MnTV(salpn)l ,(p-O)(p-OH)' and [MnW(salpn)l, ( ~ a l p n ) ] ~ ( p - to O )give ~ (p-OH),t was shown by X-ray absorption spectroscopy to result in an increase in the Mn-Mn distance of 0.10 A (282, 283). The structural effect of protonation on the Mniv(p-O)zcore is considerably less pronounced than alkylation of the corresponding MnY1(p-0)2complexes due to rearrangement of the Jahn-Teller axis in the Mn"' case. A new structure type for Mnbv(p-O)zSchiff-base complexes was reported for the complex [MnlV(salen)]z(p-O)z ligand (234, 284,285). In these complexes (Fig. 51), the salen ligand bridges the two manganese ions.
1
FIG.51. [Mn1v(salen)]2(~-0)2. [Reproduced with permission from (284). Copyright 1998 Elsevier Science.]
368
LAW, CAUDLE, AND PECORARO
FIG.52. ([Mn'"(tacn)ll(~-O)~(~-Ol)} (left) and [Mn'"(ta~n)l&-O)~ (right). [Reproduced with permission from (286)and (287). Copyright 1988 and 1990 the American Chemical Society, respectively.]
There are two unique examples of triply bridged Mny dimers that deserve comment at this point. The first is the [MnTV(tacn)lz(p-O), com lex with three bridging 0x0 ligands and a Mn-Mn distance of 2.3 (Fig. 52) the shortest Mn-Mn distance of any binuclear complex t o date (286). The second is {[Mn'v(tacn)lz(p-O)~~p-Oz)} (Fig. 52), which is the only example of a binuclear Mn complex having a p-1,2 bridging peroxide ligand and a Mn-Mn distance of 2.531 A (287). These two complexes may be considered as MniYp-O), cores with either an additional p-0x0 or the p-1,2-peroxo added as a third ligand. Because of the additional ligand, the MniV(p-O),core is puckered in contrast to simple MnY(p-O):!cores. The importance of the latter structure is that it may be the best model to date for the initial manganeseoxygen species formed immediately upon 0-0 bond formation in the critical S4-So transition in photosynthetic oxygen evolution. There is a large and important class of binuclear manganese complexes having at least one p-carboxylate bridge in conjunction with one or two p-0x0 bridges. This class of p-oxo-p-carboxylato complexes is important not just in the context of binuclear manganese enzymes such as catalase [which has recently been shown to contain such a core (12511, but as analogs for iron oxo-carboxylato complexes, which are very important in the chemistry of nonheme iron redox enzymes. The simplest example in this class of manganese compounds is represented by the p-oxo-p-carboxylato system [Mn"'(bispicen)lz(pU-O)(p~,~OAc) with a Mn-Mn separation of 3.28 A (288) (Fig. 53). This core motif has been structurally characterized only in the dimanganese(II1) oxidation state (289). Although no examples of crystallographically characterized Mn;I(p-OH)(OAc) cores exist, the topologically similar MniII(salampn)(OAc)complex (290) has a MniII(p-OR)
w
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
369
FIG. 53. [Mn"'(bi~picen)],(~-O)(~,,~-OAc) (left) and [Mn"'(tmima)]z(~-O)(~,,~-OAcj (right). [Reproduced with permission from (288) and (289). Copyright 1994 and 1993 the American Chemical Society, respectively.]
(OAc) core with a Mn-Mn separation of 3.55 A. The catalase mimic Mni1(2-OHbenzimpn)(OAc)(Fig. 541,has the same core in the lower oxidation state (291) with a nearly identical Mn-Mn separation, 3.54 A. The more common members of the class of p-0x0-p-carboxylato complexes have three bridging ligands instead of two. The Mn,(p-O), (p-RC0,) core may be considered as a Mnz(p-O), core with a carboxylate spanning the Mn-Mn vector (292-294). However, the presence of the additional carboxylate bridge induces the Mn,(p-O), unit to be (Fig. 55) has a bent. For example, [Mn~~bpy)(H,0~12(p-O),~p,,~-OAc~3+ dihedral angle between MnYpO), planes of 162" and a Mn-Mn distance of 2.64 A (295),which is somewhat shorter than Mn-Mn separations in typical Mnr(p-O), cores. Reduced derivatives containing core have similar structural parameters the Mn"'Mnrv(p-0),(p~,~-OAc)
FIG.54. Mn~Y2-OHbenzimpn)(OAcj.[Reproduced with permission from (291). Copyright 1994 the American Chemical Society.]
370
LAW, CAUDLE, AND PECORARO
P
FIG. 55. [Mn'v(bpy)(H,0)],(~-O)~(~~,3-OA~)3+. [Reproduced with permission from (295). Copyright 1994 the American Chemical Society.]
(296). No bis-p-hydroxo-p-carboxylato cores have been isolated. complexes However, several examples of Mn~1'(p-OR)2(p1,3-OAc) have been prepared and characterized (266). Mn$'(2-OHsalpn) (MeOH)2(p-OMe)(pl,3-OA~) is typical, with a Mn-Mn separation of 2.9 8 (266). The p-oxo-bis-p-carboxylato structure type, Mn,(p-O)(p-OAc),with two syn-syn carboxylate groups and a single p-0x0 ligand has been structurally characterized with a number of supporting ligands, but only in the dimanganese"' oxidation state. The Mn-Mn distance in these compounds is in the range of 3.1-3.2 8, as would be expected for the single 0x0 bridge. This core topology is particularly interesting because it is the only p-oxo-p-carboxylato motif to have been isolated in three different protonation states: Mn;"(p-O)(p-OAc),, Mn"Mn"' (p-OH)(p-OAc),, and Mnf(p-OHz)(p-OAc)z(297-299). Representative members of this series are illustrated in Fig. 56. It is noteworthy that each protonation is accompanied by a decrease in oxidation state and a consistent increase in Mn-Mn separation from 3.1 to 3.4 to 3.6-3.8 8. This is consistent with the decreased donor ability of the single-atom bridge with each successive protonation, which promotes longer bond lengths as well as stabilizing the lower dimer oxidation states. Dimanganese model complexes having only carboxylate bridges are known primarily in the lower manganese oxidation states, and complexes supported by only a single p-carboxylate unit are particularly rare (300-302). However, there are examples of dimanganese(I1) complexes bridged by a single carboxylate. Mn;*(RCO2)core has the bridging carboxylate in a syn-anti configuration with a long Mn-Mn distance of 5.6 8, Scheme 14,and complexes in which the carboxylate bridge is in the anti-anti configuration have exceptionally long 5.9to 6.0-8 Mn separations. Bis-p-carboxylato complexes, Mnr*(p-RC02)2,
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
371
FIG.56. Examples of MnZ(p-O)(p-RC02)2 cores with varying protonation of the "p-0" bridge. Upper left: [Mn"(N,N,N',N'-tetramethyl~ethylenediamine)~IZ(0Ac)~(p-OAc),(pOHZ). Right: {IMn"Mn"'(Me,tacn)Z1(p-(CH,),COO)z(p-OH)}+Z. Lower left: [Mn"' (bipy)CIZlZ(p-OA~)Z(p-O). [Reproduced with permission from (472),(298), and (473),respectively. Copyright 1992 American Chemical Society, copyright 1996 WILEY-VCH Verlag, and copyright 1993 American Chemical Society, respectively.]
have much shorter 4.3-A Mn-Mn separations with a syn-syn configuration for both p-carboxylate groups. Addition of a third carboxylate group, Mn11(p-RC02)3, yields a Mn-Mn distance of 3.6-4.0 A (303). Mn
Mn
\
0d,
R
syn-syn
Mn
Mn- O y O- Mn
\
0y o ' M n R syn-anti
R anti-anti
SCHEME 14.
Other unsupported Mnf'(p-OR)2 cores have Mn-Mn separations of 3.0-3.2 A. Complexes having one supported p-alkoxo and one unsup-
372
LAW, CAUDLE, AND PECORARO
ported p-methoxo ligand can have Mn-Mn distances of 3.0 A or lower (266,267). Alkoxo-p-carboxylato complexes are topologically similar to their bis-p-0x0 analogs, however, the Mn-Mn distances tend to be considerably longer in the p-alkoxo complexes than in corresponding p-0x0 derivatives. For example, p-alkoxo-p-carboxylato complexes, Mnp (p-OR)(p-RCO,), have typical Mn-Mn seearations of about 3.5 A. This is considerably longer than the 3.3-A separation observed for Mni1'(p-O)(p-RCOz). Mnz(p-OR),( -RC02)complexes have Mn-Mn distances ranging from 2.9 to 3.0 which is about 0.40 longer than the oxo-carboxylato-bridged analog. Carboxylato-bis-p-alkoxo complexes having the MniYp-OR)(p-RCO,), core have even longer 3.3- to 3.4-A separations (304)which extend to as much as 3.5 A upon oxidation to the MnllMnll'(p-OR)(p-RCOz)zoxidation level (305, 306). This compares favorably with 3.4 A for the MniI(p-OH)(p-RCO,)(297,298) complex recently prepared, demonstrating that alkylation of p-oxobridged dimanganese cores is a good structural analog for protonation. p-Phenoxide-bridged dimanganese clusters are structurally similar to their alkoxide-bridged counterparts (304-306). However, an additional factor not present in alkoxo-bridged complexes is the possibility of rotation about the Ar-0 bond in phenoxo complexes (238, 307, 308). Because the donor ability of the phenoxide ligand is dependent on the orientation of the metal ion with respect to the aromatic plane, we might expect that this would be manifested in structural changes upon Ar-0 rotation in p-phenoxide-bridged manganese dimers. In fact, when Mn!Yp-ArO), dimers are examined, it is observed that as the dihedral angle between the MnzOzplane is increased from nearly zero to nearly go", the Mn-Mn distance increases from 3.26 .k (304) to 3.42 A (307).This is consistent with the phenolate oxygen being a weaker donor when the metal ion acceptor is oriented perpendicularly to the aromatic plane. There are a limited number of N-bridged or imidazole-bridged dimanganese complexes with biologically relevant supporting ligands (3091, consisting of azide-bridged complexes. Azide ligation is often used as a model for peroxide ligation to manganese complexes due to the greater stability of azide complexes, which facilitates structural characterization. h i d e may bridge either in a p-1,l fashion, yielding Mn-Mn separations of 3.4-3.5 & or as an p-1,3 bridge to give a 3.5-A separation (310, 311). With respect to imidazole-bridged complexes, a complex with the ligand DCBI, 4,5-dicarboxyimidazole, was with a Mn-Mn recently reported, [Mn'"(3,5-di-t-b~tylsalpn)]~(DCBI), separation of 6.18 A (Fig. 57) (312).
[
A
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
373
FIG. 57. [Mnrv(3,5-di-t-busalpn)I,DCBI. [Reproduced with permission from (312). Copyright 1997 the American Chemical Society.]
C . TRINUCLEAR Trimanganese clusters (and clusters of higher nuclearity) can usually be considered as assemblies consisting of the binuclear units discussed in the previous section. There are no known trimanganese enzymes. However, it has been proposed in the past that the fourmanganese active site in the OEC may actually be composed of a trinuclear and mononuclear manganese site (discussed earlier). The only trimanganese complexes having only p-0x0 bridges between the manganese ions are those of the Mn3(p-0I4class (292, 313, 314). These consist of a MnZ(p-O),core that is spanned by a Mn(p-O), unit (Fig. 58). The topology of these complexes is reminiscent of the Mnz
FIG. 58. Two examples of trimeric structures with Mn0,-bridged Mn2(p-0)2cores. [Reproduced with permission from (292)and (313).Copyright 1992 and 1990 the American Chemical Society, respectively.]
374
LAW, CAUDLE, AND PECORARO
0
0
N N
N
N 6 N FIG.59. The trimeric (w3-O)(p-peroxocomplex). The other ligands are acetates and diethylenetriamine. [Reproducedwith permission from (317). Copyright 1988 the American Chemical Society.]
(p-0)2(p-RCOz)core, with the Mn(p-O), unit replacing the RCOz unit of the latter. There are several reported structures of the Mn3(p3-O)unit supported by additional p-carboxylate bridges (315, 316). The Mni11(p3O)(MeC02)6complex has each manganese ion bridged by two carboxylate and one p3-0x0group structurally analogous to the Mndp-0) (p-RCO,), complexes. The Mn-Mn separation in Mni1'(p3-O)(MeC0z)6 3.35 8, is longer than in Mnz(p-O)(RCO,),and is more reminiscent of the distance in the Mn2(p-OH)(p-RCOz),complexes. A unique member of this class of p 3 - 0 ~ 0trimanganese complexes has a single pl,z-peroxo ligand bridging two of the manganese ions (317) (Fig. 59). Another has been produced with a ps-OH and a bridging catecholate (318). Trinuclear manganese motifs are common among carboxylate comcore has found been with variplexes. The Mn~11Mn11(p-OR)z(p-RC02)4 ous Schiff-base-containing supporting ligands (269, 319, 469). These cores are characterized by two terminal Mn" ions, which "encapsulate" a central Mn"' ion with Mn-Mn neighboring distances of 3.5 8, which is consistent with the MnllMnlll(p-OR)(p-RCOz)zdimanganese cores having an identical bridging motif. The p-hydroxo-bridged analog, MnH'Mn"(p-OH),(p-RCO,), , has a similar structure with a somewhat shorter 3.4-A Mn-Mn separation (Fig. 60) (320). Univalent Mnil trimers having only bridging carboxylate units, Mni1(p-RC02)6, exhibit a Mn-Mn bridging motif in which one of the three bridging carboxylate undergoes a carboxylate shift in which one of the carboxylate oxygens bridges two Mn" ions (Fig. 61) (321). Therefore, these core units are more precisely defined as Mn&1(pl,3-RC02)4(pl,l-RCO~)
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
375
FIG.60. A Mn"iNnli/Mn'ii trimer with a Mn~"Mn'1(p-OH)2(RCOs)4 motif. [Reproduced with permission from (320).Copyright 1994 the American Chemical Society.]
with a characteristic Mn-Mn separation of 3.5-3.6 A. The p-phenolato-bridged trimanganese(I1) complexes exhibit a similar carboxylcore with interate shift, giving the Mn~l(ArO)a(pl,~-RCO,),(pl,,-RCO,), manganese separations of 3.27 A (Fig. 62) (322).
FIG.61. Mn" trimer with a Mni'(RC02)tibridging motif that includes bridging p,,.Icarboxylates and pl,,-carboxylates. [Reproduced with permission from (321). Copyright 1990 Wiley-VCH Verlag.]
376
LAW,CAUDLE, AND PECORARO
-N
FIG. 62. Mn~1(ArO)2(~Ll,s-RC02)2(CL1,~-RC02)2 core. [MnP(5-NOzsalim)2(OAc)ll. [Reproduced with permission from (322).Copyright 1995 the American Chemical Society.]
D. TETRANUCLEAR As with binuclear manganese complexes, several recurring tetramanganese core structures have been synthesized with different supporting ligand sets (44, 323, 324). The simplest tetranuclear manganese cluster is the cubane core, Mn404, which has tetrahedral symmetry in the absence of distortion. A generalized scheme of a cubane core is shown in Scheme 15. This core is arranged in such a way that two 0x0 bridges exist between any two manganese ions so that all of the manganese ions are equivalent. A number of mixed-valent distorted cubane core structures are known in which the manganese distances are not all equivalent (325,326).This core was proposed in the past as a model for the manganese cofactor in photosynthetic oxygen evolution, so it has been synthesized with a number of supporting ligand architectures. One of the 0x0 groups in the cubane core can be replaced with a carboxylate donor to form the asymmetric Mn403(PhCOJ complex, in which the four manganese are no longer equivalent and Mn-Mn distances of 2.8 and 3.3 A are observed. This complex is unique among tetramanganese cores in exhibiting pl,land p1,3 bridging by a single carboxylate group. An 0x0 group may also be replaced by a p3-chloride ligand to yield the distorted cubane, MII,(O)~C~ (Fig. 63) (327,328).In this mixed-valent Mnil'MnTV complex the three Mn11'Mnw02 faces have Mn-Mn separations of 2.8 A, as expected for bis-p-oxo-bridged Mn"'MnW complexes. However, the MniI' (O)(Cl) faces have an average Mn-Mn distance of 3.3 A. Alkoxidecontaining cubane analogs (329),Mn,(OR),, have also been prepared and shown to be topologically similar to the conventional oxide-containing cores with somewhat longer Mn separations.
377
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
Mn40nX, Core Structures
0-
MnM ,/n0 -
0-Mn-
0 Adamantane
0
'Fused-Open' Cubane
Ih
0
Mn-
1
\O+Mn 0- -Mn
\I \I Mn-
0
A Selection of Cubane Core Structural Types
'Butterfly" Type Cores
SCHEME 15
Addition of two oxide donors yields the adamantane complexes, MQO,, which also have 0x0 bridges separating all of the manganese (Scheme 15) (330). The difference between the cubane and adamantane core is the presence of two oxygen atoms, leading to the proposition that an adamantane-to-cubane conversion in the OEC may lead to oxygen evolution (31, 199, 200). Another Mn406cluster tetranuclear cluster is the "fused open cubane," in which a Mn406 cluster is formed that is the equivalent of two face-sharing Mn304 cubanes, with one manganese site vacant (329) (Scheme 15). Loss of two oxide bridges from a cubane core yields the Mn402core complexes, commonly called butterfly clusters, which have an MnzOzcore as their foundation and two p3-0x0 groups connecting to the two pendant Mn ions lying above the MnzOzplane (Scheme 15) (331, 332). A centrosymmetric Mn402core has also been observed as well. In these complexes, two p3-0x0groups from MnzOzbridge to two additional manga-
378
LAW, CAUDLE, AND PECORARO
FIG.63. A manganese cubane cluster with a corner oxygen atom replaced by a chloride. [Reproduced with permission from (327). Copyright 1993 the American Chemical Society.]
nese ions, but they are now disposed on opposite faces of the MnzOz plane. There are a number of tetramanganese clusters that cannot be considered as merely distorted members of the basic classes, but that may also be relevant in biological manganese chemistry. For example, a highly symmetric tetramanganese complex is formed with the ligand 2-OHpicpn (333).This complex, an efficient catalyst for the catalase-like disproportionation of hydrogen peroxide, is an Mn4L4structure in which each pair of manganese ions is bridged by a single alkoxide bridge covalently attached to the picpn unit (Fig. 64). Each manganese is separated from its neighbor by 3.7 A. In this complex all of the Mn-Mn linkages are identical, so it is not precisely a dimerof-dimers model. However, the complex M r ~ ~ O ~ ( t p h pcontains n ) ~ + two MnzOzcores cofacially oriented and linked by two alkoxide bridges covalently bound to the tphpn ligand (334)(Fig. 64). In this case, two different Mn-Mn linkages support the cluster, and this is accurately considered as a dimer-of-dimers structure. A variation on this theme is the dimer-of-dimers complex [Mn,O,( tmdp)]%+, which also has two MnzOzdimers, but in this case linked by methylene spacers of the tmdp ligand. This complex has a rather short 2.6-A Mn-Mn distance across the dioxo core, but a long 5.9-A Mn-Mn distance across the methylene spacer connecting the two dimers (335)(Fig. 65).
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
379
FIG.64. Two tetramers. Left, is the [Mn,0,(tphpn)214atetramer. Right, [Mni1(2-OH picpn)J tphpn = N,N ,N',N,-tetrakis~pyridylmethyl~-1,3-diaminopropan-2-01. 20Hpicpn is the Schiff-base condensation product of pyridine-2-carboxaldehydewith 1,3-diaminopropan-2-01, [Reproduced with permission from (334)and (333),respectively. Copyright 1991 and 1996 the American Chemical Society, respectively.]
A number of chainlike tetramanganese complexes may also be relevant to biological manganese chemistry. The Mn,O,(bpy)i+ complex (336) has the same core stoichiometry as the adamantane complexes (Fig. 66). However, this complex contains four Mn" ions in a linear configuration in which each pair of neighboring manganese ions is bridged by two 0x0 bridges with a Mn-Mn distance of 2.7 A. Longer Mn-Mn distances of 4.9 and 6.4 A are observed between the 1,3 and 1,4 manganese ions. A topologically similar complex, Mn4(2-OHPh PhenI6, was prepared having six phenoxide instead of oxide bridges (337).The Mn-l-Mn-2 distance here is 3.4 A, characteristic of dimanganese diphenoxide complexes, in which the Ar dihedral angle with the Mn202plane is nearly 90". Mn902(tphpn)z(Hz0)2(CF~SO~~~+ has a Mn,(0)2(RO)2 core with two manganese linked to a Mnz02 core through alkoxo bridges (Fig. 67) (338). IV. Physical Properties
Hydrated manganese in acidic aqueous solution exists in two oxidation states, Mn" and Mn"], with a potential of + 1.51 volts separating them (339). Therefore, Mn"' in solution is a strong oxidant. This
380
LAW, CAUDLE, AND PECORARO
FIG.65. A dimer of dimers. Two ( p - 0 2 - ) ,bridged Mn"'MnN dimers are bridged to one another via the ligand. Water molecules are also bound to the manganese of this complex. [Reproduced with permission from (335).Copyright 1994 the Chemical Society of Japan.]
arises from the strong driving force to form the stable half-filled d5 subshell of Mn" from the d4 Mn"' ion. The MnWstate, which is a d3 ion, exists as the dioxide MnOz or as oligomeric forms thereof. Permanganate ion, having a do Mnwlcenter, is diamagnetic. Mononuclear metal-ligand complexes with the Mn"-MnV oxidation states are all known provided suitable chelating ligands are provided, and polynuclear complexes with Mn1I-MnWare known as well. Consequently, to understand the spectroscopy of polynuclear complexes, it is necessary to begin by examination of the electronic properties of
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
381
FIG.66. The first coordination sphere of the complex Mn106(bpy)Q+ showing the core structure. Matching subscripts on the nitrogens designate that they originated from the same bipy molecule. [Reproduced with permission from (336). Copyright 1994 the American Chemical Society.]
manganese ions in these oxidation states. Mn" in an octahedral ligand field has a ground state that has spin S = 5/2 but is an orbital singlet. Therefore, no spin-orbit coupling arises from the ground state of Mn", and S is a good quantum number for describing
FIG.67. The linear chain structure with the tphpn ligand. The two inner manganese are involved in MnzOzcore and are then bridged to the outer two manganese ions by ligand alkoxides. [Reproduced with permission from (338).Copyright 1990 the American Chemical Society.]
382
LAW, CAUDLE, AND PECORARO
the spin levels of this ion. The same is true for the d3 MnIVion, which has an orbital singlet 4A2ground state in an octahedral field. However, the ground state of the octahedral d4 Mn"' ion is the orbital doublet 5E, which gives rise to substantial spin-orbit coupling, L S. Therefore, S is not a good quantum number for describing Mn"' electronic properties, and the value L + S is used. The MnVcomplexes thus far isolated are square pyramidal and diamagnetic. This suggests a very strong axial splitting in the t2gorbital set leading to a singlet 'A ground state in these complexes.
A. ELECTRONIC SPECTROSCOPY Because the most intense bands in the visible spectroscopy of manganese compounds are charge-transfer bands involving the bound ligands (i.e., metal-to-ligand or ligand-to-metal charge transfer), the visible spectra of polynuclear complexes tend to be very similar to those of mononuclear complexes with the same ligands. Consequently, little information on the structure of polynuclear complexes is contained in their electronic spectra. Exceptions to this rule are certain localized mixed-valence complexes that exhibit intervalence charge transfer bands (ITS)formally involving electron transfer between nonequivalent metal sites in the complex. Because this can only occur in polynuclear complexes that are valence localized in the ground state, valence-delocalized complexes cannot exhibit IT bands. An example of an IT transition at 1170 nm ( E = 270 M-'.cm-l) is found in the [Mn11Mn111salamp21complex, whose structure is shown in Fig. 68,
FIG.68. Core structure of [Mn"Mn"'(~almp)~]., a diagram of the ligand and the binding motif for one ligand to the manganese are shown above the core structure. [Reproduced with permission from (308).Copyright 1991 the American Chemical Society.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
383
along with a representation of the ligand (308).The appearance of IT bands is rationalized based on Marcus theory. Electron transfer between two coupled metal centers, MAand MB, is rapid compared to molecular vibrations, resulting in a low-lying excited state in which an electron formerly residing in an orbital of predominantly MAcharacter has been promoted to an orbital of predominantly MB character. This excited state may relax via a nonradiative mechanism wherein the ligand environment adjusts along a vibrational normal mode to return to the ground state configuration. Therefore, the energy and lifetime of IT excited states can be strongly influenced by the vibrational structure of the polynuclear complex. UV-vis spectroscopy proved particularly enlightening with respect to the catalase system (23, 30, 132). The spectra of model complexes bridged by carboxylate and 0x0 bridges were quite similar to those observed for the catalase. This led researchers to the proposal that such a motif might also be present, a hypothesis that has been borne out by the crystal structure of the catalase (125). B. MAGNETISM Temperature-dependent magnetic susceptibility is often used in the study of polynuclear metal complexes to determine the magnitude of spin exchange energetics between two metal centers in the same molecule (340, 341). This magnitude is called J and is a measure of the energy difference between successively higher molecular spin states when a molecule is placed in a magnetic field. The temperature dependence of the molar magnetic susceptibility x,,, arises from two sources. The first is zero-field splitting, which is measured by D and is observed in mononuclear complexes as well as polynuclear complexes. The primary factor contributing to D in manganese complexes is coupling of the orbital angular momentum and spin angular momentum vectors, and is defined by L . S. Octahedral Mn" has a t&eE configuration with a 6A1ground state, and MnN has a t!,e: configuration with a 4Azground state. Because both of these ground states are orbitally singlet and have no net orbital angular momentum, D will typically be on the order of 0.10 cm-' or less in complexes of these ions (339).Therefore, in polynuclear complexes of these ions, S is a good quantum number and the magnetic exchange interaction between the ions can be described by using a spin-only Hamiltonian. However, Mn"' has a t!,e; configuration with a 5E ground state, which is an orbital doublet that will have a substantial contribution to the overall magnetic moment arising from the net orbital angular mo-
384
LAW, CAUDLE, AND PECORARO
mentum. As a result, Mn"' systems may have D on the order of a few cm-'. For such a system, S is not a good quantum number to describe the energy of the state, and a spin-only description of the magnetic exchange may not apply. The low-temperature susceptibility of Mn"' complexes are often dominated by the thermal depopulation of the higher spin states giving a diamagnetic ground state at low temperature. The second source for the temperature dependence of xm is spinspin (S1* S z )coupling or spin exchange, which is unique to polynuclear complexes. S1.Sz coupling arises when the spins on two metal centers, MI and M2, are correlated. The two spins may be coupled antiferromagnetically so that the ground spin state S for the system is S = S , - Sz. On the other hand, the spins may be coupled ferromagnetically to give a ground spin state S = Sl + S z . We will discuss S1* S z coupling in the context of binuclear systems with the understanding that the general principles applied in the binuclear case also apply to higher polynuclear complexes. Characterized MnH systems are typically weakly coupled, with 5 values on the order of 10 cm-' or less (30).The orbitally single 'Al ion is therefore a good example of a system whose temperature-dependent magnetic susceptibility is well treated by the spin-only formalism due to the lack of L .S coupling from the Mn" ground state. The spin-only Hamiltonian is given by H = SIJSzwhere 5 is the spincoupling constant with units of energy and Sl and Sz are the spin operators. The eigenvalues E for the system HW = EW are given by E = J [ S ( S + 1) - Sl(Sl+ 1) - Sz(S2+ l)]and the spin energy levels are given by 0, 2J,6J, 125,205, and 305. The temperature-dependent magnetic susceptibility is then a result of the depopulation of higher spin states as the temperature is decreased, according to the Boltzmann distribution of states, i.e., P, = A,exp(E,/kT)LU,exp(E,/kT), where En and A, are the energy and degeneracy of state n and P, is its relative population. For an antiferromagnetically coupled system, the ground state is diamagnetic and the magnetic moment should approach zero as the temperature is lowered. Ferromagnetic coupling gives rise to a paramagnetic S = 5 ground state so that the low-temperature magnetic moment will approach 11 BM per binuclear complex. Binuclear Mn'" complexes may also be treated by a spin-only Hamiltonian with eigenvalues E = 0, 25, 6J, and 125. The ground state is also a spin singlet in this case, but the high-temperature moment should approach 5.4 BMhinuclear complex due to equivalent population of all of the spin states.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
385
The Mn"' ion has an 5E ground state that is orbitally nondegenerate. Even so, the spin-only contribution to the temperature-dependent magnetic susceptibility in strongly coupled Mn"MnT", Mn"' 2 , and Mn"'MnIV complexes may be treated in the same way as for the Mnil case discussed earlier. For cases in which J 4 D, D can be considered as a perturbation on the spin-only energy levels and the spin-only formalism is usually adequate to explain the magnetic data and to calculate J . For systems with extremely weak coupling (i.e., J < D,J may be considered as a perturbation on the D-split levels, and the magnetism will be dominated by D. However, when J and D effects are of the same order of magnitude, treatment of the temperaturedependent susceptibility is complicated by the fact that neither S nor L is a good quantum number for describing the spin exchange. A full treatment of magnetic exchange involving both spin and orbital contributions is quite involved and beyond the scope of this discussion (340, 341). Great care must be exercised in the treatment of data on weakly coupled systems containing Mn"' in order to accurately assign which effect is responsible for the observed temperature-dependent magnetic data. In many cases, one cannot uniquely fit the magnetic data to a single set of J , D, and g values.
C. EPR SPECTROSCOPY Electron paramagnetic resonance spectroscopy is one of the primary tools in studying the electronic structure of polynuclear complexes (341). Whereas magnetic susceptibility studies are capable of detecting electronic interactions as small as a wavenumber (discussed earlier), the EPR spectrum of a polynuclear complex may be sensitive to intramolecular exchange couplings as small as 0.001 cm-l even at room temperature. Additionally, the 55Mnnucleus has a nuclear spin ( I ) of 5/2. For a mononuclear system, six hypefine lines are predicted. However, for a multinuclear system containing n Mn ions, as many as 6" lines may be observed. This gives rise to a rich EPR spectroscopy among polynuclear manganese complexes, which has been used extensively in the study of biological manganese systems and model complexes. Mononuclear Mn" and Mn" complexes exhibit characteristic EPR spectra that are, for the most part, insensitive to the ligand environment around the metal. However, the EPR spectra of polynuclear complexes are strongly dependent on the metal ligation environment, the nature of bridging ligands, and the degree to which the metal
386
LAW, CAUDLE, AND PECORARO
ions are coupled. We will examine some of the relevant binuclear and polynuclear cases in more detail.
1. Binuclear Complexes Mononuclear Mn" complexes almost always exhibit a single derivative signal with a crossover at g = 2 and often having a well-defined six-line Mn hyperfine splitting superimposed on it, as expected from an S = 5/2 spin system with only slight spin-orbit coupling. However, two antiferromagnetically coupled Mn" ions coupled antiferromagnetically would be expected to be EPR silent at low temperature owing to the ground-state spin singlet. Binuclear MnP complexes often do exhibit an EPR spectrum due to thermal population of the S = 1 through S = 5 spin states. Although these are non-Kramer's systems with integer spin, which often do not show X-band EPR spectra in mononuclear complexes due to zero-field effects, such effects are minimal in Mn" due to the 6A1ground state, and there are a number of examples of EPR spectra of Mn" dimers (279, 438, 439, 474, 476). Binuclear mixed-valent Mn'lMnlll complexes nearly always exhibit an EPR signal because the ground state and all thermal excited spin states are Kramer's spin systems. The form of the spectrum depends strongly on the magnitude of J and on the zero-field splitting imparted by the Mn"' ion. Strongly coupled systems typically exhibit a signal at g = 2 with a multiline hyperfine splitting from two Mn ions; more weakly coupled systems show signals withg > 2 due to transitions involving zero-field split levels (305, 475). Unlike Mnil dimeric systems, MnH' complexes almost never show EPR spectra in the X-band region due to zero-field splitting that puts allowed transitions out of range of X-band energy (9.5 GHz). However, strongly-coupled ( J = 100 cm-' ) mixed-valence Mn"'Mn'' systems show very characteristic EPR spectra at g = 2, consisting of between 12 and 19 well-resolved Mn hyperfine lines, consistent with an S = 1/2 ground state. A typical such spectrum, for M"'MnIV (p-O),L,, is shown in Fig. 69. Although 36 lines are theoretically possible for two coupled manganese ions, between 12 and 19 lines are usually observed in practice due to overlap of some of the hyperfine lines. The g = 2 signal is the only one typically observed for strongly coupled (~-0 complexes )~ (140), but Mn"'Mn" systems such as the Mn1I1MnW ( J = 10 cm-') may also exmore weakly coupled Mn"'MnTVsystems a lower-field signal at higher temperatures that is associated hibit with thermal population of a zero-field split S = 312 excited state (202), a feature also often observed in the EPR spectra of mononuclear MnIVcomplexes, which are also S = 3/2 ions (140).
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
11 2200
I
I
2600
I
I
3000
I
I
3400
I
I
3800
I
387
I
4200
G(H4 FIG. 69. A Mn"'Mn'" exchange coupled EPR spectrum, from [Mn"'Mn"(bpza)p (p-O)2]3'.bpza = 2-(2'-pyridyl)benzimidazole.LReproduced with permission from (465). Copyright 1994 Elsevier Science.]
Strongly coupled binuclear Mn;" complexes have never been shown to exhibit X-band EPR spectra due to the low-energy singlet ground state. Because stabilization of the high-valent Mn'" requires very strong oxyanionic ligands, nearly all such MnJ" complexes have at least one bridging 0x0 or hydroxo ligand, and they may contain additional bridges such as carboxylate ions, all of which contribute to a generally strong coupling in such systems. Consequently, there are few examples of weakly coupled MnF systems to make generalizations about the presence or lack of EPR spectroscopy in these systems. However, we might expect that because ground-state spin-orbit contributions to zero-field splitting are absent in MnN ions, higher spin triplet and quintet states might be EPR active here just as they sometimes are in the MnH complexes. A recent example of an EPR signal arising from a weakly coupled ( J = -2.3 cm-') Mnk" dimer was observed for the complex [Mn'"(di-t-butylsalpn)lzDCBI (Fig. 70) (312). 2. Polynuclear Complexes Several factors unique to polynuclear systems make the interpretation of their spectra more complicated than for the binuclear cases. One such factor is that a given manganese ion may couple to two or more other ions with different exchange magnitudes, J , and J z . The interpretation of such spectra is often simplified by assuming that Jz
388
LAW, CAUDLE, AND PECORARO
A 5.5K B 50K
I
0
I
I
I
I
I
I
1000 2000 3000 4000 5000 6000 7C Field (gauss)
FIG.70. X-band EPR spectra of the weakly coupled [Mn1"(3,5-di-t-busa1pn)l2DCBI complex. [Reproduced with permission from (312).Copyright 1997 the American Chemical Society.]
is much smaller than J 1 and can either be neglected or treated as a perturbation on the spin splitting by J,. On the other hand, J , may be exactly equal to Jz (in a highly symmetric trinuclear species for example), thus reducing the spin exchange problem to a simple manifold of states treatment as for binuclear species. However, when J1is of the same order of magnitude as J2,then the spin splitting is complex and cannot be treated by a simple manifold of states. A second factor complicating the EPR spectroscopy of polynuclear species is the fact that the possible number of Mn hyperfine splittings increases as 6" where n is the number of manganese ions contributing to the hyperfine interaction. Furthermore, the hyperfine splitting constant A decreases as lln. Therefore, EPR spectra of polynuclear complexes may exhibit broadened signals due t o a large number of unresolved hyperfine splittings, which may not be distinguishable from broadening from other sources such as slow tumbling or anisotropy. 3. EPR of Biological Manganese Complexes
EPR is one of the most useful techniques brought to bear in the study of polynuclear manganese centers in biology. For example, it was the hyperfine splitting in the CW EPR spectrum of spinach thylakoid membranes that first showed conclusively that the oxygen evolving complex in PS I1 consisted of a multinuclear cluster of coupled manganese ions (160, 161). The EPR from the S, oxidation state of active PS I1 centers has also been the primary signal for tracking the redox cycling of the OEC during oxygen evolution by PS I1 (33, 42, 46). There have also been numerous attempts to establish the nuclear
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
389
and oxidation levels of the manganese cluster. Brudvig, for example, used variable-temperature EPR data to support a tetranuclear cluster (342, 343). Several other groups have variously assigned these EPR spectra to two dimers (3441, dimers of dimers, and tetramers (345). More recently, Dismukes has argued that the Sz spectra result from a tetranuclear cluster composed of 3 Mn"' and 1 MnWions (346).To obtain a good fit of the data, the Mn"' ions are predicted to be five coordinate with an inverted order for the &orbital energies. The Sz signal has also yielded information on the ligand environment (187, 347) through 15N labeling and observation of the 15N hyperfine splitting with ENDOR spectroscopy, and on the binding of substrate and inhibitors (348) to the OEC and model complexes (203, 349). Very recently, two new EPR signals have been observed. The Sosignal was shown to be multiline in the presence of methanol (170-172). Using parallel-mode EPR spectroscopy, Britt has shown a beautiful multiline signal for the S, oxidation level (174). Interpretations of these new signals hold great promise for understanding this enigmatic cluster's structure. Enzymes containing binuclear Mn sites are especially amenable to structural elucidation using EPR spectroscopy. The EPR spectra of various oxidation levels in the binuclear manganese catalase were used to assign the low-valent active form as a weakly coupled MnL) center and the "superoxidized inactive form as a strongly coupled MnlIIMnWcenters (7, 23, 136, 138-140, 350, 351). These assignments were based in part on a comparison with the EPR spectra of crystallographically characterized model compounds. The arginase active site was shown to be a binuclear Mn" center by the shape of the spectra and the number and spacing of Mn hyperfine couplings (7). The structures proposed for both the manganese-containing catalase and arginase from these studies were essentially confirmed by subsequent X-ray crystal structure analysis (8). D. X-RAYA~SORPTION SPECTROSCOPY Although it is an unconventional spectroscopic technique, X-ray absorption spectroscopy (XAS) has been the primary tool in the elucidation of the structural features of biological multinuclear Mn centers for which high-resolution crystallographic data are not yet available (23, 24, 134). This is typically done by comparing features observed in the XAS of the biological complex with similar features found in the spectra of crystallographically characterized Mn model complexes. We defer an extended discussion of the theory and principles of XAS
390
LAW,CAUDLE, AND PECORARO
to several reviews on the subject of XAS in biological and model manganese complexes (24, 134). However, a few basics will suffice to understand the information that is contained in XAS of polynuclear complexes. Unlike UV and visible electronic spectroscopy, which involve promotion of outer-shell valence electrons, X-ray absorption involves the promotion of core electrons to higher energy using X-ray frequencies (24,134).As such, X-ray absorption is a high-energy technique. There are really two techniques that come under the heading of X-ray absorption. X-ray Absorption Near Edge Structure, XANES, involves promotion of a 1s electron to the 3d level in the case of manganese. (This is for the K shell; L edges with core electrons from the n = 2 shell are now being examined for models and enzyme.) This gives rise to an absorption “edge” whose energy is very characteristic of the oxidation state of the manganese. As a general rule, manganese in higher oxidation states exhibits absorption edges at higher energies, as might be expected from the tighter binding of core electrons in higher oxidation states. In fact, the edge energy seems to correlate best to the average number and bond lengths of ligands bound to the metal. In polynuclear complexes, shifts in the edge energy are correlated with changes in the average oxidation state of the manganese ions. This complicates the use of XANES in higher nuclearity complexes because one-electron oxidation results in a smaller change in the average oxidation state as complex nuclearity increases. Even so, this technique has been used extensively in tracking the oxidation states of the manganese cluster in PS I1 upon S-state advancements. The progressive edge shift upon increasing oxidation state for polynuclear manganese complexes is nicely illustrated by the XANES spectra of the isostructural series Mn2(2-OHsalpn)i(n = -2 to +1)in four successive oxidation states (Fig. 71) (467). As the X-ray energy is increased beyond that required for promotion of the core electron, ejection of core electrons into the continuum occurs. This ejected electron propagates from the Mn center until it encounters another atom from which it can be back-scattered. The interference of back-scattered waves with propagating waves leads to an interference pattern that is manifested as an oscillation in the X-ray absorption pattern. Fourier transformation of this oscillating spectrum from the frequency domain to the distance domain gives a new spectrum whose abscissa contains information on the distance between the target atom (i.e., the Mn center) and the back-scattering atoms. This second technique is called Extended X-ray Absorption Fine Structure, E M S , and has been the only spectroscopic tech-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
6531
6538
6545
6552
391
6559
Energy (ev)
FIG.71. XANES spectra of the family of complexes IMn(20Hsalpn)l" (n = -2, - 1, 0, +l), and the oxidation states on the manganese go from Mn"Mn" to Mnl'lMnl". Noticeable changes occur in the edge energy with each increase in the oxidation state of the manganese. [Reproduced with permission of the author, from (4671.1
nique to yield direct direct structural information about the manganese cluster in PS 11. The XANES of the Mn catalase provided the first definitive proof that this enzyme cycles between the Mnil and MnkI' oxidation levels (137, 352) The extent of catalase activity correlated with the proporenzyme; however, samples with Mn*" quantitation of Mnf or tively showed reduced activity. The EXAFS of the Mn catalases have been less informative because of the Mn-Mn separations in the reduced, active enzymes (135).Nevertheless, EXAF'S of the "superoxidized" enzyme demonstrated that the MnlIIMnl" enzyme has a Mn-Mn separation of ~ 2 . 7 A, which is consistent with a di-p,-oxo core (135).Subsequent spectroscopic analysis confirmed that a "diamond core" with a bridging syn,syn acetate formed the enzyme active site ( 9 ) . While interpretation of the XAS for the Mn catalase has been relatively straightforward, the application of XANES and EXAFS to the OEC has led to several significant disagreements among workers in this field (24,28,33,46, 48, 134, 154). Klein and Sauer first collected XAS data on the OEC and provided the first detailed structural and oxidation-state information about the Mn cluster (353, 354). The XANES indicated that the S , and S , enzyme forms contained significant amounts of Mn"' and Mn'". At the same time, both S states could be examined by EXAFS and exhibited short Mn-Mn vectors of 2.7 A (154-159), which is consistent with a high-valent di-pa-oxocore, for example the structure in Figs. 49 and 50. Because the S , state contains a Kramer's spin ladder (i.e., nonintegral spin system) as shown by EPR spectroscopy, the S , state was assigned as MniI'MnY and the
392
LAW, CAUDLE, AND PECORARO
S2 state was assigned as Mn"'Mn?. All of the groups using XAS have agreed upon these oxidation levels and general structural details (24, 33,154, 156, 158, 159). However, there are members of the EPR community that suggest that the oxidation levels should be adjusted downward by two oxidation states (346).The justification for this is that low-coordination-number Mn"' ions may have aberrantly high edge energies. If this is true, fits of the EPR spectra would be consistent with XAS data. However, investigations of many model compounds have not supported the idea that XANES edges are this sensitive to coordination number changes (24,33,134). One area of significant controversy is whether the Mn center is oxidized upon each S-state transition (24, 33) Klein and Sauer argued that upon forming S3,the edge energy did not increase sufficiently to support manganese oxidation (355).However, the &-state EPR spectrum disappears upon additional oxidation to S3,which indicates that the oxidizing equivalent must be in close proximity to the manganese cluster (46).These observations have led some investigators to suggest that the substrate (water) is partially oxidized or that a protein residue near the active center is oxidized. Significant excitement was generated when Boussac and Rutherford formed a new EPR signal in EGTA (Ca-depleted) samples (this signal was later shown to occur with acetate addition) (176, 356). The broad radical signal was assigned to histidine radical. However, Britt has shown recently that this signal originates from tyrosine Y;, which is the intermediate oxidized charge carrier of the enzyme (177, 178) Subsequent investigation has placed this tyrosine at <5 A from the manganese cluster. Thus, there is no longer additional evidence suggesting a nonmanganese site of oxidation. In fact, groups have suggested that the Klein and Sauer analysis is flawed based on the experimental conditions or by the fitting of the edge energies (24).Unfortunately, there still is no definitive resolution to this issue. At present, our bias is to assign oxidation to manganese at each S-state transition; however, it is possible that our views will evolve with new data. Another area of significant controversy has been the interpretation of the higher scattering in the EXAF'S. Klein and Sauer originally reported a Mn-X scatterer at ~ 3 . A3 (357).This feature was irreproducible (3551,and only when better sample preparation and detectors became available did this scatterer appear reproducibly (24,33, 156, 158, 358). Some of the contribution to this peak comes from Mn-C scattering; however, this interaction alone cannot explain the intensity of this feature (24).For this reason, the 3.3-A peak is attributed to Mn-Mn or Mn-Ca scattering. There has been great discussion as
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
393
to the identity of the scatterer, and definitive proof of the presence of calcium has not been published (477). However, there are recent reports at scientific meetings suggesting that Mn can be seen at 3.4 A in Sr2+-treatedsamples. If these withstand further scrutiny, they would provide strong evidence support for the close proximity of Ca to Mn clusters. The other cofactor often neglected in discussions of the OEC is chloride. Again, EXAFS data have been obtained to address whether or not chloride is bound to the manganese cluster (33, 358, 359). These experiments are technically very difficult. There is the possibility that chloride (or bromide in bromide-substituted samples) may be observed in the EXAFS spectrum. However, we believe that it is unlikely that one can resolve this issue, because the signal-to-noise ratio may not be sufficient (156)to observe only one chloride per four manganese ions. Very recently, chloride EXAFS of manganese model complexes has been reported (360).This new application of EXAFS may be able to resolve this issue. XAS studies of the So state have also been obtained by several groups (24, 33). These are also technically demanding experiments because the S, state is the dark stable-oxidation-state level of the OEC. Chemical reductants are used to generate So,often with light flashes required for reoxidation. Therefore, it is difficult to know precisely the amount of Sothat is formed and also whether any of the reaction centers have lost manganese by treatment with reductants. Given these caveats, it appears that the Sostate in some preparations may contain Mn" (24, 33, 361). This would suggest a total oxidation assignment of Mn"Mn"'MnF. An alternative interpretation is that there is only Mn"' and MnIVpresent in So, giving a cluster assignment of Mn&llMnw (24, 362). It may also be possible that different methods of generating So may give different relative oxidation states for the manganese in the cluster. Perhaps evaluation of the SoEPR spectrum will allow definitive interpretation of the XANES spectra. V. Reactivity
The structures and properties of manganese complexes have provided much useful information on the possible structures and metalmetal or metal-ligand interactions of manganese ions in biological systems. How enzymes carry out their respective transformations is another integral aspect for understanding the overall whole of manganese-containing redox enzymes. Model complexes have been probed
394
LAW, CAUDLE, AND PECORARO
over the years to ascertain how they mimic the reactivity of manganese-containing enzymes, with much of this work focused on the redox-active manganese enzymes, catalase, MnSOD, and the OEC. Further delving into such chemistry often unearths unexpected results and may lead to beneficial new chemistries. The following segment on the reactivity of manganese complexes will be broken into two sections. The first will cover some of the areas that compose research into the reactivity of manganese-based systems as applied to a specific synthetic organic transformation. The second segment will concentrate on systems designed specifically to model the reactivity of particular enzyme systems.
A. ORGANIC TRANSFORMATIONS Manganese has a rich history within the field of organic synthesis (363,364).For example, the permanganate anion has long been used as an oxidant to produce a variety of products (363).Manganese"' acetate also has been extensive explored over the years for the initiation of free radical reactions that lead to carbon-carbon bond formation. These topics have been reviewed and will not be presented further here (363,364).Manganese chemistry, however, has made an impact in other areas as well, notably the asymmetric epoxidation of alkenes. 1. Epoxidation
a. Salen-Based Systems (Kochi-Jacobsen-Katsuki epoxidation chemistry) One prime example of the application of manganese complexes t o organic synthesis is the burgeoning field of transition metalepoxidation catalysts based on the salen2-ligand (235, 365-367). Recent work in this area has emphasized the stereospecific generation of enantiomeric epoxides, in good yields and with high enantiomeric excesses. Such chemistry is of interest in light of the exquisite control exhibited by nature in producing such stereocenter-bearing oxidation products. Thus, one goal in exploring biomimetic oxidations is to achieve the same level of regio- and stereoselectivity as can be achieved through enzymological transformations. Furthermore, stereoselective epoxidation catalysis promises the ability to produce chiral products that are intermediates in or final products in stereospecific syntheses of use in fields such as medicinal chemistry. Since the early 199Os, manganese salen-based systems have been designed that have shown promise for the preparation of asymmetric epoxides. As is the style in organic chemistry, this important transformation has been named after key researchers in its development, and will hereaf-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
395
ter be referred to as Kochi-Jacobsen-Katsuki epoxidation chemistry. In general, previously described chemistry does not produce epoxidized products in enantiomerically pure form (235, 365-3671, especially if the epoxidized olefin's structure does not promote the formation of one enantiomer over another. Thus, many enantiomeric epoxides are often prepared as racemic mixtures. Therefore, it was necessary to design a system that would facilitate stereocontrol by orienting the substrate with respect to the oxygen atom to be transferred from the epoxidizing reagent to the substrate. Preferably, this 0x0 transfer would occur with large enantiomeric excesses. This goal has been achieved in part by the Mn-salen-based systems (235,365367). Epoxidation chemistry with a variety of metal complexes has been explored over the years. Historically, the Mn-salen catalysts grew out of research focused on understanding the reactivity of cytochrome Pd5" (368, 369). This cytochrome is employed in many enzymes that catalyze a variety of oxidations of organic substrates, including the epoxidation of alkenes (370). Usually these reactions are accomplished with complete stereochemical control. Interest in modeling this reactivity has led to a rich literature on biomimetic oxidations (365, 367369,371-375). Because cytochrome P450is a heme enzyme, the initial studies were conducted with iron porphyrin model complexes. Groves showed that one could carry out epoxidations with these model complexes in the presence of an added oxidant, iodosylbenzene (368,376378). Further studies showed that manganese porphyrin systems were even more efficient epoxidation catalysts (375, 378). Studies on these systems also contributed to the development of the oxygen rebound mechanism for oxygen atom transfer from high-valent metal0x0 porphyrins to substrate. Since the early catalysis work, a variety of MnII'porphyrin systems have been prepared with various modifications designed to improve the stereochemical control of the catalysts (discussed earlier). Due to the syntheses required to obtain such complexes, simpler, more easily prepared compounds would be desirable (235). The related Mn"'(sa1en)X family of complexes, where X is frequently a chloride anion, satisfies these requirements. A crystallographically characterized example of this type of complex is shown in Fig. 29. In 1986, Kochi and co-workers (379) reported on the epoxidation capabilities of a series of Mn"(salenlC1 complexes prepared from ringsubstituted salicylaldehyde derivatives and ethylenediamine. Their work established that these complexes were effective catalysts for the epoxidation of alkenes. Kochi was unable to crystallographically char-
396
LAW, CAUDLE, AND PECORARO
acterize the active oxygen atom transfer species, but based on the product evidence and analogy to the porphyrin systems, a MnV=O intermediate was invoked as the oxygen transfer agent. Comparison to other metal-oxo compounds capable of epoxidizing alkenes also lent support t o this proposal. Earlier, Kochi's group had explored other M"+=O complexes, such as (salen)Crv=O (380, 381). This complex could be crystallographically characterized and was shown to catalytically epoxihze some alkenes (380, 381). The Mn-salen systems showed behavior similar to that of the chromium system, in particular with the appearance of a new UV-vis band at -530 nm in the presence of the oxidant iodosylbenzene. In the chromium system, iodosylbenzene produces the Crv=O product with the concomitant change in the W-vis spectrum. Although the recently structurally characterized MnV=O complexes (255, 256) (Fig. 39) are not competent with respect to epoxidation chemistry, they do lend additional support to the existence of the MnV=0 intermediate during epoxidation. Kochi's group used spectroscopy, kinetics, and isotope labeling to explore these epoxidations and proposed a mechanism for the reaction (379). Studies of the new W-vis spectrum suggested that it was not the catalytically active compound and that it resulted from the initial oxidized manganese product (salen)MnV=0 condensing with a [Mn"Ysalen)l+ to give a (salen)MnTV-O-MnTV(salen)complex. This new complex slowly disproportionates to reform [Mn"'(salen)]+ and (salen)MnV=O. This behavior is consistent with observations made earlier for porphyrin chemistry. A recent article by Feichtinger and Plattner (382) provides additional proof for the existence of MnV=O intermediates in these systems via mass spectrometry, in which they observe mass peaks that are most consistent with the presence of MnV=O moieties. With regard to actual epoxidation, Kochi proposed -R intermediate, wherein a radithe formation of an Mn-0-C-C. cal is positioned on the carbon p from the manganese, as shown in Scheme 16. This intermediate is still invoked for the function of these R
R
\=I
! -9
Mn(V)L
R
&
-+ [Mn(lll)L]+
MnL SCHEME
+A R
R
16.
catalysts (discussed later) and is supported by the correlation of the nature of the substituent on the 5-position of the ligand's phenyl rings with the reaction rates. The more electron-withdrawing groups tended to reduce the selectivity of the complex so that it reacted at
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
397
nearly equivalent rates with a variety of substrates. Furthermore, complexes with the more electron-withdrawing groups were capable of oxidizing cyclohexane, which also suggested that a "radical-like" intermediate forms. The presence of more electron-withdrawing groups at this position also led to higher epoxide yields. This initial report also included isotopic labeling studies designed to ascertain the source of the transferred oxygen atom. When run in dry solvent with 180-labelediodosylbenzene (75% label), the norbornylene oxide product was labeled (75% label) with I S 0 . In a reaction with '*0-iodosylbenzene and 3.5 molar equivalents of added H2160-enrichedwater, the amount of l80in the product was reduced by 32%. When the labels are reversed, 160-labelediodosylbenzene, and 3.5 equivalents of H21S0 are added, then 26% of the product shows lS0incorporation. If the amount of H2180is increased to 10 equivalents, then 74% of the product shows l80incorporation. In addition, under these conditions, the oxygen of the iodosylbenzene should not be able to exchange with the water present. These results suggest, that in the normal system run in the presence of water, the 0x0 of the MnV=O may be exchanged. Such exchange behavior was reported by Kochi for the Cr systems (381). Exchange of the oxygen atom in a MnV=O complex in H21B0 has been more recently reported (256). Other groups have since picked up on this theme and have begun to explore the potential of these complexes to catalyze the asymmetric epoxidation of alkenes (374).Notable among these are the Jacobsen (383-385) and Katsuki research groups (386-388). Thus, there now exist a wide variety of complexes based on the Mn"'(sa1en)X paradigm from numerous research groups that have been reported to be competent epoxidation catalysts. These have been predominantly applied to epoxidizing cis-alkene substrates (235, 365-367). The logic behind this choice is based on unfavorable steric effects with the manganese complex. More recently, however, some trans and tetrasubstituted alkenes have been successfully epoxidized, with good enantiomeric excesses observed (367). The broad range of complexes has yielded a broad range of results as well, with overall yields that range from a few percent of desired products to well over 90%, with enantiomeric excesses that cover a comparable range (235, 365-367). Thus, with the right catalyst, a chemist may easily control the stereochemistry of a desired alkene and do so in high yield as well. A more recent report from Jacobsen's laboratories relates the ability of cobalt-salen catalysts to produce enantiomerically pure epoxides in aqueous systems with excellent stereocontrol (389).Due to the obvious breadth of this field, with its burgeoning number of catalysts and the varied applicability based on substrate, we will not attempt to assess the
398
LAW,CAUDLE, AND PECORARO
effectiveness of the numerous catalysts that have been reported. Instead, we will narrow our focus to an overview of this type of reactivity. For coverage in more depth, the reader is directed to the various reviews that have been published over the past few years (235, 365367). To conduct epoxidation chemistry, the precursor Mn"'(sa1en)X complex must be oxidized to the MnV=O state. This can be accomplished by a variety of reagents (235, 365-367). Much of the initial work in this field utilized iodosylbenzene; however, the drawbacks of this reagent have spurred intense efforts to identify more soluble and cleaner oxygen atom donors. NaOCl has emerged as the oxidant of choice in several studies, in which the reactions are often run in a two-phase format, with the substrate and Mn"' catalyst in an organic layer (often CH,Cl,) and with an aqueous layer bearing the sodium hypochlorite. Hydrogen peroxide, in large excesses, has seen limited use, whereas rn-CPBA (rn-chloroperbenzoic acid) has also been utilized at low temperatures in purely organic phase reactions. The results in systems using hydrogen peroxide are often enhanced by the addition of an N-alkylimidazole. Some newer work has employed atmospheric dioxygen as the oxidant, which is an appealing reagent for any process that might be considered for industrial scale applications. In general, the work in this field supports the notion that MnV=O is the active reagent for the epoxidation of alkenes, with the substrate presumably approaching in a side-on manner to the MnV=O moiety (235,365-367). By what mechanism, however, does the actual oxygen atom transfer occur? The side-on approach was initially proposed to account for the observed reactivity of unfunctionalized cis alkenes over trans alkenes with porphyrins (376,377). In the case of the salen systems, this route of approach has been lent support by the addition of steric bulk to these complexes and the suggested similarity to the porphyrin systems (235,365-367). The functionalization of the 3- and 5-positions of the phenyl rings of the ligand has further refined the proposed approach of the alkene as occurring across the catalyst's ligand backbone. The exact mechanism of the transfer of the oxygen atom, however, has not been fully worked out (367,390).At least two proposals have been put forward for the mechanism involved for these reactions. Arguments for both of these mechanistic positions appeared back-to-back in Angewandte Chemie articles in 1997 (391,392). The first mechanistic proposal involves the side-on approach of the alkene to the catalyst. The radical intermediate, the Mn-O-CC.-R (discussed earlier) then forms (Scheme 16). This @-carbonradical then combines with the oxygen atom, presumably leading to con-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
399
comitant product release and reduction of the manganese to the Mn"' oxidation state. Some earlier work by Jacobsen has been taken to support the potential for a concerted process in some cases, based on the epoxidation of trans-2-phenyl-1-vinylcyclopropane(393). The cyclopropane ring of this substrate should be sensitive to a radical intermediate, to give ring-opened products. Such products were not observed. Otherwise, the epoxidation product distributions generally support the notion of the radical intermediate proposed by Kochi (379). Furthermore, the epoxidation product distributions are most consistent with there having been the potential for free rotation about a previously existing alkene C-C axis before closure of the epoxide ring. Oxygen atom transfer via a concerted addition across the alkene bond would not be expected to yield products consistent with rotation about the previous alkene bond (379). A supporting experiment by Kochi et al. showed that epoxides, for example (Z)-P-methylstyrene oxide, added to a n active epoxidation system in the presence of another alkene substrate are not themselves isomerized (379). This result indicates that the observed isomerizations occurred during the oxygen transfer steps, and not after the epoxide ring had formed. An alternative viewpoint is based on the earlier work of Sharpless and co-workers from their research on oxygen transfer reactions with metal-oxo catalysts (394), for which the existence of metallaoxetane rings forming during oxygen transfer was proposed. In 1995, a proposal, based on calculations using Macromodel/MM3 (3951, suggested that such a mechanism might function in the Mn"'(sa1en) systems, as well. This mechanistic proposal (392, 395) holds that the olefin approaches the LMnV=O complex, followed by formation of the fourmembered metallaoxetane ring. One donor atom of the salen ligand is displaced, presumably a phenolate donor, into the site trans to the former 0x0 oxygen (Scheme 17). This ring then opens into a radical M-0-C-C. -R intermediate identical to that invoked for the radical mechanism and closes to release epoxide and the [Mn"'LI
0
o L R-[Mn(lll)L]+
+A
MnL
R
SCHEME 17.
R
400
LAW, CAUDLE, AND PECORARO
precursor complex. Again, bond rotation would be feasible for the M -0-C -C -R intermediate, leading to the observed mixture of products. The common factor in both of these proposals is the transient existence of M-O-C-C.-R. Finally, exploration of nonporphyrin manganese-containing systems that are competent in epoxidizing alkenes are not limited to salen-based systems (396).Tris(acetylacetonato)Mn(III) in the presence of pivaldehyde and dioxygen was observed to enantioselectively epoxidize cis-alkenes. The dioxygedpivaldehyde system has also been tested with Mn"'(salen), but appears to be less efficient overall than systems utilizing sodium hypochlorite (396). An interesting system (249) based on a [MnTV(Me3tacn)(OMe)l(PF6) complex has also been reported. Water-soluble alkene-bearing substrates were readily epoxidized by this system, in which the MnN complex is proposed to activate hydrogen peroxide a t room temperature in the aqueous solution; however, this complex is not reported to do this in an enantioselective manner.
-
b. Porphyrins The preparation of manganese-containing porphyrin systems for use as epoxidation catalysts has been intensively studied (372, 373, 375, 397). Research in Mn"'-porphyrin epoxidation chemistry predated the current generation of the salen systems (discussed earlier). The Mn"'-porphyrin systems do not appear to exhibit the same degree of broad applicability as the Mn"'(sa1en)X systems. For porphyrin complexes to function well with respect to generating an enantiomeric excess of one product, they often require that modifications be made to the porphyrin ligand to create pockets or other structures that lead to the enantioselective epoxidation of a substrate (372, 373, 375,397). Often these modifications require extensive syntheses to produce an elaborate chiral porphyrin. This is one drawback to the porphyrin systems compared with the relatively more easily synthesized salen ligands. A second drawback is that these complexes may not be as robust as the salen-based system and have been reported to decompose due to oxidation of the porphyrin (235,365-367). To combat this latter problem, olefin epoxidations with Mn-porphyrin catalysts are often run with a large excess of substrate (235, 365). This is less efficient and potentially quite costly when expensive substrates are to be epoxidized. Overall, the porphyrin systems give results with test substrates, for example styrene, that are on a par with the results reported for the Mn(sa1en) systems. They have also been shown to be functional with reagents other than iodosylbenzene such as sodium hypochlorite and dioxygen (365, 372, 373, 375, 397). An
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
401
example of a crystallized MnIII-porphyrin system is shown in Fig. 72 (398). This actual complex as shown has been utilized in aziridination reactions (discussed later), which it proved effective and catalytic. This same porphyin with a MnlI1-C1 core, instead of Mn"'(0H) (CH,OH), has been applied to epoxidation catalysis and has been shown to be one of the more efficient porphyrin epoxidation catalysts (399). The mechanisms proposed for manganese-porphyrin-catalyzed epoxidation of olefins are similar in several respects to that proposed for salen systems (365, 4001,and this was utilized initially to provide insight into the Mn-salen systems. In the case of porphyrins, a MnV=O moiety is also formed by reaction with the secondary oxidant, for example iodosyl benzene, and the olefin is then believed to approach in a side-on manner. Various intermediates have been proposed for the epoxidation of the olefin, including ones in which the alkene is polarized to give "positive" and "radical" ends, and those shown earlier for the salen systems. The concept of the metallaoxetane (371, 372) was employed to explain isomerization of some of the olefinic substrates during conversion to the epoxide products, but now
FIG. 72. Structure of the Mn'"(P)(OH)(CH,~OH).P = [5,10,15,20-tetrakis (1,2,3,4,5,6,7,8-octahydro-1,4 : 5,8-dimethanoanthracen-9-yl)porphyrin12. This complex has been tested a s a catalytic aziridination catalyst (398). This same porphyrin ligand in a n Mn"'PC1 complex has been tested a s a n epoxidation catalyst. It is one of the more efficient porphyrin epoxidation catalysts (399). [Reproduced with permission from (398). Copyright 1997 the Royal Society of Chemistry.]
402
LAW, CAUDLE, AND PECORARO
is regarded as unlikely (400). Thus, the reaction mechanism for the porphyrin systems also remains quite controversial. Porphyrin chemistry is not limited to epoxidation chemistry. There has been considerable recent interest in the toxicology of peroxynitrite, ONOO- (71, 73, 74,401).This dangerous compound is known to nitrate aromatic amino acids and has been implicated in processes and disease states as far ranging as amyotrophic lateral sclerosis (ALS) (402), Parkinson's disease (71, 72), and cell apoptosis. Presumably peroxynitrite is formed by the rapid condensation of superoxide with nitric oxide. It is now known that peroxynitrite can be efficiently destroyed by amphiphilic manganese porphyrins in liposomal assemblies (403). The process is believed to proceed by the reaction of Mn'" (porphyrin) and ONOO- to give Mnv=O(porphyrin) and nitrite, NO,. The nitrite can then react rapidly to give a MnI"(porphyrin) and NO2".Subsequent electron transfer to biological reductants destroys the NO2",and regenerates the catalyst (403,404). 2. Nitrido Group Transferlhiridination A new entrant on the high-valent manganese reagent scene are the manganese-nitrido complexes (251-253). Only a handful of these have been structurally characterized to date, but their potential as a nitrogen-transfer reagents is being explored. The initial work in this field again resulted from experiments with porphyrin-based systems. In 1983, Groves and co-workers generated a Mnv-nitrido compound and showed that in the presence of trifluoroacetic acid anhydride, this compound could transfer a nitrogen to produce an aziridine on cyclooctene (405), the nitrogen analog of an epoxide. The nitrogen atom of this ring bears a trifluoroacetyl moiety. One promising recent example of a catalytic porphyrin based aziridination system was reported in 1997 (Fig. 72) (398). This compound produced aziridination products from alkene precursors, for example styrene, with the aziridine nitrogen bearing a tosyl group from the co-oxidant (tosyliminoiodo)benzene, PhINTs. Up to 480 turnovers were reported, with reported yields ranging from 44-76% and enantiomeric excesses between 44-66%. The newer nonporphyrin Mn-nitrido systems react in a stoichiometric manner and are not catalytic like the MnV=O systems. Catalytic systems are now a focus of research with the newer systems (253). Despite the current lack of catalytic turnover, MnVnitrido systems promise to yield new, useful and easier pathways to a number of amine or aziridine products. Applications of these complexes to the amination of carbohydrates may allow these compounds
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
403
to be more readily synthesized in reactions that employ milder reagents than might otherwise be utilized (253). Initial studies have shown that the transfer of the nitrido nitrogen atom moiety to a n organic substrate from a nonporphyrin MnVnitrido complex is feasible (251). Both the complex with the underivatized salen2--ligand complex, as well as the tetramethylsalen2-, saltmen2-, with four methyl groups appended to the two backbone carbons of the ligand, were prepared. Figure 37 shows the crystal structure of this complex. The latter was employed in nitrogen-transfer reactions due to its greater solubility in organic solvents and was found to be a competent nitrogen-transfer agent to a variety of silylated enol substrates in the presence of trifluoroacetic acid anhydride. This reaction produced a-aminoketone products, with the product amine acetylated with a trifluoroacetyl group. Studies on these systems continue, with the ultimate goal being the production of a complex that is competent for asymmetric nitrogen transfer, presumably in a manner similar to the Kochi-Jacobsen-Katsuki epoxidation format. Another interesting result in this field is the report that the nitrogen of a (sa1en)Mn"GN to another Mn"'(5-MeOsalchxn)Cl complex, where chxn represents 1,2-diaminocyclohexane (406). This process was shown to be reversible by the addition of the Mn"'(sa1en)Cl to a solution of (5-MeOsalcxhn)Mnv=N, wherupon the N3- could be transferred "back" to regenerate the "original" (salen)Mnv= N complex. Related work has shown that a nitrido can be transferred from a MnV nitrido(p0rphyrin) complex to a CrIYporphyrin) complex to yield [Mn"'(porphyrin)l+ and Crvnitrido(porphyrin) (407). A note on the synthesis of Mnv-nitrido complexes seems warranted at this point. There now exist several reports of first-row transitionelement nitrido complexes (4081, of which a growing number are manganese complexes. There are also a broad range of syntheses available. Manganese-nitrido complexes have been reported via the irradiation of manganese-azido precursors in several cases, although the yields vary (258,405,409).Other techniques employ ammonia or ammonium in the presence of a chemical oxidant (251-253). The work of Carierra et al. (251) with Mn"Ysa1en) derivatives employed ammonium hydroxide and sodium hypochlorite to generate the nitrido complex. In newer complexes with bidentate ligands that expand this class of compounds, the nitrido complex is prepared via a synthesis that utilizes N-bromosuccinamide as the oxidant with the addition of gaseous ammonia to the Mn-compledN-bromosuccinamidesolution at low temperature (253).
404
LAW, CAUDLE, AND PECORARO
3. Chlorination
Many routes for the chlorination of alkenes exist, and one new entry in recent years has been manganese-based chlorinating reagents. Many of these systems have been prepared by Bellesia and co-workers (410414). These systems in general produce trans-chlorinated alkenes. In only one system has the proposed halogenating complex been structurally characterized (245). Most of the systems explored for the chlorination of alkenes with manganese have involved high-valent manganese compounds or the known oxidant Mn"' acetate. The systems reported by Bellesia et al. (410-412) have been tested against a substantial body of alkenecontaining substrates. One system that these researchers devised was based on a mixture of Mn11102and Me3SiC1in THF (410, 411), and a later system was devised that employed MnrV02,Mn"C12, and acetyl chloride in DMF (412).This group proposed that a transient "MnIVC12 or "Mn*I1C13,,acts as the reactive halogenating reagent in some of their systems. They also showed that such a Mn-based system could chlorinate ketones in the a-position (415)under milder conditions than earlier work (416).Two other similar systems have also been reported, the first based on Mn"'acetate, CaC12 and acetyl chloride (417) and the other on KMn04 and oxalyl chloride (418, 419). A variety of mechanisms may be functioning for these different systems. In the systems of Bellesia and the Mn"' acetate system of Donnelly et al., it was proposed that these reactions occurred via a nonchain radical mechanism (411, 417). The product distributions observed are most consistent with chlorination by a radical pathway and not by an ionic pathway. This is best demonstrated by norbornylene (420-422). The products of chlorination for these systems were trans-2,3-dichloronorbornaneand a much smaller amount of cis-2,3dichloronorbornane, and no other products were observed. The latter compounds would result from the generation of a nonclassical carbocation during chlorination. The lack of such products speaks against an ionic intermediate, and the observed products are consistent with chlorination having occurred via a radical-based halogenation. A mechanism involving radical character in this system is given credence by the known chemistry of Mn"' acetate, which is known to initiate radical-based chemistry (363, 364). In 1997, Marko and co-workers (423) followed up on their initial report of the KMn04 and oxalyl chloride system. This new system consists of KMn04, Me3SiC1, benzytriethylammonium chloride with dichloromethane as the solvent. It gives good yields of dichlorinated alkenes. The permanganate anion is brought into solution by the ben-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
405
zyltriethylammonium chloride, and the addition of four molar equivalents of MesSiCl produces the halogenating reagent, which they propose to be a dichlorobis(trimethy1silol)manganesecomplex. None of these systems presents an initial, structurally characterized halogenating agent. Recently we reported that the MnTV(salpn)C12 complex (245),which had been crystallographically characterized previously (Fig. 33), acted as a new reagent capable of halogenating alkenes in a trans manner. In this case a stoichiometry of two dichloro complexes to one alkene was required, and the final manganese product has been identified as Mn"Ysa1pn)Cl. The products formed with this reagent are consistent with a nonchain radical mechanism. B. ENZYME MODEL SYSTEMS 1. Manganese Peroxidase The reactivity of the manganese peroxidase (MnP) system has been explored with model compounds. The chemistry of MnP is the oneelectron oxidation of the lignin substrate that initiates decomposition 'of the lignin structure. Some of the initial functional modeling chemistry in this field established the ability of Mn"' complexes to initiate the decomposition of lignin analogs, for example, the decomposition of dimeric lignin models in the presence of Mnl" and a-hydroxy acids (95, 103).The products from these experiments involving Mn"' lactate and a variety of other a-hydroxy acids showed that Mn"' was indeed capable of initiating the one-electron oxidation of lignin. Another study with the lignin analog, and test substrate for Lip and MnP, vanillylacetone also showed that these systems were capable of such oxidations (424).Furthermore, the lactate was proposed to stabilize the Mn"' to prevent generation of Mn oxides before the manganese ion could act on the lignin, yet would not produce a complex so stable as to prevent the Mn"' lactate complex from oxidizing lignin. More recent work has shown that lignin-degrading fungi produce oxalate, which is required for the enzyme to function. Based on these model studies and data, it has been proposed that oxalate may stabilize the Mn"' to prevent loss of Mn"' by disproportion to Mn" and MnIVO2. In these studies, the manganese complex was generated in situ from added manganese and ligand. In 1991, Saadeh et al. reported the preparation and isolation of complexes with a-hydroxy acids, for example lactic acid, 2-hydroxyisobutyric acid (HIB) and 2-hydroxy2-ethylbutyric acid (HEB) (425).The latter two ligands led to Mnw complexes with 0, first coordination spheres, whereas Mn"' complexes were prepared and characterized with lactic acid. The crystal structure of the complex with HIB shows it to be two MnwL2units that are
406
LAW, CAUDLE, AND PECORARO
bound to a single sodium through the facial array of hydroxy oxygens in the complex. The monomer is shown in Fig. 34. Among the other complexes prepared were Mn"' dimers with chelated and bridging lactates or malonates (425). These isolated and well-characterized complexes were then utilized to explore the ability of such a-hydroxy acid-manganese complexes to oxidize the lignin analog vanillylacetone. A scheme representing the Mn"' system is shown in Scheme 18 (425).It was found that the Mn'" Vanillylacetone
+
-
2 H 2 0 + 4 Mn3+
do
HO
Pyruvaldehyde
+
Vanillin
+
4Mn2+ + 4 H +
OCHa
+ 0
SCHEME 18.
complex with HIB and the Mn"'-lactate complexes were both capable of oxidizing this organic substrate. The oxidation product, pyruvaldehyde, is consistent with the oxidation of the substrate via successive one-electron oxidations by manganese (425).The reaction is proposed to begin via the generation of a phenoxy radical, which then rearranges to form a ketone from the phenol alcohol and places a radical alpha to the ketone. Two successive Mn"'-initiated oxidations lead to the formation of pyruvaldehyde. A final oxidation of the remaining substrate radical, followed by incorporation of the second molecule of water and a rearrangement, yields vanillin. Furthermore, Mn" was identified by EPR spectroscopy in these reaction mixtures. This lends additional strong support to the one-electron oxidant role for manganese in MnP. A second area of MnP model chemistry has dealt with Mn" oxidation by Fe-porphyrin systems. An early example suggested that Mn" in the presence of excess pyrophosphate could be catalytically oxidized to Mn"' by an Fe-porphyrin system. In these experiments, sulfonated porphyrins were utilized, and the co-oxidant was potassium monopersulfate. Co-substrates, such as veratryl alcohol, were reported to enhance the rate of Mn" oxidation (468). Another interesting reactivity-related system is the recent report of an iron-containing porphyrin that has been functionalized with propi-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
407
onate arms, patterned from the crystallographic data for MnP (426). This system was designed to study electron transfer from the iron/ porphyrin to the manganese ion. The porphyrin bears pentafluorophenyl groups on three of its meso positions, and the fourth meso position is modified by the addition of an alkyl arm that terminates in a 6-(methylamino)-2,2'-bipyridine functional group that is intended to be the site of manganese binding. Among other characterization data, EPR spectra confirm the presence of Fe"' and Mn" in the final product, and the apparent lack of magnetic interactions between the two ions suggests that they are not coupled t o one another. In the presence of the oxidant pentafluoroiodosylbenzene, the Mn" EPR signal was observed to slowly disappear. The addition of 2,6-dimethylphenol, a standard test substrate for MnP, led to the formation of the oxidation product 2,2',6,6'-tetramethyldiphenoquinone and the reappearance of the Mn" EPR signal. This appears to be fully reversible and was shown to function in a catalytic manner. In control experiments, neither Mn'I bound to the iron-free modified porphyrin nor Mn1Ybipy),C12 were oxidized by pentafluoroiodosylbenzene. In the presence of a (chloro)[tetrakis(pentafluorophenyl)porphyrinliron complex and pentafluoroiodosylbenzene, MnYbipy),Cl:,was observed to be oxidized. These results are consistent with the petafluoroiodosylbenzene oxidation of the ironlporphyrin portion of the complex with ensuing electron transfer to oxidize the bound Mn". The Mn"' thus produced would then oxidize the substrate. 2. Manganese Superoxide Dismutase
The various complexes that have been tested for their ability to mimic MnSOD form another set of systems that have been the focus of reactivity studies. One emphasis of research in this area has been the potential for the development of pharmaceutical agents that are good SOD mimics. This research is driven by the proposal that much of the damage following a shortage of oxygen due to a stroke or heart attack may be the result of a build-up of superoxide and its hazardous by-products (discussed earlier). The goal, therefore, is to design agents to scavenge superoxide and to thereby prevent an appreciable build-up of this molecule. Much of this work has centered on manganese, because this element is less likely to be able to catalyze other side reactions, such as Fenton-like chemistry (221,222, 427). The manganese complexes that have been prepared to date cover a range of structural types and ligands. Due to the self-dismutation of superoxide (2.0-3.2 X lo5 M-I. 5-l ) (52, 631, these complexes need to be quite efficient to qualify as catalysts of superoxide dismutation. Several complexes have proven to be competent MnSOD mimics. Un-
408
LAW,CAUDLE, AND PECORARO
fortunately, one of the main techniques once utilized t o establish SOD activity suffers from potential interference, so early examples of catalytic MnSOD mimics may not actually be catalytic in nature. Thus, from all that has been published, it is clear that there are roughly two classes of compounds: those that are efficient and catalytic mimics for MnSOD, and those that interact with superoxide but are not necessarily catalytic in their interaction. In 1993 Weiss, Riley, and co-workers reported a study on purported SOD mimics by stopped-flow UV-vis spectroscopy (428) in which they assessed reactivity by following the decay of the superoxide absorption at 245 nm. Two of the earlier techniques that had been used to assess SOD activity included observation by W-vis spectroscopy of the oxidation of nitroblue tetrazolium (NBT) (68) or the oxidation of a cytochrome c by superoxide (52). Both systems used superoxide from an in situ generator, frequently xanthine oxidase, wherein the complex being analyzed was compared to a calibrated oxidation of the chromophore alone and in the presence of MnSOD. The direct observation of the decrease in the superoxide signal with time by UV-vis is also possible, and superoxide may be introduced as a solution (428) or generated, in some cases, by pulsed radiolysis (79,801. In these direct observation experiments, the rate of decay of superoxide in the presence of the complex is compared to the rates of decay of superoxide alone and in the presence of one unit of activity of MnSOD. In all cases, the systems are usually referenced, or calibrated, against the same set of conditions with MnSOD. Due to interactions with cytochrome c with components of assay mixtures other than superoxide, false readings of activity were often observed for some early SOD mimics. The NBT, stopped-flow, or pulsed radiolysis techniques have tended to provide the more accurate answers on the ability of reputed MnSOD mimics. To be considered active in any manner with respect to the decay of superoxide in the stopped-flow analyses, Weiss et al. stated that compounds based on their analyses needed to exhibit keat values in excess of 105.6M-’ * s-l (428). The complex of manganese with desferrioxamine B (429432), reported to be a MnSOD mimic in 1987, illustrates this point. In the initial studies, the cytochrome c assay was applied, and this compound was reported to be a catalytic MnSOD mimic. Follow-up studies by Riley and co-workers by stopped-flow techniques showed that this complex interacts at best on a stoichiometric level with superoxide, but it could not be considered to be a catalyst and was termed “inactive” (428). This occurred due to interactions with cytochrome c wherein the manganese complex interfered with the reduction of cytochrome c , leading in effect to a false positive. Thus, although some
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
409
systems may no longer be considered to be catalytic, many of these still may exhibit interactions with superoxide. Despite earlier systems that eventually proved not to be catalytic, catalytic MnSOD mimics have been successfully prepared with several examples appearing since 1990. One of the first appeared in 1993 by Kitajima and co-workers (208) and is based on the ligand hydrotris(3,5-diisopropylpyrazol-l-yl)borate,HB-i-Pr. The two fivecoordinate MnSOD mimics that were prepared in this report are the Mn" complex with one HB-i-Pr ligand and one bidentate benzoate, and another Mn" complex composed of one HB-i-Pr, one unidentate benzoate, and one equivalent of 3,5-diisopropylpyrazole.The crystal structure of the latter is shown in Fig. 13. These complexes yielded IC,, values of 0.75 and 0.80 pmol-dm-3, respectively. These values were determined by observing the inhibition of the reduction of nitroblue tetrazolium by superoxide in a calibrated system using xanthine/xanthine oxidase as a superoxide generator. The IC,, value implies that a concentration of that particular reagent exhibits activity equivalent to 50% of that observed for one unit of actual MnSOD enzyme under the specific conditions of the experiment in question. For comparison, a n iron complex tested by Weiss et ul. (428) that was shown to be "active" in the presence of superoxide exhibited a k,,, value of 2.15 X lo6 M-'. s-l, and a n ICsovalue of 2.1 pmol .dm-3. Riley, Weiss, and co-workers have prepared a series of MnSOD mimics using macrocyclic ligands (221,222,427).The first report with several simple macrocycles indicated that the seven-coordinate Mnl' complexes with the 15- and 16-member pentaaza macrocycles and axial chlorides were efficient catalysts for the dismutation of superoxide (Fig. 20) (221). At pH 8.1, the pentaaza and hexaaza complexes exhibited kcatvalues of 2.2 X lo7 M-' * s-l and 1 x lo6 M-'. s-l, respectively. At a physiological pH, pH = 7.4, only the complex with the 15-member ring was active. The complex with the 16-member ring proved to be quite pH sensitive. At pH 7.4, the efficiency of the functional complex increased to a kcat value of 4.1 X lo7 M-' s-' . This initial work was followed up by modification of the periphery of the macrocyclic ligand to create a family of new MnSOD mimics (222). These complexes were then screened by stopped-flow UV-vis spectroscopy. All of the complexes tested catalytically decomposed superoxide, with kcatvalues ranging from 1.4 to 9.1 X lo7 M-' . s-' a t pH 7.4. The best of these was dichloro(truns-2,3-cyclohexano-1,4,7,10,13-pentaazacyclopentadecane)manganese(II). A third system is based on Mn"'-Schiff-base complexes. There are two reports of reactivity of such systems with superoxide. The first, by Matsushita and Shono, appeared in 1981 (433). In this study, the
410
LAW, CAUDLE, AND PECORARO
authors observed that in the presence of superoxide, Mn"'LC1 complexes, where L is a Schiff-base ligand such as salen2-, and in KOd complex ratios <2-3, oxygenated products were observed for complexes with Mn(IIUI1) reduction potentials of about -0.19 V or lower vs. a Hg pool (approximately 0.21 V vs. SCE). Complexes with more positive reduction potentials, ranging from -0.18 to 0.00 V vs. the Hg pool, were reduced, but oxygenated products were not observed. In 1993, Baudry et al. examined Mn"'(salen)-based systems with ligands that had been designed for the Mn"'(sa1en)Cl epoxidation systems (434).The activity of these complexes considerably varied, from rates that were nearly undetectable to rates indicative of fairly efficient MnSOD mimics. One of these Mn"'(salenjC1 complexes, with the salen ligand comprised of a 1,2-diaminocyclohexane backbone and 2-hydroxy-3-tert-butyl-5-methoxy-benzaldehyde, gave a n ICb0value of 0.32 Fmol. dm-3, or 3200 U/mM. The activity in this system was measured against the reduction of the nitroblue tetrazolium absorption. In 1997 it was reported that underivatized Mn"'(salen)Cl, referred to as EUK-8, was also a competent superoxide scavenger (435). A Mn" complex with a n all-nitrogen first-coordination sphere was reported in 1996 to be a competent MnSOD mimic (436).In this case the complex is composed of two meridonally binding 2,6-bis(benzimidazol-2-ylj-pyridine ligands. When tested against the NBT system, this complex was reported t o exhibit a n LCs0value of 0.72 pmol dm-3. This level of activity is comparable to that observed for the other systems tested by the nitroblue tetrazolium system. Most manganese complexes that have been reported to react with superoxide, however, do not react catalytically. Several have been reported by various groups in attempts to generate catalytic systems, but these only show stoichiometric interactions with superoxide. Examples of this group of complexes includes the recently reported seven-coordinate Mn" complex, with the same ligand and a structure similar to the complex shown in Fig. 25 (229). In conclusion, systems have been generated that do interact catalytically with superoxide and function as MnSOD mimics. The potential for errant readings of activity with the indirect assays of SOD mimics shows the need for careful control experiments, and perhaps the wider application of direct techniques for the determination of MnSOD mimicry to prevent false positives and to provide readily compared kcatvalues. 3. Catalase
Work on the catalase frontier has focused on the synthesis of binuclear manganese complexes that are capable of catalyzing the dispro-
411
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
portionation of hydrogen peroxide. This field has grown considerably over the past few years, and a large number of catalase mimics are now known. Catalase and catalase model systems have been reviewed previously ( 3 , 5 , 2 3 , 2 5 , 2 6 , 2 9 , 3 0 , 4 3 6 In ) . general, two types of catalase chemistry may be mimicked by manganese model complexes. The first is the reactivity of the catalase itself. The second is the catalaselike reaction thought to occur in the OEC between the Soand S2states of that enzyme (437).Several model systems have been prepared, and a selection of these will be described in some detail. A table of rate data, Table 111, is presented so that these systems may be more readily compared. A few systems not discussed here are also presented in the table. One of the earliest functional mimics of catalase was reported by Mathur et al. in 1987 (438). This system employed the ligand N , N , N ,N' - tetrakis - (a- methylenebenzimidazolato)- 1,3- diaminopropane-2-01. The complex formed with this ligand is composed of one ligand, with the alkoxide bridging the two manganese(I1) ions, with a proposed bridging chloride and two terminal chlorides, one on each TABLE I11 CATALASE ACTIVITY: ENZYMES AND MODELSYSTEMS" Enzyme or complex Thermus thermophilus catalase
260,000
200,000 26,000 250 150 22 13.2 5.5 5.4 4.4 1.2 [Mn~'(20H-benzimpn)(p-OAc)(n-BuOH)I(CI04), 1.04 (BaL+,Ca2),[Mn~llMn"(p-O)(p-OH)(p-OAc)L(tapn),l 0.79 lMneLh(p-C,H,COL)Z(NCS)I 0.26 LMn"'Mn'"(bpg' )l(p-O),I(C10,) 0.013 [Mn"'(tetraphenyIporphynn)l 0.0063 [MnY H,O)&C10, )1
Lactobacillus plantarum catalase Thermoleophilum album catalase 1Mn"'(sa1pn)(p-0)l2 [MnW-OHpicpn)J,(CIO,), IMn"(2-OHsalpn)l, [Mn"lMnl\(p-Oj&-OAc)(tacn)( bipy)(MeOH)1(C104)1 [ Mn"lMnl"(p-0 )2(p-OAc)(tacn )( 0Ac)z l(CIO,)L [Mn]" (anthracenediporphyrin)lC12 lMngLh (p-3,4-N01benzoate)1(C10,)
3.2
X
lo6
5.7 x 1 0 5 1.7 X lo6 1 x 10J 70 350
131, 295, 461, 462 23 129 446 333 440 294 294 443 463 439 447 441, 442 464 443 440
" T h e rates observed for these systems are for conditions which vary from study to study, whereas the k,,, values are presumably maximal rates with saturating hydrogen peroxide. " L = the macrocyclic 2 : 2 Schiff base condensation of 2,6-diformyl-4-methylphenol with N,Nbis(2-aminomethyl)methylamine. bpg = bis(N,N-(2-pyridylmethyllglycine. 1
412
LAW,CAUDLE, AND PECORARO
manganese. This proposed structure is similar to that of the related structure shown in Fig. 53. In the absence of crystallographic data, EPR spectroscopy was used to suggesk the binuclear structure, with a Mn-Mn distance on the order of 3.3 A. When reacted in a two-phase system of complex dissolved in butanol and a hydrogen peroxidebearing water layer, catalytic dioxygen evolution was observed. Over the course of the reaction, a Mn$ dimer built up, which was proposed to be an oxo-bridged product. This product could be independently prepared for comparison to the product of the catalase-like reaction. As the reaction continued, the pH rose, and the loss of reactivity over time was attributed to this increase in pH, which led to complex decomposition. The proposed mechanism for this catalase mimic involved the initial formation of an oxo-bridged Mn$ complex from the Mn!j complex and the formation of one molecule of water. The second step involved the reaction of this new Mn"'-O-Mn"' compound with hydrogen peroxide to reduce the complex to the Mn# state, with the concurrent formation of a second molecule of water and the release of dioxygen. The progress of the reaction was followed by collecting the evolved dioxygen. In 1994, the chemistry of this system was revisited with an analogous complex (291,439),one in which the chlorides had been replaced by a bridging acetate and two nonbinding perchlorates. One of the manganese in this structure is five-coordinate, and the other is six coordinate, with a butanol occupying the sixth coordination site. The crystal structure of this complex gives credence to the structure that had been proposed for the catalyst with chlorides (438).The exploration of this system led to a new mechanistic proposal. In this case, the lag time for observed catalase activity was reduced by the addition of 2% water to a methanolic solution of the catalyst. They proposed that the water displaces the bridging acetate; thus, a molecule of hydrogen peroxide may more easily displace the terminal waters than the bridging acetate to bridge the two manganese ions itself. The 0-0 bond of the peroxide was then proposed to be cleaved, producing two Mn"'-OH moieties. This intermediate reacts with a second molecule of hydrogen peroxide to release dioxygen and to regenerate the Mn" complex. The terminal hydroxides were protonated during this second step to form two molecules of water, thus completing the disproportionation of hydrogen peroxide (Scheme 19). In 1992, Wieghardt and co-workers reported two manganese catalase mimics based on triazacyclononane ligands (294). These were asymmetric complexes, with one half bearing an N,N',N"-trimethyl1,4,7-triazacyclononane molecule bound to a manganese that was bridged via two p-0x0 bridges and one p-carboxylato bridge to either a manganese bearing a bipyridine ligand and a terminal water or
413
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
2H 2 4
H202
SCHEME 19. Pronosed mechanism of hydrogen peroxide disproportionation by [Mn"(2-OHbenzimpn)(p-OA~~l~~+. [Reproduced with permission from (439). Copyright 1994 the American Chemical Society.]
methanol or one coordinated only by additional acetates. The former complex is shown in Fig. 73. The oxidation state of these systems is Mnl*IMnN,and presumably they cycle to Mn"MnlI1. In sodium acetatebuffered solutions, approximately 1300 turnovers were realized during the first three minutes of these reactions, with rate values of 13.2 and 5.5 s-l, respectively. Overall, both of these catalysts were five orders of magnitude slower than the catalase enzyme. In 1993, Gelasco and Pecoraro (273) reported initial data for a series of complexes based on the 2-OHsalpn ligand, N,N'-(salicylideniminel- 1,3-diaminopropan-2-01(Fig. 48). These complexes are effective catalase mimics. With all of the biologically relevant oxidation states of the catalase enzymes, MnP, MnIIMnIII, Mniu, and MnlIIMnTV, having been prepared (279). This is the first series of manganese complexes to exhibit this range of oxidation states with the same ligand.
FIG. 73. Structure of [(Me,tacn)Mnl"~p-O~~(p-OAc~Mn"'(OAc~,l. [Reproduced with permission from (294). Copyright 1992 the Royal Society of Chemistry.]
414
LAW, CAUDLE, AND PECORARO
Furthermore, this family of complexes exhibited EPR spectra that modeled those observed for the catalase enzymes (26, 279). The [Mn(2-OHsalpn)12complexes disproportionate H202by shuttling between the MniI and the MnkI' oxidation levels. Under appropriate conditions in the presence of hydrogen peroxide, either the Mnf or the Mn;" complex may be produced cleanly. The observed rate for the disproportionation ranged from 4.2 to 21.9 s-l, depending upon the derivatization of the complex's phenolate moieties. The rate-limiting step was determined to be the reduction of the Mn;" dimer to Mnf. Thus, complexes with the more electron-withdrawing ligands proved to be the most efficient catalysts in this system, because this Mn;" oxidation state was less stabilized. The complexes with the more electrondonating ligands, in contrast, tended to bind peroxide better. In reactions with hydrogen peroxide (4401, members of this family of complexes exhibit saturation kinetics. It was observed that the nature of the substituents on the phenolates affected the kcat and K,,, values of the complex with respect to hydrogen peroxide disproportionation. It has been proposed (440) that this system involves terminal binding of the hydroperoxide anion, HOO-, to a [Mn"'(20Hsalpn)12 dimer, in a manner similar to that observed in crystal structures of these compounds, with, for example, methanol or THF (Fig. 44). In this case, the hydrogen peroxide is deprotonated by a base equivalent and is bound to the dimer. One of the alkoxide bridges shifts to become a terminal instead of a briding ligand to open a coordination site for HOO- (Scheme 20). A second base equivalent then deprotonates this hydroperoxide monoanion unit, the dimer is reduced to Mnf, and dioxygen forms and is released. A second equivalent of hydrogen peroxide then bridges the Mnkl dimer, followed by reoxidation of the dimer with the concomitant reduction of this second equivalent of hydrogen peroxide to form two equivalents of hydroxide. These hydroxide molecules are base equivalents as described earlier, so that the two water molecules of hydrogen peroxide disproportionation are formed. Isotope-labeling studies showed that the dioxygen evolved arose from the same molecule of hydrogen peroxide. Okawa and co-workers have published several papers in recent years on binuclear catalase mimics based on the pentadentate or septadentate binucleating ligands, an example of which is shown in Scheme 21 (441, 442). These ligands have been used to prepare a variety of binuclear MnH complexes. Their ability to catalyze the disporportionation of hydrogen peroxide varies, but is generally low, on the order of 0.79 s-l. A functional scheme for the disproportionation of hydrogen peroxide has been proposed for these complexes that invokes a Mn'"=O intermediate (Scheme 21).
415
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
SCHEME 20. Proposed mechanism of hydrogen peroxide disproportionation by the [Mn(S-OHsalpn)j, system. [Reproduced with permission from ( 4 4 0 ) .Copyright 1998 the American Chemical Society.]
Another unique dimeric system was first reported by Naruta and Maruyama in 1991 (443-445). In this system, two Mn"'-porphyrin molecules are linked by an anthracene spacer instead of the o-phenylene spacer seen in Fig. 75 (discussed later). Their studies showed this H
H\
,L g(,,,) . 2y 6 H202
Mn(II)Mn(II) Complex
H202
7-O;
&,)
H
0 I I Mn(LI1) Mn(II1)
k4 (" N
'\
OH
"1 Y\
12H2
02-1
P-7
H
H
0
0
in(N)
hnw)
I
I
0
0
7 in(IV) Mn(N) II H202
SCHEME 21. A proposed mechanism of hydrogen peroxide disproportionation by Okawa, et al.'s system. [Reproduced with permission from (466). Copyright 1995 the Royal Chemical Society.]
416
LAW, CAUDLE, AND PECORARO
complex to be a competent catalase mimic, with a rate of 5.4 s-' observed for the disproportionation of hydrogen peroxide. It was proposed that this system functioned by the formation of two MnIV-OH porphyrin groups or high-valent Mn-oxo porphyrin groups in the presence of HzOzand imidazole, which likely acts as a requisite base. Hydrogen peroxide will then react with this intermediate to yield dioxygen, two molecules of water, and the reduced form of the complex. Isotopic labeling studies showed that isotope mixing was not observed in 1: 1 mixtures of l60and la0hydrogen peroxide, This was followed by an article in 1997 (445), wherein Naruta et al. prepared a dimer with MnIvporphyrin(OMe)(OMeor OH). The analysis of this system indicated that catalase activity occurred via approach of hydrogen peroxide to the pocket of the dimer, followed by deprotonation and concomitant reduction of the dimer and oxidation of the hydrogen peroxide to dioxygen. Their proposed catalytic cycle is shown in Scheme 22. Similar systems have been used to generate dioxygen from water (discussed later). 8:
8:
I
O2
i
B:
1
py;+, c; B: I
B:
8:
I
Mn(IV)
H?'
B:
-
~ - 0
yn(lv)
-
H202
Mn(lV)I
B:
0: SCHEME 22. A proposed mechanism of hydrogen peroxide disproportionation by Naruta, et al.'s system. [Reproduced with permission from (445).Copyright 1997 Elsevier Science.]
The most efficient catalase mimic to date is the [Mnrvsalpn(p-O)la complex (Fig. 50) (244, 446). This complex cycles between Mn;" and Mn"'Mn"'. The latter oxidation state is represented by two monomeric units, as exhibited by ligand-scrambling experiments. This was determined in catalase mimicry experiments by utilizing two complexes with differing ligand derivatization. When they were reacted with hy-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
417
drogen peroxide and isolated, mass spectrometry showed that the resulting material contained manganese dimers with both derivatized ligands in one molecule. In experiments without added hydrogen peroxide, no ligand mixing was observed. The ability to form these [MrP salpn(p-0)I2complexes from [Mn"'LI+ precursors also supported this supposition. The dioxygen produced during the reaction with hydrogen peroxide arises from the same molecule of hydrogen peroxide, based upon isotope-labeling studies with labeled hydrogen peroxide. The reoxidation step of this complex's catalytic cycle involves the incorporation of a hydrogen peroxide molecule into the compound in the form of the new 0x0 bridges. The rate a t which this complex is capable of disproportionating hydrogen peroxide is 250 s-l, and saturation kinetics are observed. A proposed catalytic cycle for this system is presented in Scheme 23. Although, this complex shuttles to a higher oxi-
t
H20
0
I
LMn(lll)L +
HO
/Mn(l'')L
SCHEME 23. A proposed mechanism of hydrogen peroxide disproportionation by [Mn'V(salpn)(p-O)]p.[Reproduced with permission from (27). Copyright 1992 WILEYVCH Verlag.1
418
LAW, CAUDLE, AND PECORARO
dation state than is proposed to exist in the catalase, its structural similarities t o the OEC make it an excellent model for the catalaselike reaction of the OEC that occurs between the So and S2 states of that enzyme. Another class of catalase mimics are comprised of tetranuclear complexes, of which at least two examples are known. The first example was reported in 1990 by Stibrany and Gorun (447). In this complex (Fig. 741, two manganese ions are bridged by the alkoxide from the ligand 1,3-diamino-2-hydroxypropane-N,N,N, W-tetraacetic acid. The tetramer is then formed when these dimeric units are bridged t o one another by two acetates and by one 0x0 and one hydroxo bridge, which hydrogen bond. The overall oxidation state of this compound is Mn"Mnbl'. This complex was shown to have a rate of hydrogen peroxide disproportionation on the order of 1.04 s-l. A second tetramer was reported by Gelasco et al. in 1995 (333). In this complex, four
FIG.74. Structure of ~ M n ~ J ' M n 1 1 ~ ~ - O ~ ~ ~ - O H ~ ~ ~anion; - O A cL) ,=~ L1,3-diamino),12~ propan-2-ol-N,N,N',N,-tetraaceticacid. CaZ' or Ba2+were reported as counterions. [Reproduced with permission from (447).Copyright 1990 Wiley-VCH Verlag.]
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
419
manganese ions are bridged by the ligand N , N -( picoly1ideneimino)1,3-di-aminopropan-2-01, 2-OHpicpn. This complex appears approximately box shaped (Fig. 63), with one ligand lying along each edge of the box. The oxidation state of each manganese in this complex is Mn”. This complex disproportionates hydrogen peroxide a t a rate of 150 s-’, making it the second most efficient catalase mimic yet reported. It also exhibits saturation kinetics. 4 . Water Oxidation
Dioxygen is at the heart of many biological processes, and without the generation of dioxygen by green plants and algae, the world be a much different place. The need for a catalase or a n MnSOD results from an aerobic existence, for example. The overall oxidation of two molecules of water to form dioxygen is a four-electron process. Although the oxidation of the OEC itself occurs by four single-electron events, it has not been determined whether the final oxidation of water to dioxygen is a four-electron process or one that occurs in smaller steps, either by sequential one-electron oxidations or by 2 twoelectron processes. One newer and intriguing proposal has suggested that this may occur via hydrogen atom abstraction from water molecules bound to a n OEC manganese ion (182). That would eventually produce a higher-valent manganese/oxygen moiety such as Mn” =0 prior to a final step that forms and releases dioxygen (Schemes 12 and 13). Water oxidation by a variety of inorganic systems (26) and considerations, such as thermodynamics, for the oxidation of water to dioxygen by the OEC have been recently reviewed (37).Model chemistry related to the viability of hydrogen atom abstraction is discussed in the Section II.G, “The Oxygen-Evolving Complex.” We will focus on well-defined manganese model complexes that have been implicated in water oxidation as potential models for the function of the OEC. Although the OEC’s tetranuclear manganese cluster is most likely not light-activated in and of itself, with respect to oxygen evolution, models have been prepared in which manganese complexes oxidize water upon irradiation by light. The classical example is the photochemical decomposition of permanganate, MnOl-, a process first reported in 1913 (448). Dioxygen could be generated by irradiating either of two ligand-metal charge-transfer bands at 311 nm or 546 nm. The 31l-nm band is the more efficient of the two. The release of dioxygen was established to be pH independent in the range from slightly basic to slightly acidic. Isotopic labeling studies with labeled oxygen atoms suggested that the oxygen atoms in the product dioxygen originated on the same permanganate anion. Furthermore, an intermedi-
420
LAW, CAUDLE, AND PECORARO
ate for this decomposition was proposed that will react with oxidizable organic substrates that were stable in the presence of permanganate prior to irradiation of the reaction mixture. In these experiments, dioxygen evolution is curtailed, and the rate of permanganate consumption increases. These data suggested that an intermediate existed during this decomposition process. It was proposed to bear, in part, a Mnv-peroxide moiety. This system is principally of historical interest, in that it served as an early functional model for the OEC. The data for the OEC suggest, however, that such a process does not occur there. Other manganese-containing molecules have been synthesized since that time that have also been claimed to oxidize water upon irradiation in solution. Two such systems have been prepared by McAuliffe and co-workers (449-451).These complexes were prepared with the sale+ ligand, and characterized as [Mn111L(H20)12(C10,)z compounds wherein the crystal structure shows two planar LMn"' moieties bridged by two water molecules. When an aqueous solution of this complex was irradiated, these authors reported that the dioxygen concentration in the solution rose, as measured by an oxygen electrode. Furthermore, they reported this to be a stoichiometric amount of dioxygen. This reaction was conducted in the presence of hydrogen atom accepting quinone molecules, which were reported to be reduced to hydroquinone products. The authors proposed that the dioxygen produced comes from the bridging water molecules, because conducting this reaction in anhydrous ethanol did not appear to noticeably retard the rate of dioxygen production. The productive region of light was found to be the range of 450 to 600 nm, with a maximal rate for dioxygen evolution at 590 nm. The final manganese complex from the More recently, process is proposed to be in the form [(Mn111L)201. McAuliffe and co-workers have prepared a system in which two Schiff-base ligands are connected via two naphthyl moieties (451). This dimeric product also binds water and has been suggested to oxidize water to dioxygen upon irradiation, too. Because the tetranuclear cluster of the OEC is most likely not photoactivated in and of itself during turnover, manganese systems that oxidize water without being irradiated first are of interest with respect to understanding the function of this complex enzyme. One system that oxidizes water was prepared and studied by Matsushita and co-workers (452,453). These complexes are mononuclear MnmL2C12 complexes, where L is a bidentate Schiff base comprised of 5-nitrosalicylaldehyde and an alkylmonoamine, with structures that should be similar to Mnl"(salpn)Clz (Fig. 33). These researchers showed that
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
42 1
these complexes are capable of oxidizing water to dioxygen in a noncatalytic process. The complexes of this type exhibit high reduction potentials, all near 1.0 V (454).Water oxidation was proposed to occur in the manner shown in Scheme 24, wherein two molecules of water
SCHEME 24.
would be oxidized to form dioxygen, and two Mn"L2 compounds would be produced. This evolution of dioxygen from water was confirmed by the observation of 180-enriched dioxygen following the introduction of isotopically labeled water into a reaction mixture. Titration by AgNO, indicated that four equivalents of C1- were retained in solution, indicating that chloride oxidation had not occurred. The observed pH changes in unbuffered reaction mixtures were consistent with the evolution of four equivalents of proton. The Mn products were confirmed as Mn" complexes; however, no yield was listed. The yield of dioxygen was only half the maximal possible yield, at neutral pH. Naruta et al. published a report of dioxygen evolution from water that was catalyzed by a dimeric-linked manganese-porphyrin dimer complex (4551, in a system similar to that employed for their catalase mimic (Fig. 75). This complex is composed of two triphenylporphyrins linked via an o-phenylene bridge. With this series of complexes, they were able to show oxidation of water to dioxygen. This was accomplished by preparing a solution of the dimeric porphyrin complex in 5% v/v HzO in acetonitrile in the presence of Bu,NOH. In dry acetonitrile, reversible electrochemistry was observed, but in the aqueous acetonitrile mixture, a n irreversible current discharge was observed a t L 1.2 V vs. Ag/Ag+.Dioxygen evolution was observed over a voltage range of 1.2 to 2.0 V; however, in the absence of the complex, anodic oxidation of the watedacetonitrile mixture did not evolve dioxygen up to +2.3 V. The rate of dioxygen evolution was found to correlate with the concentration of the manganese complex in solution, and up to 9.2 turnovers were achieved. Under the same conditions, the monomeric analogs of these dimeric complexes did not evolve dioxygen. By mass spectrometry, the authors observed that a 1: 1 mixture of HZl6Oand HZI80in the reaction led to a 1: 2 : 1 mixture of lSOz,160180, and 180z, respectively. Finally, they reported that 3.7 electrons were involved
42 2
LAW, CAUDLE, AND PECORARO
12+
1 Ar = 4-1BuC6H1
2 Ar = 2,4.6-Me3C6H2 3 Ar=C6FS
FIG.75. Representation of the linked porphyrin catalyst utilized for water oxidation. Replacement of the o-phenylene linker with a n anthracene linker generates the catalase mimic reported in the catalase segment (443).The authors have noted that reactivity can be altered by the nature of the linker (444).[Reproduced with permission from (455).Copyright 1994 Wiley-VCH Verlag.1
in this reaction overall, as determined by a rotating disk electrode and a Levich plot. Two mechanistic proposals could be made for this dioxygen evolution. The first invokes 2 two-electron processes. Because similar dimers had been shown to disproportionate hydrogen peroxide (443445) (discussed earlier), this proposal would rely on the formation of hydrogen peroxide by a dimer, followed by disproportionation of that hydrogen peroxide by one of the dimers in solution. The second proposal invokes a four-electron process. The authors found no evidence of hydrogen peroxide evolution in their system in the absence of a manganese dimer, either by addition of a dimer to a water/acetonitrile solution after anodic oxidation or by analysis by a hydrogen-peroxidesensitive electrode. Thus, they proposed that a four-electron process was operative. Their mechanism involves the coordination of two hydroxides to the manganese of the dimer on the inside of the dimer cavity. Oxidation to MnV=O moieties is then followed by formation of a Mnw-O-O-Mnw compound. They proposed that coordination of hydroxide to the manganese could then occur with loss of the p1,2peroxo bridge as dioxygen to return the dimer to the initial Mni*'oxidation state. A more recent system in which water oxidation has been observed via a manganese complex (456) is a manganese system based on the ligand dipicolinate (dpa). When [Mn**'(dpa)2]reacts with oxone, PO-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
423
tassium peroxymonosulfate, a rather strong oxidant, the generation of dioxygen, and the formation of MnOl- was observed. These reactions were carried out in buffered aqueous solutions, over a pH 3-6 range, with a n increasing rate of permanganate anion production noted for higher pH levels. The Mn"' complex appears to be a precursor to a reactive intermediate, which has been proposed to be a bis(p-0" )-bridged dimer with a terminal water on each of the manganese, perhaps with a structure similar t o that shown in Fig. 55. This green intermediate complex exhibited a Mnl*'MnTV 16-line EPR signal and had a distinctive UV-vis spectrum, which converted isosbestically to that of Mn04 during the course of the reaction. Below pH 3.5, this system catalytically generated dioxygen. Finally, in control experiments, this system did not appear to generate dioxygen from hydrogen peroxide formed via oxone hydrolysis in the presence of other Lewis acids, such as Ni"/sulfate. This also implied that a redox process involving the manganese complex was producing the observed dioxygen. To follow up on these observations, some additional ligand systems were screened, including 2,2 '-bipyridine (bpy), 1,lO-phenanthroline (phen), picolinate (pic), and 2.2' : 6',2"-terpyridine (terpy). In the cases with the bidentate ligands, fewer turnovers leading to dioxygen evolution compared to the dpa were observed. In the case of the tridentate terpy, many more turnovers were achieved and no appreciable concentration of Mn04- was observed. The authors concluded that the meridionally coordinating tridentate ligands were ideal for promoting the reactivity that they observed, and they attributed the robust nature of the terpy system to a lesser degree of ligand dissociation versus dpa, pic, or bipy (456) because the ligand must come off of the manganese ion before it may be converted to a permanganate. Two mechanistic proposals for this process were suggested. The first requires oxygen atom transfer and the production of a n MnV=O moiety; the second is based on the hydrolysis of a manganese-bound peroxymonosulfate anion to yield a Mn-0-0moiety. I n the first case, MnV=O would combine with either a second equivalent of MnV=O or a n oxone molecule in a n overall four-electron process. In the second, a two-electron process leads to the release of dioxygen from the Mn-0-0complex. Neither of these possibilities has yet to be ruled out, and the authors state that they are following up with isotope-labeling studies to ascertain what the reaction mechanism may be. Some preliminary evidence does, however, suggest that oxygen transfer may be involved. By utilizing dimethyldioxirane, a reagent capable of oxygen atom transfer, but not the transfer of a per-
424
LAW, CAUDLE, AND PECORARO
0x0 group, in the presence of Mn" and terpy, the authors observed the production of dioxygen. In conclusion, the reactivity of manganese complexes has proved to be a fertile area for research. It has provided new insights into the function of manganese-containing redox enzymes, and it has provided new synthetic techniques to the field of organic chemistry. VI. Conclusion
The biological applications of manganese are numerous and quite varied. Although most of these biological roles utilize manganese in structural or hydrolytic systems, some very striking redox systems are also known. The hydrolytic roles are, of course, not to be overlooked with respect to the widespread and critical arginase enzyme, which exhibits a strict requirement for two Mn" ions. The ribonuclease hydrolases of retroviral reverse transcriptase are another key nonredox application-for example, the ribonuclease hydrolase from HIV-1. The manganese-containing redox enzymes are quite unique. Most of these deal with dioxygen metabolism in one form or another. One key enzyme is the OEC, which generates the dioxygen required by aerobic life forms to exist by catalytically oxidizing water. New evidence now suggests that this enzyme may soon join the ranks of other known metalloenzymes that employ radical protein residues to complete their catalytic cycle. Catalase disproportionates hydrogen peroxide, and the MnSOD catalyzes the dismutation of superoxide. Both are critical for protecting organisms from oxidative damage, one of the hazards of aerobic life. Meanwhile, the newly discovered dioxygenases operate on aromatic-bearing cis-diol functionalities and open these rings by catalyzing the incorporation of dioxygen into them. All of these enzymes are driven by the oxophilicity of manganese, which is important in defining the chemistry that occurs. Another interesting feature of several of these enzymes is their iron analogs. MnSOD and Mn-dioxygenase have direct iron enzyme analogs, which in the case of SOD, are known to have nearly identical active sites, and in some cases (but not all) the metal from one may be added to the other with retention of enzyme activity. In addition, MnP bears analogy to the Lip enzymes, which rely solely on heme, although new evidence shows that they, too, may be capable of oxidizing manganese. However, one notable exception stands out, and that is the OEC. This enzyme is functional only with manganese. With so many man-
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
425
ganese enzymes that bear similarities to enzymes that conduct the same or similar reaction with different metals at their active site, why is manganese so strictly required? If the newer proposals for the function of the OEC are correct, then an answer may emerge that suggests that manganese allows the proton-coupled oxidations necessary for water oxidation to occur at a potential at which the oxidizing equivalents may be easily stored. Furthermore, there are sufficient readily available oxidation states available to allow the sequential oxidations to occur, yet despite this stability, once the high-valent MnV=O is reached, it is not so stable as to prevent the formation of dioxygen. Finally, with the acknowledged similarities that currently exist between several manganese enzymes and counterparts utilizing other metal ions, it seems quite likely that other redox-active manganese enzymes will be discovered in the future. The discussions of data gathering on enzyme systems and model chemistry, both structural and functional, as presented in this article will provide a foundation for exploring current and as yet undiscovered manganese enzymes.
ACKNOWLEDGMENTS We wish to acknowledge the NIH for funding our research endeavors (GM 39406 to V.L.P.). We also wish to thank Professor G. T. Babcock for providing us with a copy of Scheme 12 and Professor V. V. Barynin for the provision of data on the crystallographic details of the T. thermophilus enzyme.
IN PROOF NoTE ADDED
A recent article on the T. thermophilus manganese-containing catalase enzyme presents data that indicates that the structure of the active site is affected by pH. The experiments were conducted over the pH range 6.6 to 9.8. The authors propose that a t high pH a Mn"12(p-O)2(p-OAc) form predominates, while a t low pH a Mn"'2(p-O)(pOAc) form predominates. The latter form has only a single 0x0-bridge with a proposed hydroxide and water ligand to fill the remaining coordination sites on the Mn"' ions, one site per Mn"' (478).
REFERENCES 1. Wedler, F. C. In "Manganese in Health and Disease"; Klimis-Tavantzis, D. J., Ed.; CRC Press, Inc., Ann Arbor, 1994, 1. 2. Scrutton, M. C. In "Manganese in Metabolism and Enzyme Function"; Schramm, V. L., and Wedler, F. C., Eds.; Academic Press, Inc., New York, 1986, 147.
426
LAW, CAUDLE, AND PECORARO
Dismukes, G. C. Chem. Reu. 1996,96, 2909. Wilcox, D. Chem. Reu. 1996,96, 2435. Christianson, D. W. Prog. Biophys. Mol. Biol. 1997, 67, 217. Reczkowski, R. S.; Ash, D. E. J . Am. Chem. SOC.1992,114, 10992. Khangulov, S. V.; Pessiki, P. J.; Barynin, V. V.; Ash, D. E.; Dismukes, G. C. Biochemistry 1995,34, 2015. 8. Kanyo, Z. F.; Scolnick, L. R.; Ash, D. E.; Christianson, D. W. Nature 1996, 383,
3. 4. 5. 6. 7.
554.
9. Stemmler, T. L.; Sossong, Jr., T. M.; Goldstein, J. I.; Ash, D. E.; Elgren, T. E.; Kurtz, Jr., D. M.; Penner-Hahn, J. E. Biochemistry 1997, 36, 9847. 10. Sossong, Jr., T. M.; Khangulov, S. V.; Cavalli, R. C.; Soprano, D. R.; Dismukes, G. C.; Ash, D. E. JBZC 1997,2, 433. 11. Davies, 11, J . F.; Hostomska, Z.; Hostomsky, Z.; Jordan, S. R.; Matthews, D. A. Science 1991, 252, 88. 12. Cowan, J . A. JBZC 1997,2, 168. 13. Strater, N.; Lipscomb, W. N.; Klabunde, T.; Krebs, B. Angew. Chem., Znt. Ed. Engl. 1996,35, 2024. 14. Hardman, K. D.; Ainsworth, C. F. Biochemistry 1972, 11, 4910. 15. Hardman, K. D.; Agarwal, R. C.; Freiser, M. J . J. Mol. Biol. 1982, 157, 69. 16. “The Lectins: Properties, Functions, and Applications in Biology and Medicine”; Liener, I. E., Sharon, N., and Goldstein, I. J., Eds.; Academic Press: New York, 1986.
17. Sharon, N.; Lis, H. “Lectins”; Chapman and Hall: New York, 1989. 18. Naismith, J. H.; Emmerich, C.; Habash, J.; Harrop, S. J.; Helliwell, J . R.; Hunter, W. N.; Raftery, J.; Kalb, A. J.; Yariv, Y. Acta Cryst. 1994,050, 847. 19. Antanaitas, B. C.; Chasteen, N. D.; Freedman, J. H.; Koenig, S. H.; Lilienthal, H. R.; Peisach, J.; Brewer, C. F. Biochemistry 1987,26, 7932. 20. Reed, G. H.; Cohn, M. J . J . Biol. Chem. 1970,245, 662. 21. Derewenda, Z.; Yariv, J.; Helliwell, J. R.; Kalb (Gilboa), J. R.; Dodson, E. J.; Papiz, M. Z.; Wan, T.; Campell, EMBO J. 1989, 8, 2189. 22. Deacon, A.; Gleichmann, T.; Kalb (Gilboa), A. J.;Price, H.; Raftery, J.; Bradbrook, G.; Yariv, J.; Helliwell, J. R. J . Chem. Soc., Faraday Trans. 1997, 4305. 23. Penner-Hahn, J. E. In “Manganese Redox Enzymes”; Pecoraro, V. L., Ed.; VCH Publishers, Inc.: New York, 1992, 29. 24. Penner-Hahn, J. E. In “Structure and Bonding”; Hill, H. A. O., Sadler, P. J., and Thomson, A. J., Eds.; Springer Verlag: Berlin, 1998,90, 1. 25. Pecoraro, V. L.; Gelasco, A.; Baldwin, M. J . In “Mechanistic Bioinorganic Chemistry’’; Thorp, H. H., and Pecoraro, V. L., Eds.; ACS Books: Washington, DC, 1995, p. 265. 26. Pecoraro, V. L.; Baldwin, M. J.; Gelasco, A. Chem. Reu. 1994, 94, 807. 27. Pecoraro, V. L. In “Manganese Redox Enzymes”; Pecoraro, V. L., Ed.; VCH Publishers: New York, 1992, 197. 28. “Manganese Redox Enzymes”; Pecoraro, V. L., Ed.; VCH Publishers, Inc.: New York, 1992. 29. Vincent, J. B.; Christou, G. In “Advances in Inorganic Chemistry”; Sykes, A. G., Ed.; Academic Press, Inc.: New York, 1989,33, 197. 30. Wieghardt, K. Angew. Chem., Znt. Ed. Engl. 1989,28, 1153. 31. Wieghardt, K. Angew. Chem., Znt. Ed. Engl. 1994,33, 725. 32. Yachandra, V. K.; DeRose, V. J.; Latimer, M. J.; Mukerji, I.; Sauer, K.; Klein, M. P. Science 1993,260, 675.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
427
33. Yachandra, V. K.; Sauer, K.; Klein, M. P. Chem. Rev. 1996,96, 2927. 34. Manchanda, R.; Brudvig, G. W.; Crabtree, R. H. Caord. Chem. Rev. 1995,144, 1. 35. Yocum, C. F. In “Manganese Redox Enzymes”; Pecoraro, V. L., Ed.; VCH: New York, 1992, 71. 36. Thorp, H. H.; Brudvig, G. W. New J. Chem. 1991, 15, 479. 37. Ruttinger, W. F.; Campana, C.; Dismukes, G. C. J. Am. Chem. SOC.1997, 119, 6670. 38. Pecoraro, V. L.; Gelasco, A,; Baldwin, M. J. In “Bioinorganic Chemistry: An Inorganic Perspective of Life”; K. D. P., Ed.; Kluwer Academic Publishers: Amsterdam, 1995,459, 287. 39. Ghanotakis, D. F.; Yocum, C. F. Annu. Rev. Plant Phys. Plant Mol. Biol. 1990, 41, 255. 40. Pecoraro, V. L. Photochem. Photobiol. 1988,48, 249. 41. Diner, B. A.; Babcock, G. T. In “Oxygenic Photosynthesis: The Light Reactions”; Ort, D. R., and Yocum, C. F., Eds.; Kluwer Academic Publishers: Boston, 1996, 213. 42. Debus, R. J . Biochim. Biophys. Acta 1992,1102, 269. 43. de Paula, J. C.; Beck, W. F.; Brudvig, G. W. New J. Chem. 1987, 11, 103. 44. Christou, G. Ace. Chem. Res. 1989,22, 328. 45. Brudvig, G. W.; Crabtree, R. H. In “Progress in Inorganic Chemistry”; Lippard, S. J., Ed.; John Wiley & Sons: New York, 1989, 37, 99. 46. Britt, R. D. In “Oxygenic Photosynthesis: The Light Reaction”; Ort, D. R., and Yocum, C. F., Eds.; Kluwer Academic Publishers: Netherlands, 1996, 137. 47. Babcock, G. T.; Espe, M.; Hoganson, C.; Lydakis-Simantiris, N.; McCracken, J.; Shi, W.; Styring, S.; Tommos, C.; Warncke, K. Acta Chem. Scand. 1997, 51, 533. 48. “Oxygenic Photosynthesis: The Light Reactions”; Ort, D. R., and Yocum. C. F., Eds.; Kluwer Academic Publishers: Boston, 1996. 49. Hage, R.; Iburg, J . E.; Kerschner, J.; Koek, J . H.; Lempers, E. L. M.; Martens, R. J.; Racherla, U. S.; Russell, S. W.; Swarthoff, T.; van Vliet, M. R. P.; Warnaar, J . B.; van der Wolf, L.; Krijnen, B. Nature 1994,269, 637. 50. Armstrong, W. H. In “Manganese Redox Enzymes”; Pecoraro, V. L., Ed.; VCH Publishers: New York, 1992, 261. 51. “Photosynthesis: From Light to Biosphere”; Mathis, P., Ed.; Kluwer Academic Publishers: Amsterdam, The Netherlands, 1995. 52. Keele, Jr., B. B.; McCord, J . M.; Fridovich, I. J. Biol. Chem. 1970, 245, 6176. 53. Ludwig, M. L.; Pattridge, K. A,; Stallings, W. C. In “Manganese in Metabolism and Enzyme Function”; Schramm, V. L., and Wedler, F. C., Eds.; Academic Press, Inc.: New York, 1986, 405. 54. Stallings, W. C.; Pattridge, K. A.; Strong, R. K.; Ludwig, M. L. J. Biol. Chem. 1984,259, 10695. 55. Stallings, W. C.; Pattridge, K. A.; Strong, R. K.; Ludwig, M. L. J. B i d Chem. 1985,260, 16424. 56. Stallings, W. C.; Powers, T. B.; Pattridge, K. A,; Fee, J . A,; Ludwig, M. L. Proc. Natl. Acad. Sci. U.S.A. 1988.80, 3884. 57. Tierney, D. L.; Fee, J . A.; Ludwig, M. L.; Penner-Hahn, J . E. Biochemistry 1995, 34, 1661. 58. Lah, M. S.; Dixon, M. M.; Pattridge, K. A,; Stallings, W. C.; Fee, J . A.; Ludwig, M. L. Biochemistry 1995,34, 1646. 59. Barra, D.; Schinina, M. E.; Bannister, W. H.; Bannister, J . V.; Bossa, F. J . B i d . Chem. 1987,262, 1001.
428
LAW, CAUDLE, AND P E C O W O
60. Harris, J. I.; Auffret, A. D.; Northrop, F. D.; Walker, J. E. Eur. J. Biochem. 1980, 106, 297. 61. Beyer, W. F.; Fridovich, I. J. Biol. Chem. 1991,263, 303. 62. McCord, J. M.; Fridovich, I. J. Biol. Chem. 1969, 244, 6049. 63. Valentine, J, S. In “Bioinorganic Chemistry”; Bertini, I., Gray, H. B., Lippard, S. J., and Valentine, J. S., Eds.; University Science Press: Mill Valley, California, 1994. 64. Youn, H.-D.; Youn, H.; Lee, J.-W.; Yim, Y.-I.; Lee, J. K.; Hah, Y. C.; Kang, 53.-0. Arch. Biochem. Biophys. 1996,334, 341. 65. Fridovich, I. J. Biol. Chem. 1989,264, 7761. 66. McCord, J. M.; Fridovich, I. J. Biol. Chem. 1968,243, 5753. 67. Morel, F.; Doussiere, J.; Vignas, P. V. Eur. J. Biochem. 1991,201, 523. 68. Beauchamp, C.; Fridovich, I. Anal. Biochem. 1971,44, 276. 69. Davies, K. J. A. In “Free Radicah and Oxidative Stress: Environment, Food and Drug Additives”; Rice-Evans, C., Halliwell, B., and Lunt, G. G., Eds.; Portland Press: London, 1994, 61, 1. 70. Afanas’ev, I. B. “Superoxide Ion: Chemistry and Biological Implications”; CRC Press, Inc.: Boca Raton, 1989. 71. Darley-Usmar, V.; Wiseman, H.; Halliwell, B. FEBS Lett. 1995,369, 131. 72. Jenner, P.; Olanow, C. W. Neurology 1996,47, S161. 73. Dawson, V. L.; Dawson, T. M. J. Chem. Neuroanat. 1996,10, 179. 74. Beckman, J. S.; Beckman, T. W.; Chen, J.; Marshall, P. M.; Freeman, B. A. Proc. Natl. Acad Sci. U.S.A. 1990,87, 1620. 75. Borgstahl, G. E. 0.; Parge, H. E.; Hickey, M. J.; Beyer, W. F.; Hallewell, R. A,; Tainer, J. A. Cell 1992, 71, 107. 76. Ludwig, M. L.; Metzger, A. L.; Pattridge, K. A.; Stallings, W. C. J. Mol. Biol. 1991, 219, 335. 77. Fee, J. A.; Shapiro, E. R.; Moss, T. H. J. Biol. Chem. 1976,251, 6157. 78. Bull, C.; Neiderhoffer, E. C.; Yoshida, T.; Fee, J. A. J. Am. Chem. SOC. 1991, 113, 4069. 79. McAdam, M. E.; Fox, R. A,; Lavelle, F.; Fielden, E. M. Biochem. J. 1977, 165, 71. 80. McAdam, M. E.; Fox, R. A,; Lavelle, F.; Fielden, E. M. Biochem. J. 1977, 165, 81. 81. Osman, R.; Basch, H. J. Am. Chem. SOC.1984, 106, 5710. 82. Pick, M.; Rabani, J.; Yost, F.; Fridovich, I. J. Am. Chem. SOC.1974, 96, 7329. 83. Valentine, J. S.; Quinn, A. E. fnorg. Chem. 1976, 15, 1997. 84. Bull, C.; Fee, J. A. J. Am. Chem. SOC.1985, 107, 3295. 85. Whiting, A. K.; Boldt, Y. R.; Hendrich, M. P.; Wackett, L. P.; Que, L. Biochemistry 1996,35, 160. 86. Que, Jr., L.; Widom, J.; Crawford, R. L. J. Biol. Chem. 1981,256, 10941. 87. Que, L. J. In “Iron Carriers and Iron Proteins”; Loehr, T. M., Ed.; VCH Publishers, Inc.: New York, 1989,5, 467. 88. Lipscomb, J. D.; Orville, A. M. In “Metal Ions in Biological Systems”; Sigel, H., and Sigel, A., Eds.; Marcel Dekker, Inc.: New York, 1992,28, 243. 89. Que, Jr., L.; Ryn, H. Chem. Rev. lB96,96, 2607. 90. Boldt, Y. R.; Sadowsky, M. J.; Ellis, L. B. M.; Que, Jr., L.; Wackett, L. P. J. Bacterial. 1995, 177, 1225. 91. Han, S.; Eltis, L. D.; Timmis. K. N.; Muchmore, S. W.; Bolin, J. T. Science 1995, 270, 976. 92. Boldt, Y. R.; Whiting, A. K.; Wagner, M. L.; Sadowsky, M. J.; Que, Jr., L.; Wackett, L. P. Biochemistry 1997, 36, 2147.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
429
93. Shu, L. J.; Chiou, Y. M.; Orville, A. M.; Miller, M. A,; Lipscomb, J . D.; Que, Jr., L. Biochemistry 1995, 34, 6649. 94. Glenn, J. K.; Gold, M. H. Arch. Biochem. Biophys. 1985,242, 329. 95. Glenn, J. K.; Akileswaran, L.; Gold, M. H. Arch. Biochem. Biophys. 1986,251, 688. 96. Sundaramoorthy, M.; Kishi, K.; Gold, M. H.; Poulos, T. L. J. Biol. Chem. 1994, 269, 32759. 97. Sarkanen, K. V.; Ludwig, C. H. “Lignins: Occurrence, Formation, Structure and Reactions”; Wiley-Interscience: New York, 1971. 98. Gold, M. H.; Wariishi, H.; Valli, K. In “Biocatalysis in Agricultural Biotechnology”; Whitaker, J . R., and Sonnet, P. E., Eds.; American Chemical Society: Washington, DC, 1989, 127. 99. Crawford, R. L. “Lignin Biodegradation and Transformation”; John Wiley & Sons: New York, 1981. 100. Eggert, C.; Temp, U.; Eriksson, K.-E. L. FEBS Lett. 1997, 407, 89. 101. Youn, H.-D.; Yung, C. H.; Sa-Ouk, K. FEMS Microbiol. Lett. 1995, 132, 183. 102. Hammel, K. E. In “Degradation of Environmental Pollutants by Microorganisms and Their Metalloenzymes”; Sigel, H., and Sigel, A., Eds.; Marcel Dekker, Inc.: New York, 1992. 103. Wariishi, H.; Valli, K.; Gold, M. H. Biochemsitry 1989, 28, 6017. 104. Wariishi, H.; Valli, K.; Gold, M. H. J. Biol. Chem. 1992, 267, 23688. 105. Kuan, I.-C.; Johnson, K. A,; Tien, M. J . Biol. Chem. 1993,268, 20064. 106. Kuan, I.-C.; Tien, M. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 1242. 107. Khindaria, A,; Barr, D. P.; Aust, S. D. Biochemistry 1995, 34, 7773. 108. Timofeevski, S. L.; Aust, S. D. Biochem. Biophys. Res. Commun. 1997,239, 645. 109. Zapanta, L. S.; Tien, M. J. Biotech. 1997, 53, 93. 110. Pasczcynski, A,; Huyhn, V.-B.; Crawford, R. L. Arch. Biochem. Biophys. 1986, 244, 750. 111. Auling, G.; Follman, H. In “Metal Ions in Biological Systems”; Sigel, H., and Sigel, A,, Eds.; Marcel Dekker, Inc.: New York, 1994, 30, 131. 112. Stubbe, J . J. Biol. Chem. 1990, 265, 5329. 113. Sjoberg, B.-M. Structure 1994, 2, 793. 114. Sjoberg, B.-M. In “Structure and Bonding”; Hill, H. A. O., Sadler, P. J., and Thomson, A. J., Eds.; Springer Verlag: Berlin, 1997, 88, 139. 115. Mao, S. S.; Holler, T. P.; Yu, G. X.; Bollinger, Jr., J . M.; Booker, S.; Johnston, M. I.; Stubbe, J. Biochemistry 1992, 31, 9733. 116. Reichard, P. Science 1993,260, 1773. 117. Schimpf-Weiland, G.; Follman, H.; Auling, G. Biochem. Biophys. Res. Commun. 1981.102, 1276. 118. Willing, A.; Follmann, H.; Auling, G. Eur. J. Biochem. 1988, 170, 603. 119. Uhlin, U.; Eklund, H. J . Mol. Biol. 1996, 262, 358. 120. Uhlin, U.; Eklund, H. Nature 1994,370, 533. 121. Nordlund, P.; Sjoberg, B.-M.; Eklund, H. Nature 1990,345, 593. 122. Nordlund, P.; Eklund, H. J. Mol. Biol. 1993, 232, 123. 123. Willing, A,; Follmann, H.; Auling, G. Eur. J. Biochem. 1988, 175, 167. 124. Atta, M.; Nordlund, P.; Aberg, A,; Eklund, H.; Fontecave, M. J . B i d . Chem. 1992, 267, 20682. 125. Barynin, V. V.; Hempstead, P. D.; Vagin, A. A.; Antonyuk, S. V.; Melik-Adamyn, W. R.; Lamzin, V. S.; Harrison, P. M.; Artymiuk, P. J . J. Inorg. Biochem. 1997, 67, 196. 126. Griepenburg, U.; Lapmann, G.; Auling, G. Free Rad. Res. 1996,263, 473.
430
LAW,CAUDLE, AND PECORARO
Beyer, Jr., W. F.; Fridovich, I. Biochemistry 1985,24, 6460. Kono, Y.; Fridovich, I. J. Biol. Chem. 1983,258, 6015. Algood, G. S.; Perry, J. J. J. Bacteriol. 1986, 168, 563. Barynin, V. V.; Grebenko, A. I. Dokl. Akad. Nauk. S.S.S.R 1986,286, 461. Barynin, V. V.; Vagin, A. A.; Melik-Adamyan, V. R.; Grebenko, A. I.; Khangulov, S. V.; Popov, A. N.; Andrianova, M. E.; Vainshtein, B. K. Dolk. Akad. Nauk. S.S.S.R 1986,288, 877. 132. Sheats, J. E.; Czernuszewicz, R. S.; Dismukes, G. C.; Rheingold, A. L.; Petrouleas, V.; Stubbe, J.; Armstrong, W. H.; Beer, R. H.; Lippard, S. J. J. Am. Chem. SOC.
127. 128. 129. 130. 131,
1987,109, 1435. 133. Wieghardt, K.; Bossek, U.; Ventur, D.; Weiss, J . J. Chem. SOC.,Chem. Commun. 1985,347. 134. Riggs-Gelasco, P. J.; Stemmler, T.; Penner-Hahn, J . E. Coord. Chem. Rev. 1995, 144, 245. 135. Waldo, G. S.; Yu, S. Y.; Penner-Hahn, J. P. J. Am. Chem. SOC.1992, 114, 5869. 136. Fronko, R. M.; Penner-Hahn, J . E. J.Am. Chem. SOC.1988,110, 7554. 137. Waldo, G. S.; Penner-Hahn, J. E. Biochemistry 1995,34, 1507. 138. Khangulov, S. V.; Voyevodskaya, N. V.; Varynin, V. V.; Grebenko, A. I.; MelikAdamyan, V. R. Biofizika 1987,32, 960. 139. Khangulov, S. V.; Goldfeld, M. G.; Gerasimenko, V. V.; Andreeva, N. E.; Barynin, V. V.; Grebenko, A. I. J. Inorg. Biochem. 1990,40, 279. 140. Zheng, M.; Khangulov, S. V.; Dismukes, G. C.; Barynin, V. V. Inorg. Chem. 1994, 33, 382. 141. Kurtz, Jr., D. M. JBIC 1997,2, 159. 142. Diesenhofer, J.; Epp, 0.;Miki, K.; Huber, R.; Michel, H. Nature 1985,318, 618. 143. Michel, J.; Diesenhofer, J. Biochemistry 1988,27, 1. 144. Ort, D. R.; Yocum, C. F. In “Oxygenic Photosynthesis: The Light Reactions”; Ort, D. R., and Yocum, C. F., Eds.; Kluwer Academic Publishers: Boston, 1996, 1. 145. Adelroth, P.; Lindberg, K.; Andreasson, L.-E. Biochemistry 1995,34, 9021. 146. Yocum, C. F. Biochim. Biophys. Acta 1991,1059, 1. 147. Lindberg, K.; Andreasson, L.-E. Biochemistry 1996,35, 14259. 148. Lindberg, K.; Vanngard, T.; Andrbasson, L. E. Photosynth. Res. 1993,38, 401. 149. Barry, B. A.; Babcock, G. T. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 7099. 150. Gerken, S.; Brettel, K.; Schlodder, E.; Witt, H. T. FEBS Lett. 1988,237, 69. 151. Joliot, P.; Barbieri, G.; Chabaud, R. Photochem. Photobiol. 1969, 10, 309. 152. “Bioenergetics in Photosynthesis”; Joliot, P., Kok, B., and Govindjee, Eds.; Academic Press: New York, 1975. 153. Kok, B.; Forbush, B.; McGloid, M. Photochem. Photobiol. 1970, 11, 457. 154. Sauer, K.; Yachandra, V. K.; Britt, R. D.; Klein, M. P. In “Manganese Redox Enzymes”; Pecoraro, V. L., Ed.; VCH Publishers: New York, 1992. 155. Yachandra, V. K.; Guiles, R. D.; McDermott, A. E.; Cole, J . L.; Britt, R. D.; Dexheimer, S. L.; Sauer, K.; Klein, M. P. Biochemistry 1987,26, 5974. 156. Penner-Hahn, J. E.; Fronko, R. M.; Pecoraro, V. L.; Yocum, C. F.; Betts, S. D.; Bowlby, N. R. J.Am. Chem. SOC.1990,112, 2549. 157. MacLachlan, D. J.; Hallahan, B. J.; Ruffle, S. V.; Nugent, J. H. A,; Evans, M. C. W.; Strange, R. W.; Hasnain, S. S. Biochem. J. 1992,285, 569. 158. George, G. N.; Prince, R. C.; Cramer, S. P. Science 1989, 10, 789. 159. DeRose, V. J.; Mukerji, I.; Latimer, M. J.; Yachandra, V. K.; Sauer, K.; Klein, M. P. J . Am. Chem. SOC.1994, 116, 5239. 160. Dismukes, G. C.; Siderer, Y. FEBS Lett. 1980, 121, 78.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
431
161. Dismukes, G. C.; Siderer, Y. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 274. 162. Cole, J.; Yachandra, V. K.; Guiles, R. D.; McDermott, A. E.; Britt, R. D.; Dexheimer, S . L.; Sauer, K.; Klein, M. P. Biochim. Biophys. Acta 1987, 890, 395. 163. Hansson, 0.;Aasa, R.; Vaenngaard, T. Biophys. J. 1987,51, 825. 164. Casey, J. L.; Sauer, K. Biochim. Biophys. Acta 1984, 767, 21. 165. Zimmermann, J.-L.; Rutherford, A. W. Biochim. Biophys. Acta 1984, 767, 160. 166. Beck, W. F.; de Paula, J . C.; Brudvig, G. W. J. Am. Chem. SOC.1986, 108, 4018. 167. Kim, D. H.; Britt, R. D.; Klein, M. P.; Sauer, K. Biochemistry 1992, 31, 541. 168. Kim, D. H.; Britt, R. D.; Klein, M. P.; Sauer. K. J. Am. Chem. SOC.1990,112, 9389. 169. Haddy, A.; Dunham, W. R.; Sands, R. H.; Aasa, R. Biochim. Biophys. Acta 1992, 1099, 25. 170. Messinger, J.; Robblee, J. H.; Yu, W. 0.; Sauer, K.; Yachandra, V. K.; Klein, M. P. J. Am. Chem. SOC.1997, 119, 11349. 171. Messinger, J.; Nugent, J . H. A.; Evans, M. C. W. Biochemistry 1997, 36, 11055. 172. h r l i n g , K. A.; Peterson, S.; Styring, S. Biochemistry 1997, 36, 13148. 173. Dexheimer, S. L.; Klein, M. P. J . Am. Chem. SOC. 1992, 114, 2821. 274. Campbell, K. A,; Peloquin, J . M.; Pham, D. P.; Debus, R. J.; Britt, R. D. J. Am. Chem. SOC.1998, 120, 447. 175. Yachandra, V. K.; Guiles, R. D.; McDermott, A. E.; Cole, J . L.; DeRose, V. J.; Zimmermann, J. L.; Sauer, K.; Klein, M. P. Physica B (Amsterdam) 1989,158, 78. 176. Boussac, A,; Zimmermann, J.-L.; Rutherford, A. W.; Lavergne, J . Nature 1990, 347, 303. 177. Gilchrist, M. L.; Ball, J. A,; Randall, D. W.; Britt, R. D. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 9545. 178. Tang, X.-S.; Randall, D. W.; Force, D. A,; Diner, B. A,; Britt, R. D. J . Am. Chem. SOC.1996, 118, 7638. 179. Babcock, G. T.; Barry, B. A.; Debus, R. J.; Hoganson, C. W.; Atamian, M.; McIntosh, L.; Sithole, I.; Yocum, C. F. Biochemistry 1989,28, 9557. 180. Force, D. A,; Randall, D. W.; Britt, R. D. Biochemistry 1997,36, 12062. 181. Tommos, C.; Tang, X.-S.; Warncke, K.; Hoganson, C. W.; Styring, S.; McCracken, 1995, 117, 10325. J.; Diner, B. A,; Babcock, G. T. J. Am. Chem. SOC. 182. Tommos, C.; Babcock, G. T. Acc. Chem. Res. 1998,31, 18. 183. Fontecave, M.; Pierre, J. C. Bull. SOC.Chim. Fr. 1996, 133, 653. 184. Hoganson, C. W.; Lydakis-Simantiris, N.; Tang, X.-S.; Tommos, C.; Warncke, K.; Babcock, G . T.; Diner, B. A.; McCracken, J.; Styring, S. Photosyn. Res. 1995, 46, 177. 185. Babcock, G. T. In “Photosynthesis: From Light to Biosphere”; Mathis, P., Ed.; Kluwer Academic Publishers: Netherlands, 1995,II, 209. 186. Ahlbrink, R.; Haumann, M.; Cherepanov, D.; Bogershausen, 0.;Mulkidjanian, A.; Junge, W. Biochemistry 1998,37, 1131. 187. Tang, X.-S.; Diner, B. A.; Larsen, B. S.; Gilchrist, J.; M. L.; Lorigan, G. A,; Britt, R. D. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 704. 188. Wincencjusz, H.; Van Gorkom, H.; Yocum, C. F. Biochemistry 1997,36, 3663. 189. Fine, P. L.; Frascb, W. D. Biochemistry 1992,31, 12204. 190. Sandusky, P. 0.;Yocum, C. F. FEBS Lett. 1983,162, 339. 191. Latimer, M. J.; DeRose, V. J.; Mukerji, I.; Yachandra, V. K.; Sauer, K.; Klein, M. P. Biochemistry 1995,34, 10898. 192. Tso, J.; Sivaraja, M.; Dismukes, G. C. Biochemistry 1991, 30, 4734. 193. Rutherford, A. W. Trends Biochem. Sci. 1989, 14, 227. 194. Ghanotakis, D. F.; Babcock, G. T.; Yocum, C. F. FEBS Lett. 1984, 167, 127.
432
LAW, CAUDLE, AND PECORAFtO
195. DeRose, V. J.; Latimer, M. J.; Zimmerman, J. L.; Mukerji, I.; Yachandra, V. K.; Sauer, K.; Klein, M. P. Chem. Phys. 1995,194, 443. 196. Sandusky, P. 0.;Yocum, C. F. Biochim. Biophys. Acta 1986,849, 85. 197. Radmer, R.; Ollinger, 0. FEBS Lett. 1986, 195, 285. 198. Messinger, J.; Badger, M.; Wydrzynski, T. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 3209. 199. Brudvig, G. W.; Crabtree, R. H. Proc. Natl. Acad. Sci. U.S.A. 1986,83, 4586. 200. Vincent, J. B.; Christou, G. Inorg. Chzm. Acta 1987, 136, L41. 201. Kirk, M. L.; Chan, M. K.; Armstrong, W. H.; Solomon, E. I. J . Am. Chem. SOC. 1992,114, 10432. 202. Larson, E.; Haddy, A.; Kirk, M. L.; Sands, R. H.; Hatfield, W. E.; Pecoraro, V. L. J. Am. Chem. SOC.1992, 114, 6263. 203. Randall, D. W.; Gelasco, A,; Caudle, M. T.; Pecoraro, V. L.; Britt, R. D. J. Am. Chem. SOC.1997,119, 4481. 204. Baldwin, M. J.; Pecoraro, V. L. J. Am. Chem. SOC.1996, 118, 11325. 205. Caudle, M. T.; Pecoraro, V. L. J. Am. Chem. SOC.1997, 119, 3415. 206. Gardner, K. A.; Mayer, J. M. Science 1995,269, 1849. 207. Blomberg, M. R. A.; Siegbahn, P. E. M.; Styring, S.; Babcock, G. T.; Akermark, B.; Korall, P. J . Am. Chem. SOC.1997,119, 8285. 208. Kitajima, N.; Osawa, M.; Tamura, N.; Moro-oka, Y.; Hirano, T.; Masaaki, H.; Nagano, T. Inorg. Chem. 1993,32, 1879. 209. Kitajima, N.; Tolman, W. B. In “Progress in Inorganic Chemistry”; Karlin, K. D., Ed.; John Wiley & Sons, Inc.: New York, 1995,43, 419. 210. Trofimenko, S. Chem. Reu. 1993,93, 943. 211. Lah, M. S.; Chun, H. Inorg. Chem. 1997,36, 1782. 212. Oki, A. R.; Bommarreddy, P. R.; Zhang, H.; Hosmane, N. Inorg. Chim. Acta 1996, 231, 109. 213. Adams, H.; Bailey, N. A.; Crane, J. D.; Fenton, D. E.; Latour, J.-M.; Williams, J . M. J. Chem. Soc., Dalton Trans. 1990, 1727. 214. Di Vaira, M.; Mani, F. J. Chem. SOC.,Dalton Trans. 1990, 191. 215. Roesky, H. W.; Scholz, M.; Noltemeyer, M. Chem. Ber. 1990, 123, 2303. 216. Holleman, S. R.; Parker, 0. J.; Breneman, G. L. Acta. Cryst. 1994, C50, 867. 217. Andruh, M.; Huebner, K.; Noltemeyer, M.; Roseky, H.; W., 2. Naturforsch. 1993, 48B, 591. 218. Stolz, P.; Saak, W.; Strasdeit, H.; Pohl, S. Z. Natursforsch. 1989, 44B, 632. 219. Chen, X.; Long, G.; Willett, R. D.; Hawks, T.; Molnar, S.; Brewer, K. Acta Cryst. 1996,1252, 1924. 220. Glerup, J.; Goodson, P. A,; Hodgson, D. J.; Michelsen, K.; Nielsen, K. M.; Weihe, H. Inorg. Chem. 1992,31, 4611. 221. Riley, D. P.; Weiss, R. H. J. Am. Chem. SOC.1994, 116, 387. 222. Riley, D. P.; Henke, S. L.; Lemon, P. J.; Weiss, R. H.; Neumann, W. L.; Rivers, W.; Aston, J. K. W.; Sample, K. R.; Rahman, H.; Ling, C. S.; Shieh, J. J.; Busch, D. H.; Szulbinski, W. Inorg. Chem. 1996, 35, 5213. 223. Burns, J. H.; Lumetta, G. J . Acta Cryst. 1991, C47, 2069. 224. Deng, Y.; Burns, J. H.; Moyer, B. A. Inorg. Chem. 1995,34, 209. 225. Othman, A. H.; Lee, K.-L.; Fun, H.-K.; Yip, B.-C. Acta Cryst. 1996, C52, 602. 226. Chan, C.-W.; Che, C.-M.; Peng, S.-M. Polyhedron 1993, 12, 2169. 227. Brooker, S.; McKee, V. Acta Cryst. 1993, C49, 441. 228. Bencini, A.; Biannchi, A.; Dapporto, P.; Garcia-Espafia, E.; Marcelino, V.; Micheloni, M.; Paoletti, P.; Paola, P. Inorg. Chem. 1990,29, 1716.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
433
229. Deroche, A.; Morgenstern-Badarau, I.; Cesario, M.; Guilhem, J.; Keita, B.; Nadjo, L.; Houee-Levin, C. J . Am. Chem. SOC.1996, 118, 4567. 230. Di Vaira, M.; Mani, F.; Stoppioni, P. J. Chem. SOC.,Dalton Trans. 1992, 1127. 231. Hagen, K. S. Angew. Chem., Int. Ed. Engl. 1992,31, 764. 232. Gultneh, Y.; Farooq, A.; Karlin, K.; Liu, D. S.; Zubieta, J. Inorg. Chim. Acta 1993, 211, 171. 233. Knof, U.; Weyhermueller, T.; Wolter, T.; Wieghardt, K. J . Chem. SOC.,Chem. Commun. 1993, 726. 234. Bermejo, M. R.; Castineiras, A,; Garcia-Monteagudo, J . C.; Rey, M.; Sousa, A,;
235. 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251.
252. 253. 254. 255.
256. 257. 258.
Watkinson, M.; McAuliffe, C. A.; Pritchard, R. G.; Beddoes, R. L. J . Chem. SOC., Dalton Trans. 1996,2935. Jacobsen, E. N. I n “Catalytic Asymmetric Synthesis”; Ojima, I., Ed.; VCH Publishers, Inc.: New York, 1993, 159. Pospisil; Jacobsen, E. N. Chem. Eur. J . 1996,2, 974. Pecoraro, V. L.; Butler, W. M. Acta Cryst. 1986, C42, 1151. Schake, A. R.; Schmitt, E. A,; Conti, A. J.; Streib, W. E.; Huffman, J. C.; Hendrickson, D. N.; Christou, G. Inorg. Chem. 1991, 30, 3192. Shirin, Z.; Young, Jr., V. G.; Borovik, A. S. J . Chem. Soc., Chem. Commun. 1997, 1967. Aurangzeb, N.; Hulme, C. E.; McAuliffe, C. A,; Pritchard, R. G.; Watkinson, M.; Bermejo, M. R.; Sousa, A. J. Chem. SOC.,Chem. Commun. 1994, 2193. Neves, A.; Vencato, I.; Erthal, S. M. D. Inorg. Chzm. Acta 1997,262, 77. Eichhorn, D. M.; Armstrong, W. H. J. Chem. SOC.,Chem. Commun. 1992, 85. Kitajima, N.; Komatsuzaki, H.; Hikichi, S.; Osawa, M.; Moro-oka, Y. J. Am. Chem. SOC. 1994, 116, 11596. Larson, E. J.;Pecoraro, V. L. J. Am. Chern. SOC.1991, 113, 3810. Law, N. A,; Machonkin, T. E.; McGorman, J . P.; Larson, E. J.; Kampf, J. W.; Pecoraro, V. L. J. Chem. SOC.,Chem. Comrnun. 1995, 2015. Saadeh, S. M.; Lah, M. S.; Pecoraro, V. L. Inorg. Chem. 1991.30, 8 . Chandra, S. K.; Chakravorty, A. Inorg. Chem. 1992,31, 760. Chaudhuri, P.; Wieghardt, K. In “Progress in Inorganic Chememistry”; Lippard, S. J., Ed.; Wiley Interscience, 1987,35, 329. Quee-Smith, V. C.; DelPizzo, L.; Jureller, S. H.; Kerschner, J. L. Inorg. Chern. 1996,35, 6461. Schlager, 0.; Wieghardt, K.; Nuber, B. Inorg. Chern. 1996,34, 6456. Du Bois, J.; Hong, J.; Carreira, E. M.; Day, M. W. J. Am. Chem. SOC.1996, 118, 915. Du Bois, J.; Tomooka, C. S.; Hong, J.: Carreira, E. M. Acc. Chem. Res. 1997, 30, 364. Du Bois, J.; Tomooka, C. S.; Hong, J.; Carreira, E. M.; Day, M. W. Angew. Chem., Int. Ed. Engl. 1997,36, 1645. Buchler, J.; Dreher, W. C.; Lay, K.-L.; Lee, Y. J. A.; Scheidt, W. R. Inorg. Chem. 1983, 22, 888. Collins, T. J.;Gordon-Wylie, S. W. J . Am. Chem. SOC. 1989, 111, 4511. Collins, T. J.; Powell, R. D.; Slebodnick, C . ; Uffelman, E. S. J . Am. Chem. SOC. 1990, 112, 899. Fackler, N. L. P.; Zhang, S.; OHalloran, T. V. J. Am. Chenz. SOC.1996, 118, 481. Niemann, A,; Bossek, U.; Haselhorst, G.; Wieghardt, K.; Nuber, B. Inorg. Chem. 1996, 35, 906.
434
LAW, CAUDLE, AND PECORARO
259. Tetard, D.; Rabion, A,; Verlhac, J.-B.; Guilbem, J . J. Chem. SOC.,Chem. Commun. 1995,531. 260. Cheng, B.; Cukiernik, F.; Fries, P.; Marchon, J.-C.; Scheidt, W. R. Inorg. Chem. 1995,34, 4627. 261. Cheng, B.; Fries, P. H.; Marchon, J.-C.; Scheidt, W. R. Inorg. Chem. 1996,35, 1024. 262. Kipke, C. A,; Scott, M. J.; Ghodes, J. W.; Armstrong, W. H. Inorg. Chem. 1990, 29, 2193. 263. Kitajima, N.; Osawa, M.; Tanaka, M.; Moro-oka, Y. J. Am. Chem. SOC. 1991, 113, 8952. 264. Schardt, B. C.; Hollander, F. J.; Hill, C. L. J. Am. Chem. SOC.1982, 104, 3964. 265. Schardt, B. C.; Hollander, F. J.; Hill, C. L. J. Chem. SOC.,Chem. Commun. 1981, 765. 266. Mikuriya, M.; Yamato, Y.; Tokii, T. Bull. Chern. SOC.Jpn. 1992, 65, 2624. 267. Mikuriya, M.; Yamato, Y.; Tokii, T. Inorg. Chim. Actu 1991, 181, 1. 268. Nishida, Y.; Oshino, N.; Tokii, T. 2. Nuturforsch. 1988,43B, 472. 269. Bonadies, J. A.; Kirk, M. L.; Lah, M. S.; Kessissoglou, D. P.; Hatfield, W. E.; Pecoraro, V. L. Inorg. Chem. 1989,28, 2037. 270. Larson, E.; Haddy, A,; Kirk, M. L.; Sands, R. H.; Hatfield, W. E.; Pecoraro, V. L. J. Am. Chem. SOC. 1992,114, 6263. 271. Kitajima, N.; Singh, U. P.; Amagai, H.; Osawa, M.; Moro-oka, Y. J. Am. Chem. SOC. 1991, 113, 7757. 272. Glerup, J.; Goodson, P. A.; Hazell, A.; Hazell, R.; Hodgson, D. J.; McKenzie, C. J.; Michelsen, K.; Rychlewska, U.; Toftlund, H. Inorg. Chem. 1994,33, 4105. 273. Gelasco, A.; Pecoraro, V. L. J. Am. Chem. SOC.1993, 115, 7928. 274. Zhang, Z. Y.; Brouca-Cabarreq, C.; Hemmert, C.; Dahan, F.; Tuchagues, J. P. J. Chem. SOC.,Dalton Trans. 1995, 1453. 275. Jensen, A. F.; Su, Z.; Hansen, N. K.; Larsen, F. K. Inorg. Chem. 1995,34, 4244. 276. Manchanda, R.; Brudvig, G. W.; de Gala, S.; Crabtree, R. H. Inorg. Chem. 1994, 33, 5157. 277. Frapart, Y.-M.; Boussac, A,; Albach, R.; Anxolabehere-Mallart, E.; Deiroisse, M.; Verlhac, J.-B.; Blondin, G.; Girerd, J.-J.; Guilhem, J.; Cesario, M.; Rutherford, A. W.; Lexa, D. J. Am. Chem. SOC. 1996,118, 2669. 278. Brewer, K. J.; Calvin, M.; Lumpkin, R. S.; Otvos, J. W.; Spreer, L. 0. Inorg. Chem. 1989,28, 4446. 279. Gelasco, A.; Kirk, M. L.; Kampf, J. W.; Pecoraro, V. L. Inorg. Chem. 1997,36, 1829. 280. Larson, E.; Lah, M. S.; Li, X.; Bonadies, J . A.; Pecoraro, V. L. Inorg. Chem. 1992, 31, 373. 281. Gohdes, J. W.; Armstrong, W. H. Inorg. Chem. 1992,31, 368. 282. Larson, E. J.; Riggs, P. J.; Penner-Hahn, J. E.; Pecoraro, V. L. J. Chem. SOC., Chem. Commun. 1992, 102. 283. Baldwin, M. J.; Stemmler, T. L.; Riggs-Gelasco, P. J.; Kirk, M. L.; Penner-Hahn, J. E.; Pecoraro, V. L. J. Am. Chem. SOC.1994, 116, 11349. 284. Torayama, H.; Nishide, T.; Asada, H.; Fujiwara, M.; Matsushita, T. Polyhedron 1998, 17, 105. 285. Torayama, H.; Nishide, T.; Asada, H.; Fujiwara, M.; Matsushita, T. Chem. Lett. 1996,387. 286. Wieghardt, K.; Bossek, U.; Nuber, B.; Weiss, J.; Bonvoisin, J.; Corbella, M.; Vitols, S. E.; Girerd, J. J. J. Am. Chem. SOC.1988, 110, 7398. 287. Bossek, U.; Weyhermiiller, T.; Wieghardt, K.; Nuber, B.; Weiss, J. J. Am. Chem. SOC.1990, 112, 6387.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
435
288. Arulsamy, N.; Glerup, J.; Hazell, A.; Hodgson, D. J.; McKenzie, C. J.; Toftlund, H. Znorg. Chem. 1994,33, 3023. 289. Oberhausen, K. J.; 0. B. R. J.; Richardson, J. F.; Buchanan, R. M.; Costa, R.; Latour, J.-M.; Tsai, H.-L.; Hendrickson, D. N. Znorg. Chem. 1993,32, 4561. 290. Bertoncello, K.; Fallon, G. D.; Murray, K. S. Polyhedron 1990, 9, 2867. 291. Pessiki, P. J.; Khangulov, S. V.; Ho, D. M.; Dismukes, G. C. J. Am. Chem. SOC. 1994,116, 891. 292. Pal, S.; Armstrong, W. H. Znorg. Chem. 1992, 31, 5417. 293. Wieghardt, K.; Bossek, U.; Zsoinai, L.; Huttner, G.; Blondin, G.; Girerd, J.-J.; Babonneau, F. J. Chem. SOC.,Chem. Comm.un. 1987, 651. 294. Bossek, U.; Saher, M.; Weyhermuller, T.; Wieghardt, K. J . Chem. SOC.Chem. Commun. 1992, 1780. 295. Reddy, K. R.; Rajasekharan, M. V.; Padhye, S.; Dahan, F.; Tuchagues, J.-P. Znorg. Chem. 1994,33, 428. 296. Bashkin, J. S.; Schake, A. R.; Vincent, J. B.; Chang, H.-R.; Li, Q.; Huffman, J . C.; Chem. Commun. 1988, 700. Christou, G.; Hendrickson, D. N. J. Chem. SOC., 297. Caneschi, A,; Ferraro, F.; Gatteschi, D.; Melandria, M. C.; Rey, P.; Sessoli, R. Angew. Chem., Znt. Ed. Engl. 1989,28, 1365. 298. Bossek, U.; Hummel, H.; Weyhermuller, T.; Wieghardt, K.; Russell, S.; van der Wolf, L.; Kolb, U. Angew. Chem., Znt. Ed. Engl. 1996, 35, 1552. 299. Coucouvanis, D.; Reynolds, R. A,; Dunham, W. R. J. Am.. Chem. SOC.1995, 117, 7570.
300. Nepveu, F.; Gaultier, N.; Korber, N.; Jaud, J.; Castan, P. J. Chem. SOC., Dalton Truns. 1995,4005. 301. Turner, P.; Gunter, M. J.; Hambley, T. W.; White, A. H.; Skelton, B. W. Znorg. Chem. 1992,31, 2295. 302. Chen, X.-M.; Tong, Y.-X.; Xu, Z.-T.; Mak, C. W. T. J. Chem. SOC.,Dalton Trans. 1995, 4001. 303. Menage, S.; Vitols, S. E.; Bergerat, P.; Codjovi, E.; Girerd, J.-J.; Guillot, M.; Solans, x.; Calvet, T. Znorg. Chern. 1991, 30, 2666. 304. Hodgson, D. J.; Schwartz, B. J.; Sorrell, T. N. Znorg. Chem. 1989,28, 2226. 305. Chang, H.-R.; Diril, H.; Nilges, M. J.; Zhang, X.; Potenza, J . A.; Schugar, H. J.; Hendrickson, D. N.; Isied, S. S. J. Am. Chem. SOC.1988, 110, 625. 306. Diril, H.; Chang, H.-R.; Zhang, X.; Larsen, S. J. Am. Chem. SOC.1987, 109, 6207. 307. Wesolek, M.; Meyer, D.; Osborn, J. A,; De Cian, A.; Fischer, J.; Derory, A.; Legoll, P.; Drillon, M. Angew. Chem., Int. Ed. Engl. 1994, 33, 1592. 308. Yu, S. B.; Wang, C. P.; Day, E. P.; Holm, R. H. Znorg. Chem. 1991, 30, 4067. 309. Ishimura, Y.; Inoue, K.; Koga, N.; Iwamura, H. Chem. Left. 1994, 1693. 310. Cortes, R.; Lezama, L.; Pizzaro, J. L.; Arriotura, M. I.; Solans, X.; Rojo, T. Angew. Chem., Znt. Ed. Engl. 1994, 33, 2488. 311. Cortes, R.; Pizzaro, J. L.; Lezama, L.; Arriotura, M. I.; Rojo, T. Inorg. Chem. 1994, 33, 2697. 312. Caudle, M. T.; Kampf, J. W.; Kirk, M. L.; Rasmussen, P. G.; Pecoraro, V. L. J . Am. Chem. SOC. 1997,119, 9297. 313. Auger, N.; Girerd, J.-J. J. Am. Chem. SOC.1990, 112, 448. 314. Pal, S.; Chan, M. K.; Armstrong, W. H. J . Am. Chem. SOC. 1992, 114, 6398. 315. Vincent, J. B.; Chang, H.-R.; Folting, K.; Huffman, J. C.; Christou, G.; Hendrickson, D. N. J . Am. Chem. SOC.1987,109, 5703. 316. Bakie, A. R. E.; Hursthouse, M. B.; New, L.; Thomton, P.; Whit, P. G. J . Chem. SOC.,Chem. Comnzun. 1980, 684.
436
LAW, CAUDLE, AND PECORARO
31 7. Bhula, R.; Gainsford, G. J.;Weatherburn, D. C. J. Am. Chem. SOC.1988,110, 7550. 318. Reynolds, R. A.; Yu, W. 0.; Dunham, W. R.; Coucouvanis, D. Inorg. Chem. 1996, 35, 2721. 319. Kessissoglou, D. P.; Kirk, M. L.; Bender, C. A,; Lah, M. S.; Pecoraro, V. L. J. Chem. SOC.,Chem. Commun. 1989, 84. 320. Kitajima, N.; Osawa, M.; Imai, S.; Fujisawa, R; Moro-oka, Y. Znorg. Chem. 1994, 33, 4613. 321. Rardin, R. L.; Bino, A.; Poganiuch, P.; Tolman, W. B.; Liu, S.; Lippard, S. J. Angew. Chem., Znt. Ed. Engl. 1990,29, 812. 322. Baldwin, M. J.; Kampf, J. W.; Kirk, M. L.; Pecoraro, V. L. Inorg. Chem. 1995, 34, 5252. 323. Wang, S.; Tsai, H.-L.; Hagen, K. S.; Hendrickson, D. N.; Christou, G. J. Am. Chem. SOC.1994, 116, 8376. 324. Hendrickson, D. N.; Christou, G.; Schmitt, E. A.; Libby, E.; Bashkin, J. S.; Wang, S.; Tsai, H. L.; Vincent, J. B.; Boyd, P. D. W.; Huffman, J. C.; Folting, K.; Li, &. Y.; Streib, W. E. J. Am. Chem. SOC.1992, 114, 2455. 325. Aubin, S. M. J.; Wemple, M. W.; Adams, D. M.; Tsai, H.-L.; Christou, G.; Hendrickson, D. N. J . Am. Chem. SOC.1996, 118, 7746. 326. Mikuriya, M.; Hashimoto, Y.; Kawamori, A. Chem. Lett. 1995, 1095. 327. Wemple, M. W.; Tsai, H.-L.; Folting, K.; Hendrickson, D. N.; Christou, G. Inorg. Chem. 1993,32, 2025. 328. Wang, S.; Tsai, H.-L.; Libby, E.; Folting, K.; Streib, W. E.; Hendrickson, D. N.; Christou, G. Znorg. Chem. 1996,35, 7578. 329. Brooker, S.; McKee, V.; Shepard, W. B. J. Chem. SOC.,Dalton Trans. 1987, 2555. 330. Wieghardt, K.; Bossek, U.; Gebert, W. Angew. Chem., Znt. Ed. Engl. 1983,22, 328. 331. Bashkin, J. S.; Chang, H.-R.; Streib, W. E.; Huffman, J . C.; Hendrickson, D. N.; Christou, G. J. Am. Chem. SOC.1987, 109, 6502. 332. Bouwman, E.; Bocar, M. A.; Libby, E.; Huffman, J. C.; Folting, K.; Christou, G. Inorg. Chem. 1992,31, 5185. 333. Gelasco, A,; Askenas, A.; Pecoraro, V. L. Inorg. Chem. 1996,35, 1419. 334. Chan, M. K.; Armstrong, W. H. J. Am. Chem. SOC.1991,113, 5055. 335. Kawasaki, H.; Kusumoki, M.; Hayashi, Y.; Suzuki, M.; Munezawa, K.; Suenaga, M.; Senda, H.; Uehara, A. Bull. Chem. SOC.Jpn. 1994, 67, 1310. 336. Philouze, C.; Blondin, G.; Girerd, J.-J.; Guilhem, J.; Pascard, C.; Lexa, D. J. Am. Chem. SOC.1994,116, 8557. 337. Jeffery, J. C.; Thornton, P.; Ward, M. D. Znorg. Chem. 1994,33, 3612. 338. Chan, M. K.; Armstrong, W. H. J. Am. Chem. SOC.1990,112, 4985. 339. Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry (John Wiley & Sons, New York, 1988). 340. Weil, J. A.; Bolton, J . R.; Wertz, J. E. Electron Paramagnetic Resonance: Elementary Theory and Practical Applications (John Wiley & Sons, New York, 1994). 341. Bencini, A.; Gatteschi, D. Electron Paramagnetic Resonance of Exchange Coupled Systems (Springer-Verlag, New York, 1989). 342. de Paula, J. C.; Beck, W. F.; Brudvig, G. W. J. Am. Chem. SOC.1986,108, 4002. 343. de Paula, J. C.; Brudvig, G. W. J . Am. Chem. SOC.1986,107, 2643. 344. h r l i n g , K.; Pace, R. J . Biophys. J. 1995, 68, 2081. 345. Bonvoisin, J.; Blondin, G.; Girerd, J. J.; Zimmermann, J. L. Biophys. J. 1992, 61, 1076. 346. Zheng, M.; Dismukes, G. C. Znorg. Chem. 1996,35, 3307.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
437
347. Tang, X. S.; Gilchrist, M. L.; Lorigan, G. A,; Larson, B.; Britt, R. D.; Diner, B. A. J. Am. Chem. SOC.1993, 115, 2382. 348. Britt, R. D.; Zimmermann, J.-L.; Sauer, K.; Klein, M. P. J. Am. Chem. SOC.1989, 111, 3522. 349. Randall, D. W.; Sturgeon, B. E.; Ball, J . A.; Lorigan, G. A,; Chan, M. K.; Klein, M. P.; Armstrong, W. H.; Britt, R. D. J. Am. Chem. SOC.1996, 117, 11780. 350. Khangulov, S. V.; Voyevodskaya, N. V.; Varynin, V. V.; Grebenko, A. I.; MelikAdamyan, V. R. Biofizika 1987,32, 1044. 351. Khangulov, S.; Sivaraja, M.; Barynin, V. V.; Dismukes, G. C. Biochemistry 1993, 32, 4912. 352. Waldo, G. S.; Fronko, R. M.; Penner-Hahn, J . E. Biochemistry 1991, 30, 10486. 353. Kirby, J. A,; Robertson, A. S.; Smith, J . P.; Thompson, A. C.; Cooper, S. R.; Klein, M. P. J. Am. Chem. SOC.1981, 103, 5528. 354. Kirby, J . A.; Goodin, D. B.; Wydrzynski, T.; Robertson, A. S.; Klein, M. P. J . Am. Chem. SOC.1981,103, 5537. 355. Guiles, R. D.; Zimmermann, J.-L.; McDermott, A. E.; Yachandra, V. K.; Cole, J . L.; Dexheimer, S. L.; Britt, R. D.; Wieghardt, K.; Bossek, U.; Sauer, K.; Klein, M. P. Biochemistry 1990,29, 471. 356. Berthomieu, C.; Boussac, A. Biochemistry 1996,34, 1541. 357. Guiles, R. D.; Yachandra, V. K.; McDermott, A. E.; Cole, J. L.; Dexheimer, S. L.; Britt, R. D.; Sauer, K.; Klein, M. P. Biochemistry 1990,29, 486. 358. Klein, M. P.; Sauer, K.; Yachandra, V. K. Photosyn. Res. 1993,38, 265. 359. Yachandra, V. K.; DeRose, V. J.; Latimer, M. J.; Mukerji, I.; Sauer, K.; Klein, M. P. Photochem. Photobiol. 1991,53, 98s. 360. Rompel, A,; Andrews, J. C.; Cinco, R. M.; Wemple, M. W.; Christou, G.; Law, N. A,; Pecoraro, V. L.; Sauer, K.; Yachandra, V. K.; Klein, M. P. J . Am. Chem. SOC.1997, 119, 4465. 361. Roelofs, T. A.; Liang, W.; Latimer, M. J.; Cinco, R. M.; Rompel, A.; Andrews, J. C.; Sauer, K.; Yachandra, V. K.; Klein, M. P. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 3335. 362. Riggs-Gelasco, P. J.; Mei, R.; Yocum, C. F.; Penner-Hahn, J. E. J . Am. Chem. SOC. 1996,118, 2387. 363. Amdt, D. In “Manganese Compounds as Oxidizing Agents in Organic Chemistry”; Lee, D. G., Ed.; Open Court Publishing Company: La Salle, IL, 1981. 364. Snider, B. B. Chem. Rev. 1996,96, 339. 365. Jacobsen, E. N. In “Comprehensive Organometallic Chemistry 11: A Review of the Literature 1982-1994”; Hegedus, L. S., Ed.; Elsevier Science, Inc.: Tarrytown, NY, 1995,12, 1097. 366. Katsuki, T. Coord. Chem. Reu. 1995, 140, 189. 367. Katsuki, T. J. Mol. Catal. A-Chem. 1996, 113, 87. 368. Groves, J. T.; Nemo, T. E.; Myers, R. S. J . Am. Chem. SOC.1979, 101, 1032. 369. Groves, J . T.; Haushalter, R. C.; Nakamura, M.; Nemo, T. E.; E. B. J., J. Am. Chem. SOC.1981,103, 2884. 370. “Cytochrome P450: Structure, Mechanism, and Biochemistry”; Ortiz de Montellano, P. R., Ed.; Plenum Press: New York, 1995. 371. Jorgensen, K. A. Chem. Rev. 1989,89, 431. 372. Jorgensen, K. A.; Schiott, B. Chem. Rev. 1990, 90, 1483. 373. Meunier, B. Chem. Rev. 1992,92, 1411. 374. Meunier, B., Ed., J . Mol. Catal. A-Chem. 1996, 113.
438
LAW, CAUDLE, AND PECORARO
375. Stern, M. K.; Groves, J . T. In “Manganese Redox Enzymes”; Pecoraro, V. L., Ed.; VCH Publishers, Inc.: New York, 1992, 233. 376. Groves, J. T.; Stern, M. K. J. A m . Chem. SOC.1983, 105, 5791. 377. Groves, J . T.; Nemo, T. E. J. Am. Chem. SOC.1983,105, 5786. 378. Groves, J. T.; Stern, M. K. J. Am. Chem. SOC.1987, 109, 3812. 379. Srinivasan, K.; Michaud, P.; Kochi, J. K. J. Am. Chem. SOC.1986, 108, 2309. 380. Samsel, E. G.; Srinivasan, K.; Kochi, J . K. J. Am. Chem. SOC.1986, 107, 7606. 381. Srinivasan, K.; Kochi, J. K. Inorg. Chem. 1985,24, 4671. 382. Feichtinger, D.; Plattner, D. A. Angew. Chem., Int. Ed. Engl. 1997,36, 1718. 383. Zhang, W.; Loebach, J . L.; Wilson, S. R.; Jacobsen, E. N. J. Am. Chem. SOC.1990, 112, 2801. 384. Jacobsen, E. N.; Zhang, W.; Muci, A. R.; Ecker, J . R.; Deng, L. J. Am. Chem. SOC. 1991,113, 7063. 385. Jacobsen, E. N.; Zhang, W.; Guler, M. L. J. Am. Chem. SOC.1991, 113, 6703. 386. hie, R.; Ito, Y.; Katsuki, T. Synlett 1991, 3, 265. 387. Irie, R.; Noda. K.; Ito, Y.; Matsumoto, N.; Katsuki, T. Tetrahedron Lett. 1990, 31, 7345. 388. Irie, R.; Noda, K.; Ito, Y.; Katsuki, T. Tetrahedron Lett. 1991, 32, 1055. 389. Tokunaga, M.; Larrow, J. F.; Fumistoshi, K.; Jacobsen, E. N. Science 1997,277, 936. 390. Linker, T. Angew. Chem., Znt. Ed. Engl. 1997,36, 2060. 391. Finney, N. S.; Pospisil, P. J.; Chang, S.; Palucki, M.; Konsler, R. G.; Hansen, K. B.; Jacobsen, E. N. Angew. Chem., Int. Ed. Engl. 1997,36, 1720. 392. Linde, C.; Arnold, M.; Norrby, P.-0.; Akermark, B. Angew. Chem., Int. Ed. Engl. 1997,36, 1723. 393. Fu, H.; Look, G. C.; Zhang, E. N.; Jacobsen, E. N.; Wong, C.-H. J . Org. Chem. 1991,56, 6497. 394. Sharpless, K. B.; Teranish:, A. Y.; Backvall, J.-E. J. Am. Chem. SOC.1977,99, 3120. 395. Norrby, P.-0.; Linde, C.; Akermark, J . J. Am. Chem. SOC.1995, 117, 11035. 396. Yamada, T.; Imagawa, K.; Mukaiyama, T. Chem. Lett. 1992, 2109. 397. Collman, J. P.; Zhang, X.; Lee, V. J.; Uffelman, E. S.; Braumann, J. Science 1993, 261, 1404. 398. Lai, T.-S.; Kwong, H.-L.; Che, C.-M.; Peng, S.-M. J. Chem. SOC.,Chem. Commun. 1997,2373. 399. Halterman, R. L.; Jan, S.-T. J . Org. Chem. 1991, 56, 5253. 400. Groves, J . T.; Han, Y.-Z. In “Cytochrome P450: Structure, Mechanism, and Biochemistry”; Ortiz de Montellano, P. R., Ed.; Plenum Press: New York, 1996, 3. 401. Beckman, J. S. Chem. Res. Toxicol. 1996, 9, 836. 402. Beal, M. F. Neuroscientist 1997,3, 21. 403. Hunt, J . A.; Lee, J. B.; Groves, J. T. Chem. Biol. 1997, 4, 845. 404. Marla, S. S.; Lee, J.; Groves, J . T. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 14423. 405. Groves, J . T.; Takahashi, T.; Butler, W. M. Inorg. Chem. 1983,22, 884. 406. Chang, C. J.; Low, D. W.; Gray, H. B. Znorg. Chem. 1997,36, 270. 407. Bottomley, L. A.; Neeley, F. L. Znorg. Chem. 1997,36, 5435. 408. Dehnicke, K.; Strahle, J. Angew. Chem., Int. Ed. Engl. 1992,31, 955. 409. Groves, J . T.; Takahashi, T. J . Am. Chem. SOC.1983, 105, 2073. 410. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. J. Chem. Res. ( S ) 1989, 360. 411. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. J. Chem. Res. ( S )1989, 108. 412. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. Synth. Commun. 1991,21, 489. 413. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. J. Chem. Res. ( S )1991, 284.
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
439
414. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. Gazz. Chim. Italiana 1993, 123, 289. 415. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. J . Chem. Res. ( S )1990, 188. 416. Kurosawa, K.; Yamaguchi, K. Bull. Chem. Soc. Jpn. 1981,54, 1757. 417. Donnelly, K. D.; Fristad, W. E.; Gellerman, B. J.; Peterson, J. R.; Selle, B. J. Tetrahedron Lett. 1984,25, 607. 418. Marko, I. E.; Richardson, P. F. Tetrahedron Lett. 1991, 32, 1831. 419. Richardson, P. F.; Marko, I. E. Synlett, 1991, 733. 420. Uemara, S.; Okazaki, H.; Onoe, A.; Okano, M. Bull. Chem. Soc. Jpn. 1978, 51, 3568. 421. Tanner, D. D.; Gidley, G. C. J . Org. Chem. 1968, 33, 38. 422. Poutsma, M. L. J. Am. Chem. Soc. 1965,87, 4285. 423. Marko, I. E.; Richardson, P. R.; Bailey, M.; Maguire, A. R.; Coughlan, N. Tetrahedron Lett. 1997, 38, 2339. 424. Huynh, V.-B.; Crawford, R. L. FEMS Microbinl. Lett. 1985,28,119. 425. Saadeh, S. M. Ph.D. Thesis, University of Michigan, 1992. 426. Policar, C.; Artaud, I.; Mansuy, D. Znorg. Chem. 1996,35, 210. 427. Riley, D. P.; Lennon, P. J.; Neumann, W. L.; Weiss, R. H. J. Am. Chem. Soc. 1997, 119, 6522. 428. Weiss, R. H.; Flickinger, A. G.; Rivers, W. J.; Hardy, M. M.; Aston, K. W.; Ryan, U. S.; Riley, D. P. J. Biol. Chem. 1993,268, 23049. 429. Beyer, W. F. J.; Fridovich, I. Arch. Biochem. Biophys. 1989, 271, 149. 430. Beyer, W. F. J.;Fridovich, I. Methods E n q m . 1990, 186, 242. 431. Darr, D.; Zarilla, K. A,; Fridovich, 1. Arch. Biochem. Biophys. 1987,258, 351. 432. Rush, J . D.; Maskos, Z.; Koppenol, W. H. Arch. Biochem. Biophys. 1991,289, 97. 433. Matsushita,T.; Shono, T. Bull. Chem. Soc. Jpn. 1981, 54, 3743. 434. Baudry, M.; Etienne, S.; Bruce, A,; Palucki, M.; Jacobsen, E.; Malfroy, B. Biochem. Biophys. Res. Commun. 1993, 192, 964. 435. Doctorow, S. R.; Huffman, K.; Marcus, C . B.; Musleh, W.; Bruce, A.; Baudry, M.; Malfroy, B. Academic Press: New York, 1997,38, 247. 436. Rajan, R.; Rajaram, R.; Nair, B. U.; Ramasami, T.; Mandal, S. K. J . Chem. Soc., Dalton Trans. 1996, 2019. 437. Frasch, W. D.; Mei, R. Biochim.. Biophys. Acta 1987, 891, 8. 438. Mathur, P.; Crowder, M.; Dismukes, G. C. J . Am. Chem. Soc. 1987,109, 5227. 439. Pessiki, P. J.; Dismukes, G. C. J. Am. Chem. Soc. 1994, 116, 898. 440. Gelasco, A,; Bensiek, S.; Pecoraro, V. L. Znorg. Chem. 1998, 37, 3301. 441. Sakiyama, H.; Okawa, H.; Isobe, R. J. Chem. Soc., Chem. Commun. 1993,882. 442. Sakiyama, H.; Tamaki, H.; Kodera, M.; Matsumoto, N.; Okawa, H. J. Chem. Sac., Dalton Trans. 1993, 591. 443. Naruta, Y.; Maruyama, K. J. A m . Chem. Soc. 1991, 113, 3595. 444. Naruta, Y.; Sasayama, M. J . Chem. Soc., Chem. Commun. 1994, 2667. 445. Naruta, Y.; Sasayama, M.-A.; Ichihara, K. J . Mol. Cat. A 1997, 117, 115. 446. Larson, E. J.; Pecoraro, V. L. J . Am. Chem. SOC.1991, 113, 7809. 447. Stibrany, R. T.; Gorun, S. M. Angew. Chem., Int. Ed. Engl. 1990, 29, 1156. 448. Mathews, J. H.; Dewey, L. H. J. Phys. Chem. 1913, 17, 211. 449. Ashmawy, F. M.; McAuliffe, C . A,; Parish, R. V.; Tames, J. J . Chem. Soc., Chem. Cornmun. 1984, 14. 450. Ashmawy, F. M.; McAuliffe, C. A. J . Chenz. Soc., Dalton Trans. 1985, 1391. 451. Watkinson, M.; Whiting, A,; McAuliffe, C . A. J . Chem. Soc., Chem. Commun. 1994, 2141.
440 452. 453. 454. 455. 456. 457. 458. 459. 460. 461. 462.
463. 464.
LAW, CAUDLE, AND PECORARO
Fujiwara, M.; Matsushita, T.; Shono, T. Polyhedron 1985,4, 1895. Matsushita, T.; Fujiwara, M.; Shono, T. Chem. Lett. 1981, 631. Matsushita, T.; Hirata, Y.; Shono, T. Bull. Chem. SOC.Jpn. 1982,55, 108. Naruta, Y.; Sasayama, M.; Sasaki, T. Angew. Chem., Znt. Ed. Engl. 1994,33, 1839. Limburg, J.; Brudvig, G. W.; Crabtree, R. H. J . Am. Chem. SOC.1997,119, 2761. Lind, J.; Shen, X.; Ericksen, T. E.; Merenyi, G. J. Am. Chem. SOC.1990,112, 479. McMillen, D. F.; Golden, D. M. Ann. Rev. Phys. Chem. 1982,33, 493. Thorp, H. H.; Sarneski, J. E.; Brudvig, G. W.; Crabtree, R. H. J. Am. Chem. SOC. 1989, I l l , 9249. Cooper, S. R.; Calvin, M. J. Am. Chem. SOC.1977,99, 6623. Shank, M.; Barynin, V.; Dismukes, G. C. Biochemistry 1994,33, 15433. Vainshtein, B. K.; Melik-Adamyan, W. R.; Barynin, V. V.; Vagin, A. A. In “Progress in Bioorganic Chemistry and Molecular Biology”; Ovchinnikov, Y. A,, Ed.; Elsevier Science Publishers: New York, 1984, 117. Nagata, T.; Mizukami, J. J. Chem. SOC.,Dalton Trans. 1995, 2825. Nishida, Y.; Akamatsu, T.; Kazuyoshi, T.; Sakamoto, M. Polyhedron 1994, 13,
2251. 465. Dave, B. C.; Czernuszewicz, R. S. Znorg. Chin. Acta. 1994,227, 33. 466. Higuchi, C.; Sakiyama, H.; Okawa, H.; Fenton, D. E. J. Chem. Soc., Dalton Trans. 1995,4015. 467. Riggs-Gelasco, P. J. PbD. Thesis, University of Michigan, 1995. 468. Defrance, S.; Meunier, B.: Seris, J.-L. New J. Chem. 1992, 16, 1015, 469. Tangoulis, V.; Malamatari, D. A.: Soulti, K.; Stergiou, V.; Raptopoulou, C. P.; Terzis, A.; Kabanos, T. A.; Kessissoglou, D. P. Znorg. Chem. 1996,35, 4974. 470. Cheng, B.; Scheidt, W. R. Acta Cryst. 1996, C52, 361. 471. Stephenson, G. R. In “Advanced Asymmetric Synthesis”; Stephenson, G. R., Ed.; Chapman and Hail: New York, 1996,367. 472. Vincent, J. B.; Tsai, H.-L.; Blackman, A. G.; Wang, S.; Boyd, P. D. W.; Folting, K.; Huffman, J. C.; Lobovsky, E. B.; Hendrickson, D. N.; Christou, G. J. Am. Chem. SOC.1993, 115, 12353. 473. Yu, S.-B.;Lippard, S. J.; Shewky, I.; Bino, A. Znorg. Chem. 1992,31, 3502. 474. Mikuriya, M.; Fujii, T.; Tokii, T.; Kawamori, A. Bull. Chem. SOC.Jpn. 1993, 66, 1675. 475. D i d , H.; Chang, H.-R.; Nilges, M. J.; Zhang, X.; Potenza, J. A.; Schugar, H. J.; Isied, S. S.; Hendrickson, D. N. J. Am. Chem. SOC.1989, 111, 5102. 476. Kessissoglou, D. P.; Butler, W. M.; Pecoraro, V. L. Znorg. Chem. 1987,26, 495. 477. Riggs-Gelasco, P. J.; Rui, M.; Ghanotakis, D. F.; Yocum, C. F.; Penner-Hahn, J. E. J . Am. Chem. SOC.1996,118, 2400. 478. Michaud-Soret, I.; Jacquamet, L.; Debaecker-Petit, N.; Le Pape, L.; Barynin, V. V.; Latour, J.-M. Znorg. Chem. 1998,37, 3874.
ADVANCES IN INORGANIC CHEMISTRY. VOL. 46
CALCIUM-6INDING PROTEINS BRYAN E. FINN and TORBJORN DRAKENBERG Division of Physical Chemistry 2, Lund University, S-22100 Lund, Sweden
I. Introduction 11. Intracellular EF-Hand Calcium-Binding Proteins A. Introduction B. Calmodulin C. Troponin C D. S l O O proteins 111. Calcium-Mediated Membrane-Binding Proteins A. Introduction B. Recoverin C. Annexins D. C2 Domains E. y-Carboxyglutamic Acid Sites IV. Extracellular Calcium-Binding Proteins A. Introduction B. Conantokins C. EGF-like Domains D. Serine Proteases E. a-Lactalbumin and Lysozyme F. Cadherins G. SPARC V. Summary References
I. Introduction
Calcium is of widespread and fundamental importance in biochemistry, for the calcium ion functions as a second messenger, that is, one whose signals are propagated by proteins specifically evolved for this purpose. It is regarded as one of the most important bioinorganic elements (1). Apart from its relevance to biological solid-state materials, Ca2' has key roles in many physiological processes. The Ca2+ion has a radius of 100 pm and can be compared to Mg2+(72 pm) with its 441 Copyright 0 1999 by Academic Press All rights of repioduct~onin any farm reserved 0898.8638199 $25 00
442
FINN AND D-NBERG
strict coordination requirements, having coordination numbers from 6 to 8. In its biological coordination chemistry it has a preference for 0-donor ligands from aminoacid side-chain carboxylate or hydroxyl groups, peptide carbonyl oxygens, or HaO. The hexaaqua ion is very labile and has an HaO exchange “off’ rate constant of 5 10’ s-’ (2). This could in a simplistic way be thought of as an upper limit for the “on” rate of Ca2+to its binding site. Higher on rates are possible, however, when outer-sphere complexes are favorable and especially when the binding site is highly negatively charged as discussed in (3,4). A review on calcium-binding proteins is a quite difficult task. During work on this review we have found that there are now an immense number of such proteins, making some selection necessary. Even for those proteins selected we have no claims of completeness in our coverage. We have therefore divided the proteins (or more correctly protein domains), into those that are (a) intracellular, (b) calcium mediated membrane bound, and (c) extracellular. The intracelM a r calcium-binding proteins are dominated by the EF-hand family (defined later). This can be clearly seen from the recent book “Guidebook to the Calcium Binding Proteins” (5),in which the EF-hand proteins occupy 180 of a total of 230 pages. Likewise, in the section on intracellular proteins in this review we have chosen to concentrate on the EF-hand family. In the section on calcium-mediated membrane-binding proteins we have chosen to be more varied and discuss proteins interacting with membranes from the inside as well as from the outside. In this section it might be more proper to talk about protein domains than proteins. In general it is the location of the calcium-binding domain that has been used to define whether it is intra- or extracellular or membrane bound. For example, the cadherins are treated as extracellular in this review even though they are truly membrane proteins, and the various domains from the blood coagulation factors are treated as membrane binding (Gla domains) as well as extracellular (EGF and serine protease domains). We hope this will not be too confusing for those used to other definitions; however, in this review focused on the calcium ion we found this division natural. We are aware that several important calcium-binding proteins have been neglected and that our selection of proteins is clearly biased by our own particular interests. We have also included some odd proteins such as conantokins and neglected more important ones such as phospholipase Aa.
CALCIUM-BINDING PROTEINS
443
(I. lntracellular EF-Hand Calcium-Binding Proteins
A. INTRODUCTION The EF hand is by far the most common motif for intracellular calcium-binding proteins. Briefly, it is an approximately 30 amino acidlong peptide chain composed of a central calcium-binding loop flanked by two alpha helixes. The name EF-hand was coined by Kretsinger ( 6 ) and refers to the resemblance of the E and F helices of parvalbumin to a hand with the thumb and index finger extended. The calcium ion is bound by oxygen ligands contributed mostly from carboxylic acid side chains from aspartic or glutamic acid. Often one ligand is contributed by water, and in rare cases, as discussed later, the backbone carbonyl oxygens can contribute ligands. The sequence of a typical EF-hand is shown in Fig. 1 and a stick figure of its calciumbinding loop is shown in Fig. 2. The ligating residues are relatively highly conserved and have been used t o identify almost a thousand EF-hands in over 150 proteins (7). However, of those identified, approximately a third are believed to have lost their ability to bind calcium tightly, that is to say, with a dissociation constant of 10 ' to lo-* M as is typical for these proteins. As expected, those that have
Helix
E F L E
D L T
W L
v
Helix
K K I S
FIG.1. A schematic drawing of the second EF-hand calcium-binding subdomain from calbindin Dgk (residues 45-74). Ligands from side-chain carboxyls are indicated in solid lines; the ligand from the backbone carbonyl is indicated by a dashed line.
444
FINN AND DRAKENBERG
FIG.2. A stick figure of the second EF-hand binding loop from calbindin Dgk (residues 54-65). The carboxylic side chains and backbone carbonyls that ligand the calcium ion are indicated in black.
lost the ability to bind calcium are also those whose sequences vary the most from the canonical EF-hand sequence. A second feature that is conserved among the EF-hand proteins is the arrangement of hydrophobic residues on the interior faces of the two helices and the central loop. These residues are clustered as a hydrophobic surface and help to explain another key feature of EFhand motifs. This feature is that except on rare occasions, EF-hands occur in pairs with their hydrophobic surfaces joined together and the calcium loops paired in a small P-sheet interaction. The two EF-hands in the pair share a pseudo-axis of rotation centered between the four helices. This four-helix bundle arrangement is very stable and is observed even in non-calcium-binding proteins such as hemerythrin (8) and cytochrome BEG2 (9).For example, the stability of calbindin Dgk,a small calcium-binding protein composed of a single EF-hand pair, in its calcium-loaded state is so high that it cannot be denatured by urea or heating to 100°C (10).
CALCIUM-BINDING PROTEINS
445
Although EF-hands tend to occur in pairs, proteins can contain anywhere from 1 to 8 EF-hand motifs with the most common number being 4. A survey of a large number of known EF-hand calciumbinding proteins has been presented by Nakayama, et al. (7). The functions of EF-hand proteins can be divided into two categories, signaling and buffering, and the distinction between the two groups is determined by their structural response to calcium binding. In the case of signaling proteins, the binding of calcium induces a conformational change, which renders the protein in its active form. Once active, the protein is able to bind a target protein or peptide to carry out its function. The classical example of this type is calmodulin. The activation process is tightly regulated by the concentration of calcium in the cell, which is M in a resting cell but can increase by several orders of magnitude as a result of calcium intake from outside the cell or release from internal stores. In the case of the buffering proteins, calcium binding has little or no significant effect on the structure of the protein and the function of the protein is most likely limited to its calcium-binding role. The best studied example of the latter is calbindin D9k. The majority of the EF-hand family of calcium-binding proteins occur within the cell. However, there are several recent cases in which EF-hand motifs have been identified in proteins expressed outside of the cell or on the cell surface. These are discussed in Sections I11 and IV.In this section we will concentrate our discussions on intracellular calcium-binding proteins and give some examples of the best-known and most studied examples of such proteins. This is by no means an exhaustive list of EF-hand proteins, only a consideration of some examples that have been extensively studied from a structural perspective. Excellent introductions to the entire EF-hand family by Kawasaki and Kretsingar ( 1 1 ) and Celio ( 5 ) are recommended to those interested in other members of this fascinating family of proteins.
B. CALMODULIN That the EF-hand proteins are often referred to as the calmodulin superfamily of calcium-binding proteins is no coincidence. Though the term “EF-hand was coined for parvalbumin (61, calmodulin is the most well characterized of the EF-hand proteins with respect to every aspect from molecular structure to physiological role. It serves a central role as a calcium-sensitive second messenger and is found in all eukaryotic organisms (12). Rather than having a single target as is
446
FINN AND DRAKENBERG
common for signaling proteins, calmodulin has over 100 targets and thus regulates a wide range of cellular function including signal transduction, DNA synthesis, secretion, motility and cell division ( I 1, 12).It is composed of a single chain of 148 amino acids with a molecular mass of approximately 18 kDa. It contains four EF-hand calciumbinding subdomains that are arranged pairwise in two domains separated by a central linker. All four sites are active and bind calcium with high affinity, with the carboxy-terminal EF-hand pair having slightly higher affinity (KD= MI than the amino-terminal pair (KD= M). The first structure of calcium-loaded calmodulin from rat was determined by X-ray crystallography in 1985 (13)and was later refined to high resolution in 1988 (14). These structures reveal a surprising architecture for the protein in that the two EF-hand pair domains are separated by a long, apparently rigid helix (Fig. 3). This “dumbbell” shape of calmodulin was surprising because biochemical data indicated that the target binding site was evenly distributed between the two domains, but its target peptides were obviously too small to span the distance between the two domains. In addition, small-angle X-ray scattering and neutron scattering experiments indicated that the solution form of calmodulin was much more compact than that determined by X-ray crystallography (15,161. Further X-ray structures of calmodulin from human (171, Drosophila (18) and Paramecium (19) spp. also demonstrated the presence of a central helix. It was not until a high-resolution structure of calmodulin in solution was obtained by nuclear magnetic resonance (NMR) spectroscopy that the apparent paradox of calmodulin’s structure was resolved. Al-
FIG.3. A ribbon diagram of active (Cazt human calmodulin (1 7). The central flexible linker separating the two EF-hand pair domains is indicated in dark gray.
CALCIUM-BINDING PROTEINS
447
though there was good agreement between the structures determined by X-ray crystallography and NMR spectroscopy, the central linker was found to be nonhelical in solution (20). Furthermore, measurements of the dynamics of the backbone using NMR relaxation data showed that this linker was more flexible than the surrounding helices (21). Target binding by calmodulin is an interesting case of molecular recognition. Although calmodulin binds a large number of target proteins, it binds with high affinity, often in the nanomolar range. Therefore, one might have expected that they share a common calmodulin binding sequence. However, the many calmodulin recognition sequences share little sequence homology and only resemble one another in the general properties of being 14 to 26 amino acids long, hydrophobic, and basic (22). Both NMR spectroscopy and X-ray crystallography played key roles in understanding the target binding of calmodulin. In this case, the two methods show much more agreement. Three high-resolution structures of complexes between calmodulin and a target protein have been published, two by X-ray crystallography on complexes with peptides with the smooth muscle light-chain kinase (smMLCK) (23)and calmodulin-dependent protein kinase I1 (CaMKII) (241, and one by NMR spectroscopy on the complex with a peptide from skeletal muscle light-chain kinase (skMLCK) (251. This latter structure is shown in Fig. 4. The smMLCK and skMLCK complex structures resemble one another, as is expected because the peptides share a degree of sequence homology. The main features of the complex are two hydrophobic anchors a t opposite ends of the peptide that interact with hydrophobic residues in the calmodulin-binding pocket. In addition, electrostatic interactions between the basic residues of the peptide with glutamic acid residues in both domains of calmodulin are observed. In the case of CaMKII, the hydrophobic anchors are present, but are closer to one another in the peptide sequence. Calmodulin adapts its binding to this sequence in two ways. First, because the two domains are connected by a flexible tether, they are able to rearrange with respect to one another and the target peptide so as best to accommodate the target peptide’s hydrophobic amino acid pattern. The second adaptation occurs within the two binding clefts themselves. These clefts contain a larger than usual proportion of methionine residues that have a long side chain, thus giving them more degrees of freedom to adjust their conformations as necessary to accommodate the target peptide.
448
FINN AND DRAKENBERG
FIG.4. A ribbon diagram of the binding of skMLCK peptide to calmodulin (25).The peptide and calcium ions are indicated in light gray and calmodulin in dark gray.
Although the structures of calmodulin bound to target peptides have yielded much information about the complexes and their plasticity, the mechanism by which binding of calmodulin to the target proteins activates them remains to be definitively demonstrated. This will require the structure of a calmodulin-protein complex. However, much evidence available now allows us to infer the mechanism of calmodulin-dependent activation. The strongest hypothesis is the pseudosubstrate or autoinhibition hypothesis, which was first suggested by Kemp et al., who noted the similarity of the substrate and autoinhibitory region of smMLCK (26). Subsequently, it was demonstrated that removal of the autoinhibitory region by proteolysis could activate smMLCK in the absence of calmodulin (27).A recent structure of the autoinhibited form of calmodulin-dependent protein kinase I (CaMKI) clearly shows that its autoinhibitory region lies in the substrate binding site and that part of the calmodulin binding segment protrudes out, thus potentially able to bind calmodulin (28). Though the vast majority of target proteins bind to calciumactivated calmodulin, there are cases in which calmodulin has been
CALCIUM-BINDING PROTEINS
449
FIG.4. (continued)
reported to bind proteins in the absence of calcium. However, in most cases they are believed to be affected by changes in calcium concentration. One example is that of neuromodulin, which binds calmodulin tighter in the absence of calcium at low ionic strength. The physiological significance of this binding is unclear because it is much weaker at physiological salt concentrations. Other evidence suggests that neuromodulin may act to bind and sequester calmodulin a t the plasma membrane in a calcium-dependent manner (29,301. The mechanism of calcium-induced activation of calmodulin has been and continues to be a debated topic. Since the first X-ray crystal structure of calcium-loaded calmodulin (13), models have been proposed for the structure of the apo form of calmodulin (32) based on
450
FINN AND DRAKENBERG
the structure of related troponin C, which was first crystallized with only two of four calcium-binding sites loaded with calcium (32, 33). Briefly, the model predicted that the hydrophobic binding sites of calcium-calmodulin are closed and sequestered from solvent in the apo form. The general features of this model were confirmed ten years after the crystal structure of the calcium-loaded from by several solution structures of apo-calmodulin determined by NMR (34-36). The conformational change observed for the carboxy-terminal domain is shown in Fig. 5 . Additionally, this structure demonstrated the autonomy of the two domains with respect to their conformational response to calcium binding being independent of the linker connecting the two domains (37).
C. TROPONIN C Troponin C (TnC) is a calcium-binding protein that is part of the troponin complex localized on the thin filament of muscle fibers. The other two components of the complex are troponin I (TnI) and troponin T (TnT), which together with TnC are arranged in a heterotrimer. Two isoforms are known, the fast skeletal muscle (sTnC) and the cardiac and slow skeletal muscle (cTnC) forms. The structure of TnC is similar to that of calmodulin, with four E F hands organized pairwise in two domains. Indeed, our understanding of these two proteins has evolved in parallel, with a good deal of what has been learned about troponin C being applicable to calmodulin and vice versa. Like cal-
FIG.5. A space-filling drawing of the conformational change in the TRzC domain of calmodulin (37). The hydrophobic side chains lining the half of the target binding site in the TRzC domain are indicated in white with the backbone and remainder of the side chains in gray.
CALCIUM-BINDING PROTEINS
45 1
modulin, all four calcium-binding sites in sTnC are active, with the carboxy-terminal sites being tighter than the N-terminal pair (KD lo-* M vs. KD M). However, in cTnC the first site is inactive. Unlike calmodulin, which binds to a large array of target proteins, TnC has one very specific function, namely the cyclic binding and release of TnI in a calcium-dependent manner, which in turn regulates muscle contraction. The structure of sTnC was first determined by X-ray crystallography and remarkably in a half-saturated (Ca2+)2 state (32, 33). These initial structures formed the basis for modeling of the mechanism of calcium-induced activity of troponin C (38) as well as the calciuminduced activation of calmodulin (31).Studies of the solution structures of sTnC added the structures of the (Ca2’l4 calcium-saturated form (39)and the apo and calcium-saturated forms of the N-terminal, regulatory, domain (40,411to the already known half-saturated form. Together, these paint a complete picture of the structural changes induced by calcium binding on troponin C and have led to a model for troponin C’s role in muscle contraction. The model predicts that in the resting state, only the C-terminal domain of sTnC is calcium loaded and bound to TnI a t its N-terminal domain. Actin is also bound to TnI in an inhibitory complex. Upon binding of calcium to TnC’s N-terminal domain, a hydrophobic patch is opened and binds to the inhibitory and C-terminal domains of TnI, thus releasing actin from TnI. The actin is thus available for binding to myosin, which leads to the muscle contraction. Support for another part of this model was provided by studies of the interaction of sTnC and its binding target domain on TnI by NMR (421,which confirmed that TnI interacts with the regulatory domain of TnC. Further details of the mechanism of action of troponin C have also been provided by solution structural studies, including mutants of troponin C aimed a t dissecting the role of individual residues in the process. The final bidentate glutamate ligand in the calcium-binding loops has been a n especially attractive target in troponin C (43, 44) as well as calmodulin (45, 46). These studies have demonstrated these to be key residues in the calciuminduced activation process. Whether they are the only keys or part of an interplay between a number of calcium-ligand and other interactions is a subject of continuing study.
-
-
D. S l O O PROTEINS The S l O O proteins make up a distinct subfamily within the EFhand family. They share structural characteristics of being small
452
FINN AND DRAKENBERG
(9-12 kDa), being acidic, containing only 2 EF-hands, and having a homology of between 25 and 40%. They also share one feature that is unique among calcium-binding EF-hand proteins, that is, the insertion of two residues in the first EF-hand, which alters the conformation of the calcium-binding loop such that the ligands are contributed from the peptide backbone carbonyl oxygens instead of the side-chain carboxyl groups as in a standard EF-hand. This variation is known as a “pseudo-EF-hand.” Beyond this, they differ widely with respect to expression, activity, oligomer state, and conformational response to calcium binding. 1. Calbindin Dgk
Calbindin Dgk,also known as ICBP and SlOOD, is expressed highly in the intestine, placenta, and uterus, and has been suggested to be involved in calcium transport, though a definitive demonstration of its function remains to be accomplished. It has been shown to interact with the calmodulin-binding domain of the plasma membrane calcium pump (47),though it is unclear if this interaction plays a significant functional role. It is a unique member of the ,9100 family in that it is monomeric whereas most members of this subfamily are homo- or heterodimeric. From a structural perspective, calbindin Dgk is the most extensively studied of the SlOO subfamily. It is precisely its small size as well as its stability that makes it an attractive target for structural studies. The first X-ray structure of calcium-loaded calbindin Dgk was determined in 1981 (48) and was later refined to 2.3 A (49),and then to 1.6 (50). Solution structures have also been determined for several forms of bovine calbindin Dgkincluding (Ca2+12(51), apo (52, 531, (Cd2+I1(54) and (Cd2+I2.In addition, the structure of the porcine protein has been determined (55) by NMR methods. These studies when taken together show another unique feature of calbindin Dgk:the relatively small effect that calcium has on the conformation of the protein. On going from the apo to calcium-loaded state, only a minor adjustment of the secondary structure is observed, and this is limited to the second EF hand. The structure of the singly loaded (Cd2+Ilstate showed that most of these changes occur upon binding of the first ion, which binds in the second, C-terminal binding site (54). This lack of a large structural response to calcium binding, which is typical for the calcium sensors such as calmodulin and troponin C, indicates that calbindin Dgk)s function is probably not one of signaling but more likely one of regulating calcium homeostasis.
A
CALCIUM-BINDING PROTEINS
453
Although ion binding has little effect on the structure of calbindin, it has been shown to have a large effect on the dynamics and stability of the protein. Calbindin Dgkhas served as a convenient model system for studies of protein dynamics by NMR and other techniques. Results from studies of hydrogen-deuterium exchange (56, 57) have shown that the structure of calbindin can be significantly stabilized by ion binding. The effect is greatest in the binding loops, but it is also propagated through hydrogen-bonding networks in the helices. Similar studies on the fast (nano- to picosecond) time-scale dynamics by spinrelaxation techniques (58, 59) have shown that ion binding exerts an effect here as well but is localized to the C-terminal binding loop, which is more dynamic in the apo than in the calcium-loaded form. Structural studies of calbindin Dgkhave also yielded several unique observations of a more general nature. The first was the observation that the purified protein appeared heterogeneous in isoelectricfocusing experiments. This heterogeneity was traced to an isoaspartyl linkage formed by deamidation of an Asn-Gly pair (60).A second apparent heterogeneity in the protein preparation was identified as well. However, instead of being a configurational heterogeneity it was found to be a conformational heterogeneity traced to the cis-trans isomerization of a proline in the linker between the two EF hands (61).This result, later confirmed by crystallographic studies GO), was one of the first direct demonstrations that such isomers of proline occur at equilibrium. Calbindin Dgkhas also proved to be a convenient model system for studies of calcium binding and cooperativity (62).Because it has only two calcium ions, it presents a simpler system than those proteins that have four or more calcium binding sites.
2. sloop
Sloop occurs most commonly as SlOOB, a pp dimer, but can form heterodimers (ap). It is one of the better characterized members of the S l O O family. It was the first protein given the designation S100, which was given to Sloop when it was first isolated and means, simply, that it is partially soluble in 100% saturated ammonium sulfate (63). Thus, though the name S l O O does not indicate any functional aspect, it has since been used by some to designate this entire family of proteins due to their homology to Sloop and common chromosomal localization (64). The exact function of Sloop is not known. However, it has been implicated in a number of cellular functions such as cell growth,
454
FINN AND DRAKENBERG
apoptosis, and energy metabolism. It is most abundant in the nervous system, especially the glial cells, but it can also be found in other tissues such as the skin, testis and cartilage (65). I t has also been shown to exhibit extracellular functions related to neuronal and glial cell growth and proliferation (66). Sloop contains two cysteine residues (68 and 84),and the SlOOB dimer is thus able to form a covalently linked dimer. Although the disulfide linkage is not necessary for dimerization, it has been shown to be essential for its extracellular function and some intracellular functions (67). Calcium binding causes a conformational change in SlOOB, which produces an increase in hydrophobic surface area, similar to calmodulin and troponin C. However, the details of this conformational change are only now beginning to come to light. The high-resolution structure of SlOOB was first determined in apo form for both the bovine (68) and rat (69) proteins. The rat SlOOB structure is shown in Fig. 6. These structures showed that the protein is composed of a dimer of two regular EF-hand pairs as expected. The dimer interface is an X-type four-helix bundle made up of helices 1, 4, l', and 4'. However, the structures differed significantly from one another in several important respects. First, the position of helix 3 in the second EF hand in each monomer differs enormously. Second, a small extra helix in the linker between helices 2 and 3 is reported for the bovine structure but not observed in the rat structure. Some suggest that these differences may be due to modeling rather than actual difference in the structures of the proteins from different species (70),but independent evidence will be required to resolve this question satisfactorily. Several structures of the calcium-bound form of SlOOB determined by X-ray crystallography of the bovine protein (71), and by NMR spectroscopy of the rat (72) and human (73) proteins, have been recently determined. The structures confirm several features predicted earlier (69, 70), namely that the major conformational changes are in the linker between the two EF hands, helix 3, and the second calciumbinding loop. The dimer interface is essentially unchanged. Evidence has also been recently obtained by NMR chemical shift analysis on the interaction of calcium-activated SlOOB and target proteins CapZ (74) and p53 (75). Here again, there are differences in the reported binding sites on SlOOB, which may be due to differences in the binding peptides. 3. Calcyclin
Calcyclin, also called S100A6 (64) or CaBP (761, is a dimeric protein similar to Sloop in many respects. It is found in various tissues in-
CALCIUM-BINDING PROTEINS
455
FIG.6 . A ribbon diagram of the structure of SlOOB a s determined by Drohat et al. (69).The two members of the dimer are shaded differently to distinguish them. Helixes 3 and 3', those believed to be involved in conformational change and target binding, are located a t the top. Helixes 4 and 4' lie a t the dimer interface perpendicular to the plane of the figure; helixes 1 and 1' lie under these parallel to the plane of the figure.
cluding lung, heart, platelets, and muscle. It undergoes a conformational change upon calcium binding and exposes a hydrophobic patch, which is most probably the target binding site. As was the case for calmodulin and troponin C, the structural characterizations of calcyclin and SlOOB have been closely linked. A solution structure of apo-calcyclin was determined by NMR spectroscopy (771, and the protein was found to be a symmetrical dimer whose interface was composed of helices 1, 4, l', and 4',as was later found for S100B. However, there are differences between calcyclin and the two reported apo-S100B structures (68, 691, especially in the region of helix 3. These differences may be genuine, but helix 3 is the least well determined segment of the apo-calcyclin structure, and a higher-resolution structure will be needed to settle the question of how different they really are. Characterization of the calcium-loaded form of calcyclin
456
FINN AND DRAKENBERG
has also been done by NMR methods (78). As opposed to SlOOB, only subtle changes in the conformation of calcyclin are observed upon calcium binding, and these are similar to those observed in calbindin Dgk.This result is surprising and in apparent conflict with calcyclin’s role as a calcium sensor and data that indicate that a hydrophobic surface is opened upon calcium binding (76, 79, 80). The functional role and interactions with target proteins remain an area of intense study. One target protein has been identified and is expressed predominantly in the brain, similar to calcyclin itself (81,82). This suggests that calcyclin may act as a calcium-dependent neuronal signaling protein.
111. Calcium-Mediated Membrane-Binding Proteins
A. INTRODUCTION Calcium-mediated membrane binding of proteins occurs both inside cells and outside. At first one might think that these events should be very different in these two different environments, with calcium concentrations in pM and mM, respectively. It is, however, not necessarily so. Two extreme models for the calcium-mediated membrane binding may be envisaged. The first one, exemplified by recoverin, is when calcium binding to the protein causes a structural change that, somewhere else, exposes a hydrophobic patch that may bind to the membrane without any direct involvement of the calcium ion. In the case of recoverin the hydrophobic patch is in fact a myristoyl chain attached to the N terminus of the protein. The other extreme case would be where the calcium ions form a link between the protein and membrane surface. This has for some time been a working model for the binding of Gla-rich domains from blood coagulation proteins to phosphatidyl serine-containing membranes. More recently, however, another, more specific, model for this interaction has emerged, as will be discussed later. In our opinion the mode with the calcium ion as a glue between two negatively charged surfaces will probably not be found because there will be no specificity in such an interaction. However, there are certainly examples in which the calcium ion is coordinated to a protein as well as to a phospholipid head group. We will now discuss a few cases of calcium-mediated protein binding to membranes.
CALCIUM-BINDING PROTEINS
457
B. RECOVERIN Recoverin is a protein of the EF-hand type described in Section 11. However, as opposed to most EF-hand proteins, it is a membraneassociated protein that binds the intracellular face of membranes in a calcium-dependent manner. The protein is composed of four E F hands arranged pairwise, with only two of the four sites able to bind calcium (83). It is a member of the neuron-specific calcium sensor (NCS) family of proteins, which includes closely homologous proteins such as Smodulin and visinin, and less homologous proteins such as the vilip proteins or frequenin. The membrane binding is mediated by a n acyl group, usually myristoyl, covalently attached to the amino terminus. In the apo form this myristoyl group is sequestered from solvent and released upon the binding of two equivalents of calcium (84). The structure of the unmyristoylated recoverin (85) was determined by X-ray crystallography and was followed by structures of the myristoylated apo form (86) and myristoylated calcium form (87). Together these structures paint the full pictures of the molecular transition between the inactive apo form and the active calcium form (Fig. 7). Briefly, in the apo state, the myristoyl group is buried in a deep hydrophobic pocket predominantly located in the amino-terminal EFhand pair but with some contacts to the first E F hand in the carboxyterminal domain. Upon calcium binding the hydrophobic pocket is closed and the myristoyl group is forced out. This is accompanied by a 45 degree rotation of the two EF-hand pair domains relative to one another. This type of calcium-induced conformational change is faintly reminiscent of that observed in calmodulin and troponin C but of opposite direction, (i.e., calcium binding closes a hydrophobic binding site in recoverin whereas it opens one in calmodulin). Functionally, recoverin acts by binding to rhodopsin kinase and inhibits rhodopsin phosphorylation (88).It has also been observed to bind rod outer segment (ROS) membranes (89). The interaction with rhodopsin kinase does not require amino-terminal acylation, but binding to ROS membranes is acylation dependent, suggesting two distinct binding sites a n d o r binding modes on recoverin. C. ANNEXINS The annexins form a group of mainly intracellular proteins with a Ca2+-dependentbinding to phospholipids. No clear physiological role of the annexins has so far been defined, but a wide range of biological functions have been suggested (90-97). Amino acid sequence analysis
CALCIUM-BINDING PROTEINS
459
indicates that the Ca2+coordination in the annexin sites differs from that of the EF hand, even though both structures are dominated by an antiparallel four-helix bundle. The crystal structures have been determined for annexin V from humans (98-1011, chickens (102),and rats (103, 104) as well as for annexin I from humans (105). These structures all contain four domains with multiple Ca”-binding sites. The loop between helixes A and B contains a double site (AB and AB’),and the loop between helixes C and D contains a single site (CD). In the structure described by Seaton and co-workers (103, 104) the four domains have very similar AB loops, RMSD = 0.39 A, and the Ca2+coordination is identical for all domains. The AB site carries a fully coordinated Ca2+whereas the Ca2’ in the AB‘ site has fewer ligands and only half the occupancy of the AB site. It therefore appears clear that the AB site is the stronger one of these two sites. The fact that the AB’ site is only partially occupied in the crystal is in agreement with Ca2+-bindingstudies that have shown that the Ca”binding is weak in the absence of phospholipid membranes compared to what it is in the presence of membranes, mM and pM, respectively (106-111). In the presence of C a Z 4annexin , V binds to phosphatidylserine (PS) and phosphatidylethanolamine (PE)-containing membranes, and shows a clear preference for PS (112). Annexin V, however, binds very poorly t o phosphatidylcholine (PC) or phosphatidylinositol (PI), so charge is not the only determinant for this binding, because PE and PC are neutral whereas PS and PI are negatively charged at physiological pH. In a recent time-resolved fluorescence study (113)it was shown that the environment of the sole tryptophan in annexin V changes upon Ca2+binding in the presence of phospholipid vesicles. The changes were interpreted to indicate that the tryptophan side chain moved from a hydrophobic to a more hydrophilic environment (i.e., from the hydrophobic core of the protein to the interface between the protein and the membrane). This interpretation is in good agreement with crystal structures of annexin V in the presence of glycerophosphoserine (GPS) or glycerophosphoethanolamine (GPE). These crystals were obtained from soaking annexin crystals with either GPS or GPE solutions and were found to bind only one GPS/GPE per annexin, namely to domain 3 (104). Due to the high degree of similarity of the four domains and the observa-
FIG.7. A schematic diagram of the conformational change and extrusion of the myristoyl group from the N-terminal domain of recoverin upon calcium binding followed by insertion into the membrane.
460
FINN AND DRAKENBERG
tion by high-resolution electron microscopy that the annexin molecules are flattened on the membrane surface (114, 115),it has been assumed that in solution all four domains will bind to the surface in a similar way (104). In both complexes a phosphoryl oxygen coordinates to the CaZt on the Al3 site (Fig. 8) replacing a water molecule. This will place the tryptophan side chain clearly inside the polar head group of the lipid and in level with the bond connecting the hydrocarbon chains with the glycerol moiety. These crystal structures will also explain the preference of PS over PE for annexin binding. The serine carboxylate coordinates the AEi’ CaZt and the serine amino group forms hydrogen bonds to both the side chain of Thr187 and the main chain of Glu226. None of these interactions are present in the PE complex, and for the bulky PI and PC, which bind annexin poorly, repulsive steric interaction may be important. The crystal structure of annexin V/GPS clearly indicates that two Ca2+in each domain are simultaneously coordinated to the protein and the lipid head group. This is the only direct evidence so far for this kind of interaction that has also been discussed for the membrane binding of C2 and Gla domains (discussed later). It is, however, also clear that there are many interactions that contribute to the binding (e.g., all four domains have a hydrophobic residue in the position analogous to the tryptophane in domain 3). D. C~DOMAINS The Ca2+-bindingC2 domain, first identified in various isoforms of mammalian Ca2+-dependentprotein kinase C (PKC) (116-118) is ca. 130 amino acids long. Because it was found that PKCs lacking the second domain did not show any Ca2+regulation, it was proposed that the C2 domain was responsible for the CaZt regulation of PKC. Similar domains have been identified in various proteins such as synaptotagmin (119-123), cytosolic phospholipase Az (cPLAz)(124-126), and phosphoinositide-specific phospholipase C, IP-PLC (127-135), among others. [For a recent review see (136).] It has been shown that a single C2 domain from synaptotagmin will bind to phospholipid membranes in a CaZt-dependent manner (119, 120). It has also been shown that this binding depends on the phospholipid composition (121). C2 domains from all the studied synaptotagmins (II-VI) bound to negatively charged phospholipids (phosphatidylserine, PS, and phosphatidylinositol, PI) in a CaZtdependent manner. All but synaptotagmin IV were also found to bind to vesicles with a 1: 1 mixture of negative and neutral phospholipids.
CALCIUM-BINDING PROTEINS
46 1
FIG. 8. Stereo view of the phospholipid binding site in the third domain of rat annexin V, including calcium ions in sites AB and AB’, ( a ) with bound glycerophosphoserine (GPS), and (b) with bound glycerophosphoethanolamine (GPE). GPS and GPE a s well as the calcium ions are drawn dark. The oxygens in the protein fragment are drawn gray. (Redrawn from Swairjo et al. 1995).
462
FINN AND DRAKENBERG
None bound to only neutral phospholipid vesicles. The difference between IV and the other was localized to an Asp-to-Ser substitution and verified by mutant studies. Falke and co-workers (137) have shown that an isolated C2 domain from cPLA2binds Ca2' and binds to phosphatidylcholine vesicles in a Ca2+-dependentmanner. The domain binds two Ca2+in the absence as well as in the presence of phospholipid vesicles; however, the binding affinity increases from KD = 24 p M in the absence of vesicles to KD = 3 pM in the presence of vesicles. In both instances the binding appears to be cooperative, resulting in a steep response to changes in Ca2' concentration. The binding of Ca2' to the C2 domain also increased the tryptophan fluorescence from Trp71, showing a conformation change caused by Ca2+binding. The fluorescence increase was twice as large in the presence of vesicles as in the absence, showing that the tryptophan environment changes upon membrane binding. Falke and co-workers also studied the kinetics of the Ca2+binding. It was thus found that in the phospholipid free state there is a single off rate for the two Ca2+.This shows that either the two off rates are the same or the off rate for C2"Ca is much faster than that for C2*Ca2. The latter could well be imagined for a system with strong cooperative Ca2+binding and has been observed for EF-hand proteins (62, 138).In the presence of vesicles, however, the Ca2+off rate is much slower and there are clearly two off rates differing by a factor of 10 (Fig. 9). Very similar rates were obtained by measuring either the Ca2' off-rate directly, the conformational change caused by the Ca2+release, or the release of the C2 domain from the phospholipid surface. This therefore shows that the C2 domain with a single Ca2+is still bound to the membrane surface, at least transiently. The equilibrium measurements, however, indicate that under conditions where, on average, one Ca2' is bound per C2 domain, about half of the C2 domains will bind two Ca2+and the other half will be in the apo form. It would be interesting to confirm this with a stopped-flow experiment designed in such a way that the final state will have a Ca2+/C2domain ratio of 1: 1. The Ca2'-dependent binding of protein kinase C to phospholipid membranes has been shown to be specific for negatively charged phospholipids, and it has also been shown that Ca2+binding depends on the presence of negatively charged phospholipids (116). The binding of PKC to phospholipid membrane depends not only on Ca2+but also on the presence of diacylglycerol (117). This effect is very specific for PS-containing micelles. Even though PKC binds to phosphatidic acidcontaining micelles, this binding is not affected by diacylglycerol. Fur-
463
CALCIUM-BINDING PROTEINS
-0.5
-0.25 ~
Lu K
u
I
,
I
,
,
,
,
,
,
,
,
,
1
,
8
1.25
C 1.00
0
1
2
3
4
Time (sec)
F ~ G9.. Kinetics of CaL+dissociation from the cPLA, C2 domain in the absence and presence of phosphatidylcholine vesicles. (a) Stopped-flow measurement of the Cad+off rate using the fluorescent calcium indicator Quin-2. (b) Conformational changes triggered by Ca2+removal from the C2 domain monitored using the intrinsic fluorescence of Trp71. (c) Membrane release of the C2 domain followed by protein-to-membrane FRET. (From Nalefski et al. 1997 with permission.)
464
FINN AND DRAKENBERG
thermore, only phosphatidyl-L-serine, and not phosphatidyl-D-serine, will cause the high-affinity binding of PKC to diacylglycerolcontaining membranes (118). The structure of the N-terminal C2 domain of synaptotagmin has been solved to a resolution of 1.9 A for both the apo and Ca2+forms (139). The structure is described as an eight-stranded @-sandwich with a conserved four-stranded motif named a C, key. The Ca2+-binding site is found at the top of the @-sheetwith ligands from four Asps in the loops connecting @2/@3and @6/@7(Fig. 10). A second possible but not occupied Ca2+-bindingsite was identified in the other end of the molecule. Ca2+binding to this domain caused only minor structural changes, indicating that the site is preformed. The structure of a C2 domain in an intact protein has also been determined for phosphoinositide-specific phospholipase C-8, (133-135). The structure of PI-PLCS1 comprises an N-terminal PH domain, a catalytic core, and a C-terminal C2 domain. Only the C2 domain will be further
FIG. 10. Structure of the synaptotagmin C2 domain showing the calcium binding site a t the top and ligands for a potential calcium binding site a t the bottom. (From Sutton et al. 1995 with permission.)
CALCIUM-BINDING PROTEINS
465
discussed here. From a comparison of the structure of C2 domains from synaptotagmin and PI-PLCG1, it is apparent that there are a t least two different topological variants of the C2 domains (135).It is possible to go from the synaptotagmin topology to the PI-PLC topology by connecting the C terminus of /3-strand 8 with the N terminus of /3-strand 1 and cutting the loop between /3-strands 1 and 2 open. This, however, appears to have no effect on the calcium-binding regions that are in the opposite end of the structure. Using lanthanum as a calcium analog showed that three loops form the metal-ion-binding sites with two adjacent metal ions. For another crystal form of PI-PLC, with samarium as a replacement for calcium, the same two sites were found (134).A comparison of the apo and lanthanide structures revealed only minor changes in the metal-ion-binding loops; however, the maximum distance between metal-ion-binding loops 1 and 2 changed from 12 A in the apo form to 21 A in the Sm3+form (Fig. 11).Even though the functional role of the C2 domain in PI-PLC is still unknown, mutational studies have shown that it is essential for catalytic activity (130, 140). The calcium-mediated binding of C2 domains to lipid membranes is still not well understood. The calcium-bridging model is to us very unattractive because it does not explain the specificity for PS. More studies are certainly needed, and our expectations are that a more specific membrane-binding model will eventually emerge, as it has for the annexins and the GLA domain in the vitamin K-dependent serum proteins (discussed later). N649
t‘c..
FIG.11. Stereo view of C2 jaws of the apo and Sm” forms of PLC-81. The apo form is shown with dark backbone and light side chains, and the Sm’+form with light backbone and dark side chains. (Redrawn with permission from Grobler et al. 1996.)
466
FINN AND DRAKENBERG
E. y-CARBOXYGLUTAMIC ACIDSITES The amino acid y-carboxyglutamic acid (Gla) is not one coded by DNA, but glutamic acid is posttranslationally carboxylated. Gla contains a malonic acid moiety with an affinity for Ca2’ of Kb = 30 M-’. This is too weak to be of any biological relevance by itself. However, many of the modified amino acids often appear together, though not necessarily in sequence, as in the case of the Gla module of blood coagulation factors, in which there are ca. 10. Robertson et al. (141) have shown, by using 43CaNMR, that calcium binding to a Gla-Gla moiety is sufficiently strong to be relevant under physiological conditions (KO = 0.6 mM). Likewise Colpitts and Castellino (142) have found a KD = 1.6 mM for calcium binding to the peptide CysIleGlaGlaIleCys but with a stoichiometry of two peptides per calcium ion. Later, we will discuss the Gla domain from the blood coagulation system at some length and, to a minor degree in Chapter 4, the Glacontaining conotoxins, whereas for the bone Gla and matrix Gla proteins (143)the reader is referred to older reviews (144). The Gla-containing proteins of the blood coagulation system are all modular with the Gla-domain N terminal. Factors VII, IX, and X and protein C form a group with the same domain structure. The Gla domain is followed by two EGF-like domains, of which the first one also binds calcium (see Section IV.C), and a serine protease domain, also with a calcium-binding site (see Section 1V.D). Prothrombin and protein S have somewhat different domain structures (145, 146). I n prothrombin the Gla domain is followed by a hexapeptide with a disulfide loop, two kringle domains, and the C-terminal serine protease domain. In protein S the Gla domain is followed by the “thrombin-sensitive” loop, four EGF-like domains, and the C-terminal domain that is homologous to plasma steroid hormone-binding proteins. The DNA sequence for the Gla domain show that there should be 9-12 Glu residues. These Glu residues are all carboxylated to Gla by a vitamin K-dependent carboxylase. This modification is a prerequisite for calcium binding and biological activity. The first attempts to characterize the calcium binding properties of Gla domains were made in the early 1970s (147-149). These studies were made before the Gla residue had been identified (250, 151). Since these early works, the calcium-binding properties of various Gla domains have been reported on several occasions. In most cases equilibrium dialysis experiments with 45Cahave been used, but spectroscopic techniques as well as calorimetry and ion-selective electrodes have also been em-
CALCIUM-BINDING PROTEINS
467
ployed (152-163). Equilibrium studies have indicated that all vitamin K-dependent plasma proteins bind 6 to 12 calcium ions (159). The low affinity of these calcium binding sites (KDbetween 0.1 and 2 mM) and their large number have made the determinations difficult. In two recent studies (164, 165) on synthetic peptides with sequences corresponding to the N-terminal domain(s1 of factor IX and protein C it has been shown that it is not sufficient with only the Gla domain [factor IX(1-42) and protein C(1-3811 for full calcium binding. However, including the amino acids from the hydrophobic stack [factor IX(1-471, protein C(1-48)I seems to restore both calcium and lipid binding up to the level in intact proteins. Even though the sequences of these peptides are 50% identical, there seem to be some important differences regarding calcium and lipid binding of the truncated peptides. Removing the five C-terminal residues from factor IX(1-47) drastically decreases calcium binding as well as the binding to phospholipid vesicles, whereas for protein C( 1-38) the calcium binding has been reduced merely by a factor of two to three, and full binding to phospholipid vesicles can be obtained by a modest increase in the calcium concentration. It is a t present not clear what causes this difference, but it seems to reside in the first 30 amino acids, because a peptide with the first 30 amino acids from the human protein C sequence and the remaining eight residues from the factor VII sequence showed a calcium-binding pattern similar to the 38-mer with the protein C sequence (142). Culpitts et al. (1661, using synthetic protein C peptides selectively 13Clabeled on individual Gla residues, have been able to show that Gla 6, 16, 25, and 26 (factor X numbering) participate in the stronger calcium binding. Even though it seems clear that it is sufficient with the Gla domain and the hydrophobic stack to restore full calcium-binding activity, the apo form of these peptides appears essentially structureless (1671, whereas the fragment from factor X (also comprising the N-terminal EGF domain) has a more ordered structure (168).It has also been shown recently that the dynamic properties of Trp42 in the apo form of factor X peptides critically depend on the length of the peptide (Drakenberg, unpublished results). The correlation time of Trp42 is lengthened ca. threefold (from 1.5 t o 4.5 ns) by extending the 46-amino acid peptide with either the EGF domain or five amino acids from the C-terminal helix from calbindin Dgk.It is thus clear that even though the Gla domains of factor IX and protein C do not need the EGF domain for full calcium-binding properties, the apo-Gla domain of factor X needs the EGF domain to stabilize the structure. The structure of the apo form of the Gla domain from factor X, in the
468
FINN AND DRAKENBERG
GlaEGF fragment, has all the secondary structure found in the calcium form (discussed later); however, the relative orientations of the structural elements differ, especially in the N terminus. In this structure the Gla residues are all, as could be expected, exposed to the solvent, and the hydrophobic residues Phe4, Leu5, and Val8 form a cluster facing the interior of the Gla domain [Fig. 12(a)l. Pheh -7
4
Leu5
Val8
IY
Gla7
& Gla14
aPO Ca2+ FIG.12. Comparison of the energy-minimized average NMR structure of the Gla domain from factor X with model-built calcium-loaded Gla domain (based on the X-ray structure of prothrombin). (a) Location of residues Phe4, Leu5 and Val 8. (b) Location of Gla residues 6,7, 16, 19,20,26, and 29.The essential residues 7, 16, 20,26,and 29 are shaded dark.
CALCIUM-BINDING PROTEINS
469
The structure of calcium-saturated Gla domains has been determined both in the crystal state by X-ray crystallography for prothrombin fragment 1 (169)and for factor VII in complex with tissue factor (170)and in solution by NMR for the N-terminal47 amino acids from factor IX (171, 172). These three structures are essentially the same and are shown in Fig. 12(b). The main secondary structure elements of the structures are three helixes (14-17, 25-31, and 36-47) and a n N-terminal loop. Seven Ca2* are identified in the structure, and they have coordination numbers ranging from three to seven. Five of these Ca2+ are chelated between Gla6, Gla7, Gla16, Gla25, Gla26, and Gla29, rendering four of them inaccessible to water, In the NMR structure determination there is no direct information on the position of the Ca2+;however, in a refinement they used a genetic algorithm to identify the locations of the Ca2+(172).The identified Ca2+positions agreed well with the crystallographically determined positions and were subsequently used in the refinement of the solution structure. The two crystal structures of the Gla domains are very similar, even though they are in completely different settings, with a n RMSD of 0.63 A for 40 a-carbons. A similar comparison of the solution structure of factor !X and the prothrombin crystal structure reveals an RMSD of 1.3 A for residues 12-47. The backbone conformation for the 11 N-terminal residues differs significantly between the solution and crystal structures. All Gla domains except that in human factor IX have a n Ala as the N terminus. In factor IX it is a Tyr.This larger residue cannot be accommodated within the crystal structure, which may explain the difference between the crystal and solution structures in this region. However, the unusual solvent mixture with 3 M urea and 2.5 M guanidine hydrochloride used in the NMR study may also cause some structural changes. Based on these structures and the structure of the apo-Gla domain from factor X determined by NMR (1731, we can now speculate about how the Gla domains are bound to phospholipid membranes (122, 174, 175). It can be assumed that the conserved hydrophobic residues, Phe4, Leu5, and Val8 (human factor X numbering) are important for the membrane binding because they are exposed on calcium binding (Fig. 12). Also, based on site-specific mutagenesis studies of Leu5, it has been suggested that this side chain is bound into the phospholipid membrane. The hydrophobic binding is certainly not sufficient to explain all experimental data on the membrane binding. For optimal binding the membrane must contain negatively charged head groups, phosphatidylserine. This clearly indicates that also electrostatic effects are important for the membrane binding, and Li et al. (122)argue that some
470
FINN AND DFtAKENBERG
of the peripheral calcium ions in the Gla domain as well as the side chains of Lys3, Arg9, LyslO, and Argl5 in human prothrombin may be directly involved in the interaction with the membrane. However, only for residues 9 and 15 is the charge conserved in the Gla domains, so we may argue that they are strong candidates for direct interaction with the negatively charged lipid head groups.
IV. Extracellular Calcium-Binding Proteins
A. INTRODUCTION The total concentration of calcium in the blood plasma is ca. 2 mM, and about half of it is bound to proteins, mainly serum albumin. This high calcium concentration is typical for all extracellular fluids, in stark contrast to the very low free calcium concentration in resting cells. One might therefore be led to believe that there is no specific function of calcium in the extracellular fluids. This is certainly not true. It has been known for almost a century that calcium is critical for blood coagulation, and it is also well known that calcium is a major component in our skeleton. It is, however, also obvious that the requirement for a protein to be calcium binding in a milieu with free calcium at 1 mM is very different from one with calcium at p M levels.
B. CONANTOKINS Cone snails produce biologically active peptides to paralyze their prey. The peptides are known as conotoxins (176-1 78). Most of them are small and stabilized by disulfide bonds. Two unusual members of the conotoxin family are conantokin G from Conus geographica (179) and conantokin T from Conus tulipa (180),which lack disulfide bonds but are rich in y-carboxyglutamic acid residues: con-G: GlyGluGlaGlaLeuGlnGlaAsnGlnGlaLeuIle ArgGlaLysSerAsn con-T: G l y G l u G l a G l a T y r G l n L y s M e t L e u G l a A s n L e u A r g a
The role of the Gla residues is not yet fully understood, but a synthetic conantokin peptide with Glu instead of Gla has been shown to be inactive. However, it is still not clear whether Ca2+is essential for biological functioning of the conantokin peptides (181, 182).There appears to be no thorough study of the calcium ion binding to these peptides, though CD spectroscopy has been used to study the change in helical content as a function of calcium concentration (183). The
CALCIUM-BINDING PROTEINS
47 1
two peptides behave quite differently. con-G had almost no helical structure in the calcium-free state but it increased to ca. 50% in the presence of 9 mM Ca”. con-T on the other hand already had roughly 50% helix without Ca2+added, and this did not change upon addition of calcium. The same effect was found with Cu2- whereas Mg2+and Zn2’ were found to introduce almost 70% helix in con-G (Fig. 13). Another interesting observation is that ca. half of the calcium-induced helix in con-G remains in the presence of excess EDTA. The authors have interpreted this as being due to a very strong calcium ion binding to con-G; however, we also have to consider that EDTA has been shown to interact with various proteins in a calcium-dependent manner (184). We therefore find a more likely explanation that the Ca2 EDTA binds t o con-G and introduces some helical structure. Skjaerbaek et al. (183)have also determined the solution structure of con-G in the presence of calcium, and it appears to be more helical than in the absence of calcium. However, the resolution is not sufficient to define binding sites for calcium, making the presence of a site strong enough to compete with EDTA for calcium less likely. Rigby et al. have determined the solution structure of con-G to high resolution in the absence (185)as well as in the presence of CaL+(186).The peptide is well structured under both conditions, with backbone RMSDs of 0.8 and 0.6 A, respectively. As shown in Fig. 14, there is a pronounced structural change upon CaL+binding, with the formation of a straight a-helix from Gla3 to Lysl5. In this structure Gla residues 3, 7, 10, and 14 become aligned in a linear array on one side of the helix. The Ca”, which cannot be “seen” by NMR, were localized with a genetic algorithm (172).The best model was obtained with 4 Cal’lpeptide. All the Ca2 were found to have 3 or 4 ligands from the Gla side chains. This arrangement has some similarity to the Ca’ binding to the Gla domains in blood coagulation proteins (170, 1871, in which there also is a linear array of Ca2- coordinated to Gla side chains. In both cases the calcium binding will also expose some hydrophobic amino acids t o become solvent accessible.
C. EGF-LIKEDOMAINS The epidermal growth factor (EGFblike domains are approximately 45 amino acids long and contain six cysteine residues that are paired in a characteristic manner, 1 to 3, 2 to 4, and 5 to 6, with a double-stranded 0-sheet as the main structural feature. The EGF domain has been found in a wide variety of proteins including these involved in blood coagulation, fibrinolysis, neuronal development, and
472
FINN AND DRAKENBERG 4 10‘
-
A
n
3 10‘-
d
l i 0
+-. ---i)-
-+-
J
No Cac’2 76 UM CaC12 0.38 mM CaC12
1.5 rnM CnC12 3.0 rnM CaCIz -74.5 mM CaC12 -6.5 mM CaCn +9.2 mM Cam
-u-
E 210‘ a
-&-
Q, -1 ld
!
-210‘
I
I
I
6
“Eu
I
-4 10’
180
I
210
220
230
240
250
200
210
220
230
240
250
6 lo4
-
1
200
n
4 10‘
0
E
-0 2
10‘
“E
2
a,
Q, -2
104
-4 10‘
190
FIG. 13. CD spectra of con-G and con-T under different conditions. (a) CD spectra of con-G a t different CaC1, concentrations. (b) CD spectra of con-T a t different CaCl, concentrations. ( c ) CD spectra of con-G with different metal ions. (d) CD spectra of conT with different metal ions. (e) CD spectra of con-G with fixed CaClz concentration and different EDTA concentrations. (From Skjaerbaek et al. 1997 with permission.)
473
CALCIUM-BINDING PROTEINS
- 1
D
-210‘ I90
200
210
220
230
240
250
3 lo‘
-8
n
d
I
210.
a
t 10‘
--
4 o mM caaz + 3.0 mM EDTA 4 0 mM Cam + 20.0 nw mTA
“E 0
2 a
v ~
+ 05 mM EDTA
0
iro4
-2 l @ 180
200
210
220
230
240
250
Wavelength (nm) FIG.13 (continued)
cell adhesion (188).A subset of these domains has a hydroxylated aspartic acid (Hya) or hydroxylated asparagine (Hyn), most prominently found in vitamin K-dependent plasma proteins with a n identified consensus sequence of Cys-X-Hya/Hyn-X-X-X-X-TyrPhe-X-Cys-XCys (189, 190). The consensus sequence for Asp/Asn hydroxylation has also been found in EGF domains from, for example, fibrillin, which has 46 EGF domains whereof 43 have the consensus sequence for hydroxylation (191-193); fibulin-1, in which four of nine EGF domains have the consensus sequence (194-197); and the Notch receptor, in which 21 of 36 EGF domains have the consensus sequence for hydroxylation (198).[For a review see (1881.1 The presence of the hydroxylated amino acid has also been correlated with the presence of a Ca2+-bindingsite with a KDin the range of 10 to 100 FM (199204). Studies of Ca”-binding properties of a protein fragment from protein C containing the two EGF domains demonstrated that this fragment contains one Ca2+-bindingsite (205-207). This site was fi-
474
FINN AND DRAKENBERG
C Gla 14
Gla 10
Ca2+
4
N N FIG.14. Calcium-induced transition in c o n 4 with the apo form to the left and the calcium form to the right. (From Rigby et al. 1997 with permission.)
nally localized to the N-terminal EGF domain in protein C as well as in factors IX and X (206,208,209). The isolated EGF domains have been shown to bind Ca2+with low affinity (KD, 0.5-5 mM) (206,208-212); however, these domains usually appear a t least in pairs and often several in tandem or as in the vitamin K-dependent serum proteins with a Gla domain bound N-terminally. It has been shown that when the N-terminal domain is intact, the Ca2+affinity to the next domain has increased by approximately 10-fold, to KD= 0.1 mM, making these sites essentially saturated under physiological conditions (162,213,214).Similarly, Handford and co-workers (211, 215) have found that a recombinant fragment of fibrillin comprising a pair of EGF domains has a high affinity (KD= 0.35 mM) and a low affinity (KD= 9.2 mM) site. The increased affinity to the C-terminal domain is explained as caused by the donation of a ligand from the N-terminal EGF domain to the site of the C-terminal EGF domain. The Ca2+binding to the high-affinity site is in good agreement with a value of 0.25 mM reported as the average affinity to a proteolytic fragment of fibrillin containing seven intact EGF domains (293). Protein S has four EGF domains in tandem, and it has been shown that the Ca2+affinity to this protein is much stronger (KD, to MI than for other EGF-containing
CALCIUM-BINDING PROTEINS
475
proteins studied so far (216).There is no obvious sequence variation to explain this extremely strong Ca2+binding, and the isolated EGF domains from protein S do not display any extreme Ca2+affinity (217). However, fragments containing various numbers of EGF domains from protein S showed a very interesting Ca2+affinity (218). Extremely strong Ca2+binding was thus found for fragments containing EGF1-4 and EGF2-4 (Kr, < in physiological salt) whereas hundredfold higher Kd)s where found for fragments not containing EGF4. Another noteworthy observation is that EGF3 has a 350-fold higher affinity for calcium in fragment EGF2-4 than in fragment EGF1-3. It has been shown on several occasions that the Ca2+affinity to an EGF domain is reduced significantly if the protein segment N-terminal to it is removed. This may be an effect of stabilization of the site, either by having a calcium ligand in this part of the protein or just by stabilizing the conformation of the calcium site. The data reported on the site in EGF3 of protein S demonstrates an influence on Ca2+ binding from a domain on the C-terminal side. This indicates that the three-dimensional structures of these fragments are such that there are direct interactions between nonadjacent domains and not a linear array of domains. Another protein displaying an unusually high calcium binding to EGF domains is Notch-l, a fundamental receptor for cell fate decisions (219). This protein has 36 EGF domains, of which 21 have the consensus sequence for calcium binding. Calcium binding studies to the recombinant fragments comprising EGFll-12 and EGF10-13 have shown that this protein has an calcium affinity in the low pM range and thus somewhere in between protein S and the other calcium-binding EGF domains. Although the role of AsplAsn hydroxylation on calcium binding has been a matter of debate, there have been very few attempts to resolve this question. Morita and Kisiel (200) compared the calcium affinity of bovine and human factor IX. The calcium affinities were shown to be similar even though the degree of hydroxylation is different for these two proteins. The conclusion was uncertain, however, due to interference with other calcium-binding sites in the proteins. Later Selander Sunnerhagen et al. (210) showed, using single EGF domains, that even though there is a small effect due to hydroxylation, it is not significant, so what might be the role of the hydroxylation is still an open question. The structures of several isolated EGF domains have been determined by NMR, the first one more than 10 years ago (220-232). These structures are all quite similar, revealing a domain structure consisting of two relatively independent subdomains. The N terminal
476
FINN AND DRAKENBERG
constitutes about two thirds of the domain, and its major structural motif is a &sheet with two or three strands. The C-terminal third of the domain has been described as a small antiparallel double-hairpin structure (231) or as two loops connected by a short p-sheet and a flexible C terminus (224).The three-dimensional structures of EGF domains have also, more recently, been determined by X-ray crystallography (170, 233-236). Only a few of these studies deal with EGF domains of the Ca2+-bindingtype (170,235,236).In the following we will restrict the discussion to the EGF domains of the Ca“-binding type (cbEGF). The solution structure of the N-terminal cbEGF domains from factor IX (fIX-EGFN)and factor X (fX-EGFN), both containing Hya, in the absence of calcium, has been determined by means of 2D NMR (227-229). The structure of fX-EGFNhas also been determined in the presence of calcium (230). Even though the calcium ion cannot be “seen” by NMR, the calcium site can be readily identified (Fig. 15). In the calcium form of ~X-EGFN there is a cavity of appropriate size for a calcium ion that is lined with oxygens oriented toward the center. There are five well-defined ligands, two side-chain carboxylate groups (Asp46 and Hya631, one side-chain carbonyl group (Gln49), and two 45
64
50
70
FIG.15. Stereo view of the calcium-binding site of fX-EGFN. Amino acids unambiguously assigned as calcium ligands are labeled and the coordinating oxygens are marked with filled symbols (Gly47, Gln49, Hya63, and Gly64). Asp46, a potential ligand, is marked with an open circle.
CALCIUM-BINDING PROTEINS
477
backbone carbonyls (Gly47 and Gly64). This was later confirmed in two crystallographic studies on the N-terminal EGF domain of factor IX (235) and for the N-terminal EGF domain in factor VII in complex with tissue factor (170). Furthermore, the side chain of the Hya residue is rotated in such a way that the hydroxyl group cannot be a ligand to the calcium ion (230). The changes in the NMR spectra caused by calcium binding are localized to residues 46-51 and 62-68, and are not very dramatic. A comparison of the structures of the apo and calcium forms of fX-EGFN shows no major differences. There are, however, some minor differences. The N terminus is more well defined in the calcium form and has moved slightly toward the main p-sheet. The turn connecting the two strands in this p-sheet is bent somewhat toward the N terminus. These conclusions are confirmed by the X-ray crystal structure of fIX-EGFN with calcium (2351, and the backbone RMS deviation between the two structures is 1.2 A (237). There are now also available structures on the Gla-EGF pair from factor X (168)and on a n EGF pair from fibrillin (2371,both in solution and solved by NMR in the presence of calcium. Information on the function of calcium in this kind of structure is still very limited. Only for the Gla-EGF pair is information available for both the apo and calcium-loaded states. Even though the structures of the individual domains are reasonably well defined, the relative orientation, which is probably more interesting, is not. It is, however, from NMR as well as SAXS data (1681, clear that the relative orientation of the two domains is calcium dependent. In the apo state there is a more or less linear arrangement of the two domains, whereas in the calcium form there is a more compact structure, with the two domains oriented almost perpendicular to each other (Fig. 16). This orientation, however, does not agree with what was found in the crystal state for the complex between factor VII and tissue factor (170),where the Gla and the N-terminal EGF domain are arranged in a more or less linear manner. We have to consider, though, that these are different proteins, though quite similar, each with its own specific interactions. Based on the crystal structure of the isolated cbEGF domain of factor IX,it was speculated that an N-terminal domain might donate a ligand to the calcium site in the following domain (235).In none of the structures determined so far with two domains and a calcium-binding site in the C-terminal domain has such a ligand donation been observed. More recently Downing et al. (237)have found from sequence analysis of domain pairs indications that they can be divided into two classes based solely on the number of residues linking the two do-
478
FINN AND DRAKENBERG
(b) FIG.16. Stereo representation of the secondary structure elements and relative orientation of the domains in the Gla-EGF domain pair. (a) with 1 equivalent of CaZ’ added and (b) in the absence of Ca”. (From Sunnerhagen et al. 1996 with permission.)
mains. Sequence alignment of the two classes of pairs was applied to define the consensus sequences as shown in Fig. 17. They also suggest that the Class I1 pairs are likely to adopt a different conformation and that a conserved carboxylatelcarboxamide side chain may donate a ligand to the following domain, as is observed in the crystal structure of fIX-EGFN. The confirmation of this hypothesis has to await the structure determination of such a domain pair. This hypotheClass I EGF1: Dx D/N ECxxxxxxCxxxxxCxNxxGSY/F xCxCxxG Y/F xxxxxxxxC EGFB: xD IN D/N ECxxxxxxCxxxxxCxNxxGSYIF xCxCxxG Y/F xxxxxxxxC Class I1 EGF1: xx DM xCxxxPCxNG G/A xCxxxxxx Y/F xCxCxxG Y/F XGXXC EGFS: xx D/N W D/N E D CxxxPCxNG GIA xCx DM x IN xx Y/F xCxCxxG Y/F xGxxC FIG. 17. Consensus sequences for class I an? class I1 EGF-cbEGF and cbEGFcbEGF pairs. x indicates that there is no preferred amino acid and letters separated by / indicates that both amino acids are commonly present.
479
CALCIUM-BINDING PROTEINS
sis may explain the high calcium affinity in the fragments from Notch studied by Rand et al. (219)with KDin the low pM range even in the presence of physiological salt concentration. For protein S on the other hand, this does not offer any explanation for the high affinity because the protein S pairs with high affinity belong to Class I(218).
D. SERINE PROTEASES
It has long been known that some serine proteases contain one or two calcium-binding sites. It has thus been shown that trypsinogen has two sites, one of low affinity in the activation peptide and one with a higher affinity (KD= 1.6 . MI that is also present in the active enzyme (238, 239). The structure of trypsi? with one bound calcium has been determined to a resolution of 1.8 A, and the calcium ligands could be identified (Fig. 18) (239). The calcium ion is coordinated by six ligands a t the edges of an octahedron. Four ligands are from a loop of the protein, Glu70-Glu80, where Glu70 and Glu80 are coordinating with their side-chain carboxylate groups and Am72 and Val75 are coordinating with the backbone carbonyls. The two remaining ligands are water molecules that are also hydrogen bonded to Glu70 and Glu77, respectively. Sites similar to this archetype can be inferred from the sequences of chymotrypsin; blood coagulation factors VII, IX, and X; and protein C (184,240-243). For trypsin, chymotrypsin, and their zymogenes it has been shown, using 43CaNMR and stopped-flow experiments, that the Ca2 exchange is fairly slow even though the binding affinity is modest (Fig. 19) (184). The on rate of the calcium ion was thus found to be 105-10b s two to three orders GLU70 ASN72
GLU70
R
ASN72
GLUE0
GLU80
FIG.18. Stereo view of the calcium-binding loop in trypsin including the calcium ion and internal water molecules.
480
FINN AND DRAKENBERG
40
20
0
-20
FIG.19. '"a NMR spectra of (a) 2 mM tosyl-trypsin and 2.5 mM Ca2' at pH 6.5 and 24OC; (b) 1 mM tosyl-trypsinigen and 1.3 mM Ca2+at pH 6.2 and 24°C.
of magnitude slower than what could be expected for a diffusionlimited process. Also, the off rate from chymotrypsin(0gen) is at least one order of magnitude faster than that from trypsin(ogen). These findings are in agreement with the reported higher flexibility in the calcium-binding site in chymotrypsin apparent in the crystal structure. However, the crystal structures do not offer any explanation of the slow calcium-ion exchange because the binding site is close to the surface and there are two water ligands. The calcium-binding sites in the protease domains of coagulation factors VII, IX, and X and protein C all have calcium affinities comparable to that in trypsin (241, 244-246). In the crystal structures of factors IX and X, no calcium ions could be identified, most likely due to the crystallization conditions because the calcium-binding site is present though not occupied (234). The crystal structure of factor VII in complex with tissue factor shows the presence of calcium in the
CALCIUM-BINDING PROTEINS
48 1
serine protease domain with a structure very similar to the one in trypsin. Calcium binding to the serine protease domain of protein C has a pronounced effect on the activation of protein C by thrombin. This activation is very slow in solution and is reduced even further in the presence of calcium. However, if protein C is activated by the thrombin-thrombomodulin complex, the activation is rapid and is increased in the presence of calcium (234). Calcium binding to the serine protease domain of coagulation factor VII is important both for amidolytic activity and for interaction with the tissue factor (247). The subtilisins all have a calcium-binding site whose occupation protects the enzyme against autolysis (248). This site is very different from the one in trypsin, however. The ligands come from three different parts of the protein, and the calcium binding seems to bring the N terminus close to the 75-81 and 40-42 loops with coordination to the side chains of Gln2 and Asp41, and to the peptide carbonyl of Leu75, Asn77, Gly79, and Val81 (249-251).
E. Q-LACTALBUMIN AND LYSOZYMES a-Lactalbumins (aLACs) are milk proteins that play a n important role in the biosynthesis of the milk sugar lactose. By binding to the enzyme galactosyltransferase (GT), the lactose synthase complex is formed. This binding is reversible, and aLAC functions as a regulatory subunit. GT alone is unable to catalyze the synthesis of lactose a t physiological glucose concentrations due to its low affinity for glucose. In the lactose synthase complex the affinity for glucose has increased 1000-fold and lactose can be produced (252-254). On the other hand are the lysozymes (LZs), lytic enzymes that catalyze the degradation of peptidoglycans. From the high degree of homology between aLAC and LZ as revealed by amino acid sequences (255,2561, intron-exon arrangements (257, 2581, and three-dimensional structure (259, 2601, it has been postulated that these functionally different proteins have evolved from a common ancestral molecule. It is recognized that all known aLACs possess one high-affinity Ca2+-binding site, and crystal structures have revealed that the Ca2+-binding site in aLAC is formed by the side-chain carboxylate groups from three aspartate residues (Asp82, Asp87, and Asp€%), which are conserved in all aLACs. Also, the backbone carbonyls from Lys79 and Asp84 contribute to the ligation of Ca2+as well as two water molecules (259, 260). Ca2+ binding to lysozymes from horse and pigeon with the conserved aspartate residues required for Ca2' binding in aLAC was initially less well defined. Even though both lysozymes
482
FINN AND DRAKENBERG
FIG. 20. Stereo view of a n overlay of the structures of aLAC from human, guineapig, and goat. (From Pike et al. 1996 with permission.)
were shown to be Ca2'-binding proteins (261-263) the crystal structure of the same proteins showed no electron density for Ca2+(264). On the other hand a mutant of human LZ in which the two missing coordinating Asp residues were introduced showed both Ca2+binding and a Ca2' in the crystal, with a geometry analogous to that in human and baboon aLAC (265). In agreement with the Ca2'-binding studies of LZ from horse and pigeon (261-263) a 43CaNMR study of CaZt binding to the two lysozymes and bovine and human aLAC showed very similar behavior, strong Ca2+binding for one site, and the same symmetry of the site as revealed by the quadrupole coupling constant of the bound 43Ca2t.Furthermore, the off rate of the bound Ca2+was shown t o be slow for all four proteins (266, 267). The Ca2+-bindingsite in a-lactalbumin has some similarity with the EF-hand binding sites in that it consists of a helix-loop-helix motif, with all protein ligands contained in the loop. However, the pattern of ligands is not the same, and the site in aLAC has been named an "elbow" to differentiate it from the EF hand (260,268).Another difference between these two kinds of Ca2+-bindingsites is that the elbow is a single site whereas the EF hands appear in pairs (discussed earlier). Crystal structures of several aLACs are available (260,269,270) and are all very similar. The structure is divided into a large domain ( a domain) and a small domain ( p domain) by a cleft (Fig. 20) (270).
CALCIUM-BINDING PROTEINS
483
The Ca2+-bindingsite is in the interface between the two domains. The only variation between the various structures occurs for residues 101-110. In the structure from human and recombinant bovine aLAC, residues 105-1 10 adopt a distorted a-helical conformation, whereas in goat and guinea-pig aLAC these residues form a loop (270).The difference may be due to different crystallization conditions and may indicate that this region is flexible in solution (271). Mutations in the flexible loop indicate that this is important for the interaction with GT. Recently there have also appeared studies on the thermodynamics of the Ca2+-loadedpartially unfolded state (2721, the importance of the disulfide bonds (2731, and the functional role of calcium-binding residues (274).Furthermore, aLAC has been a model for protein folding because it forms a molten globule state during folding (275, 276), which will not be discussed here. F. CADHERINS The cadherins form a family of cell-cell adhesion receptors (277, 278). They are transmembrane glycoproteins with typically five tandemly repeated extracellular domains (CAD repeats) (279, 280). The adhesive action of cadherins depends on the presence of Ca2+(278). A detailed study of Ca2+binding to E-cadherin domains has shown that two CAD repeats are needed for Ca2+binding (2811. It was shown that the domain pair comprising the two N-terminal domains, ECAD12, bind Ca2' strongly compared to the extracellular Ca2+concentration, whereas individual domains showed no or very weak Ca2+affinity. A crystal structure solved for the N-terminal domain of N cadherin shows that the polypeptide fold includes seven &strands arranged in two @sheets with the N and C termini a t opposite ends of the domain, making a linear arrangement of the five domains possible. A metalion binding site occupied by an Ybs' has been identified with two bidentate ligands from G l u l l and Glu69 (282).The solution structure of the N-terminal domain of E cadherin in the presence of Ca2+has been determined by NMR. Even though the Ca2+cannot be "seen" by NMR, Ca2+ligands were inferred from Ca2+-dependentchemical shift effects. In this way G l u l l , Glu69, and Asp100 were suggested as ligands (280, 2831, in agreement with the crystal data (282) in which two water molecules could also be identified as ligands, still resulting in an incomplete coordination of the metal ion. Recently the crystal structure of the two N-terminal domains of E cadherin was solved (284). The general location of the Ca2+-bindingsite suggested in the
484
FINN AND DRAKENBERG
earlier studies is confirmed; however, it is now shown that in each domain-domain interface there are three Ca2+.Furthermore, the structure reveals a twofold symmetric dimer. The dimeric structure is in good agreement with solution state studies, which have shown a Ca2+-dependentdimerization of the same protein fragment (2811. A model based on the assumption that the relative orientation of the two domains in ECAD12 is also used for the other linking regions in a 240-A-long extracellular region. This is in good agreement with a value of 220 A obtained by electron microscopy (285) in the presence of calcium. G. SPARC The protein SPARC (Secreted Protein Acidic and Rich in Cysteine) is also known as BM-40 and osteocalcin (286-288). It was originally isolated from bone but is found in other matrix-producing tissues such as ligaments. It is also found in corticosteroid-secreting tissues such as the adrenal cortex; in epithelia in, for example, the gut and skin, and in platelets (289). It is a multidomain protein that is unique among extracellular proteins in that it possesses an EF-hand pair motif in the last of its three domains. The first two domains consist of a variable acidic domain and a follistatin-like module. It shares this domain organization with other proteins such as the brain protein SC1 (290) and the retinal protein QR1 (291). In addition to calcium, it has been shown to bind a number of extracellular proteins in uitro, including collagen, plasminogen, and albumin (289). It is believed to be a modulator of cell-matrix interactions. The structure of the calcium-binding domain of SPARC was determined by X-ray crystallography to a resolution of 2.0 A (292). The structure of the EF-hand pair resembled those found in intracellular proteins but had several unique features. In the first E F hand, there was a single residue insertion, which alters the fold of the calciumbinding loop such that a cis-peptide bond is formed and a backbone carbonyl acts as a calcium ligand. In the second EF hand, a disulfide bond is found, which may act to stabilize the calcium-binding loop. Finally, rather than having the hydrophobic interior of the EF-hand pair available for binding of target proteins, an amphiphilic aminoterminal helix binds in the cleft. The structure of the calcium-binding domain together with the second follistatin-like domain has also been determined (293). The effect of calcium binding on the structure of SPARC has not been determined at molecular resolution. However, what is known
CALCIUM-BINDING PROTEINS
485
about the calcium binding is that the EF-hand pair contains one highaffinity and one low-affinity site (294).Calcium binding induces a conformational change resulting in an increase in helical structure as monitored by circular dichroism (295). Curiously, this apparent increase in regular structure does not alter the stability of the protein as measured by chemical denaturation.
V. Summary
In this review of calcium-binding proteins, we have not attempted to give an exhaustive summary of all calcium-binding proteins, but have considered some of the best examples, especially those for which high-resolution structural data are available. However, even this limited subset of the hundreds of calcium-binding proteins identified to date demonstrates the myriad ways in which nature uses calcium to regulate biological processes. It also demonstrates how finely tuned each protein is with respect to its environmental parameters such as calcium concentration as well as rates and magnitudes of changes in thia concentration. It also shows the many ways in which the calciumbinding proteins are adapted to the particular target, whether it be a single protein, many proteins, membranes, or simply calcium itself. Despite the wealth of information available to us, much is left to learn about the functions and mechanisms of the calcium-binding proteins. Readers interested in further information are encouraged to explore the articles and reviews listed in the references.
REFERENCES 1. Forsen, S.; Kordel, J . In “Bioinorganic Chemistry”; Bertini, H. B. C . I., Lippard, J. S., and Valentine, J. S., Eds.; University Science Books, 1994. 2. Lincoln, S. F.; Merbach, A. E. In “Advances in Inorganic Chemistry”; Sykes, G. A,, Ed.; Academic Press: New York, 1995, 1. 3. Johansson, C.; Brodin, P.; Grundstrom, T.; Forsen, S.; Drakenberg, T. Eur. J . Biochem. 1991,202,1283. 4. Johansson, C. Thesis; Lund University, Lund, 1993. 5. “Guidebook to the Calcium Binding Proteins”; Celio, M. R., Ed.; Oxford University Press: New York, 1996. 6. Kretsinger, R. M.; Nockolds, C . B. J. B i d . Chem. 1973, 248, 3313. 7. Nakayama, S.; Moncrief, N. D.; Kretsinger, R. H. J. Mol. Euol. 1992,34, 416. 8. Holmes, M. A,; Stenkamp, R. E. J. Mol. Blol. 1991, 220, 723. 9. Mathews, F. S.; Bethge, P. H.; Czenvinski, E. W. J. Biol. Chem. 1979,254, 1699.
486
FINN AND DRAKENBERG
10. Wendt, B.; Hofmann, T.; Martin, S. R.; Bayley, P.; Brodin, P.; Grundstrom, T.; Thulin, E.; Linse, S.; Forsen, S. Eur. J . Biochem. 1988, 175, 439. 11. Kawasaki, H.; Kretsinger, R. H. “Calcium-Binding Proteins.” Academic Press: New York, 1994. 12. Cohen, P.; Klee, C. B. In “Molecular Aspects of Cellular Regulation”; Klee, C. B., Ed.; Elsevier: Amsterdam, 1988. 13. Babu, Y. S.; Sack, J. S.; Greenbough, T. C.; Bugg, C. E.; Means, A. R.; Cook, W. J. Nature 1985,315, 37. 14. Babu, Y. S.; Bugg, C. E.; Cook, W. J. J. Mol. Biol. 1988,204, 191. 15. Seaton, B. A,; Head, J . F.; Richards, F. M. Biochemistry 1985,24, 6740. 16. Heidorn, D. B.; Trewhella, J . Biochemistry 1988,27, 909. 17. Chattopadhyaya, R.; Meador, W. E.; Means, A. R.; Quiocho, F. A. J. Mol. Biol. 1992,228, 1177. 18. Taylor, D. A.; Sack, 3. S.; Maune, J . F.; Beckingham, K.; Quiocho, F. A. J. Biol. Chem. 1991,266, 21375. 19. Rao, S. T.; Wu, S.; Satyshur, K. A.; Ling, K. Y.; Kung, C.; Sundaralingam, M. Protein Sci. 1993, 2, 436. 20. Ikura, M.; Spera, S.; Barbato, G.; Kay, L. E.; Krinks, M.; Bax, A. Biochemistry 1991,30, 9216. 21. Barbato, G.; Ikura, M.; Kay, L. E.; Pastor, R. W.; Bax, A. Biochemistry 1992, 31, 5269. 22. ONeil, K. T.; DeGrado, W. F. Trends Biochem. Sci. 1990, 15, 59. 23. Meador, W. E.; Means, A. R.; Quiocho, F. A. Science 1992,257, 1251. 24. Meador, W. E.; Means, A. R.; Quiocho, F. A. Science 1993,262, 1718. 25. Ikura, M.; Clore, G. M.; Gronenborn, A. M.; Zhu, G.; Klee, C. B.; Bax, A. Science 1992,256, 632. 26. Kemp, B. E.; Pearson, R. B.; Guerriero, J. V.; Bagchi, I. C.; Means, A. R. J . Biol. Chem. 1987,262, 2542. 27. Pearson, R. B.; Wettenhall, R. E. H.; Means, A. R.; Hartshorne, D. J.; Kemp, B. E. Science 1988,241, 970. 28. Goldberg, J.; Nairn, A. C.; Kuriyan, J. Cell 1996, 84, 875. 29. Liu, Y.; Storm, D. R. Trends Pharmacol. Sci. 1990, 11, 107. 30. Gamby, C.; Waage, M. C.; Allen, R. G.; Baizer, L. J. Biol. Chem. 1996,271, 26698. 31. Strynadka, N. C. J.; James, M. N. G. Proteins: Struct. Func. Genet. 1988,3, 1. 32. Herzberg, 0.;James, M. N. G. Nature 1985,313, 653. 33. Herzberg, 0.;James, M. N. G. J. Mol. Biol. 1988, 203, 761. 34. Zhang, M.; Tanaka, T.; Ikura, M. Nut. Struct. Biol. 1995,2, 758. 35. Kubinowa, H.; Tjandra, N.; Grzesiek, S.; Ren, S.; Klee, C. B.; Bax, A. Nut. Struct. Biol. 1995, 2, 768. 36. Finn, B. E.; Forsen, S. Structure 1995,3, 7. 37. Finn, B. E.; Evenas, J.; Drakenberg, T.; Waltho, J. P.; Thulin, E.; Forsen, S. Nut. Struct. Biol. 1996,2, 777. 38. Herzberg, 0.;Mvult, J.; James, N. G. J. B i d . Chem. 1986,261, 2638. 39. Slupsky, C. M.; Sykes, B. D. Biochemistry 1995,34, 15953. 40. Findlay, W. A.; Sonnichsen, F. D.; Sykes, B. D. J . Biol. Chem. 1994,269, 6773. 41. Gagne, S. M.; Tsuda, S.; Li, M. X.; Smillie, L. B.; Sykes, B. D. Nut. Struct. Biol. 1995, 2, 784. 42. McKay, R. T.; Tripet, B. P.; Hodges, R. S.; Sykes, B. D. J. Biol. Chem. 1997, 272, 28494.
CALCIUM-BINDING PROTEINS
487
43. Li, M. X.; Gagne, S. M.; Spyracopoulos, L.; Kloks, C. P.; Audette, G.; Chandra, M.; Solaro, R. J.; Smillie, L. B.; Sykes, B. D. Biochemistry 1997, 36, 12519. 44. Gagne, S.; Li, M.; Sykes, B. D. Bioch,emistry 1997, 36, 4386. 45. Beckingham, K. J. Biol. Chem. 1991,266, 6027. 46. Evanas, J.; Thulin, E.; Malmendal, A,; Forsen, S.; Carlstrom, G. Biochemistry 1997,36, 3448. 47. James, P.; Vorherr, T.; Thulin, E.; Forsen, S.; Carafoli, E. FEBS Lett. 1991,278, 155.
48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65.
Szebenyi, D. M. E.; Obendorf, S. K.; Moffat, K. Nature 1981,294, 327. Szebenyi, D. M. E.; Moffat, K. J. Biol. Chem. 1986,261, 8761. Svensson, L. A.; Thulin, E.; Forsen, S. J. Mol. Biol. 1992, 223, 601. Kordel, J.; Skelton, N. J.; Akke, M.; Chazin, W. J . J . Mol. Biol. 1993, 231, 711. Skelton, N. J.; Kordel, J.; Akke, M.; Forsen, S.; Chazin, W. J . Nut. Struct. Biol. 1994, 1, 239. Skelton, J . N.; Kordel, J.; Chazin, W. J. J . Mol. Biol. 1995,249, 441. Akke, M.; Forsen, S.; Chazin, W. J . J. Mol. Biol., in press (1995). Akke, M.; Drakenberg, T.; Chazin, W. J . Biochemistry 1992,31, 1011. Linse, S.; Teleman, 0.;Drakenberg, T. Biochemistry 1990,29, 5925. Skelton, N. J.; Kordel, J.; Akke, M.; Chazin, W. J . J. Mol. Biol. 1992, 227, 1100. Akke, M.; Skelton, N. J.; Kordel, J.; Palmer, A. D.; Chazin, W. J . Biochemistry 1993,32, 9832. Kordel, J.; Skelton, N. J.; Akke, M.; Palmer, A. D.; Chazin, W. J. Biochemistry 1992,31, 4856. Chazin, W. J.; Kordel, J.; Thulin, E.; Hofmann, T.; Drakenberg, T.; Forsen, S. Biochemistry 1989,28, 8646. Chazin, W. E.; Kordel, J.; Drakenberg, T.; Thulin, E.; Brodin, P.; Grundstrom, T.; Forsen, S. Proc. Nutl. Acud. Sci. U.S.A. 1989, 86, 2195. Linse, S.; Forsen, S. Adu. Second Messenger Phosphoprotein Res. 1995, 30, 89. Moore, B. W. Biochem. Biophys. Res. Commun. 1965, 19, 739. Schafer, B. W.; Wicki, R.; Engelkamp, D.; Mattei, M. G.; Heizmann, C. W. Genomics 1995,25, 638. Zimmer, D. B.; Cornwall, E. H.; Landar, A,; Song, W. Bruin Res. Bull. 1995, 37,
417. 66. Barger, S. W.; Wolchok, S. R.; Van Eldik, L. J. Biochim. Biophys. Actu 1992, 1160, 105. 67. Landar, A,; Hall, T. L.; Cornwall, E. H.; Correia, J . J.; Drohat, A. C.; Weber, D. J.; Zimmer, D. B. Biochim. Biophys. Actu 1997, 1343, 117. 68. Kilby, P. M.; Van Eldik, L. J.; Roberts, G. C. K. Structure 1996,4, 1041. 69. Drohat, A. C.; Amburgey, J . C.; Abildgaard, F.; Stanch, M. R.; Baldisseri, D.; Weber, D. J. Biochemistry 1996,35, 11577. 70. Groves, P.; Finn, B. E.; Kuznicki, J.; Forsen, S. FEBS Lett. 1998,421, 175. 71. Matsumura, H.; Shiba, T.; Inoue, T.; Harada, S.; Kai, Y. Structure 1998, 6, 233. 72. Drohat, A. C.; Baldisseri, D. M.; Rustandi, R. R.; Weber, D. J. Biochemistry 1998, 37, 2729. 73. Smith, S.; Shaw, G. S. Structure 1998, 6, 211. 74. Kilby, P. M.; Van Eldik, L. J.; Roberts, G. C. K. Prot. Sci. 1997, 6, 2494. 75. Rustandi, R. R.; Drohat, A. C.; Baldisseri, D. M.; Wilder, P. T.; Weber, D. J. Biochemistry 1998,37, 1951. 76. Kuznicki, J ; Filipek, A. Biochem. J. 1987, 247, 663.
488
FINN AND DRAKENBERG
77. Potts, B. C.; Smith, J.; Akke, M.; Macke, T. J.; Okazaki, K.; Hidaka, H.; Case, D. A,; Chazin, W. J. Nut. Struct. Biol. 1996,2, 790. 78. Sastry, M.; Ketchum, R. R.; Crescenzi, 0.; Weber, C.; Lubienski, M. J.; Hikada, H., Chazin, W. J. Structure 1998, 6, 223. 79. Filipek, A,; Kuznicki, J. Actu Biochim. Pol. 1993,40, 171. 80. Pedrocchi, M.; Schafer, B. W.; Durussel, I.; Cox, J. A.; Heizmann, C. W. Biochemistry 1994,33, 6732. 81. Filipek, A,; Wojda, U.Biochem. J. 1993,320, 585. 82. Filipek, A,; Kuznicki, J. J . Neurochem. 1998, 70, 1793. 83. Ames, J. B.; Porumb, T.; Tanaka, T.; Ikura, M.; Stryer, L. J. Biol. Chem. 1995, 270, 4526. 84. Ames, J. B.; Tanaka, T.; Ikura, M.; Stryer, L. J. Biol. Chem. 1996,270, 30909. 85. Flaherty, K. M.; Zozulya, S.; Stryer, L.; McKay, D. B. Cell 1993, 75. 86. Tanaka, T.; Ames, J. B.; Harvey, T. S.; Stryer, L.; Ikura, M. Nature 1995,376, 444. 87. Ames, J. B.; Ishima, R.; Tanaka, T.; Gordon, J. I.; Stryer, L.; Ikura, M. Nature 1997,389, 198. 88. Chen, C.-K.; Inglese, J.; Lefkowitz, R. J.; Hurley, J. B. J. Biol. Chem. 1994, 2780, 18060. 89. Dizhoor, A. M.; Chen, C.-K.; Olshevskaya, E.; Sinelnikova, V. V.; Phillipov, P.; Hurley, J. B. Science 1993, 259, 829. 90. Liemann, S.; Lewit-Bentley, A. Structure 1995, 3, 233. 91. Swairjo, M. A.; Roberts, M. F.; Campos, M.-B.; Dedman, J. R.; Seaton, B. A. Biochemistry 1994,33, 10944. 92. Raynal, P.; Pollard, H. B. Biophys. Biochim. Actu 1994, 1197, 63. 93. Davidson, F. F.; Lister, M. D.; Dennis, E. A. J. Biol. Chem. 1990,265, 5602. 94. Russo Mane, F. In “The Annexins”; Moss, S. E., Ed.; Portland Press: London, 1992,35-46. 95. Andree, H. A. M.; Stuart, M. C. A.; Hermens, W. T.; Reutelingsperger, C. P. M.; Hemker, H. C.; Frederick, P. M.; Willems, G. M. J. Biol. Chem. 1992,267, 17907. 96. Pollard, H. B.; Guy, H. R.; Arispe, N.; de la Fuente, M.; Lee, G.; Rojas, E. M.; Pollard, J. R.; Srivastava, M.; Zhang Keck, Z. Y.; Merezhinskaya, N.; Caohuy, H.; Burns, A. L.; Rojas, E. Biophys. J. 1992,62, 15. 97. Demange, P.; Voges, D.; Benz, J.; Liemann, S.; Gottig, P.; Berendes, R.; Burger, A.; Huber, R. Trends Biochem. 1994, 19, 272. 98. Huber, R.; Romisch, J.; Pbques, E.-P. EMBO J . 1990,9, 3867. 99. Huber, R.; Schneider, M.; Mayr, I.; Romisch, J.; Pbques, E.-P. FEBS Lett. 1990, 275, 15. 100. Lewit-Bentley, A.; Morere, S.; Huber, R.; Bodo, G. Eur. J. Biochem. 1992,210, 73. 101. Sopkova, J.; Renouard, M.; Lewit-Bentley, A. J. Mol. Biol. 1993,234, 816. 102. Bewley, M. C.; Boustead, C. M.; Walker, J. H.; Waller, D. A,; Huber, R. Biochernistry 1993,32, 3923. 103. Concha, N. 0.; Head, J. F.; Kaetzel, M. A,; Dedman, J. R.; Seaton, B. A. Science 1993,261, 1321. 104. Swairjo, M. A.; Concha, N. 0.; Kaetzel, M. A.; Dedman, J. R.; Seaton, B. A. Nut. Struct. Biol. 1995,2, 968. 205. Weng, X.; Luecke, H.; Song, I. S.; Kang, D. K.; Kim, S.-H.; Huber, R. Prot. Sci. 1993,2, 448. 106. Shadle, P. J.; Gerke, V.; Weber, K. J. Biol. Chem. 1986, 260, 16354. 107. Schlaepfer, D. D.; Mehlman, T.; Burgess, W. H.; Haigler, H. T. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 6078.
CALCIUM-BINDING PROTEINS
108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129.
489
Schlaepfer, D. D.; Haigler, H. T. J . Biol. Chem. 1987,262, 6931. Haigler, H. T.; Schlaepfer, D. D.; Burgess, W. T. J . Biol. Chem. 1987,262, 6921. Bazzi, M. D.; Nelsestuen, G. L. Biochemistry 1991, 30, 971. Tait, J. F.; Gibson, D.; Fujikawa, K. J . Biol. Chem. 1989, 264, 7944. Blackwood, R. A.; Ernst, J. D. Biochem. J . 1990,266, 195. Follenius-Wund, A.; Piemont, E.; Freyssinet, J.-M.; Gerard, D.; Pigault, C. Biochem. Biophys. Res. Commun. 1997,234, 111. Voges, D.; Berendes, R.; Burger, A.; Demange, P.; Baumeister, W.; Huber, R. J. Mol. Biol. 1994,238, 199. Pigault, C.; Follenius-Wund, A,; Schmutz, M.; Freyssinet, J.-M.; Brisson, A. J. Mol. Biol. 1994,236, 199. Bazzi, M. D.; Nelsestuen, G. L. Biochemistry 1990,29, 7624. Orr,J. W.; Newton, A. C. Biochemistry 1992,31, 4660. Newton, A. C.; Keranen, L. M. Biochemistry 1994,33, 6651. Davletov, B. A,; Suhof, T. A. J . Biol. Chem. 1993, 268, 26386. Chapman, E. R.; Jahn, R. J. Biol. Chem. 1994,269, 5735. Fukuda, M.; Kojima, T.; Mikoshiba, K. J. Biol. Chem. 1996,271, 8430. Li, L.; Darden, T.; Foley, C.; Hiskey, R.; Pedersen, L. Prot. Sci. 1996,4, 2341. Ullrich, B.; Li, C.; Zhang, J. Z.; McMahon, H.; Andersson, R. G. W.; Geppert, M.; Sudhof, T. C. Neuron 1994,13, 1281. Clark, J . D.; Lin, L.-L.; Kriz, R. W.; Ramesha, C. S.; Sultzman, L. A,; Lin, A. Y.; Milona, N.; Knopf, J. L. Cell 1991, 65, 1043. Nafelski, E. A,; Sultzman, L. A,; Martin, D. M.; Kriz, R. W.; Towler, P. S.; Knopf, J. L.; Clark, J. D. J. Biol. Chem. 1994,269, 18239. Kramer, R. M.; Sharp, J. D. FEBS Lett. 1997,410, 49. Takenawa, T.; Homma, Y.; Emori, Y. Methods Enzymol. 1991, 197. Rhee, S. G.; Choi, D. J. J. Biol. Chem. 1993, 267, 12393. Rebecchi, M.; Boguslavsky, V.; Boguslavsky, L.; McLaughlin, S. Biochemistry
1992,31, 12748. 130. Ellis, M. V.; Carne, A,; Katan, M. Eur. J . Biochem. 1993,213, 339. 131. Yagisawa, H.; Hirata, M.; Kanematsu, T.; Watanabe, Y.; Ozaki, S.; Sakuma, K.; Tanaka, H.; Yabutan, N.; Kamata, H.; Hirata, H.; Nojima, H. J. Biol. Chem. 1994, 269, 20179. 132. Cheng, H.-F.; Jiang, M.-J.; Chen, C.-L.; Liu, S.-M.; Wong, L.-P.; Lomasney, J. W.; Kmg, K. J. Biol. Chem. 1996, 270, 5495. 133. Essen, L.-0.; Perisic, 0.; Cheung, R.; Katan, M.; Williams, R. L. Nature 1996, 380, 595. 134. Grobler, J . A,; Essen, L.-0.; Williams, R. L.; Hurley, J. H.; Nut. S t r u t . Biol. 1996, 3, 788. 135. Essen, L.-0.; Perisic, 0.; Lynch, D. E.; Katan, M.; Williams, R. L. Biochemistry 1997.36, 2753. 136. Nalefski E. A,; Falke, J. J . Prot. Sci. 1996, 5, 2375. 137. Nalefski, E. A,; Slazas, M. M.; Falke, J. J. Biochemistry 1997, 36, 12011. 138. Falke, J. J.; Drake, S. K.; Hazard, A. L.; Peerse, 0. B. Q.Reu. Biophys. 1994, 27, 219. 139. Sutton, R. B.; Davletov, B. A,; Berghuis, A. M.; Sudhof, T. C.; Sprang, S. R. Cell 1996, 80, 929. 140. Cifuentes, M. E.; Honkanen, L.; Rebecchi, M. J . J . Biol. Chem. 1993,268, 1158611593. 141. Robertson, P. J.; Hiskey, R. G.; Koehler, K. A. J . Biol. Chem. 1978,253, 5880.
490
FINN AND DRAKENBERG
142. Colpitts T. L.; Castellino, F. J. Znt. J. Peptide Protein Res. 1993, 41, 567. 143. Atkinson, R. A,; Evans, J. S.; Hauschka, P. V.; Levine, B. A,; Meats, R.; Triffitt, J. T.; Virdi, A. S.; Williams, R. J. P. Eur. J. Biochem. 1995,232, 515. 144. Hauschka, P. V.; Lian, J. B.; Cole, D. E.; Gundberg, C. M. Physiol. Reu. 1989, 69, 990. 145. Furie, B.; Furie, B. C. Cell 1988,53, 505. 146. Stenffo,J.; Dahlback, B. In “The Molecular Basis of Blood Diseases”; Stamatoyan-
147. 148. 149. 150.
nopoulus, G., Nienhuis, A. W., Majerus, P. W., and Varmus, H., Eds.; Saunders: Philadelphia, 1994. Stenflo, J.; Garnot, P. J. Biol. Chem. 1972, 247, 8160. Stenflo, J. J . Biol. Chem. 1973, 248, 6325. Nelsestuen, G. L.; Suttie, J. W. Biochemistry 1972, 20, 351. Stenflo, J.; Fernlund, P.; Egan, W.; Roepstorff, P. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 2730.
151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161.
Nelsestuen, G. L.; Zytkonics, T. H.; Howard, J. B. J. Biol. Chem. 1974,249, 6347. Henriksen, R. A.; Jackson, C . M. Arch. Biochen. Biophys. 1975,170, 149. Nelsestuen, G. L. J. Biol. Chem. 1976,25, 5648. Prendergast, F. G.; Mann, K. G. J . Biol. Chem. 1977,252, 840. Marsh, H. C.; Robertson, P. J.; Scott, M. E.; Koehler, K. A.; Hiskey, R. G. J . Biol. Chem. 1979,254, 10268. Strickland, D.; Castellino, F. J. Arch. Biochem. Biophys. 1980, 199, 61. Amphlett, G. W.; Kisiel, W.; Catellino, F. J. Biochemistry 1981,20, 2156. Bajaj, P. S. J. Biol. Chem. 1982, 257, 4127. Jackson, C. M. In “Current Advances in Vitamin K Research; Suttie, J. W., Ed.; Elsevier: New York, 1988. Ohlin, A.-K.; Bjork, I.; Stenflo, J . Biochemistry 1990, 29, 644. Monroe, D. M.; Deerfield, D. W. I.; Olson, D. L.; Stewart, T. N.; Treanor, R. E.; Roberts, H. R.; Hiskey, R. G.; Pedersen, L. G. Blood Coagul. Fibrinolysis 1990,
1, 633. 162. Persson, E.; Bjork, I.; Stenflo, J. J. Biol. Chem. 1991,266, 2444. 163. Sabharwal, A. K.; Padmanabhan, K.; Tulinsky, A.; Mathur, A.; Gorka, J.; Bajaj, P. J. Biol. Chem. 1997,272, 22037. 164. Jacobs, M.; Freedman, S. J.; Furie, B. C.; Furie, B. J . Biol. Chem. 1994,269, 25494. 165. Colpitts, T. L.; Castellino, F. J. Biochemistry 1994, 33, 3501. 166. Colpitts, T. L.; Prorok, M.; Castellino, F. J. Biochemistry 1995,34, 2424. 167. Freedman, S. J.; Furie, B. C.; Furie, B.; Baleja, J. D. J. Mol. Biol. 1996,270, 7980. 168. Sunnerhagen, M.; Olah, G.; Stenflo, J.; Drakenberg, T.; Trewhella, J . Biochemistry 1996,35, 1547. 169. Soriano-Garcia, M.; Park, C. H.; Tulinsky, A.; Ravichandran, K. G.; SkrzypczakJankun, E. Biochemistry 1989,28, 6805. 170. Banner, D. W.; DArcy, A.; Chene, C.; Winkler, F. K.; Guha, A,; Konigsberg, W. H.; Nemerson, Y.; Kirchhofer, D. Nature 1996,380, 41. 171. Freedman, S. J.;Furie, B. C . ; Furie, B.; Baleja, J. D. Biochemistry 1995,34, 12126. 172. Li, L.; Darden, T. A.; Freedman, S. J.; Furie, B. C.; Furie, B.; Baleja, J. D.; Smith, H.; Hiskey, R. G.; Pedersen, L. G. Biochemistry 1997,36, 2132. 173. Sunnerhagen, M.; Forsen, S.; Hoffren, A.-M.; Drakenberg, T.; Teleman, 0.;Stenflo, J. Nut. Struct. Biol. 1996, 2, 504. 174. Arni, R. K.; Padmanabhan, K.; Padmanabhan, K. P.; Wu, T. P.; Tylinsli, A. Chem. Phys. Lipid 1994, 67/68, 59. 175. Zhang L.; Castellino, F. J. J. Bid. Chem. 1994,269, 3590.
CALCIUM-BINDINGPROTEINS
49 1
176. Olivera, B. M.; Rivier, J.; Clark, C.; Ramilo, C.; Corpuz, G.; Bogadie, F.; Woodward, s.;Hillyard, D.; Cruz, L. J . Science 1990, 249, 257. 177. Olivera, B. M.; Rivier, J.; Scott, J. K.; Hillyard, D. R.; Cruz, L. J. J . Biol. Chem. 1991,266, 22067. 178. Cruz, L. J.; Ramilo, C. A.; Corpuz, G. P.; Olivera, B. M. Biol. Bull. 1992,183, 159. 179. McIntosh, J . M.; Olivera, B. M.; Cruz, L. J.; Gray, W. R. J . Biol. Chem. 1984, 259, 14343. 180. Haak, J . A.; Rivier, J.; Parks, T. N.; Mena, E. E.; Cruz, L. J.; Olivera, B. M. J . Biol. Chem. 1990, 265, 6025. 181. Chandler, P.; Pennington, M.; Maccacchini, M.-L.; Nashed, N. T.; Skolnick, P. J. Biol. Chem. 1993,268, 17173. 182. Zhou, L.-M.; Szendrei, G. I.; Fossom, L. H.; Maccecchini, M.-L.; Skolnick, P.; Otvos, L. J. J . Neurochem. 1996, 66, 620. 183. Skjaerbaek, N.; Nielsen, K. J.; Lewis, R. J.; Alewood, P.; Craik, D. J . J. Biol. Chem. 1997,272, 2291. 184. Chiancone, E.; Drakenberg, T.; Teleman, 0.; Forsen, S. J. Mol. Biol. 1985, 185, 201. 185. Rigby, A. C.; Baleja, J. D.; Furie, B. C.; Furie, B. Biochemistry 1997,36, 6906. 186. Rigby, A. C.; Baleja, J. D.; Li, L.; Pedersen, L. G.; Furie, B. C.; Furie, B. Biochemistry 1997, 36, 15677. 187. Soriano-Garcia, M.; Padamanbhan, K.; deVos, A. M.; Tulinsky, A. Biochemistry 1992,31, 2554. 188. Campbell, I. D.; Bork, P. Curr. Opin. Struct. Biol. 1993, 3, 385. 189. Stenflo, J.; Lundwall, A.; Dahlback, B. Proc. Nutl. Acad. Sci. U.S.A. 1987,84, 368. 190. Gronke, R. S.; Van Dusen, W. J.; Garsky, V. M.; Jacobs, J . W.; Sardana, M. K.; Stern, A. M.; Friedman, P. A. Proc. Nutl. Acud. Sci. U.S.A. 1989,86, 3609. 191. Pereira, L.; D’Alessio, M.; Ramirez, F.; Lynch, J. R.; Sykes, B.; Pangilinan, T.; Bonadio, J . Hum. Mol. Genet. 1993,2, 961. 192. Corson, G. M.; Chalberg, S. C.; Dietz, H. C.; Charbonneau, N. L.; Sakai, L. Y. Genomics 1993, 17, 476. 193. Glanville, R. W.; Qian, R.-Q.; McClure, D. W.; Maslen, C. L. J . Biol. Chem. 1994, 269, 26630. 194. Argraves, W. S.; Tran, H.; Burgess, W. H.; Dickerson, K. Cell Biol. 1990, I l l , 3155. 195. Pan, T. C.; Sasaki, T.; Zhabg, R. Z.; Fassler, R.; Timpl, R.; Chu, M. L. J. Cell. Biol. 1993,123, 1269. 196. Tran, M. J.; Mattei, M.; Godyna, S.; Argraves, W. S. Matrix Biol. 1997, 15, 479. 197. Tran, H.; Van Dusen, W. J.; Argraves, W. S. J . Biol. Chem. 1997,272, 22600. 198. Wharton, K. A,; Johansen, K. M.; Xu, T.; Artavanis-Tsakonas, S. Cell 1985, 43, 567. 199. Morita, T.; Isaacs, B. S.; Esmon, C. T.; Johnson, A. E. J. Biol. Chem. 1984, 259, 5698. 200. Morita, T.; Kisiel, W. Biochenz. Biophys. Res. Commun. 1985, 130, 841. 201. Sugo, T.; Bjork, I.; Holmgren, A,; Stenflo, J . J. Biol. Chem. 1984,259, 5705. 202. Sugo, T.; Dahlback, B.; Stenflo, J. J. Biol. Chem. 1986, 261, 5116. 203. Esmon, N. L.; Debault, L. E.; Esmon, C. T. J . Biol. Chem. 1983,258, 5548. 204. Viliers, C. L.; Arlaud, G . J.; Painter, R. H.; Colomb, M. G. FEBS Lett. 1980, 117, 289. 205. Ohlin A.-K.; Stenflo, J. J. Biol. Chem. 1987, 262, 13798. 206. Ohlin, A.-K.; Linse, S.; Stenflo, J . J . Biol. Chem. 1988, 263, 7411.
492
FINN AND DRAKENBERG
207. Ohlin, A.-K.; Laudes, G.; Bourdon, P.; Oppenheimer, C.; Wydro, R.; Stenflo, J. J. Biol. Chem. 1988,263, 19240. 208. Persson, E.; Selander, M.; Linse, S.; Drakenberg, T.; Stenflo, J. J. Biol. Chem. 1989,264, 16897. 209. Handford, P. A,; Baron, M.; Mayhew, M.; Willis, A.; Beesley, T.; Brownlee, G. G.; Campbell, I. D. EMBO J. 1990, 9, 475. 210. Selander Sunnerhagen, M.; Persson, E.; Dahlqvist, I.; Drakenberg, T.; Stenflo, J.; Mayhew, M.; Robin, M.; Hanford, P.; Tilley, J. W.; Campbell, I. D.; Brownlee, G. G. J. Biol. Chem. 1993,268, 23339. 211. Handford, P.; Downing, A. K.; Rao, Z.; Hewett, D. R.; Sykes, B. C.; Kielty, C. M. J. Biol. Chem. 1995,270, 6751. 212. Wu, Y.-S.; Bevilacqua, V. L. H.; Berg, J. M. Chem. Biol. 1995,2, 91. 213. Astermark, J.; Bjork, I.; Ohlin, A.-K.; Stenflo, J. J. Biol. Chem. 1991,266, 2430. 214. Valcarce, C.; Selander Sunnerhagen, M.; Tamlitz, A.-M.; Drakenberg, T.; Bjork, I.; Stenflo, J. J. Biol. Chem. 1993, 268, 26673. 215. Knott, V.; Downing, A. K.; Cardy, C. M.; Handford, P. J. Mol. Biol. 1996,255, 22. 216. Dahlback, B.; Hildebrand, B.; Linse, S. J. Biol. Chem. 1990,265, 18481. 21 7. Stenberg, Y.; Linse, S.; Drakenberg, T.; Stenflo, J . J . B i d . Chem. 1997,272, 23255. 218. Stenberg, Y.; Julenius, K.; Dahlqvist, I.; Drakenberg, T.; Stenflo, J. Eur. J. Biochem. 1997,248, 163. 219. Rand, M. D.; Lindblom, A.; Carlson, J.; Villoutreix, B. 0.; Stenflo, J . Prot. Sci. 1997, 6, 2059. 220. Cooke, R. M.; Wilkinson, A. J.; Baron, M.; Pastore, A.; Tappin, M. J.; Campbell, I. D.; Gregory, H.; Sheard, B. Nature 1987,327, 339. 221. Kohda D.; Inagaki, F. J. Biochem. 1988,103, 554. 222. Campbell, I. D.; Cooke, R. M.; Baron, M.; Harvey, T. S.; Tappin, M. J. Prog. Growth Fac. Res. 1989, 1, 13. 223. Tappin, M. J.; Cooke, R. M.; Fitton, J . E.; Campbell, I. D. Eur. J. Biochem. 1989, 179, 629. 224. Kline, T. P.; Brown, F. K.; Brown, S. C.; Jeffs, P. W.; Kopple, K. D.; Mueller, L. Biochemistry 1990,29, 7805. 225. Cooke, R. M.; Tappin, M. J.; Campbell, I. D.; Dohda, D.; Miyake, T.; Fuwa, T.; Miyazawa, T.; Inagaki, F. Eur. J . Biochem. 1990,193, 807. 226. Harvey, T. S.; Wilkinson, A. J.;Tappin, M. J.; Cooke, R. M.; Campbell, I. D. Eur. J. Biochem. 1991, 198, 555. 227. Huang, L. H.; Cheng, H.; Pardi, A.; Tam, J . P.; Sweeney, W. V. Biochemistry 1991, 30, 7402. 228. Baron, M.; Norman, D. G.; Harvey, T. S.; Handford, P. A.; Mayhew, M.; Tse, A. G. D.; Brownlee, G . G.; Campbell, I. D. Prot. Sci. 1992, 1, 81. 229. Ullner, M.; Selander, M.; Persson, E.; Stenflo, J.; Drakenberg, T.; Teleman, J. Biochemistry 1992,31, 5974. 230. Selander Sunnerhagen, M.; Ullner, M.; Persson, E.; Teleman, 0.; Stenflo, J.; Drakenberg, T. J. Biol. Chem. 1992,267, 19642. 231. Moy, F. J.; Li, Y.-C.; Rauenbuchler, P.; Winkler, M. E.; Scheraga, H. A.; Montelione, G. T. Biochemistry 1993,32, 7334. 232. Freedman, S. J.; Sanford, D. G.; Bachovchin, W. W.; Furie, B. C.; Baleja, J. D.; Furie, B. Biochemistry 1996,35, 13733. 233. Graves, B. J.; Crowther, R. L.; Chandran, C.; Rumberger, J. M.; Li, S.; Huang, K.-S.; Presky, D. H.; Familletti, P. C.; Wolitzky, B. A.; Burns, D. K. Nature 1994, 367, 532.
CALCIUM-BINDING PROTEINS
493
234. Padmanabhan, K.; Padmanabhan, K. P.; Tulinsky, A,; Park, C. H.; Bode, W.; Huber, R.; Blankenship, D. T.; Cardin, A. D.; Ksiel, W. J . Mol. Biol. 1993,232,947. 235. Rao, Z.; Handford, P.; Mayhew, M.; Knott, V.; Brownlee, G. G.; Stuart, D. Cell 1995, 82, 131. 236. Banner, D. W. Thromb. Haemostasis 1997, 78, 512. 237. Downing, A. K.; Knott, V.; Werner, J. M.; Cardy, C. M.; Campbell, I. D.; Handford, P. A. Cell, 1996, 85, 597. 238. Bode, W.; Schwager, P. FEBS Lett. 1975,56, 139. 239. Bode, W.; Schwager, P. J. Biol. Chem. 1975, 98, 693. 240. Persson, E.; Hogg, P. J.; Stenflo, J. J. Biol. Chem. 1993, 268, 22531. 241. Rezaie, A. R.; Neuenschwander, P. F.; Morrissey, J. H.; Esmon, C. T. J . Biol. Chem. 1993,268, 8176. 242. Rezaie, A. R.; Mather, T.; Sussman, F.; Esmon, C. T. J. Bid. Chem. 1994, 269, 2151. 243. Wildgoose, P.; Foster, D.; Schiodt, J.; Wiberg, F. C.; Birktoft, J. J.; Petersen, L. C. Biochemistry 1993, 32, 114. 244. Bajaj, S. P.; Sabhanval, A. K.; Gorka, J.; Birktoft, J. J. Proc. Natl. Acad. Sci. U.S.A. 1992.89, 152. 245. Rezaie, A. R.; Esmon, N. L.; Esmon, C. T. J. Biol. Chem. 1992,267, 11701. 246. Sabharwal, A. K.; Birktoft, J. J.; Gorka, J.; Wildgoose, P.; Petersen, L. C.; Bajaj, S. P. J. B i d . Chem. 1996,270, 15523. 247. Freskghrd, P.-0.; Olsen, 0. H.; Persson, E. Prot. Sci. 1996,5, 1531. 248. Brittain, H. G.; Richardson, F. S.; Martin, R. B. J. Chem. Soc. (London) 1976, 98, 8255. 249. Neidhart D. L.; Petsko, G. A. Prot. Eng. 1988, 2, 271. 250. Bode, W.; Papamokos, E.; Musil, D. Eur. J. Biochem. 1987,166, 673. 251. Bott, R.; Ultsch, M.; Kossiakoff, A.; Graycar, T.; Katz, B.; Power, S. J . B i d . Chent. 1988,263, 7895. 252. Khatra, B. S,;Herries, D. G.; Brew, K. Eur. J. Biochem. 1974,44, 537. 253. Bell, J. E.; Beyer, T. A,; Hill, R. L. J. Biol. Chem. 1976,251, 3003. 254. Brew, K.; Grobler, J. A. In “Advanced Dairy Chemistry”; Fox, P., Ed.; Elsevier Press: London, 1992, 1, 191. 255. Brew, K.; Vanamna, T. C.; Hill, R. L. J. Bid. Chem. 1967,242, 3747. 256. Brew, K.; Castellino, F. J.; Vanaman, T. C . ; Hill, R. L. J . Bid. Chem. 1979, 245, 4570. 257. Qasba, P. K.; Safaya, S. K. Nature 1984, 308, 377. 258. Hall, L.; Craig, R. K.; Edbrooke, M. R.; Campbell, P. N. Nucleic Acids Res. 1982, 10, 3503. 259. McKenzie, H. A,; Whitw, F. H. Adu. Protein Chem. 1991,41, 173. 260. Acharya, K. R.; Stuart, D. I.; Walker, N. P. C.; Lewis, M.; Philips, D. C. J. Mol. B i d . 1989,208, 99. 261. Nitta, K.; Tsuge, H.; Sugai, S.; Shimazaki, K. FEBS Lett. 1987,223, 405. 262. Nitta, K.; Tsuge, H.; Shimazaki, K.; Sugai, S. Biol. Chem.. Hoppe-Seyler 1988, 369, 671. 263. Sugai, S.; Nitta, K.; Tsuge, H. Colloq. INSERM 1989, 179, 591. 264. Tsuge, H.; Ago, H.; Noma, M.; Nitta, K.; Sugai, S.; Miyano, M. J. Biochem. (Tokyo) 1992,111, 141. 265. Inaka, K.; Kuroki, R.; Kikuchi, M.; Matsushima, M. J . Biol. Chem. 1991, 266, 20666.
494
FINN AND DRAKENBERG
266. Aramini, J . M.; Drakenberg, T.; Hiraoki, T.; Ke, H.; Nitta, K.; Vogel, H. J. Biochemistry 1992, 31, 6761. 267. Aramini, J . M.; Hiraoki, T.; Grace, M. R.; Swaddle, T. W.; Chiancone, E.; Vogel, H. J . Biochim. Biophys. Acta 1996,1293, 72. 268. Stuart, D. I.; Acharya, K. R.; Walker, N. P. C.; Smoth, S. G.; Lewis, M.; Philips, D. C. Nature 1986,324, 84. 269. Acharya, K. R.; Ren, J.; Stuart, D. J.; Philips, D. C.; Fenna, R. E. J . Mol. Biol. 1991,221, 571. 270. Pike, A. C.; Brew, K.; Acharaya, K. R. Structure 1996,4, 691. 271. Harata, K.; Muraki, M. J. Biol. Chem. 1992,267, 1419. 272. Vanderheeren, G.; Hanssens, I.; Meijberg, W.; Van Aerschot, A. Biochemistry 1996,35, 16753. 273. Wu, L. C.; Schulman, B. A.; Peng, Z.-Y.; Kim, P. S. Biochemistry 1996,35, 859. 274. Anderson, P. J.; Brooks, C. L.; Berliner, L. J. Biochemistry 1997, 36, 11648. 275. Ptitsyn, 0. B. Curr. Opin. Struct. Biol. 1995, 5, 74. 276. Privalov, P. L. J. Mol. Biol. 1996, 258, 707. 277. Chothia C.; Jones, E. Y. Annu. Rev. Biochem. 1997,66, 823. 278. Takeichi, M. Annu. Rev. Biochem. 1990,59, 237. 279, Hatta, K.; Nose, A,; Nagafuchi, A.; Tackeichi, M. J. Cell Biol. 1988, 106, 873. 280. Overduin, M.; Harvey, T. S.; Bagby, S.; Tong, K. I.; Yau, P.; Takeichi, M.; Ikura, M. Science 1995,267, 386. 281. Koch, A. W.; Pokutta, S.; Lustig, A,; Engel, J . Biochemistry 1997, 36, 7697. 282. Shapiro, L.; Fannon, A. M.; Kwong, P. D.; Thompson, A.; Lehman, M. S.; Grubel,
G.; Legrand, J.-F.; Als-Nielsen, J.; Colman, D. R.; Hendrickson, W. A. Nature 1996,374, 327. 283. Overduin, M.; Tong, K. I.; Kay, C. M.; Ikura, M. J . Biomol. NMR 1996, 7, 173. 284. Nagar, B.; Overduin, M.; Ikura, M.; Rini, J. M. Nature 1996,380, 360. 285. Pokutta, S.; Herrenknecht, K.; Kemler, R.; Engel, J . Eur. J. Biochem. 1994, 223, 1019. 286. Termine, J . D.; Kleinman, H. K.; Whitson, S. W.; Conn, K. M.; McGarvey, M. L.; Martin, G. R. Cell 1981,26, 99. 287. Mason, I. J.; Taylor, A,; Williams, J. G.; Sage, H.; Hogan, B. L. M. EMBO J. 1986, 5, 1465. 288. Dziadek, M.; Paulsson, M.; Aumailley, M.; Timpl, R. Eur. J. Biochem. 1986, 161, 455. 289. Lane, T. F.; Sage, H. FASEB J. 1994,8, 163. 290. Johnston, I. G.; Paladino, T.; Gurd, J. W.; Brown, I. R. Neuron 1990,2, 165. 291. Guermah, M.; Crisanti, P.; Langier, D.; Dezelee, P.; Bidou, L.; Pessac, B.; Calothy, G. Proc. NatL. Acad. Sci. U.S.A. 1991, 88, 4503. 292. Hohenester, E.; Maurer, P.; Hohenadi, C.; Timpl, R.; Jansonius, J. H.; Engel, J . Nut. Struct. Biol. 1996, 3, 67. 293. Hohenester, E.; Maurer, P.; Timpl, R. EMBO J. 1997, 16, 3778. 294. Pottgiesser, J.; Maurer, P.; Mayer, U.; Nischt, R.; Mann, K.; Timpl, R.; Krieg, T.; Engel, J. J. Mol. Biol. 1994, 238, 563. 295. Maurer, P.; Mayer, U.; Bruch, M.; Jeno, P.; Mann, K.; Landwehr, R.; Engel, J.; Timpl, R. Eur. J . Biochem. 1992,205, 233.
ADVANCES IN INORGANIC CHEMISTRY, VOL
46
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS MICHAEL J. DAVIES,* CHRISTEL MATHIEUt, and ALAIN PUPPOt The Heart Research Institute, Carnperdown, Sydney, New South Wales 2050, Australia t Laboratoire de Biologie Vegetale et Microbiologie,CNRS URA 1114, Universite de NiceSophia Antipolis, 06018 Nice Cedex 2, France
I. Introduction 11. Structure 111. Biological Localization IV. Reactions with Different Molecules A. Diatomic Gases B. Other Molecules V. Oxidation of Fe(I1) Leghemoglobin VI. Oxidation of Fe(II1) Leghemoglobin VII. Reduction of F e W ) = 0 Leghemoglobin A. Reaction with H200 B. Reaction with Ascorbate C. Reaction with Glutathione, Other Thiols, and Related Species D. Reaction with Other Reductants VIII. Reduction of Fe(II1) Leghemoglobin A. Enzymatic B. Nonenzymatic IX. Reaction of Globin-Derived Radicals A. Reactions with the Heme Ring B. Reactions with Low-Molecular-Weight Species C. Reactions with Other Leghemoglobin Molecules D. Reactions with Membranes References
I. Introduction
Plants release oxygen during photosynthesis and use it as the terminal electron acceptor in the respiratory chain of mitochondria. There is now evidence that they contain a monomeric oxygen-binding protein to coordinate and transport oxygen (I). The occurrence of these myoglobin-like proteins in plants was first described by Kubo in 1939, who demonstrated that the nitrogen-fixing root nodules of soy495 Copyright 'c 1999 by Academic Press. All Fights of reproduction in any form reserved. 0898-883R/99 $25.00
496
DAVIES, MATHIEU, AND PUPPO
bean and other plants contained a red-colored soluble heme protein that could reversibly bind oxygen (2). The protein was named leghemoglobin (Lb) at an early stage by analogy to the mammalian protein. The nodules from which the protein is extracted arise from the interaction between Rhizobium spp., a soil bacteria, and leguminous plants (Fig. 1).More recent studies have demonstrated the presence of similar structures in the nonleguminous plant Parasponia spp. (3). The origin of the symbiosis in legumes has been reviewed (4), and the processes that occur during the development of the relationship are reasonably well characterized (5-7). Hemoglobins and their parent genes have also been recently identified in a number of nonnodulating plants and in phylogenetically diverse plant genera including monocots (8, 9). Moreover, a new hemoglobin gene has been isolated recently from soybean (10)that is expressed in different parts of the plant (including the nodules). This contrasts with the expression of Lbs, which are active only in the infected cells of the nodule. Thus, two different types of hemoglobin are known to be present in plants, (i) a nonsymbiotic type that is widely distributed among species and (ii) a symbiotic type that is induced only on nodulation in (at least) two plant families (1). The genes that code for these two types of hemoglobins are separated into four exons by three introns: the first and third introns are in positions homologous to those of the two introns found in vertebrate a-globin and /I-globin and myoglobin genes. It has been proposed, and seems likely, that there is a common ancestral gene for plant and animal hemoglobins (1, 11-13). This review will concentrate on the properties and reactions of the symbiotic type of plant protein, the leghemoglobins. Until now, Lb has been the most extensively studied oxygen-carrying hemoprotein from plants. The main role of Lb in uiuo is now widely established (though for many years this was hotly disputed) as being to facilitate oxygen transport t o the vigorously respiring, phosphorylating, nitrogen-fixing Rhizobium microsymbionts, at a low and stable oxygen tension (around 10 nM in soybean nodules) (7, 12, 14, 15). Terminal oxidases having a K,,, for oxygen as low as 7 nM are present in the microsymbionts and are able to use the Lb-delivered oxygen to support their respiration under the oxygen-limiting conditions of the nodule (16).This low oxygen tension appears to be a compromise between the requirement for oxygen by the respiring bacteria and the damage that oxygen can inflict on the nitrogen-fixing complex; the latter is very readily inactivated at high oxygen levels (12,15).
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
497
FIG.1. Nitrogen-fixing root nodules on a soybean plant.
II. Structure
Early studies by Ellfolk demonstrated that there were several Lb isomers in soybean nodules (17). Ammonium sulfate fractionation of nodule extracts followed by anion exchange chromatography in ace-
498
DAVIES, MATHIEU, AND PUPPO
tate buffers allowed the separation and purification of two major and two minor species of soybean Lb. The major species were named Lb, and Lb,, and the minor components Lbb and Lbd. Subsequent work using isoelectric focusing, which has a greater resolving power, has shown that there are four major components (subscripts a , c 1 , c2, and c 3 ) rather than the two originally separated, and four minor ones (b, d l , d2, and d3) (18).The latter minor components are N-terminal acetylation products of the major components (19).All of these species are present in single nodules, with the relative concentrations of the different components varying with nodule development (20). This pattern of multiple gene products is widely established in other legume and nonlegume symbioses (12, 18,211. All of the multiple Lb proteins from soybean have been sequenced, and it is now known that the differences in amino acid sequences between different Lb components from a particular plant species are relatively slight. Despite discrepancies between the published sequences obtained in early studies, it has been shown in a large study of both 69 domesticated cultivars of soybean and 18 wild species that there are few (if any) differences between the same type of leghemoglobin obtained from a number of different cultivars of a single species. The published discrepancies have been ascribed to problems in the sequencing procedures (22).In all cases the proteins (molecular weight ca. 16,000) consist of a single polypeptide chain of 144-155 amino acids, and the iron protoporphyrin IX prosthetic group. The proteins are acidic with PI values in the range 5.0-5.5. The amino acid sequences of the soybean proteins, as well as many others, have a very high degree of amino acid sequence homology with vertebrate myoglobins and hemoglobins, and an even higher degree of homology of amino acid type. The sequence of soybean Lb,, which is one of the most extensively studied species, contains two histidine residues (His61 and His92) (171, which are analogous to the proximal and distal histidines of mammalian hemoglobins and myoglobins. The spacing of these residues in the sequence is also very similar. Crystal structure work on both the soybean [(23);Ellis and Freeman, unpublished data] and lupin (24-30) forms of the protein have emphasized the similar key features of the three-dimensional protein structures of the plant and mammalian proteins. There are also a number of key differences, which appear to have quite dramatic functional consequences and determine both the very high oxygen-binding affinity of this protein (discussed later) and the ability of the plant proteins to bind larger ligands at the heme site. The evolution of the plant globin gene family has been the subject of a number of studies [see, for ex-
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
499
ample, (3111 and it is now believed that the plant and animal gene families diverged a t a very early stage. This is estimated to be 900 to 1.4 billion years ago, though the derivation of these numbers requires certain assumptions about the rate of evolutionary change (9, 13,311. A number of studies have shown that the iron protoporphyrin IX group of Lb can be replaced, in uitro, by a variety of other species including mesohemin IX, deuterohemin IX, hematohemin IX, diacetyldeuterohemin IX, and other analogs. As expected, a number of these substitutions cause significant changes in the heme environment, and this is reflected in both the optical absorption and EPR spectra of these species (32,33). The apoprotein can also be reconstituted with a number of noniron porphyrins. Thus, holoproteins have been produced with both cobalt (disulfophthalocyanine and protoporphyrin IX) and zinc (protoporphyrin 1x1among others (18, 34-36); in the case of the cobalt protoporphyrin protein, information has also been obtained on the oxygen-binding capabilities of this species (18).
111. Biological Localization
After considerable controversy in the literature over the last thirty years it is now widely accepted that Lb only occurs in the cytoplasm of the host cell, and not in the peribacteroid space that surrounds the bacteroid (bounded by the peribacteroid membrane, Fig. 2) or the bacteroid itself (7, 15, 37). The nature and development of the membrane that surrounds the bacteroids have been examined in some detail because this membrane regulates the transport of materials to and from the bacteroids (e.g., dicarboxylates, metal ions, etc.) (38-41 ). The local concentration of Lb can be high (up to 2-3 mM) in very active nodules, with an average concentration over the entire nodule of ca. 300 pM (42).Although it is well established [see reviews (12, I8)l that the apoprotein is produced by the plant (43)and that the Lb assembly must occur in the host cell or on the outside the peribacteroid membrane, the origin of the protoporphyrin moiety remains unclear. It is generally proposed that heme is made by the Rhizobium symbiont (44, 451, but the possibility remains that plant organelles are a t least partly responsible for this process (46).The mechanism of control of the synthesis of the apoprotein is disputed. The observation that heme synthesis in isolated Bradyrhizobium japonicum is greatly stimulated by microaerobic conditions (45)suggests that heme export from the nascent bacteroids to the plant cell cytoplasm might be a major stimulus for apoprotein synthesis. However, the observa-
500
DAVIES, MATHIEU, AND PUPPO
FIG.2. Transmission electron micrograph (16,OOOX enlargement) through an infected root cell of a soybean plant. The subcellular organelles of the host cell [mitochondria and plasts; (a)] are present at; the periphery of the cell adjacent to the host cell wall (b). The nitrogen-fixing bacteroids (c) are kept apart from the host cell cytoplasm (the location of leghemoglobin) by the peribacteroid space (d) and the peribacteroid membrane (e), which regulates transport of materials to and from the bacteroids.
tion that ineffective nodules in some mutant forms of Bradyrhizobium japonicum deficient in heme synthesis can contain almost normal levels of apoprotein (47, 48) suggests that synthesis of the protein may not be wholly heme regulated. The assembly of holoprotein appears to precede and be independent of the presence of functional nitrogenase (37,43,49). However, the synthesis of the latter appears to depend on the former, possibly as a result of the control exerted by intact Lb on oxygen tension, which prevents damage to the nitrogenase (37). The turnover time of intact Lb is relatively short [reviewed in (37)], cf: the long half-lives of myoglobin and hemoglobin in mammals, with the major Lb components of pea nodules having turnover times of about two days (49). This is considerably less than the lifetime of the nitrogenase (50).There must therefore be continual synthesis and degradation of Lb within the host cell cytoplasm, and there is evidence for dramatic changes (as mentioned earlier) in relative concen-
50 1
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
trations of different Lb isoforms during this period. Thus, it is known that Lb I predominates over Lb IV in young but not old pea nodules (51), and Lbrpand Lb, predominate in young soybean nodules (20). The destruction of Lb is believed to occur by multiple pathways (12, 37), with the relative contributions of these different routes strongly dependent on the environmental situation; some of these processes are discussed further later. Whatever the process, there is a strong correlation between nitrogen-fixing ability and the presence of intact functional Lb, and considerable work is under way to both understand and control these factors.
IV. Reactions with Different Molecules
A. DIATOMIC GASES The extremely high oxygen-binding activity of the Fe(1I) Lb is critical for its biological function, and considerable work has been undertaken to understand the factors that determine this feature of the protein. It is now well established that the high affinity arises from larger “on” (association) rate constants and smaller “off (dissociation) rate constants (see Table I), which results in KDvalues in the range of 36 to 78 nM depending on the source of the Lb and the pH (15,21, 37,51,52). The 0, off rate constants for soybean Lb decrease by about TABLE I KINETICS AND
EQUILIBRIUM CONSTANTS FOR BINDING OF GAsEOLJS LIGANDS TO LEGHEMOGLOBINS AND OTHER MONOMERIC HEMOGLOBINS (21, 55, 56, 166, 167) Oxygen
Protein Spermwhale Mb Spermwhale Mbd Human a-chain Hb Soybean Lb Lupin Lb I Lupin Lb 11
k’a
kb
12 14 140 1600
50 120 540 320
Carbon monoxide
28 5.6 20 25
Kc
k‘“
0.51 860 11400 5.8 560 4.0 48 13 36 42 78 52
” Association rate constant; pM-’s
0.019 0.038 0.013 0.0078 0.014 0.015
37 6.6 3.25 0.6 0.33 0.015
’.
Dissociation rate constant; s-l. Equilibrium constant, kik’; nM. Mutant spermwhale myoglobin with E7 His ’ Major component of heterogeneous reaction.
’’
kh
-
Gly mutation.
Nitric oxide
k!e
kh
K
17 150 30 120
0.00012 0.00015 0.000046 0.00002
-
-
0.007 0.001 0.0015 0.0001 -
-
502
DAVIES, MATHIEU, AND PUPPO
fivefold as the pH is lowered from 7 to 4,with a pK, of 5.46. This has been ascribed to the formation of a hydrogen bond between the bound oxygen and distal histidine (52). The KD values are 11 to 24 times smaller than those for sperm whale myoglobin, which is mainly due to the higher on rate constant (21, 29). Thermodynamic quantities (AG, AH, AS) have been determined for the binding of a number of ligands including O2 and CO to soybean Lb (53), and the KD values for all eight components of soybean Lb have been determined. Though there are significant differences in these values, the variation in terms of free 0, concentration within nodules has been calculated to be a maximum of 3% (53).Thus, an earlier suggestion (54)that the large number of different forms of Lb in particular species is due to differing oxygen requirements, hence requiring Lb species of different oxygen affinities, is not strongly supported by this data. High on rate constants have also been determined in studies on a sperm whale myoglobin mutant in which the E7 distal His was mutated to a Gly. The on rate constant for this mutant is approximately the same as that of soybean Lb, though it is still smaller than for two lupin forms (21, 55, 56).This type of alteration to the heme pocket is not, however, responsible for the high affinity of Lb, as the off rate constant of this mutant is orders of magnitude greater than those for the soybean and lupin forms (55, 56). There must therefore be other factors that contribute to the very high affinity of Lb molecules for oxygen. Detailed X-ray crystallographic studies of a number of different states of lupin Lb have been reported (24-30,57),and similar studies have been undertaken on the soybean protein (23).The structures of the oxy, deoxy, CO, and NO forms of Fe(I1) lupin Lb, determined to 1.7- to 1.8-A resolution (29, 301, have allowed Fe-ligand geometries and structural changes on ligand binding to be examined. It has been found that the Lb protein has a unique feature as compared to other hemoglobins in that the proximal histidine has a remarkable rotational freedom. Thus, in the deoxyform of the protein, the imidazole is able to oscillate between two different orientations, eclipsing either a line between the N-1 and N-3 pyrrole nitrogens of the porphyrin or a line between N-2 and N-4 (29).In the oxy form, however, it is fixed in a staggered orientation. As with other heme proteins, the iron atom moves (by ca. 0.3 A) from an out-of-plane position to an in-plane position on oxygen binding, with the Fe-N bond distances remaining constant (29).The Fe-0-0 bond angle is 152", very close to that seen in the human hemoglobin a-chain (153'1, and the oxygen is hydrogen bonded to the distal histidine. In both forms of Lb the heme is ruffled
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
503
due to rotation of the pyrrole groups around the N-Fe-N bonds, in accord with earlier suggestions based on resonance Raman studies (58),though this suggestion was later disputed (59). This ruffling results in the methine bridges being out of plane by up to 0.32 A (29). It has been suggested from the crystal structures that the reason for the high on rate constant is the mobility of both the proximal and distal histidines in the deoxy form of the protein (29). The eclipsed nature of the proximal histidine would maximize steric hindrance with the pyrrole nitrogens in the deoxy form and hence minimizer T --$ p electron donation, whereas the staggered position in the oxy form would result in the opposite effect. These two factors in combination might result in a reduction of the activation energy for oxygen binding (29). Furthermore, though the distal histidine appears fixed in the crystal structure, it has been postulated that this can swing out of the heme pocket a t a rapid rate to allow ready access of the oxygen molecule (15,23,60). The kinetics of CO and NO binding to Lb have been studied, and rate constants for both association and dissociation determined under a number of different conditions (e.g., pH). The values obtained are compared with those for Mb and Hb in Table I (61). The off rate constants for CO are, unlike those for 02,independent of pH (521. In particular, such studies have been directed to determine the mode of binding of the ligand and whether this has a bent geometry as has been observed for some hemoglobins. Such an orientation is believed to result from a steric interaction with the globin and is a major factor in reducing the affinity for these poisons (30).The large flexible heme site of Lb and the rapid movement of the distal histidine out of the heme pocket might be expected to result in large on rate constants for reaction with both NO and CO. The crystal data a t 1.8-A resolution has been interpreted in terms of a bent orientation of both CO and NO with an Fe-CO angle of 160" and an Fe-NO angle of 147" (30).The former angle is in conflict with spectroscopic data obtained with myoglobin, where a linear geometry was determined (62, 631, but in accord with the angles obtained from myoglobin X-ray crystal structures (64, 65) and picosecond time-resolved infrared measurements (66).The NO conformation is in accord with the expected bent geometry, with a strong Fe-NO bond (and hence low off rate constant), and the (predicted) lengthened (2.2-A) iron-proximal histidine bond (30). The binding of NO to the Fe(I1) form of Lb (LbNO) may be of considerable biological relevance as a result of the slow off rate constant (21,30).This will result in the formation of Lb, which cannot fulfill its
504
DAVIES, MATHIEU, AND PUPPO
role as an oxygen carrier. Such reactions have long been recognized as having a potential role in nitrate-induced stress (67). It has been demonstrated that Lb extracted from plants grown in the presence of nitrate contains significant amounts of LbNO; thus, an early study reported that up to 27% of the Lb present in soybean nodules was in this form of the protein (68).The species can be readily recognized on the basis of its characteristic UV-vis absorption features (69, 70) or its intense EPR signal in the g ca. 2 region, which has highly distinctive hyperfine couplings arising from the nitrogen nucleus (68). Although it is possible that some of the NO may arise as an artifact of the isolation procedure, it may also arise from the reduction of nitrate to NO by either enzymatic or nonenzymic reactions (67, 68, 71). The former process may involve a nitrate reductase from either the plant or bacteroid. Such reactions have been characterized previously (71). It is known that a variety of reducing agents such as ascorbate or reduced flavins can also bring about this reduction. Such reactions are believed to be responsible for the formation of nitrosylmyoglobin in mammalian systems (72). Recent careful studies have confirmed that the root nodules of nitrate-fed soybean plants do indeed contain LbNO and that this species is not an artifact of the extraction procedure (69, 70). Thus, extraction of Lb from root nodules in the presence of high concentrations (100 nmol/mg protein) of nitrate added to the nodules at the time of harvest did not result in amounts of LbNO higher than those seen in plants grown in the absence of nitrate. It has also been demonstrated that the formation of LbNO requires both the presence of the Fe(I1) protein and a reducing agent. Incubation with monoiodoacetic acid, a thiol (-SH) blocking agent, repressed the formation of this species, suggesting that an enzymatic process may be involved (69). The levels of LbNO in root nodules appear to be enhanced by culturing the plants in higher concentrations of nitrate [cf: the detection of 27% LbNO by Maskell et al. for plants grown with 0.5 mM nitrate, and of 60% by Kanayama and Yamamoto using 10 mM nitrate (68, 6913 or for increasing lengths of time [cf. 60% and 86% LbNO for culture with 10 mM nitrate for 12 and 24 h respectively (6911. The formation of such LbNO complexes may have profound physiological significance, particularly at these very high nitrate levels, as the levels of LbNO in root nodules have been shown to be inversely correlated with the acetylene (ethylenebreducing activity of the nodules (a commonly used marker of their reductive capacity), and hence their ability to fix nitrogen (69). Although these experiments have determined that LbNO can arise from the growth of plants in the presence of nitrate, the situation
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
505
with plants grown in the absence of nitrate is less well established. In the work of Kanayama and Yamamoto, levels of LbNO obtained from plants grown in the absence of nitrate were reported as being less than lo%, suggesting that reduction (either enzymatic or direct) of nitrate was a major, if not the only, source of NO (69).However, very recent reports have suggested that this may not be the case. Thus, it has been reported that root nodules contain an NO synthase activity that generates NO from arginine in a manner akin to mammalian NO synthases (73), and a recent EPR study has reported the detection of signals from LbNO in frozen intact root nodules from plants grown in the complete absence of nitrate (74). The levels of this complex were found to be highest in young nodules, decreased in mature nodules, and almost completely absent in old and senescent species. These reports, if they are confirmed, together with earlier studies that have demonstrated that NO can play a role in regulating gene transcription via a sensory system involving proteins of the family (75-77), suggest that the formation and levels of NO may play a vital role in determining the nitrogen-fixing activity of root nodules. The determination of the crystal structures of these different forms of Lb have also allowed previous suggestions that the heme pocket of Lb is larger and more flexible than those of other heme proteins (15, 23, 37, 781, to be examined in detail. Superposition of the backbone carbon trace of the E and F helices of the Lb structure with that obtained from HbCO and MbCO provides direct confirmation of the larger size of the heme pocket and the altered orientation of the distal histidine residue for both lupin (30) and soybean proteins (15, 23). This greater size and flexibility also explains a further unusual property of Lb: a t slightly acidic pH values the distal His of Fe(II1) (met) Lb can move close enough to the iron atom to become a ligand (151, whereas at higher (neutral) pH values it can swing right out of the heme pocket to allow entrance of large ligands such as fatty acids, nicotinic acid, and isoquinoline to occur (30). In contrast to this larger pocket, the Lb structure shows a somewhat narrower “entrance7’with the a-carbon of the E7 (distal) His nearer to the plane of the heme ring by ca. 1.5 A compared t o hemoglobin and myoglobin, but with the imidazole ring at approximately 90” to the angle found in these other proteins (30). Early suggestions [e.g., (79)1that Lb might act to facilitate nitrogen fixation by acting as an ancillary nitrogen carrier have been shown to be unlikely. Thus, although it has been reported that N, complexes of both Fe(II1) and Fe(I1) Lb can form, these only exist to any significant extent when nitrogen gas is employed at high pressure. Thus, 1.2 at-
506
DAVIES, MATHIEU, AND PUPPO
mospheres of Nz were required to give half formation of the Fe(II1) adduct and 1.7 atmospheres for half formation of the Fe(I1) form (80). These species are unlikely to be of biological significance (37).
B. OTHERMOLECULES It has been established in a number of studies that Lb(I1) behaves in a somewhat similar manner to myoglobin and hemoglobin with respect to its binding of a number of small inorganic and organic ligands (37). Thus, it is known that Lb can form stable complexes with toxic groups such as azide, cyanide, alkylisocyanides, and fluoride in a manner similar to other heme proteins (81). The kinetics of many of these binding reactions have been examined using fast laser pulses to monitor reactions after photodissociation of the complexes, and these values have been compared with those for other heme proteins (55).In many cases the pattern of high on and slow off rate constants seen with diatomic molecules is mirrored with these organic ligands [see, for example, (8111. The formation of the fluoro species has been suggested (37) to be of use in the detection of Fe(II1) Lb in nodule slices or extracts. When used in large excess [concentrations up t o 0.1 M have been employed (2, 8211, the fluoride will displace other ligands and give rise to a characteristic (and diagnostic) W-vis absorption band at 608-610 nm (83). This complex is in a pure highspin state, which has proved useful in the detection of Lb by use of EPR spectroscopy (37). The EPR parameters of a number of other Lb-ligand complexes have been studied in some detail (68). However, unlike most mammalian heme proteins, Lb(II1) is also known to readily form complexes with a number of small carboxylic acids including acetate, propionate, butyrate, and valerate (84). The reaction with acetate has been shown to occur in a proton-dependent reaction, pK, 4.8 (84), to form a high-spin complex (85).This complex is only slowly reduced by powerful reductants such as sodium dithionite (83). The pK, coincides with that of acetic acid, and it was suggested (84)that undissociated acetic acid binds to the heme iron. However, it has also been suggested that this the pK, may arise from another ionizable group on the protein and that this group may need to be protonated before the acetate ion can ligate to the iron ion (37). These observations and those with other ligands (discussed later), form the basis of an “electrostatic gate” model of binding of ligands to Lb at high pH values (86, 871, with electronic interactions between anionic ligands and the ionized residue(s) on the protein restricting access (23). The exact nature of the residues that give rise to this gate
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
507
and its significance in viva requires further study. The early observation (84)that larger carboxylates bind to Lb(II1) almost as readily as acetate led to the suggestion that the heme pocket of Lb must be larger and more open than that of myoglobin, in which little or no binding is observed. This suggestion was subsequently borne out by X-ray crystallographic studies (23, 29). The pocket is large enough to accommodate ligands such as imidazoles (though these appear to be somewhat different complexes from those seen with myoglobin and hemoglobin), nicotinic acid (pyridine3-carboxylic acid) and various derivatives (e.g., the 5-bromo and 5fluoro species), pyridine and some substituted species, a wide variety of amines, including some long-alkyl-chain species, and isoquinoline (37, 88-90). The binding of nicotinic acid to Lb has been the subject of several studies as a result of the observation that this ligand is often bound to Lb extracted from root nodules (37, 88, 89, 92). It is known that a number of nicotinic acid derivatives will not react with Fe(II1) Lb, suggesting that the ligand makes a number of close contacts with the surrounding globin protein. Thus, amide and N-methyl derivatives of nicotinic acid do not readily bind. The binding of nicotinic acid, like that of acetate, is pH dependent, pK, 4.9,and the dissociation constants for the Fe(I1) and Fe(II1) complexes are 33 p M and 1.3 pM respectively (89, 90). The pH dependence and lack of reaction with the N-methyl derivative have been interpreted in terms of reaction of the un-ionized ring nitrogen with the heme iron and interaction of the ionized carboxyl group with a protonated residue on the apoprotein. These predictions have been borne out by the crystal structure of the soybean Lb-nicotinate complex, with the species interacting with the carboxyl substituent being either the imidazole nitrogen of the distal histidine or possibly the phenolic -OH group of Tyr30 (23).
V. Oxidation of Fe(ll) Leghemoglobin
A number of species that are known to be or may be present in root nodules can oxidize Fe(1I) Lb. These species include nitrite (92), superoxide (93), and peroxides (94). The process would be expected to be relatively facile, as the reduction potential for the Fe(II1) LbFe(I1) Lb couple has been reported to be +0.22 V a t pH 7 and +0.27 V a t pH 6 (95). Even in cases in which considerable care has been taken during the extraction process, most extracts from root nodules contain a t least traces of the Fe(II1) form of the protein. Whether this arises
508
DAVIES, MATHIEU, AND PUPPO
during the extraction process or is present in the intact root nodule has been the subject of some debate [e.g., (12, 3713. Detection of the Fe(II1) form of the protein by optical spectroscopy in intact root nodules is fraught with difficulty, due to the light scattering produced by the nodules and an incomplete knowledge of the different ligand species that might be present and their respective contributions. Early studies suggested that Fe(II1) Lb could not be detected in soybean root nodules (96). However, in a review article (97) Klucas and Becana have recently reported the results of experiments carried out on old intact nodules attached to roots and maintained under an atmosphere of either air or 100% oxygen, which have suggested that low concentrations of Fe(II1) Lb can be present. These nodules were exposed t o fluoride (presumably at high concentrations), and the absorbance at the 625-nm parent peak and at 610 nm for the fluoride complex monitored. The low initial absorbance at the former wavelength is rapidly lost, and an increase in absorbance at 610 nm is observed. These changes have been ascribed to the formation of the Fe(II1) fluoride complex (97). A similar treatment with nicotinate also resulted in a decrease in absorbance at 625 nm and an increase in absorbance at 557 and 526 nm, which correspond to the absorption maxima of the Fe(II1) nicotinate complex. Both experiments have been interpreted in terms of the prior presence of low levels of Fe(II1) Lb in the intact nodules (97). However, it is impossible to rule out the possibility that the high concentrations of these two ions, which have a high affinity for Fe(II1) Lb, are enhancing formation of this species through alteration of the redox potential. In contrast, recent EPR studies on intact nodules from soybean plants of different ages that were rapidly frozen immediately after harvesting provided no evidence for high-spin complexes of Fe(II1) Lb, which have very distinct g values of ca. 6 and 2 (74). Whether low-spin forms of the Fe(II1) protein were present was more difficult to discern due to the relatively weak nature of these features and the presence of strong absorption bands from the NO complex of Lb(I1) (also discussed later). Thus, it is still not unequivocally established whether Fe(II1) species of Lb are present at any significant levels in young and mature nodules. The situation with senescent nodules is somewhat different, and it is likely that intact Fe(II1) Lb and/or partially degraded Fe(II1) Lb are present (74, 97). Reaction of Fe(I1) Lb with superoxide radicals has been suggested as an important route to the formation of Fe(II1) Lb via Eq. (1)(93). The precise nature of the reaction remains to be fully established; it has been reported that the process is inhibited in uitro by the presence of the enzyme superoxide dismutase (SOD), which removes the super-
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
509
oxide radical [Eq. (211 (93). It should be noted, however, that 02.also appears to be able to act as a reductant for the Fe(II1) Lb, reducing it to the Fe(I1) form [Eq. (311, and that the process is ameliorated by the presence of SOD (93).The latter reaction [Fe(III) to Fe(1I)I appears to be considerably slower than the former. The significance of these 0,. dependent processes will depend on the form of the protein present initially. Current evidence is that the major form in functional nodules is the Fe(I1) form, and it is likely that the major effect of 02.generation in uiuo will be oxidative [generating inactive Fe(II1) Lbl rather than reductive (which might be protective). '
202.
+ 2H' 2H,O,
-t O2
Fe(II1)+ 0 2 ,-+Fe(I1)+ 0,.
(2)
(31
The LbO, adduct is known to undergo a slow autoxidation to the Fe(II1) form with release of O,.- (93). This process is usually slow, but the rate may be enhanced under certain conditions (e.g., pH, temperature, concentration of Lb, and the presence of certain metal ions, chelators, and anions) (93, 97). Thus, the high levels of Lb02 present in intact nodules and the drop in pH observed in old and senescent nodules [pH values of ca. 6.4 have been reported for young nodules, and ca. 5.5 in old or stressed species (98)] would be expected to enhance formation of the Fe(II1) form, with release of O,--. Incubation of soybean bacteroid preparations with nitrite (0.4 mM) has been shown to markedly inhibit ethylene reduction and concomitantly be capable of inducing the oxidation of LbOz to Fe(II1) Lb (92). The inhibition of nitrogen fixation in these bacteroids has been ascribed to both a direct inactivation of the nitrogenase and the effect of nitrite in converting LbOz into an inactive form (92). In contrast, direct evidence for (uncatalyzed) autoxidation processes in uiuo is lacking because, despite numerous efforts, significant amounts of the Fe(III)Lb have not been detected in intact root nodules. This is probably due to efficient reduction of the Fe(II1) form to active Fe(I1) Lb. In contrast t o the autoxidation of LbOz, LbNO is resistant to both autoxidation and oxidation by added oxidants such as [Fe(CN1613(68-70); it cannot therefore be readily converted back to the oxygenbinding form of Lb even by oxidationheduction cycles.
510
DAVIES, MATHIEU, AND PUPPO
Reaction of deoxy Fe(I1) Lb with small excesses of Hz02results in rapid changes in the UV-vis spectrum of the heme center, with the appearance of new absorption bands a t 416 nm ( E 97.2 mM-'. cm-') and 543 nm ( E 10.4 mM cm-'1, with a shoulder a t 575 nm (94). The second-order rate constant for this reaction is 2.24 X lo4M-' s-' (94). These changes have been ascribed to the oxidation of the Fe(I1) center by the peroxide to give a water molecule and a species that has optical properties similar to Compound 11, the Fe(IV)=O species in peroxidases:
-
Fe(I1) + H,O,
+ Fe(IV)=O
+ HzO.
(5)
Approximately 1.5 equivalents of hydrogen peroxide are required per mole of heme to obtain maximum yields of this species (94). This species is remarkably stable, with little decay over 12 h at 25°C in the absence of reducing agents (94). The corresponding species formed in the case of myoglobin and hemoglobin decay much more rapidly. The number of oxidizing equivalents present in this species has been examined by titration with reducing agents such as sodium dithionite. Two equivalents of the latter are required per mole of oxidized Lb, whereas only one equivalent was required to reduce Fe(II1) Lb (94). When similar reactions were carried out with larger excesses of HzOz, a more rapid conversion of Fe(I1) to F e W ) was observed. From the absorbance changes at 543 nm due to the Fe(1V) species, the yield is somewhat reduced with excess hydrogen peroxide above 1 :4 (Lb :H20z).Under such conditions the Fe(1V) absorbance is gradually lost due to slow reactions resulting in destruction of the heme chromophore (94). Such a n interpretation is in line with studies carried out using chelating ligands with a high affinity for iron. In these studies no significant release of iron was observed until a greater than 4 : 1 excess of hydrogen peroxide was employed (99). It should be noted that the release of iron by this method may not accurately reflect the release of iron in other situations, as such high-affinity chelating ligands may remove iron from the protein, where it would otherwise have remained bound. Thus, the assessment of heme degradation by W - v i s absorbance measurements, which reflect early stages of heme degradation, do not necessarily correlate with the release of iron as measured by other methods. When similar reactions were carried out with LbO,, the absorbance changes were much slower. Early studies with mammalian heme proteins suggested that such reactions might give rise to free hydroxyl radicals via one-electron reduction of the hydrogen peroxide ( l o o ) ,but
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
511
it is now believed that this reaction proceeds via the formation of a high-oxidation-state species similar to that described earlier, a t least when low excess peroxide levels are employed. No evidence for “free” hydroxyl radical formation has been obtained under these conditions by product analysis (99) or EPR spin trapping (101).When higher concentrations of hydrogen peroxide were employed, which can result in heme degradation with release of iron [typically a t peroxide : heme ratios > 6 : 1; LbO, appears to be less sensitive than the Fe(I1) form], some hydroxyl radicals may then be generated by the iron released (99).With low concentrations of peroxide and intact LbOz, the reaction presumably involves displacement of the dioxygen by the hydrogen peroxide and reaction as described earlier. Such changes are readily monitored using spectrophotometry by the absorption decrease a t 574 nm or the increase a t 625 nm. As with the species formed from the deoxy form of the protein, this species is very stable in the absence of reducing agents. Oxidation of ligated forms of Lb can also occur; thus, it has been shown that the Lb(I1) nicotinate can be oxidized to the Lb(II1) form by hydrogen peroxide. Further oxidation to higher oxidation states does not occur, however (94).The nature of this oxidation, and in particular the fate of the second oxidizing equivalent from the hydrogen peroxide, remains to be established.
VI. Oxidation of Fe(lll) Leghemoglobin
Reaction of H202with Fe(1II) Lb gives lower yields of Fe(IV)= 0 Lb than those seen with Fe(I1) (94, 102). The absorbance changes observed during reaction with low excesses of the peroxide were ascribed to the formation of both the Fe(IV)=O species and an additional (green) compound (102).The latter is characterized by UV-vis absorbance bands a t 410 and 630 nm (Fig. 3). This species has since been shown to be very stable and has been further isolated and characterized (see Section IX,A) (103).The nicotinate complex of Fe(II1) Lb does not undergo these reactions (94). The formation of Fe(IV)=O Lb suggests that the Fe(II1) center is oxidized by only one of the two oxidizing equivalents from the peroxide. The fate of the second equivalent has been the subject of much study. In early work, in which the ability of Fe(II1) Lb/HzOamixtures to initiate lipid peroxidation and oxidation of organic molecules was examined, it was postulated that the second oxidizing equivalent was released as a hydroxyl radical:
512
DAVIES, MATHIEU, AND PUPPO
Wavelength
(nm)
FIG.3. Formation of the additional (green) compound on reaction of soybean Fe(II1) Lb (50 pM)with HzOz(100 pM) in 25 mM KH2P04/KOHbuffer, pH 7.4, containing 0.1 mM DTPA. Visible range spectrum recorded 120 s after addition of HzOz a t 25°C. Fe(II1) + H202+ Fe(IV) = 0 + HO..
(6)
Further studies on this system and related work with the analogous mammalian heme proteins myoglobin and hemoglobin have shown that free hydroxyl radicals are not released, at least with low excesses of hydrogen peroxide, when oxidative damage to the heme and release of iron is minimal (101). These reactions therefore appear to occur in a similar manner to other peroxidase enzymes in which the initial oxidation by the peroxide involves a two-electron transfer from the heme iron to the peroxide. However, unlike peroxidases, the second oxidizing equivalent does not appear to be retained by the heme center for any significant length of time, and the second equivalent is rapidly dissipated into the surrounding protein (globin) matrix [Eq. (7); (101)l.It appears therefore that the initial reaction gives an Fe(IV)=O species and a heme (porphyrin) radical cation. The radical cation, unlike those present in classical peroxidases, reacts with the surrounding protein by what appears to be an electron-transfer process to oxidize one or more globin residues with concomitant reformation of the intact heme ring (101).Thus, within a very short period the reaction gives an Fe(IV)=0 center identical t o that formed with Fe(I1) Lb, and one or more globin radicals. This situation is similar to what occurs with myoglobin (104-110): Fe(1V)=0 (heme+.)(globin)+ Fe(IV)=0 (hemeKglobin' .).
(7)
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
5 13
Though there is now some considerable evidence that globin radicals are formed in these reactions, it is still not absolutely clear where the oxidizing equivalent resides and why. In initial (direct, room temperature) EPR studies a multiplet signal was detected on reaction of the Fe(II1) form of the protein with low excesses of hydrogen peroxide (Fig. 4) ( 1 0 1 ) .This signal was only detected shortly after initiation of the reaction. Monitoring of the intensity of the EPR signal by carrying out time-course experiments in which the magnetic field was set to that corresponding to the position of one of the absorption lines, allowed the half-life of the species to be determined as approximately 40 seconds (Fig. 5). This globin-derived species was observed with all three types of Fe(II1) Lbs tested (a, e l , and c3) and was also detected with a number of other hydroperoxides and two-electron oxidants. The former observation suggests that the minor changes in sequence and structure between the isoforms are unimportant in determining the course of the reaction; the latter observation suggests that a common mechanism occurs, with similar globin-derived radicals generated in all cases. The latter strongly suggests that the globin radical does not contain any part of the original oxidant (e.g., is not an adduct formed by addition of the hydroperoxide to an amino acid). In further experiments in which the Lb was iodinated before treatment with hydrogen peroxide (IOI),the EPR signal was not observed. The condi-
FIG. 4. EPR spectrum observed immediately after mixing soybean FeUII) Lb (250 pM) at pH 7.4. Reaction studied using a two-way stopped-flow mixing system inserted into the cavity of the EPR spectrometer. The signal is assigned to a sterically constrained tyrosine phenoxyl radical formed at position 133 (reproduced with permission from Davies, M. J.; Puppo, A. Biochem. J . 1992,281,197-201).
pM) with HaOz (250
514
DAVIES, MATHIEU, AND PUPPO
Time after mixing (s)
FIG.5. Build-up and decay with time of the EPR signal (see Fig. 4) from the tyrosine phenoxyl radical formed a t position 133 of the soybean protein on reaction of Fe(II1) Lb (250 pM) with H202 (250 p M ) a t pH 7.4 (reproduced with permission from Davies, M. J.; Puppo, A. Biochern. J. 1992,281, 197-201).
tions under which this iodination was carried out were such that only the Tyr residues in the protein ought to be affected, suggesting that the Tyr residues present in the globin are either the site of the radical species or involved in the generation of the radical. Determination of the g value of the observed signal (2.0044) and analysis of the hyperfine coupling constants suggest that the signal is a Tyr-derived species, with both the g value and the coupling pattern consistent with the presence of a Tyr phenoxyl radical (101). This species does not, however, have coupling constants identical to those of a free Tyr phenoxyl radical [see, for example, ( I l l ) ] , particularly with respect to the coupling of the hydrogens of the methylene (-CH,-) group, which attaches the aromatic ring to the backbone. The variation between these values is consistent with the surrounding globin structure forcing the radical to adopt a fixed conformation with little free rotation round the methylene-to-ring bond (112). This species also has somewhat different coupling constants with regard to the methylene hydrogens than those reported for analogous Tyr phenoxyl radicals in other proteins (113-1241, suggesting that the conformation of such radicals is not universal in proteins. In the soybean form of the protein, which was used in these experiments, there are three Tyr residues, with one of these, that at position 133, near the heme ring (Fig. 6). This was suggested to be the site of the radical on account of its proximity to the heme center at which the initial oxidation must be occurring (101). Further studies with a variety of scavenging agents demonstrated that the radical can undergo a number of reactions. Thus, its signal was lost when high per-
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
515
FKi. 6 . Position of the Tyr133 residue in soybean metLb in relation to the heme ring. Computer-generated molecular model using crystal coordinates ( 2 3 ) (Ellis and Freeman, unpublished data).
oxide concentrations were employed, a process assigned to reaction between the radical and excess hydrogen peroxide. The signal was also diminished in intensity, or lost completely, when various hemebinding agents (cyanide, azide, nicotinic acid) were used or reducing agents were added (see Section IX,B). The latter was attributed to one-electron reduction of the radical (102).The observation that the globin-derived radical reacts with agents expected to be present solely in the aqueous phase, and not expected to penetrate far into the protein structure, suggests that the radical must be present on the protein surface, or not far removed from it.
516
DAVIES, MATHIEU, AND PUPPO
Further studies with one of the lupin forms of Lb have helped elucidate the site of the radical (112). This protein contains only two Tyr residues, with only one of them conserved when compared to the soybean structure. Thus, it was hypothesized that if the radical had identical parameters and similar kinetics for formation and decay, then it could be assigned to the single conserved residue (Tyr133 in soybean and Tyr138 in lupin). This proved to be the case, with a very similar EPR signal observed with the Fe(II1) form of the lupin protein on treatment with small excesses of hydrogen peroxide (112).The signals obtained in experiments with the lupin protein were better resolved than those with the soybean protein, and this allowed a detailed analysis of the splitting pattern and coupling constants to be carried out. The spectra were computer simulated using this data, adding further support to the theory that the globin radical is indeed a Tyr-derived phenoxyl radical (112). The mechanism of formation has been postulated to involve rapid one-electron oxidation of the aromatic ring by the heme radical cation to give a Tyr radical cation [Eq. (8); (101, 112)l. Using model phenols such species undergo rapid deprotonation to give a phenoxyl radical with rate constants ca. 1O’O dm3.mol-l- 5-l (125, 126). Thus, these model studies are consistent with the formation of a radical within the globin polypeptide: Fe(1V)=0 (heme’.)(globin-Tyr-OH)4 Fe(IV)=O (heme)(globin-Tyr-OH”) + Fe(IV)= 0 (heme)(globin-Tyr-0)+ H i .
(8)
In the experiments reported in these studies, little attempt was made to quantify the extent of formation of the radical. Such studies are technically challenging, especially when the kinetics of formation of the radical are poorly defined. However, from the intensity of the signals the steady-state concentration of the radical is considerably less than that expected from the concentrations of protein and oxidant employed. Thus, it appears that the radical is not formed in a very efficient manner. This suggests that there are other routes to the loss of the second oxidizing equivalent from the heme center or there are other pathways that result in the loss of the Tyr radical. In an effort t o determine whether other radical species are generated on the globin, EPR spin-trapping experiments were carried out using the spin trap DMPO (5,5-dimethyl-l-pyrroline-N-oxide). In these experiments, unlike those carried out with myoglobin, no significant signals were observed (101).At the time the interpretation
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
517
was in terms of nonformation of additional radicals, but later work has shown this to be incorrect. The incorrect conclusion probably arose from the fact that the spin trap chosen was not ideal, with the trapping rate constants rather slow (i.e., other species formed were not trapped and hence not observed). These negative experiments did, however, allow one possible mechanism of formation of the observed radical to be ruled out. The trap reacts rapidly with HO. (K ca. 2 x 1 0 9 M - I. s - 11 to give a spin adduct with distinctive EPR parameters and a long half-life (127, 128). The absence of this wellcharacterized product suggests that free HO. are not formed (101). Similarly, the absence of alkoxyl radicals (and related species), which also give long-lived and well-characterized spin adducts, in experiments in which hydroperoxides (and similar compounds) were used to generate the radicals, rules out the involvement of these radicals (101). Subsequent studies using alternative spin traps have demonstrated that other radical species are generated on the globin when Fe(II1) Lb is incubated with low excesses of hydrogen peroxide (129). Thus, incubation in the presence of the spin trap MNP (2-methyl-2-nitrosopropane) has resulted in the detection of broad, partially anisotropic EPR signals. The partial anisotropy of these signals is believed to be due to the trapping of relatively large, slowly tumbling radical adducts. These are believed to be trapped globin-derived radicals (129). Similar experiments carried out with other spin traps, such as the nitrones PBN (a-phenyl-N-t-butylnitrone) and POBN [a-(pyridine-1oxide)-N-t-butylnitrone], also give EPR signals that have been assigned to globin-derived radicals (129).The formation of these species does not appear to be affected by the presence of dioxygen in the reaction mixture. This suggests that the radicals are not formed from an oxygen-derived species (such as a peroxyl radical) and that they do not react rapidly with oxygen. There are, however, two important caveats to these suggestions. First, it is very difficult to make such reactions completely anoxic, as oxygen may be generated from the decomposition of hydrogen peroxide. Second, the high concentrations of spin traps used in these experiments (to obtain reasonable signal-to-noise ratios) may result in reaction with dioxygen being uncompetitive. In only one case, with the nitroso spin trap DBNBS (3,5-dibromo-4-nitrosobenzene sulfonic acid), were signals observed in the absence of hydrogen peroxide. In the latter case the signals have been assigned to a novel compound arising from a chemical reaction of the heme moiety with the trap and not spin trapping. Subsequent oxidation of the initial product gives the observed signal (129).
5 18
DAVIES, MATHIEU, AND PUPPO
In the case of POBN at least two adducts were detected, and the EPR signals were almost completely isotropic in nature, in contrast with those detected with MNP and PBN. This suggests that the radicals present have considerable mobility (129). The trapping with POBN, which is a highly hydrophilic trap and partitions very poorly into hydrophobic solvents, suggests that the radicals detected are present on the outer surface of the globin. The situation with the other traps is less clear, as PBN and MNP partition readily into hydrophobic media. Unfortunately, the spectra obtained do not give extensive information as to the nature of the radicals formed. Thus, the spectra observed with PBN and MNP give little information due to their anisotropic nature, whereas the sharper lined spectra observed with POBN do not allow much information to be obtained because the adducts formed do not show a dramatic change in hyperfine coupling constants (128). However, the signals observed with MNP suggest that the added radical is carbon-centered in nature. To obtain further information about the species trapped in these experiments, enzymatic digestion experiments were performed on the spin adducts (129). This technique, which has been successfully employed to obtain information on the nature of radicals trapped from other proteins, DNA, RNA, and complex carbohydrates (130, 1311, relies on the fact that many of the trapped species are stable for considerable periods of time. The macromolecular radical adduct can therefore be treated with either nonspecific o r specific proteases that will release smaller fragments, at least some of which should still have the spin trap attached. This process may be aided by the fact that many proteolytic enzymes degrade damaged proteins more readily than the corresponding undamaged parent macromolecule (126, 132). Such fragments would be expected to tumble rapidly in solution and hence give isotropic spectra, from which further information may be obtained. When such experiments were carried out with the Lb adducts detected on treatment of Fe(II1) Lb with hydrogen peroxide in the presence of MNP, a conversion of the anisotropic signal to more isotropic features was observed. The spectra observed at the end of this process consisted of a single triplet, which is consistent with the trapping of at least one tertiary carbon-centered radical (129).Though these experiments do not allow complete identification of the protein radical, they suggest that the radicals are formed at stabilized sites k e . , tertiary rather than secondary or primary). Recent studies on the analogous myoglobin reactions have suggested that a highly resonance-stabilized radical that is trapped via the C-3 position on the indole ring (108, 1101, can be formed on the
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
5 19
side chain of Trp residues. This is a tertiary site and would be expected to be formed readily due to the ease of oxidation of such Trp side chains. Thus, it is possible that a similar species may form and become trapped in the Lb experiments described earlier, and that such a radical may be the source of the tertiary nitroxide signals seen with MNP (and possibly the signal observed with PBN and one of those detected with POBN). The detection of two distinct species with POBN as the trap suggests that, even if this speculation is correct, other types of radical must also be present. Recent studies have shown that phenoxyl species can be trapped under certain circumstances, and in particular when other decay routes (such as dimerization) are inhibited by, for example, steric factors (133).Thus, some, but not all, of these spin-adduct signals may arise from the trapping of Tyr-derived species. The additional radical species may arise via two (or more) competing damage-transfer pathways within the protein or may be species formed along a single pathway. Evidence from the analogous myoglobin experiments suggests that more than one pathway exists, with both tyrosine and tryptophan-derived species observed (104-110).Recent studies have also implicated histidinederived radicals (log),though it should be noted that some of this evidence has been obtained from the study of mutant proteins (from site-directed mutagenesis) from which key amino acids (e.g., the Tyr residues) have been removed. Such data might be misleading, for removal of a key residue may merely switch damage transfer to another pathway and hence other final sites that are not normally damaged to any great extent. Considerable work therefore still needs to be carried out in order to obtain a complete picture of the radical species that are formed in these reactions.
VII. Reduction of Fe(lV)=O Leghemoglobin
As mentioned earlier, Fe(IV)=O Lb is particularly stable, as least when compared to the analogous species detected with other heme protein systems (94). This species remains observable 12 h after its formation under favorable conditions and can be isolated from other forms of the protein on isoelectric focusing gels. However, in the presence of reducing agents or readily oxidizable compounds it undergoes further reaction, resulting in its conversion to a lower oxidation state. Some of these reactions are biologically significant, as they repair the oxidized species.
520
DAVIES, MATHIEU, AND PUPPO
A. REACTION WITH H,O, Reaction with excess hydrogen peroxide results in the oxidation of the heme ring and release of iron, though the exact mechanism of this type of reaction is unknown. It may involve oxidation of a further molecule of hydrogen peroxide (101) with consequent formation of the HOO. radical, which may then react with the heme ring, possibly at the methylene bridge sites. The latter is consistent with known degradation reactions of the heme ring with loss of the iron atom (99).
B. REACTION WITH ASCORBATE It has been reported recently that addition of ascorbate to Fe(1II) Lb immediately prior to the addition of small excesses of hydrogen peroxide results in transient formation of Fe(IV)=O Lb as monitored by UV-vis spectrophotometry (134).The situation when ascorbate is added after formation of Fe(IV)=O Lb is more complex due to the formation of additional species during the reaction of Fe(II1) Lb with the hydrogen peroxide. In the former case, with 1 mM ascorbate and 50 p M Lb, the conversion back to Fe(II1) is complete in about 2.5 minutes (Fig. 7) (134).Under these conditions a further slow reduction of the Fe(II1) species is also observed (discussed later). The observation of isobestic points is consistent with the presence of only the Fe(1V)= 0 and Fe(II1) species in the reaction system. As expected, the rate of the Fe(IV)=O to Fe(II1) conversion was found to depend on the ascorbate concentration and is associated with oxidation of the ascorbate. The stoichiometry of the ascorbate oxidized to hydrogen peroxide added is 1: 1. This ratio implies that both oxidizing equivalents in the peroxide are responsible for ascorbate oxidation and hence that the protein-derived radicals are also removed. The reduction of the Fe(IV)=O form by ascorbate has been shown to be a oneelectron transfer reaction by the use of EPR spectroscopy. In these experiments intense signals due to the ascorbyl radical were observed (134).This process may occur via reaction of ascorbate at the heme edge, which is known to be exposed to agents present in bulk solution (Fig. 8). Similar results were obtained when Fe(IV)=O Lb was generated from Fe(I1) Lb, though in this case both oxidizing equivalents from the hydrogen peroxide are believed to remain at the heme center before being removed by reaction with ascorbate.
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
52 1
1 0.9
0.6
0.7
8
9
0.6
0.5
0
$
0.4
4
0.3
0.2 0.1
0.9 0.8 0.7
8
0.6
:
0.5 0.4
4 0.3 0.i
0.: 1
Wavelength
(nm)
FIG.7. Spectrophotometric changes during the reduction of soybean Fe(IV)=O Lb by ascorbate. Fe(1II) Lb (50 p M ) was mixed with H L 0 2 (100 /*MI immediately after addition of 1 mM ascorbate a t pH 7.4 ( 2 5 mM KH,PO,/KOH buffer containing 0.1 mM DTPA). Reactions were run a t 25"C, and repetitive scans were recorded a t 30-s intervals ( a ) and then 30-min intervals (b). Data are representative of experiments carried out in triplicate [reprinted from Phytochemistq, 39, Moreau, S.; Puppo, A.; Davies, M. J.; The reactivity of ascorbate with different redox states of leghaemoglobin, pp. 1281-1286, copyright (19951, with kind permission of Elsevier Science Ltd, The Boulevard, Langford Lane, Kdlington, OX5 IGB, UKI.
c. REACTION WITH GLUTATHIONE, O T H E R THIOIS,AND RELATED SPECIES As in the case of ascorbate reactions, addition of glutathione (GSH; 1 mM) to Fe(II1) Lb (50 pM)before reaction with hydrogen peroxide results in a decreased lifetime of the Fe(IV)=O species (102).The loss
522
DAVIES, MATHIEU, AND PUPPO
Frc:. 8. The heme edge (dark shaded residues) in soybean Fe(II1) Lb is partially exposed on the surface of the protein. Computer generated molecular model using crystal coordinates (23)(Ellis and Freeman, unpublished data).
of the UV-vis absorption bands of Fe(IV)=O Lb is accompanied by the reappearance of those for Fe(II1) Lb, with good isobestic points confirming that no other species is involved (Fig. 9). The reduction to Fe(II1) Lb is accompanied by formation of oxidized glutathione [glutathione disulfide (GSSG), assayed using NADPH in the presence of GSSG reductase] in approximately stoichiometric amounts (H,O, added :GSH formed), again suggesting that both oxidizing equivalents from the hydrogen peroxide are utilized. The conversion of GSH into GSSG is also believed to occur via a one-electron (radical) process, because glutathione thiyl radicals have been detected by EPR spectroscopy. Similar behavior is observed with Fe(I1) Lb treated with small excesses of hydrogen peroxide (102). As might be predicted on the basis of these data, a number of other thiol compounds have been shown to reduce Fe(IV)=O Lb to the Fe(II1) state with concomitant formation of thiyl radicals (detected by use of EPR spin trapping using 5,5-dimethyl-l-pyrrolineN-oxide) (135). In some cases, however, other species are formed, and these have been identified from their UV-vis absorption spectra as novel sulfur species formed by nucleophilic attack on the tetrapyrrole ring by the thiol group (135).The ability of a thiol t o undergo two such competing reactions is dictated by steric and electronic characteristics
523
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS 0.60
I
0)
C
0.30
s
9
n 450
750
600
Wavelength
(nm)
FIG.9. Visible range spectral changes during the reduction of soybean Fe(IV)=O Lb by GSH. Fe(II1) Lb (50 pM) reacted with HzOz (100p M ) at pH 7.4 (25 mM KH,PO,/ KOH buffer containing 0.1 mM DTPA). Repetitive scans were recorded 1 min after addition of H202and then a t 3-min intervals (reproduced with permission from Puppo, A.; Monny, C.; Davies, M. J. Biochem. J . 1993,289, 435-438).
of the thiol. Addition of cysteine results in reduction to the Fe(II1) [and subsequently to Fe(I1) Lbl and formation of the Fe(II1) and Fe(I1) sulfur derivatives via nucleophilic reactions. The Fe(I1) sulfur derivative appears in this system to be formed from the corresponding Fe(II1) sulfur compound, suggesting that nucleophilic addition can be followed by heme reduction. The difference in behavior between cysteine and GSH may arise from the greater steric bulk of the latter, which may inhibit reaction a t the heme edge. This hypothesis is supported by the observation that the behavior of N-acetylcysteine is intermediate between the two. Cysteamine (2-mercaptoethylamine) gives extremely rapid heme reduction with formation of Fe(I1) Lb and little formation of the sulfur derivatives of Lb. 3-Mercaptopropionic acid on the other hand gives only slow reduction of Fe(IV)=O Lb, even slower reduction to Fe(I1) Lb, and some sulfur-Lb formation. This suggests that the overall charge on the reducing agent can have a major influence on the kinetics and the extent of competition between the two pathways (135). These differences have been rationalized in terms of the negatively charged gate in the globin, which gives access to the heme. Electronic interactions between the negative charge on the protein and the reductant may slow down the reduction with 3-mercaptopropionic acid and give an enhanced rate with positively charged cysteamine. In contrast the formation of sulfur-Lb derivatives always appears to be rel-
524
DAWES, MATHIEU, AND PUPPO
atively slow and is only therefore a significant competing reaction when reduction at the heme center is slow. The exposed nature of the heme edge in Lb (23)(Fig. 8) may also explain the much greater prevalence of nucleophilic processes as compared to myoglobin (1361, where such reactions appear to occur much less readily. Reduction of Fe(IV)=O Lb by disulfides does not occur to any significant extent, except in the case of oxidized lipoic acid, which brings about slow reduction of Fe(IV)= 0 Lb to Fe(II1) Lb (135). No radical species were detected by EPR spectroscopy during the reaction, though such species are very likely generated by one-electron oxidation of the disulfide. No reaction between Fe(1V)= 0 Lb and hydrogen sulfide has been observed, unlike the situation with myoglobin (137). D. REACTION WITH OTHERREDUCTANTS Other chemical reducing agents can bring about loss of Fe(IV)=O Lb. Thus, ferrocyanide can reduce Fe(IV)=O Lb to Fe(II1) Lb, though this reaction is relatively slow and complex [as judged by changes in the absorption spectra (94)l. In contrast, reaction with dithionite appears to be quantitative and has been used to estimate the concentration of Fe(IV)=O Lb (94).The Fe(IV)=O Lb does not undergo reaction with dioxygen, carbon monoxide, cyanide, or nicotinic acid (94). Similarly, it does not appear to undergo reaction with Fe(I1) Lb (i.e., synproportionation to give two molecules of Fe(II1) Lb does not occur) (138).The potential reaction of Fe(IV)=O Lb with lipids and membranes is discussed in Section IX,D. VIII. Reduction of Fe(ll1) Leghemoglobin
A. ENZYMATIC There is considerable evidence that functional root nodules possess efficient means of reducing Fe(II1) Lb back to Fe(I1) Lb. It has been shown that this occurs via a number of different mechanisms, including both enzymatic and nonenzymatic pathways [reviewed in (97, 139)l. Early studies by Appleby (140) showed that bacteroids slowly reduce Fe(II1) Lb to Fe(I1) Lb under anaerobic conditions. It was suggested that a n enzymatic process was involved. Later reports [e.g., (141, 142)l confirmed that a specific enzymatic Fe(II1) Lb reduction system is present in root nodules, and the nature of this enzymatic
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
525
action has been characterized in some detail (143). The protein responsible for the activity was initially characterized in lupin nodules but has since been shown to be present in other nodules (e.g., soybean). It is a FAD-containing species that also contains active thiol groups, but no catalytic metals (139, 143-147). The enzyme has been purified in a number of cases. There are significant differences in size and other properties of the enzyme from different species (e.g., the molecular mass varies from 60 kDa for the lupin form to 100 kDa for the soybean). In lupin nodules there are positive correlations between activity of the enzyme and symbiotic performance as measured by the rate of nitrogen fixation and the Lb content (148).The protein from soybean root nodules is present primarily in the plant fraction of the nodules (91% of the enzymatic activity), with only low levels present in the bacteroids (143). The isolated protein appears to reduce all eight forms of soybean Lb equally effectively. Several different forms of this species have been identified in soybean nodules, suggesting that there are a number of modified or different forms of the protein present (149). The activity of this enzyme requires a reductant (NADH or NADPH) and dioxygen but does not require an intermediate electron carrier. Activity is decreased if the Fe(II1) Lb is complexed with a number of small ligands such as acetate, nicotinate, and nitrate, and is also inhibited by the presence of exogenous catalase, but not by superoxide dismutase. The latter observation suggests that a peroxide-type intermediate may be formed during the catalytic cycle (97, 139). The rate of reduction does not appear to depend on the species that finally ligates Fe(I1) Lb; thus, identical rates of reduction are observed with O2 or CO ligated in the axial position. As with nonenzymatic reduction (discussed later) the enzymatic process is pH dependent. Rate constants are constant over the pH range 7.6 to 6.5, but increase at lower pH values, with a threefold increase at pH 5 . 2 (143).Further studies and partial sequencing of the enzyme have shown that this protein is similar to myoglobin reductase, and the NADH : cytochrome b5 reductase from erythrocytes (147).
B. NONENZYMATIC There are a number of low-molecular-weight species present in root nodules that might act as reductants for Fe(1II) Lb. In mature root nodules there are significant levels of NADH and NADPH (150-200 pM), free cysteine (200 pM), reduced glutathione (40-150 pM) and reduced ascorbate (1-2 mM). The exact concentrations depend on the age of the nodules and the method by which the extraction and quan-
526
DAVIES, MATHIEU, AND PUPPO
tification is carried out (97, 139, 150). The reduction catalyzed by a number of these species is pH dependent. Rate constants for the reduction of Fe(II1) Lb by high concentrations (1 mM) of NADH or NADPH were found to increase with change in pH from pH 7.0 to 5.2, with maximum rates observed at the lowest value. The reduction was O2 dependent and was inhibited by superoxide dismutase and catalase (93, 151). These results implicate superoxide radicals andor hydrogen peroxide as the reductants. However, at higher pH and physiological levels of reductant (400 pM), the rate constants are very small (151). Furthermore, only ascorbate and cysteine of these reductants appear to be able to reduce Fe(II1) Lb a t significant rates (139). A number of nonphysiological thiols also appear to be able to bring about reduction. In a number of cases, including cysteine, competing processes generating sulfur-Lb species have been shown to occur (135).If such sulfur-Lb species could be detected in intact root nodules, it might give an indication of the occurrence and importance of cysteine-mediated reduction of Fe(II1) [and Fe(1V)= 01 Lb in uiuo. Disulfides do not appear to be capable of reducing Fe(II1) Lb to Fe(I1) Lb (135). Reaction with sulfide itself appears to be different and gives rise t o a Fe(II1) Lb-sulfide complex rather than reduction or nucleophilic attack (135). The complex is, however, susceptible to reduction by strong reducing agents such as dithionite and ascorbic acid. It has also been reported (152)that incubation of Fe(II1) Lb with a large excess of ascorbic acid (10 mM) can result in the formation of LbOz together with heme degradation products due to the generation of radicals (from autoxidation of excess ascorbate). More recent work has cast doubt on the significance of this type of reaction (134). Rate constants for many of these reduction processes have been shown to depend on a number of factors including pH, the presence of metal ions, and cofactors. Thus, it has been reported that the reduction of Fe(II1) Lb by NADH can be stimulated by addition of Mn(I1) ions (139).This is thought to be due to the formation of superoxide radicals, which then reduce the Fe(II1) center. Flavins have been suggested as stimulators of Fe(II1) Lb reduction (139).Thus, in the presence of NADH or NADPH, free flavins can act as efficient reducing agents in a process that appears to be independent of mediators such as superoxide radicals (no effect of SOD). This direct reaction may be aided by the exposed nature of the heme edge of Lb (Fig. 8) (23, 29). This type of process does not require dioxygen, and may be a significant pathway in uiuo due t o the high flavin content of root nodules (139).
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
527
IX. Reactions of Globin-Derived Radicals
A number of reactions of the globin radicals have been elucidated, and these can be readily grouped into various categories.
A. REACTIONS WITH THE HEMERING During the reaction of Fe(II1) Lb with hydrogen peroxide, the yield of Fe(1V)= 0 Lb is relatively low compared t o reaction with the Fe(I1) state. However a further species is formed, which can be detected by the UV-vis absorption bands a t 410 and 630 nm (94, 102, 103). The observation that this species is produced only under conditions in which globin radicals are formed suggested that it results from subsequent reactions of the globin radical (153).The stability of this additional (green) species has allowed it to be isolated from reaction mixtures by isoelectric focusing using immobiline gels with a 5.7 to 4.7 pH gradient. The green compound has an isoelectric point of 5.45. Electroelution has allowed this species to be further characterized, and UV-vis and fluorescence spectra have been determined (153).The visible spectrum is completely different from Fe1III) Lb in that the Soret band 1410 nm) is much less intense and broader, and the charge-transfer bands are dramatically altered, with a broad absorption a t 630 nm. The product reacts with strong reducing agents such as dithionite or ascorbate (Fig. lo), which results in the formation of a product with UV-vis bands similar to those of LbOL.This reduction is slow and occurs over approximately 18 h (153). The pyridine hemochromogen spectrum of the product provides strong evidence for alterations to the heme ring but with no loss of iron (153). The UV-vis spectrum is very different from other heme degradation products such as biliverdin. SDS-PAGE analysis gives a single band with similar R f to Fe(II1) Lb, suggesting that the protein is not a fragment or dimer. The suggestion that it contains a modified heme ring (and iron), has been confirmed by EPR spectra, which demonstrate the presence of high-spin Fe(II1) in a distorted environment. The usual iron absorption a t g = 6 was not observed, but a sharp peak a t g ca. 4.3 was detected together with a broad background signal. This may be due to an envelope of signals from completely free or partially bound Fe(II1) (153). The possibility that addition of a globin radical to the heme ring gives an altered heme moiety has been confirmed by heme extraction experiments (using acidic 2-butanone) following treatment of soybean Fe(II1) Lb with a small excess of hydrogen peroxide. Measure-
528
DAVIES, MATHIEU, AND PUPPO
Wavelength (nm) FIG.10. Visible range spectral changes during the reduction of the additional (green) species formed on reaction of Fe(I1I) Lb (50 pM) with H2O2(100 pM) at pH 7.4 (25 mM KH,PO,/KOH buffer containing 0.1 mM DTPA) a t 25"C, with subsequent addition of 1 mM ascorbate. Scans were recorded 0, 1, 2, 6, and 18 h after the addition of ascorbate; spectral changes are indicated by arrows. Data are representative of experiments carried out in triplicate [reprinted from Phytochernistry, 39, Moreau, s.;Puppo, A.; Davies, M. J.; The reactivity of ascorbate with different redox states of leghaemoglobin, pp. 1281-1286, copyright (1995), with kind permission of Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington, OX5 lGB, UK].
ment of the ratio of the absorption of Soret band to that at 280 nm was used to monitor heme removal from the holoprotein. With untreated Fe(II1) Lb, heme retention was of the order of 4.4% (i.e., nearquantitative removal), whereas with the green compound heme retention was calculated as ca. 63%. The green compound is therefore believed to be a crosslinked species arising from reaction of a globin radical with the heme (153).Similar species have been demonstrated on reaction of myoglobin with hydrogen peroxide (154-156). Though the exact nature remains to be confirmed, it has been suggested that in the case of soybean the species arises via a mechanism involving reaction of the Tyr133 phenoxyl radical with either one of the vinyl groups of the heme moiety or a methylene bridge site; further reaction of the initial adduct gives the stable green compound (153).Examination of molecular models of both the lupin and soybean forms of Lb from the crystal coordinates [(23,29); (Ellis and Freeman, personal
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
529
communication)l show that the Tyr residue (138 and 133 respectively) is aligned in such a way that the phenolate oxygen atom is in close proximity to a vinyl group and methylene bridge (Fig. 6). This suggestion therefore seems very reasonable. Proteolytic digestion of the green compound and sequence analysis of the fragments, in a manner similar to that for the corresponding myoglobin heme-protein crosslinked species (154, 155), should allow the nature of this link to be confirmed. The heme-to-globin cross-linking process seems to be an inherent feature of proteins in which a radical center is generated in the vicinity of the heme. The formation of such species has been suggested as a marker of exposure to reaction with hydrogen peroxide in uivo (155, 156). It is interesting to note that it has long been known that exposure of functional root nodules to stress (e.g., absence of light) results in the formation of “green” nodules (157). The green coloration may be a result of the formation of hydrogen peroxide and its subsequent reaction with Lb. It has been reported that at least two green pigments are formed on oxidative damage to Lb both in uitro and in uiuo (12, 67, 157, 158). One of these, called cholegobin, originates from oxidative attack on the ring without loss of iron (158), whereas biliverdins arise from heme degradation with concomitant loss of both the iron and carbon monoxide (152, 159). Pigments that have these characteristics have been partially purified from senescent pea nodules (12), but no further characterization has been reported. It has also been reported that nodules from aging gram plants (Vicer arietinum) retain their pink color longer, and do not develop a green coloration until they are much older, when they are kept in a nutrient medium containing ascorbic acid (37, 160). This result is in accord with the protective effects of ascorbic acid against oxidative damage, and with the suggestion that oxidative radical-mediated damage is of importance in the aging of nodules in uiuo. Further in uitro studies have been carried out, and it has been shown that Fe(II1) Lb obtained from common bean (but not soybean) can be reduced by glycine at alkaline pH with the formation of glyoxylate and hydrogen peroxide (159).The LbO, produced during this reaction is rapidly degraded to a green product that still contains iron and has an absorption maximum at 697 nm. Although the significance of this process in uiuo remains to be established, it has obvious parallels to the radical processes already described. It is also interesting to note that the formation of the green pigment can be stimulated by pretreatment of the Lb with a proteolytic enzyme (carboxypeptidase) and is inhibited by superoxide dismutase and catalase (159).
530
DAVIES, MATHIEU, AND PUPPO
The latter observations are in accord with the formation of the pigment from a hydrogen peroxide-mediated radical process. A similar formation of green products has been observed during the reaction of Fe(II1) Lb with ascorbate (10 mM). This process again results in the transient formation of LbOz and rapid breakdown to biliverdin-like materials (152). Analysis of the products has shown that in the case of the pea and vetch Lbs, only a single major biliverdin isomer (the b form) was detected, whereas with soybean and the common bean protein, mixtures of products were obtained [ca. 30% a isomer, 50% b, and 20% d; (152)l.The difference between these Lb forms is of interest. The lack of specificity is indicative of random damage (as might be envisaged from a “free” radical type of reaction, i.e., a low-molecular-weight rapidly diffusing molecule), whereas the highly specific damage seen with pea and vetch is more suggestive of site-specific damage induced by a bound oxidant (i.e., a radical centered on a particular residue). These suggestions relating to the type of radical (either “free” or protein bound) are in accord with the observation that the formation of green products is inhibited by superoxide dismutase and catalase (152).
B. REACTIONS WITH LOW-MOLECULAR-WEIGHT SPECIES Spectroscopic evidence has been obtained for the reaction of the Tyr phenoxyl radical derived from soybean Lb with a number of low-molecular-weight species. Thus, it has been demonstrated by EPR that the globin radicals can react with ascorbic acid, cysteine, glutathione, Trolox C (a water-soluble analog of vitamin El, thiourea, salicylate, and desferal (desferrioxamine) (101,102,134).Other compounds such as 2-deoxyribose and mannitol either do not react or react slowly. In a number of cases radicals from these reactions could be detected, for example the ascorbate radical, the cysteine and glutathione thiyl radicals, the Trolox C phenoxyl radical, and the desferal nitroxide species (101, 102, 134). In the case of glutathione, complete loss of the phenoxyl radical was not observed, whereas with an equivalent concentration of cysteine it was. This suggests that access to the radical by species present in bulk solution might be limited (102). The accessibility problem is strongly supported by molecular modeling (from the crystal coordinates) of the soybean protein, which shows that the Tyr133 phenoxyl radical is only partially exposed on the protein surface (Fig. 11) (112). It should be noted that with a number of these compounds, reaction occurs with both the Fe(IV)=O heme center and the Tyr radical. Rel-
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
531
Fir;. 11. The Tyr133 residue (dark-shaded residues) in soybean Fe(II1) Lb is only partially exposed on the protein surface. Computer-generated molecular model using the crystal coordinates obtained ( 2 3 )(Ellis and Freeman, unpublished data).
atively little is known about the reactions of the other globin-derived radicals (129).However, the stoichiometric conversion of ascorbate to dehydroascorbate (1341 and reduced glutathione to GSSG (102) with respect to the amount of hydrogen peroxide added suggests either that the other globin-derived radicals are present a t low concentrations or, more probably, also react rapidly.
C. REACTIONS WITH OTHER LEGHEMOGLOBIN MOLECULES Examination by SDS-PAGE of reaction mixtures in which Fet 111) Lb was incubated with a two-fold excess of hydrogen peroxide has provided evidence for the formation of dimeric species (Fig. 12) (103). It has been shown that the native protein migrates abnormally fast and appears a t the bottom of the gel (103, 159). However a distinct band a t ca. 32 kDa is observed, with little evidence for other species. No dimerization was observed with apolb, confirming the role of the heme group in this process. Confirmation of the dimerization process was obtained using size-exclusion HPLC on columns previously calibrated with horse myoglobin ( M , ca. 16 kDa) and carbonic anhydrase ( M , 31.5 kDa). In contrast to untreated samples, in which a single band with a retention time very similar to that observed with horse
532
DAVIES, MATHIEU, AND PUPPO
FIG.12. Dimerization of soybean Lb and its inhibition by ascorbate, glutathione, and tyrosine. Aliquots (30 p g of protein each) of incubations containing untreated 1 mM soybean Lb (lane B), or Lb incubated for 2 h in the presence of 2 mM HzOzwithout (lane C) or with 5 mM ascorbate (lane E), 5 mM glutathione (lane D) or 5 mM p-tyrosine (lane F) were analyzed by SDS-PAGE. Molecular weight markers were in lane A. The abnormally fast migration of monomeric Lb was evidenced by Western blotting with polyclonal antibodies raised against purified soybean Lb, in the absence and a blank was performed with egg (lane G) or in the presence (lane H) of Haon, lysozyme (MW 14.4 kDa) (lane I). The position of the monomeric Lb is indicated by the arrow and that of the dimeric Lb by the double arrow [reprinted from Biochin. Biophys. Actu, 1251, Moreau, S.; Davies, M. J.; Puppo, A.; Reaction of ferric leghemoglobin with HzOz: formation of heme-protein cross-links and dimeric species, pp. 17-22, copyright (1995), with kind permission from Elsevier Science-NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands].
myoglobin was observed, the treated samples (two-fold excess of hydrogen peroxide over Fe(II1) Lb) gave two bands, a major band corresponding to the parent material and a weaker band that eluted with approximately the same retention time a s carbonic anhydrase, consistent with the presence of a dimer (Fig. 13) (103).Integration of the peak areas suggests that the dimer represents approximately 6% of the initial Lb. Sufficient material could be obtained from such chromatographic separations to allow further studies to be carried out on nature of the cross-linking. Unlike the situation with sperm whale myoglobin, in which dimer formation is also observed (109, 161), no evidence was obtained for the presence of dityrosine after acid hydrolysis of the dimeric material (103).Thus, unlike myoglobin the crosslinking does not appear to occur via the dimerization of two tyrosine phenoxyl radicals. This is not unexpected because the phenoxyl radical site is only slightly exposed on the surface of the protein (Fig. 11) (112),unlike the case of sperm whale myoglobin in which the Tyrl51 residue is very accessible (161). The mechanism of formation of this dimer was further investigated by examining the effect of various putative radical scavengers on the intensity of the dimer band on SDS-PAGE gels (Fig. 12). These stud-
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
533
I li
I
0
1 10
20
30
40
t
Time (min)
FIG.13. High-pressure liquid chromatography of HzOz-treated Lb. Aliquots (100 pL) of reaction mixtures containing 1 mM soybean Fe(II1) Lb incubated for 2 h in the preswere chromatographed. The trace marked (-) shows the aromatic ence of 2 mM H202 amino acid absorbance a t 280 nm; the peak marked with a n arrow is that of the dimer. Control experiments were performed with Lb alone [reprinted from Biochim. Biophys. Actu, 1251, Moreau, S.; Davies, M. J.; Puppo, A.; Reaction of Fe(II1) leghemoglobin with H202:formation of heme-protein cross-links and dimeric species, pp. 17-22, copyright (1995),with kind permission from Elsevier Science-NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands].
ies have demonstrated that ascorbate completely inhibits dimer formation, and glutathione and para-tyrosine also significantly inhibit its generation. In contrast, ortho- and rneta-tyrosine have little effect. The reason is not immediately apparent, but may be due to steric factors. The results are in accord with dimerization being a radical process, but not via dimerization of two tyrosine-derived phenoxyl radicals (103).The cross-linking may involve dimerization of the reactive radicals detected in spin-trapping experiments (129).
D. REACTIONS WITH MEMBRANES A number of studies have examined the possible role of both Fe(1V)=0 and globin radicals in the initiation of membrane peroxidation. Thus, it has been reported that although reaction of Fe(II1) Lb with hydrogen peroxide (two-fold excess) initiates peroxidation of peribacteroid membranes from French beans (as measured by the yield of malondialdehyde), similar reactions with Fe(I1) Lb and hydrogen peroxide did not give a similar effect (162). Little oxidation was ob-
534
DAVIES, MATHIEU, AND PUPPO
served in the absence of hydrogen peroxide, whereas much higher levels of oxidative breakdown products are observed with a 5- or 10-fold excess. In the latter, however, it is known that significant release of iron occurs, and it is possible that the lipid oxidation is due to the reaction of the iron with excess peroxide (99). The observation that with Fe(II1) Lb and a low excess of peroxide, damage is initiated, whereas none is observed with Fe(I1) Lb [which is a more efficient method of generating F e W ) =01, suggests that Fe(IV)=0 is not able to initiate damage to membrane fractions (162). This is not entirely unexpected because Fe(1V)=0 would not be normally expected t o have ready access to membrane lipids (or other membrane components) due to the surrounding protein. Recent studies with symbiosomes from French beans have supported the suggestion that Fe(1V)=0 alone cannot initiate lipid oxidation (163).Isolated symbiosomes exposed to Fe(II1) Lb that had been preincubated with a twofold excess of hydrogen peroxide did not appear to suffer damage as monitored by the uptake of succinate (a useful marker of symbiosome membrane integrity). Enhanced uptake of succinate was observed, however, with six-fold excesses of hydrogen peroxide when iron release is likely to have occurred (163). A number of groups have suggested that metal ions derived from Lb and/or trace transition-metal ions can play a role in initiating or propagating damage (162-1641, but the mechanism by which this occurs is poorly understood. It is possible that the release of iron from the heme protein (possibly as a result of oxidative damage) or from a build-up of trace-metal ions in old nodules by other routes, may be responsible for the catalysis of radical generation, though the details remain uncertain. Recent reports have demonstrated that low-molecular-weight chelates of transition-metal ions are present in old nodules (162, 1641, and it has been suggested that hydroxyl radicals may be a key species (164).It should be noted, however, that specific products of hydroxyl radical attack were not identified, and it is possible that the oxidation of dimethyl sulfoxide observed may be due to the formation of other radicals. Other studies have also reported that hydroxyl radicals are not involved (162).Previous studies using fluorescence quenching (165) have suggested that Lb can interact with peribacteroid membranes, possibly via specific binding sites. The presence of such “receptors” might be expected to facilitate damage induced by either protein radicals or high-oxidation-state species. In the light of the evidence reviewed here that at least some of the globin radicals formed on reaction of Fe(II1) Lb with hydrogen peroxide are present on the surface of the protein and that these species
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
535
can react both with low-molecular-weight species and other protein molecules, it was of interest to determine whether these species could transfer damage to a membrane (i.e., initiate lipid oxidation). Of particular relevance is whether the Lb-derived radicals could initiate damage to the peribacteroid membrane that surrounds the microsymbiont in uiuo (Fig. 2) and whether such reactions play a role in the loss of nitrogen-fixing ability of nodules and nodule senescence. The latter membrane appears to be more sensitive to oxidation than others (e.g., microsomes prepared from identical nodules) (1621, possibly as a result of the different compositions of these structures (e.g., the high lipid : protein ratio of the peribacteroid membrane). Incubation of purified peribacteroid membranes (PBMs) with soybean Fe(II1) Lbhydrogen peroxide and analysis of the extent of Lb dimerization by SDS-PAGE have demonstrated that the presence of the PBMs decreases (in a dose-dependent manner) the extent of protein dimerization (129) (Fig. 14). This has been interpreted as evidence for interaction of the radicals involved in the dimerization reaction with the PBMs. Further evidence for such a process has been obtained from examining the intensity of the signal (which is directly proportional to the radical concentration) of the Tyr-derived phenoxyl radical generated from soybean Fe(II1) Lb and Hz02in the presence of added PBMs by stopped-flow EPR spectroscopy. Inclusion of the PBMs (>0.3 mg proteidml) resulted in a statistically significant decrease in concentration of the phenoxyl radical, consistent with scavenging by membrane components (129). Such a process would be expected to result in the formation of one or more radical species in the membrane fractions.
FIG. 14. Quenching by peribacteroid membranes (PBMs) of the dimerization process of Fe(II1) Lb in the presence of H p 0 2 .Fe(II1) Lb (210 pM) was incubated with H,O, (420 pM) for 1 h in the absence (lane 2) or in the presence of PBMs at the following protein concentrations: 0.23 mg/mL (lane 3), 1.16 mg/mL (lane 4) and 2.33 mg/mL (lane 5). Lane 1 represents Fe(II1) Lb alone and lane 6 and 7 PBMs plus H2O2and PBMs alone, respectively (reprinted with permission from: Moreau, S.; Davies, M. J.; Mathieu, C.; Herouart, D.; Puppo, A. J. Biol. Chern., 1996,271, 32557-32562).
536
DAVIES, MATHIEU, AND PUPPO
Inclusion of PBM fractions in Fe(II1) Lbhydrogen peroxide spintrapping experiments has also provided evidence for such interactions. Thus when PBMs (ca. 3 mg proteidml) were added to an equimolar amount of soybean Fe(II1) Lb and hydrogen peroxide in the presence of the spin trap PBN, significant changes were observed in the EPR spectra as compared to control spectra in the absence of the membrane fractions (Fig. 15) (129). The additional features present are consistent with the presence of large, slowly tumbling radical adduct species and have been assigned to membrane-derived radicals.
w
FIG. 15. EPR spectra observed on reaction of soybean Fe(II1) Lb (500 pM) with HaOa (1 mM) and the spin trap PBN (46 mM) in the absence or presence of peribacteroid membranes (PBMs). (a) Incubation carried out for 2 h at 29°C in the absence of added membrane fractions; signal assigned to one (or more) large protein-derived radical adducts to the spin trap. (b) Incubation carried out for 2 h at 29°C in the presence of added membrane fractions (2.95 mg proteidml); signal assigned to one (or more) large protein-derived radical adducts to the spin trap together with a further species (arrowed features) believed to arise from a radical generated from the membrane fraction. (c) As (b) but spectrum of the organic layer obtained after extraction of the membrane fractions with an equal volume of Nz-degassed toluene; signal assigned to a t least one lipid-soluble radical adduct arising from radical induced damage to the membrane fraction (reprinted with permission from: Moreau, S.; Davies, M. J.; Mathieu, C.; Herouart, D.; Puppo, A. J. Biol. Chem., 1996,271,32557-32562).
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
537
To obtain further information on the nature of these intermediates, solvent extraction experiments were carried out. When toluene was added, the radical adducts were observed only in the organic phase. This is consistent with the presence of lipid-derived radicals (or other hydrophobic species). The signals observed from the toluene extracts are considerably more mobile than those seen in the aqueous system, suggesting that the slow tumbling is due to the formation of a lipid radical in the membrane, which is prevented from rotating rapidly by the membrane. The sharp (isotropic) nature of the signals in these organic extracts has allowed further fine structure of the EPR signals to be observed. The hyperfine coupling constants are consistent with a lipid-derived species (129). There is now considerable evidence for an interaction of the globinderived radicals with the peribacteroid membrane resulting in a transfer of the radical into the membrane (129). The membranederived radicals would be expected to participate in chain reactions in the presence of oxygen, and this would result in peroxidative damage to the membrane. Lipid peroxidation reactions are known to result in widespread alterations to membrane structure and can lead to irreversible loss of function. Such reactions may play a role in the disruption of the peribacteroid membrane during nodule senescence, when conditions for peroxide formation and subsequent radical reactions would be expected to be favorable. An obvious corollary is that species that protect against the formation of Fe(III1, F e W ) =0 and globin radicals might inhibit the reactions and hence nodule degradation. These may include nonenzymic reactions (e.g., with ascorbate, cysteine, and glutathione), all of which are present in nodules at high concentrations, and enzymatic reactions brought about by superoxide dismutase, catalase, Fe(II1) Lb reductase, and possibly other enzymes. Manipulation of the levels of activity of these species is an obvious potential strategy for increasing the lifetime and activity of the nitrogen-fixing processes. Some evidence suggests that this hypothesis is indeed correct-for example, the observation that the nodules from aging gram plants remain pink for longer periods in the presence of added ascorbate in the growth medium (160).Although significant protection by addition of exogenous materials such as ascorbate and thiols is unlikely to be of any great practical utility, the manipulation by genetic means of the enzymes that regulate the production of these low-molecular-weight materials and the concentration and activity of the protective enzymes within the host plant and bacteroid may be realistic and achieveable goals. Work toward these targets is already under way.
538
DAVIES, MATHIEU, AND PUPPO
ACKNOWLEDGMENTS Work reported in this review has been supported by the European Union Human Capital and Mobility Program (contract CHRX-CT94-0605),the Alliance Program, and the Australian Research Council. We are also extremely grateful to all our co-workers for their very positive contributions, to Prof. H. C. Freeman and Dr. P. J . Ellis (School of Chemistry, University of Sydney, Australia) for providing the coordinates of soybean Lb, prior to publication, to Dr. C. A. Appleby for stimulating advice and comments, and to Mrs. G. Van der Sype for help with the photography.
REFERENCES 1. Hardison, R. C. Proc. Natl. Acad. Sci. U.S.A. 1996,93, 5675. 2. Kubo, H. Acta Phytochim. (Tokyo) 1939,11, 195. 3. Appleby, C. A.; Tjepkema, J. D.; Trinick, M. J. Science 1983,220, 951. 4 . Sharifi, E. Biosystems 1983, 16, 269. 5. Beringer, J. E.; Brewin, N.; Johnston, A. W.; Schulman, H. M.; Hopwood, D. A. Proc. Roy. SOC.Lond. B Biol. Sci. 1979,204, 219. 6. Legocki, R. P.; Verma, D. P. Cell 1980,20, 153. 7. Bergersen, F. J. “Root Nodules of Legumes: Structure and Functions”; Research Studies Press, John Wiley and Sons Ltd.: London, 1982. 8. Tjepkema, J. D. Can. J. Bot. 1983, 63, 2924. 9. Appleby, C. A.; Bogusz, D.; Dennis, E. S.; Peacock, W. J. Plant Cell Enuiron. 1988, 11, 359. 10. Anderson, C. R.; Jensen, E. 0.; Llewellynm, D. J.; Dennis, E. S.; Peacock, W. J. Proc. Natl. Acad. Sci. U.S.A. 1996,93, 5682. 11. Hyldig-Nielsen, J.; Jensen, E. 0.;Paludan, K.; Wiborg, 0.;Garrett, R.; Jorgensen, P.; Marcker, K. A. Nucleic Acid Res. 1982, 10, 689. 12. Appleby, C. A. Annu. Rev. Plant Physiol. 1984, 35, 443. 13. Appleby, C. A.; Dennis, E. S.; Peacock, W. J . Aust. J. Syst. Biol. 1990,3, 81. 14. Wittenberg, J. B. J . Biol. Chem. 1974,249, 4057. 15. Appleby, C. A. Sci. Prog. 1992, 76, 365. 16. Zufferey, R.; Preisig, 0.; Hennecke, H.; Thony-Meyer, L. J. Biol. Chem. 1996, 271, 9114. 17. Ellfolk, N.; Sievers, G. Acta. Chem. Scand. 1972,26, 1155. 18. Fuchsman, W. H. Adu. Comp. Environ. Physiol. 1992, 13, 23. 19. Whittaker, R. G.; Lennox, S.; Appleby, C. A. Biochem. Int. 1981,3, 117. 20. Fuchsman, W. H.; Appleby, C. A. Biochim. Biophys. Acta. 1979,579, 314. 21. Gibson, Q.H.; Wittenberg, J. B.; Wittenberg, B. A,; Bogusz, D.; Appleby, C. A. J. Biol. Chem. 1989,264, 100. 22. Fuchsman, W. H. Arch. Biochem. Biophys. 1986,243, 454. 23. Ollis, D. L.; Appleby, C. A,; Colman, P. M.; Cutten, A. E.; Mitchell Guss, J.; Venkatappa, M. P.; Freeman, H. C. Aust. J . Chem. 1983,36, 451. 24. Vainshtein, B. K.; Harutyunyan, E. H.; Kuranova, I. P.; Borisov, V. V.; Sosfenov, N. I.; Pavlovsky, A. G.; Grebenko, A. I.; Konareva, N. V. Nature 1976,254, 163. 25. Vainshtein, B. K.; Andreeva, N. S. Mol. Biol. Mosk. 1977, 1 1 , 1258. 26. Arutiunian, E. G . Mol. Biol. Mosk. 1981, 15, 27.
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
539
27. Arutyunyan, E. G.; Deisenhofer, I.; Teplyakov, A. V.; Kuranova, I. P.; Obmolova, G. V.; Vainshtein, B. K. Dokl. Akad. Nauk. Ser. Biokhim. 1983,270, 732. 28. Arutyunyan, E. G.; Safanova, T. N.; Obmolova, G. V.; Teplyakov, A. V.; Popov, A. N . ; Rusakov, A. A.; Rubinsky, S. V.; Kuranova, I. P.; Vainshtein, B. K. Bioorg. Khim. 1990, 16, 293. 29. Harutyunyan, E. H.; Safonova, T. N.; Kuranova, I. P.; Popov, A. N.; Teplyakov, A. V.; Obmolova, G. V.; Rusakov, A. A,; Vainshtein, B. K.; Dodson, G. G.; Wilson, J. C.; Perutz, M. F. J. Mol. Biol. 1996,251, 104. 30. Harutyunyan, E. H.; Safonova, T. N.; Kuranova, I. P.; Popov, A. N.; Teplyakov, A. V.; Obmolova, G . V.; Vainshtein, B. K.; Dodson, G. G.; Wilson, J. C. J. Mol. B i d . 1996,264, 152. 31. Brown, G. G.; Lee, J. S.; Brisson, N.; Verma, D. P. J . Mol. Euol. 1984,21, 19. 32. Ellfolk, N.; Sievers, G. Acta. Chem. Scand. 1965, 19, 2409. 33. Perttila, U.; Sievers, G. Biochim. Biophys. Acta. 1980, 624, 316. 34. Kuranova, I. P.; Luk‘ianova, E. V. Dokl. Akad. Nauk SSSR 1983,269, 1503. 35. Dalvit, C.; Tennant, L.; Wright, P. E. J . Inorg. Biochem. 1986,28, 303. 36. Ikeda-Saito, M.; Hori, H.; Inubushi, T.; Yonetani, T. J . Biol. Chem. 1981, 256, 10267. 37. Appleby, C. A. “The Biology of Nitrogen Fixation”; Quispel, A., Ed.; North-Holland Pub. Co.: Amsterdam & Oxford, 1974, 522-554. 38. Verma, D. P.; Kazazian, V.; Zogbi, V.; Bal, A. K. J. Cell. B i d . 1978, 78, 919. 39. Udvari, M. K.; Price, G. D.; Gresshoff, P. M.; Day, D. A. FEBS Lett. 1988,231, 36. 40. Herrada, G.; Puppo, A.; Rigaud, J. J. Gen. Microbiol. 1989, 135, 3165. 41. Moreau, S.; Meyer, J.-M.; Puppo, A. FEBS Lett. 1995,361, 225. 42. Bergersen, F. J.; Appleby, C. A. Planta 1981, 152, 534. 43. Verma, D. P.; Ball, S.; Guerin, C.; Wanamaker, L. Biochemistry 1979, 18, 476. 44. OBrian, M. R.; Kirshbom, P. M.; Maier, R. J. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 8390. 45. Keithley, J. H.; Nadler, K. D. J. Bacteriol. 1983, 154, 838. 46. Dimitrijevic, L.; Puppo, A.; Trinchant, J. C.; Rigaud, J. J . Plant Physiol. 1989, 134, 642. 47. Noel, K. D.; Stacey, G.; Tandon, S. R.; Silver, L. E.; Brill, W. J. J . Bacteriol. 1982, 152, 485. 48. O’Brian, M. R.; Kirshbom, P. M.; Maier, R. J. J . Bacteriol. 1987, 169, 1089. 49. Bisseling, T.; van, S. J.; Houwaard, F. Biochim. Biophys. Acta. 1980, 610, 360. 50. Kijne, J. W. Physiol. Plant Pathol. 1976, 7, 17. 51. Uheda, E.; Syono, K. Plant Cell Physiol. 1982,23, 75. 52. Appleby, C. A.; Bradbury, J. H.; Morris, R. J.;Wittenberg, B. A.; Wittenberg, J. B.; Wright, P. E. J. B i d . Chem. 1983,258, 2254. 53. Martin, K. D.; Saari, L.; Wang, G. X.; Wang, T.; Parkhurst, L. J.; Klucas, R. V. J. Bid. Chem. 1990,265, 19588. 54. Fuchsman, W. H.; Barton, C. R.; Stein, M. M.; Thompson, J. T.; Willett, R. M. Biochem. Biophys. Res. Commun. 1976, 68, 387. 55. Rohlfs, R. J.; Olson, J. S.; Gibson, Q. H. J. Bid. Chem. 1988, 263, 1803. 56. Rohlfs, R. J.;Matthews, A. J.; Carver, T. E.; Olson, J. S.; Springer, B. A.; Egeberg, K. D.; Sligar, S. G. J . Biol. Chem. 1990,265, 3168. 57. Arutiunian, E. G. Dokl. Akad. Nauk. SSSR 1980,252, 1264. 58. Irwin, M. J.;Armstrong, R. S.; Wright, P. E. FEBS Lett. 1981, 133, 239. 59. Rousseau, D. L.; Ondrias, M. R.; LaMar, G. B.; Smith, K. M. J. Biol. Chem. 1983, 258, 1740.
540
DAVIES, MATHIEU, AND PUPPO
60. Ollis, D. L.; Wright, P. E.; Pope, J. M.; Appleby, C. A. Biochemistry 1981,20,587. Appleby, C. A. Biochemistry 1979,18,1309. 61. Fuchsman, W. H.; 62. Ray, G. B.; Li, X.-Y.; Ibers, J. A.; Sessler, J. L.; Spiro, T. G. J. Am. Chem. SOC. 1994,116,162. 63. Lim, M.; Jackson, T. A.; Anfinrud, P. A. Science 1995,269,962. 64. Quillin, M. L.; Arduini, R. M.; Olson, J. S.; Phillips, G. N. J. J . Mol. Biol. 1993, 234,140. 65. Quillin, M. L.; Tiangsheng, L.; Olson, J. S.; Phillips, G. N. J.; Yi, D.; Ikeda-Saito, M.; Regen, R.; Carlson, M.; Gibson, Q. H.; Haiying, L.; Elber, R. J. Mol. Biol. 1995, 245,416. 66. Moore, J. N.; Hansen, P. A.; Hochstrasser, R. M. Proc. Natl. Acad. Sci. U.S.A. 1988,85,5062. 67. Virtanen, A. I. Biol. Rev. 1947,22, 239. 68. Maskall, C . S.;Gibson, J. F.; Dart, P. J. Biochem. J. 1977,167,435. 69. Kanayama, Y.; Yamamoto, Y. Plant Cell Ph.ysio1. 1990,31,207. 70. Kanayama, Y.; Watanabe, I.; Yamamoto, Y. Plant Cell Physiol. 1990,31,341. 71. Dean, J. V.;Harper, J. E. Plant Physiol. 1988,82,718. 72. Fox, J. B. J.; Thomson, J. S. Biochemistry 1983,2,465. 73. Cueto, M.;Hernandez-Perera, 0.;Martin, R.; Bentura, M.; Rodrigo, J.; Lamas, S.; Golvano, M. P. FEBS. Lett. 1996,398,159. 74. Mathieu, C.; Moreau, S.; Frendo, P.; Puppo, A.; Davies, M. J . Free Rad. Biol. Med. 1998,24,1242. 75. Cebolla, A.; Palomares, A. J. Microbiologia 1994,10,371. 76. Gilles-Gonzalez, M. A.; Gonzalez, G.; Perutz, M. F.; Kiger, L.; Marden, M. C.; Poyart, C.Biochemistry 1994,33,8067. 77. Winkler, W. C.; Gonzalez, G.; Wittenberg, J. B.; Hille, R.; Dakappagari, N.; Jacob, J.; Gonzalez, L. A.; Gilles-Gonzalez, M. A. Chem. Biol. 1996,3,841. 78. Pieve, Y.V.; Atanasov, B. P.; Zhiznevskaya, G. Y.; Krasnobaeve, N. N. Dokl. Akad. Nauk. SSSR 1972,202,965. 79. Abel, K.Phytochem. 1983,2,429. 80. Ewing, G. J.; Ionescu, L. G. J. Phys. Chem. 1972,76,591. 81. Stetzkowski, F.; Cassoly, R.; Banerjee, R. J . Biol. Chem. 1979,254,11351. 82. Keilin, D.; Smith, J. D. Nature 1947,159,692. 83. Wittenberg, B. A.; Wittenberg, J. B.; Appleby, C. A. J. Biol. Chem. 1973,248,3178. 84. Ellfolk, N.Acta Chem. Scand. 1961, 15, 975. 85. Ehrenberg, A.;Ellfolk, N. Acta Chen. Scand. 1983, 17, S343. 86. Trewhella, J.; Wright, P. E. Biochem. Biophys. Res. Commun. 1979,88,713. 87. Job, D.; Zeba, B.; Puppo, A.; Rigaud, J. Eur. J. Biochem. 1980, 107,491. 88. Appleby, C. A.; Wittenberg, B. A,; Wittenberg, J. B. Proc. Natl. Acad. Sci. U.S.A. 1973,70,564. 89. Appleby, C . A.; Wittenberg, B. A.; Wittenberg, J. B. J. Biol. Chem. 1973,248,3183. 90. Appleby, C. A.; Blumberg, W. E.; Peisach, J.; Wittenberg, B. A.; Wittenberg, J. B. J. Biol. Chem. 1976,251,6090. 91. Appleby, C. A. Biochim. Biophys. Acta. 1969,189,267. 92. Rigaud, J.; Puppo, A. Biochim. Biophys. Acta. 1977,497,702. 93. Puppo, A,; Rigaud, J.; Job, D. Plant Sci. Lett. 1981,22,353. 94. Aviram, I.; Wittenberg, A.; Wittenberg, J. B. J . Biol. Chem. 1978,253, 5685. 95. Henderson, R. W.; Appleby, C. A. Biochim. Biophys. Acta. 1972,283,187. 96. Nash, D. T.; Schulman, H. M. Biochem. Biophys. Res. Commun. 1976,68,781. 97. Becana, M.; Klucas, R. V. Plant Physiol. 1992,98,1217.
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS
541
98. Pladys, D.; Barthe, P.; Rigaud, J. Plant Sci. 1988,56, 99. 99. Puppo, A.; Halliwell, B. Planta 1988, 173, 405. 100. Sadrzadeh, S. M.; Graf, E.; Panter, S. S.; Hallaway, P. E.; Eaton, J. W. J. Biol. Chem. 1984,259,14354. 101. Davies, M. J.; Puppo, A. Biochem. J. 1992,281, 197. 102. Puppo, A,; Monny, C.; Davies, M. J. Biochem. J. 1993,289, 435. 103. Moreau, S.; Davies, M. J.; Puppo, A. Biochim. Biophys. Acta. 1995, 1251, 17. 104. George, P.; Irvine, D. H. Biochem. J . 1952,52, 511. 105. King, N. K.; Looney, F. D.; Winfield, M. E. Biochim. Biophys. Actu 1967,133, 65. 106. Yonetani, T.; Schleyer, H. J . Biol. Chem. 1967, 242, 1974. 107. Davies, M. J. Biochim. Biophys. Acta 1991, 1077, 86. 108. Gunther, M. R.; Kelman, D. J.; Corbett, J. T.; Mason, R. P. J. Biol. Chem. 1995, 270, 16075. 109. Tschirret-Guth, R. A.; Ortiz de Montellano, P. R. Arch. Biochem. Biophys. 1996, 335, 93. 110. DeGray, J. A,; Gunther, M. R.; Tschirret-Guth, R.; Ortiz de Montellano, P. R.; Mason, R. P. J. Biol. Chem. 1997,272, 2359. 111. Sealy, R. C.; Harman, L.; West, P. R.; Mason, R. P. J . Am. Chem. SOC.1985, 107, 3401. 112. Davies, M. J.; Puppo, A. Biochim. Biophys. Acta 1993, 1202, 182. 113. Larsson, A.; Sjoberg, B. M. EMBO J. 1986, 5, 2037. 114. Innes, J. B.; Brudvig, G. W. Biochemistry 1989,28, 1116. 115. Fontecave, M.; Gerez, C.; Atta, M.; Jeunet, A. Biochem. Biophys. Res. Chem. 1990, 168, 659. 116. Kulmacz, R. J.; Ren, Y.; Tsai, A. L.; Palmer, G. Biochemistry 1990,29, 8760. 117. Nordlund, P.; Sjoberg, B. M.; Eklund, H. Nature 1990, 345, 593. 118. Whittaker, M. M.; Whittaker, J. W. J . Biol. Chem. 1990,265, 9610. 119. Hoganson, C. W.; Babcock, G. T. Biochemistry 1992,31, 11874. 120. Kulmacz, R. J.; Palmer, G.; Tsai, A. L. J. Lipid Mediat. 1993, 6, 145. 121. Nordlund, P.; Eklund, H. J . Mol. Biol. 1993, 232, 123. 122. Pedersen, J. 2.; Finazzi, A. A. FEBS Lett. 1999,325, 53. 123. Ormo, M.; Regnstrom, K.; Wang, Z.; Que, L. J.;Sahlin, M.; Sjoberg, B. M. J. B i d . Chem. 1995,270,6570. 124. Sun, X.; Ollagnier, S.; Schmidt, P. P.; Atta, M.; Mulliez, E.; Lepape, L.; Eliasson, R.; Graslund, A,; Fontecave, M.; Reichard, P.; Sjoberg, B. M. J. Biol. Chem. 1996, 271, 6827. 125. von Sonntag, C. “The Chemical Basis of Radiation Biology”; Taylor and Francis: London, 1987. 126. Davies, M. J.; Dean, R. T. “Radical-Mediated Protein Oxidation: From Chemistry to Medicine”; Oxford University Press: Oxford, 1997. 127. Marriott, P. R.; Perkins, M. J.; Griller, D. Can. J . Chem. 1980,58, 803. 128. Buettner, G. R. Free Rad. Biol. Med. 1987,3, 259. 129. Moreau, S.; Davies, M. J.; Mathieu, C.; Herouart, D.; Puppo, A. J. Biol. Chem. 1996,271, 32557. 130. Davies, M. J.; Gilbert, B. C.; Haywood, R. M. Free Rad. Res. Commun. 1991, 15, 111. 131. Davies, M. J . Res. Chem. Intermed. 1993, 19, 669. 132. Dean, R. T.; Fu, S.; Stocker, R.; Davies, M. J. Biochem. J . 1997,324, 1. 133. Barr, D. P.; Gunther, M. R.; Deterding, L. J.; Tomer, K. B.; Mason, R. P. J. Biol. Chem. 1996,271, 15498.
542 134. 135. 136. 137. 138.
DAVIES, MATHIEU, AND PUPPO
Moreau, S.; Puppo, A,; Davies, M. J. Phytochem. 1996,39, 1281. Puppo, A.; Davies, M. J. Biochim. Biophys. Acta 1996, 1246, 74. Romero, F. J.; Ordonez, I.; Arduini, A.; Cadenas, E. J. Biol. Chem. 1992,267,1680. Berzofsky, J. A,; Peisach, J.; Blumberg, W. E. J. Biol. Chem. 1971,246, 3367. Mathieu, C.; Swaraj, K.; Davies, M. J.; Trinchant, J.-C.; Puppo, A. Free Rad. Res.
1997,27, 165. 139. Becana, M.; Klucas, R. V. Proc. Natl. Acad. Sci. U.S.A. 1990,87, 7295. 140. Appleby, C. A, Biochim. Biophys. Acta 1969, 188, 222. 141. Kretovich, V. L.; Melik-Sarkissyan, S. S.; Bashirova, N. F.; Topunov, A. F. J. Appl. Biochem. 1982,4, 209. 142. Topunov, A. F.; Melik, S. S.; Lysenko, L. A.; Iarpilenko, G. P.; Kretovich, V. L. Biokhimiia 1980,45, 2053. 143. Saari, L. L.; Klucas, R. V. Arch. Biochem. Biophys. 1984,231, 102. 144. Puppo, A,; Rigaud, J.; Job, D.; Ricard, J.; Zeba, B. Biochim. Biophys. Acta 1980, 614,303. 145. Golubeva, L. I.; Topunov, A. F.; Talyzin, V. V.; Kretovich, V. L. Dokl. Akad. Nauk. S S S R 1987,294, 178 (Engl. Trans.). 146. Golubeva, L. I.; Topunov, A. F.; Goncharova, S. S.; Aseeva, K. B.; Kretovich, V. L. Biokhimioa 1988, 53,~1478(Engl. Trans.). 147. Ji, L.; Wood, S.; Becana, M.; Klucas, R. V. Plant Physiol. 1991,96, 32. 148. Swaraj, K.; Topunov, A. F.; Golubeva, L. I.; Kretovich, V. L. Fiziol. Rust. 1986, 33, 87. 149. Puppo, A.; Rigaud, J . FEBS. Lett. 1979, 108, 124. 150. Dalton, D. A,; Russell, S. A.; Hanus, F. J.; Pascoe, G. A.; Evans, H. J . Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 3811. 151. Saari, L. L.; Klucas, R. V. Biochim. Biophys. Acta 1987,912, 198. 152. Lehtovaara, P.; Perttila, U. Biochem. J. 1978, 176, 359. 153. Moreau, S.; Davies, M. J.; Puppo, A. Biochim. Biophys. Acta 1995, 1251, 17. 154. Catalano, C. E.; Choe, Y. S.; Ortiz de Montellano, P. R. J. Biol. Chem. 1989, 264, 10534. 155. Osawa, Y.; Williams, M. S. Free. Rad. Biol. Med. 1996,21, 35. 156. Sugiyama, K.; Highet, R. J.; Woods, A.; Cotter, R. J.; Osawa, Y. Proc. Natl. Acad. Sci. U.S.A. 1997,94, 796. 157. Virtanen, A. I. Nature 1946,155, 747. 158. Lemberg, R.; Legge, J. W. “Hematin Compounds and Bile Pigments”; Interscience Publishers Inc.: New York, 1949. 159. Lehtovaara, P. Biochem. J. 1978, 176, 351. 160. Swaraj, K.; Garg, 0. P. Plant Physiol. 1969,23, 889. 161. Tew, D.; Ortiz de Montellano, P. R. J. Biol. Chem. 1988,263, 17880. 162. Puppo, A,; Herrada, G.; Rigaud, J. Plant Physiol. 1991, 96, 826. 163. Herrada, G.; Puppo, A.; Moreau, S.; Day, D. A,; Rigaud, J. FEBS. Lett. 1993, 326, 33. 164. Becana, M.; Klucas, R. V. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 8958. 165. Zhiznevskaya, G. Y.; Borodenko, L. I.; Mekler, V. I.; Belonogova, 0. V.; Kotelnikov, A. I.; Izmailov, S. F. Russ. J. Plant Physiol. 1994,41, 69. 166. Gibson, Q. H.; Olson, J. S.; McKinnie, R. E.; Rohlfs, R. J. J. Biol. Chem. 1986, 261, 10228. 167. Olson, J. S.; Rohlfs, R. J.; Gibson, Q. H. J. Biol. Chem. 1987, 262, 12930.
INDEX
A
B
Actinides, xenon fluoride reactions, 91 Alcohol oxidation, molybdenum chloride clusters, 19 Alkenes chlorination, 115-117 chlorination, manganese complex applications, 403-405 epoxidation, manganese complex applications, 395-401 Aluminum trihydride, 121-122 Americium, xenon fluoride reactions, 91 Ammonia, fluorine gas reaction, 117 Amyotrophic lateral sclerosis (ALS),peroxynitrite implicated, 402 Annexins, calcium binding in, 457 Antimony complexes, xenon fluoride reactions, 86, 88 Apo-calcyclin, 455 Apo-calmodulin, 449-450 Arginase, 306-307 binuclearity, 325-327 EPR spectroscopy, 389 hydrogen peroxide disproportionation, 326-327 manganese requirement, strict, 307, 424 Argon chemistry, 52, 54 Aromatic compounds, dioxygenase decomposition, 312-324 Arsenic complexes, xenon fluoride reactions, 86, 88 Arteriosclerosis, peroxynitrite implicated, 311 Ascorbate, leghemoglobin reaction, Fe(IV)=O, 520-521 Aziridination, manganese complex applications, 402-403
Bacteria manganese catalase, 323-325 manganese(I1) dioxygenase, 312, 314 manganese ribonucleotide reductase, 319, 321 manganese superoxide dismutase, 3 10-3 12 BEDT-TTF-based supramolecular complexes, 195 Benzene-iodine complex, charge-transfer reaction, 146- 148 Benzothiadiazole-bridged supramolecular complexes, 214-219 Biliverdins, 529 Binuclear manganese redox enzymes, see Manganese redox enzymes, binuclear Bisnitrile-bridged supramolecular complexes, 243-246 Blood coagulation factors, 442, 466-470, 474-477,479-481 Blood plasma, calcium concentration, 470 RM-40 ISPARC), 484 Boron hydrides, reactive intermediate studies, 110 Boron nitride, reactive intermediate studies, 108-109 Bradyrhizobium .japonicum, leghemoglobin in, 499-500 t-Butylgallium sulfide, thermal decomposition, 112-113
C CaBP (Calcyclin), 454-456 Cadherins, calcium binding in, 483-484 Calbindin D9k, 452-453 stability, 444
543
544
INDEX
Calcium biological coordination chemistry, 442 biological roles, 441-442 mixed-metal active site, concanavalin A, 308 oxygen-evolving complex requirement, 328, 338-339 resting cell concentration, 470 Calcium-binding domains, 442 Calcium-binding proteins, 441-442; see also specific proteins EF hand domains, 442,443-445 EGF-like domains, 471-479 extracellular, 442, 470-485 intracellular, 442, 443-456 ligand preferences, 442 Calcium2+-dependentprotein kinase C, 460,462-463 Calcium-mediated membrane-binding proteins, 442, 456-457; see also specific proteins y-carboxyglutamic acid sites, 465-470 C2 domains, 460-465 Calcyclin, 454-456 Calixarenes, 175 Calmodulin, calcium binding, 444, 445-450 Calmodulin-dependent protein kinase I, 448 Calmodulin-dependent protein kinase 11, 447 Carbon disulfide, photodissociation, 149-150 Carbon-hydrogen bond activation, 143 Carbonyls, see Metal carbonyls; specific carbonyl complexes y-Carboxyglutamic acid sites, calciummediated membrane-binding proteins, 465-470 Cellulose decomposition, 316-3 17 CeriumW), peroxotitanium reaction, 160-161 Cesium niobium iodide complexes, 34 Cesium pseudohalide complexes, 42 Chain structure supramolecular complexes, 251-263 Channels, supramolecular complexes, 222-227,283-289, 291-292 Chemical reaction intermediates, see Reactive intermediates
Chemical reaction mechanisms, 102 Chemical reaction rate, 101-102 Chevral phases, 3, 14 Chloride, oxygen-evolving complex requirement, 328, 329, 338-339, 393 Chlorine oxide, 109-110, 158 Chlorofluorocarbons, 109 Chlorophyll a, 328-329 Chloroplasts, 328 Cholegobin, 529 Chromium(III), oxidation to chromium(V), hydrogen peroxide, 161 Chromium complexes, xenon fluoride reactions, 89 Chromium fluorides, 59 Chromium hexacarbonyl, reactive intermediates, 102-104, 134 Chromium pentacarbonyl, 106, 164-165 flash photolysis, 139-140 matrix isolation, 128-130 Chromium pentacarbonyl dihydrogen complex, 133-134, 135 Chromium-salen complexes, 396 Chymotrypsin, calcium binding, 479 Clathrates, noble gases in, 51, 54 Coagulation factors, 442, 466-470, 474477, 479-481 Cobalt complexes intercalation, 155 nitric oxide insertion, 104 xenon fluoride reactions, 89 Cobalt trifluoride, 59 Conantokins, calcium binding, 470-471 Concanavalin A, 308 Conotoxins, 470 CopperU), 174-175 Copper(1) supramolecular complexes, 174- 175 channels, 222-227,283-289,291-292 diamondoid frameworks, 240-251 graphite frameworks, 204-219 helical frameworks, 176-192 hexagonal frameworks, 204-219 hydrogen bond-assembled frameworks, 219-228 infinite-chain structures, 251-271 n-n interaction-assembled frameworks, 228-239 self-assembly pathway, 175 stereochemistry, 174-175
545
INDEX
stoichiometries, 176 sulfur-sulfur contact-assembled frameworks, 192-204 two-dimensional structures, 271-283 Copperlzinc superoxide dismutase, 310 Crown ethers, 175 Cryptands, 175 Cysteamine, leghemoglobin reaction,
523-524 Cysteine, leghemoglobin reaction, 523,
526 Cystolic phospholipase A2, 460 Cytochrome Bss2,four-helix bundle arrangement, 444 Cytochrome c oxidase, dioxygen reaction,
143 Cytochrome P,eo,395
D Diamondoid framework supramolecular complexes, 240-251 Dioxygen cytochrome c oxidase reaction, 143 manganese catalase, 322 manganese dioxygenase, 312 oxygen-evolving complex, 327-328,
330,419
Electron diffraction, reactive intermediates, 110-113 time-resolved, 149 Electronic absorption spectrophotometry,
160 Electrospray ionization mass spectrometry, 162-163 Epidermal growth factor (EGF)-likedomains, calcium-binding proteins,
471-479 Epoxidation, manganese complex applications, 394-402 4,5-Ethylenedithio-l,3-dithiole-2-thionebased supramolecular complexes,
200-204 Extended X-ray absorption fine structure (EXAF'S), 390-391 Extracellular calcium-binding proteins,
442,470-485
F Factor VII, calcium binding, 466,467,
479-480,481 Factor IX, calcium binding, 466,467,
469,474-477,479-480 Factor X, calcium binding, 466,467-468,
474-477,479-480
singlet, 122 Dioxygenases, 312;see also Manganese(I1) dioxygenase
Fast-scan electrochemical techniques, 163 Feldspar-like supramolecular complexes,
1,3-Dithiole-2-thione-4,5-dithiolate-based supramolecular complexes, 193-194 Dixenon cation, 68 DNA damage, peroxynitrite implicated, 310 synthesis, 319
Femtosecond pulses, reactive intermediate characterization, 146-148 Ferrocyanide, leghemoglobin reaction,
E Early transition metal halide clusters, see Group 5 metal halide clusters; Group 6 metal halide clusters E cadherin, calcium binding in, 483-484 EF hand domains, calcium-binding proteins, 442,443-445 Electric discharge, reactive intermediates, 118-121 Electrochemical techniques, fast-scan,
163
289-290
524 Fibrillin, calcium binding, 473,474,477 Fibulin-1, calcium binding, 473 Flash photolysis, 106,137,139-140 Flow characterization methods, reactive intermediates, 156-164 Fluorescence, 156 Fourier-transform microwave spectroscopy, pulsed-nozzle, 114-115
G Gallium trihydride, 121-122 Gas-phase reactive intermediates low pressure, 107-113 supersonic jets, 113-121
546
INDEX
Germanium complexes, xenon fluoride reactions, 85 Globin-derived radicals, leghemoglobin heme-globin crosslinking, 527-530 leghemoglobin reactions, other molecules, 531-533 low-molecular weight species reactions, 530-531 membrane reactions, 533-537 Glutamic acid, posttranslational carboxylation, 466-470 Glutathione, leghemoglobin reaction, Fe(IV)=O, 521-522 Gold, oxidative fluorination, 59 Graphite framework supramolecular complexes, 204-219 GroEL, 162-163 Group 5 metal halide clusters, 1-3, 24 charge-transfer salt complexes, 40-41 chemically rpodified surfaces, 43-44 electronic structure, 31 extended solids, 42-43, 43 halide bridges, 1-2 intercluster bridging, weak, 3 ligand substitution, 27-31 molecular structure, 31-33 oxidation states, 24 redox chemistry, 25-27 supported cluster materials, 39 synthesis, 25 Group 6 metal halide clusters, 1-4 charge-transfer salt complexes, 39-40 electronic structure, 19-20 halide bridges, 1-2 halide-chalcogenide clusters, 14-15 higher nuclearity clusters, 35-37 intercluster bridging, weak, 3 ligand chemistry, axial, 5-13 ligand chemistry, inner, 13-15 mixed halo clusters, 10 molecular structure, 20-24 photophysics, 17-19 redox chemistry, 15-17 supported cluster materials, 37-39 synthesis, 4-5
H Helical framework supramolecular complexes, 204-219
Hemerythrin, four-helix bundle arrangement, 444 Hemoglobin, see also Leghemoglobin; Myoglobin ligand binding, gases, 501-506 plant analogs, 496 Heterometallic polymeric cluster compounds, 258-259 Hexagonal framework supramolecular complexes, 204-219 HIV-1, RNase H enzyme, 307 Hydrogen atoms, scavenging, 121 Hydrogen bonding supramolecular copper(I)/silver(I) complexes, 219-228 transition metal halide clusters, 43 Hydrogen peroxide arginase disproportionation, 326-327 chromium(II1) oxidation, 161 leghemoglobin reaction, Fe(II), 510-511 leghemoglobin reaction, Fe(III), 511517,527 leghemoglobin reaction, Fe(IV)=O, 520 manganese catalase reaction, 322, 414 metLB reaction, 531-533, 534-535 Hydrolases, 306 Hydroxyl radical, 119
I ICP (Calbindin Dgk),444, 452-453 Infinite-chain supramolecular complexes, 251-271 Infrared diode laser spectroscopy, 119, 148 Infrared spectroscopy, time-resolved, 138, 139 ultrafast, 146 Inorganic polymers, 281 Intracellular calcium-binding proteins, 442,443-456 Iridium complexes, xenon fluoride reactions, 89 Iridium vinyl hydride complex, 131-132 Iron carbonyl complexes flash photolysis, 140-142 matrix isolation, 127-128 Iron dihydrides, reactive intermediates, 143-146
547
INDEX
Iron dioxygenases, 312-314 Iron protoporphyrin IX group leghemoglobin, 498, 499 manganese peroxidase, 315-318 Iron ribonucleotide reductases, 319-322, 327; see also Manganese ribonucleotide reductase Iron superoxide dismutase, 310; see also Manganese superoxide dismutase
K Ketone chlorination, manganese complex applications, 403-405 Kochi-Jacobsen-Katsuki epoxidation, 395 Krypton-carbon bonds, 52 Krypton chemistry, 52, 55-61 Krypton difluoride, 55-61 Krypton-fluorine bonds, 52, 55-61 Krypton-nitrogen bonds, 52, 57-58 Krypton-oxygen bonds, 52, 55, 57 KryptonR5recovery, 61
L a-Lactalbumin, calcium binding in, 481-483 Lactose synthase complex, 481 Lanthanides, xenon fluoride reactions, 91 Lanthanum tetrafluorides, 59,91 Lapis lazuli, 122, 123 Laser-induced fluorescence, 156 Laser magnetic resonance, 156-157 Lectins, 308 Leghemoglobin, 496 amino acid sequence homology, 498 carbon monoxide binding, 501-503 globin-derived radicals heme-globin crosslinking, 527-530 leghemoglobin reactions, other molecules, 531-533 low-molecular weight species reactions, 530-531 membrane reactions, 533-537 isomers, 497-498 localization, 500-501 nitric oxide binding, 503-505 reactions, 501-507 structure, 497-499 turnover time, 500-501
Leghemoglobin, Fe(I1) diatomic gas binding, 501-505 nitrogen complexes, 505-506 oxidation, 507-511 small ligands, 506 Leghemoglobin, Fe(II1) NADH reaction, 526 nitrogen complexes, 505-506 oxidation, 511-519 reduction, enzymatic, 524-525 reduction, nonenzymatic, 525-526 small ligands, 506-507 Leghemoglobin, Fe(IV)=0 ascorbate reaction, 520-521 glutathione reaction, 521-522 hydrogen peroxide reaction, 520 membrane reactions, 533-537 reduction, 519-524 thiols reaction, 522-524 Leguminous plants, leghemoglobin in, 496; see also Leghemoglobin Light-induced crystal oscillation, 251-254 Lignin peroxidase, 315-318 Liquid-phase reactive intermediates, see Solution-phase reactive intermediates Lupin leghemoglobin, 498, 502, 516 Lysozymes, calcium binding in, 481-482
M Magnetic susceptibility, 383-384 Main group metals, xenon fluoride reactions, 85-89 Manganese biological role, 305-306, 424-425 coordination chemistry, 309 electronic properties, 380-382 nonredox roles, 306-308 oxidation states, aqueous solution, 379-380 synthetic applications chlorination, 403-405 epoxidation, 394-402 nitrido group transfedaziridination, 402-403 Manganese catalase, 306, 309, 322-325 binuclear structure, 325-327 mimics, 410-419
548
INDEX
modeling, 344 model system reactivity, 410-419 uv-vis spectroscopy, 383 x-ray spectroscopy, 391 Manganese(I1) dioxygenase, 309, 312-314 active site, 315 modeling, 315 Manganese-nitrido complexes, 358-359, 402-403 Manganese peroxidase, 309, 315-318 model system reactivity, 405-407 Manganese-porphyrin complexes, 400-402 Manganese redox enzymes, 305-309; see also Arginase; Oxygen-evolving complex; specific enzymes binuclear, 309, 325-327 catalase mimics, 414-415 EPR spectroscopy, 386-387 manganese-manganese distances, 360 mixed-valent complexes, 365-367 structural models, 359-372 triple bridged, 368 electronic spectroscopy, 382-383 EPR spectroscopy, 385-389 magnetism, 383-385 metalloenzyme analogs, 424-425 model systems, 405-424 mononuclear, Mn(11) eight-coordinate, 350-351 EPR spectroscopy, 385 five-coordinate, 344-345 low spin complex, 351 seven-coordinate, 347-350 six-coordinate, 345-347 mononuclear, Mn(II1) five-coordinate, 352-353 Mn-peroxide moiety, 354 six-coordinate, 353-354 mononuclear, Mn(IV), 354-357 EPR spectroscopy, 385 mononuclear, MnW), 357-359 physical properties, 379-393 polynuclear, 309, 373-379 electronic spectroscopy, 382 EPR spectroscopy, 387-388 magnetism, 383-385
reactivity, 405-424 structural geometry, 315 tetranuclear, 376-379 mixed-valent complexes, 376-377 trinuclear, 373-375 x-ray absorption spectroscopy, 389-393 Manganese ribonucleotide reductase, 306, 309, 319-322 binuclearity, 325-327 hydroxyurea sensitivity, 322 Manganese-salen complexes, 394-400 Manganese superoxide dismutase, 306, 309, 310-312 active site, 315 modeling, 315, 344 model system reactivity, 407-410 monomolecular structure, 344, 347351, 354 pharmaceutical application, 315 Manganese tetrafluoride, 59 Mass spectrometry, electrospray ionization, 162-163 Matrix isolation methods, 122-131 Membrane-binding proteins, calcium-mediated, see Calcium-mediated membrane-binding proteins Membranes, leghemoglobin reactions, 533-537 Mercury complexes, xenon fluoride reactions, 90-91 Mercury titanium sulfide clusters, intercalation, 155 Metal carbonyls fluorides, 89-90 matrix isolation, 126-131 Metal cyanides, 205-206 Metals noble gas combinations, 51 oxidative fluorination, 59 self-assembly, 174-175 Methyl radical, 119 Molecular recognition, calmodulin, 447 Molybdenum chloride, 4-24, 35-44 Molybdenum halide clusters, 1-4 charge-transfer salt complexes, 39-40 electronic structure, 19-20 extended solids, 42-43 halide-chalcogenide clusters, 14-15 higher nuclearity clusters, 35-37
549
INDEX
ligand chemistry, axial, 5-13 ligand chemistry, inner, 13-15 mixed halo clusters, 10 molecular structure, 20-24 photophysics, 17-19 redox chemistry, 15-17 supported cluster materials, 37-39 synthesis, 4-5 Molybdenum hexacarbonyl, 4 matrix isolation, 128 Mononuclear manganese redox enzymes, see Manganese redox enzymes, mononuclear Myoglobin, see also Hemoglobin; Leghemoglobin function observation, 156 hydrogen peroxide reaction, 518-519 ligand binding, gases, 501-506 plant analogs, 495-496 MGY, transition metal halide clusters, see Group 6 metal halide clusters M6Y12transition metal halide clusters, see Group 5 metal halide clusters
N Network structure supramolecular complexes, 273-280 Neuromodulin, 449 Neuron-specific calcium sensor (NCS) proteins, 457 Nickel carbonyl complexes, reactive intermediates, 134-135 Nickel hexafluoride, 59 Nickel superoxide dismutase enzymes, 310 Niobium chloride clusters, 24-33 Niobium halide clusters, 1-3, 24 charge-transfer salt complexes, 40-41 chemically modified surfaces, 43-44 electronic structure, 31 extended solids, 43 ligand substitution, 27-31 molecular structure, 31-33 redox chemistry, 25-27 supported cluster materials, 39 synthesis, 25 Niobium iodide clusters, 2, 33-35 Nitrido group transfer, manganese complex applications, 402-403
Nitrogen, noble gas combinations, 51 Nitrogen dioxide, 138 Nitrogen-fixing root nodules, leghemoglobin in, 495-501, 529-530 Nitrogen oxides, 159 thermal de-NO,, 148-149 NMR spectroscopy, 161-162 Noble-gas chemistry advances in, 51-54 argon chemistry, 52, 54 krypton chemistry, 52, 55-61 radon chemistry, 52-53, 54, 91-93 review literature, 53-54 xenon chemistry, see Xenon chemistry Noble gases, 51-54 matrix isolation application, 122 reactive intermediate solution study application, 132 Notch receptor proteins, 473, 475
0 Oligophenanthrolines, 176 Oligopyridines, 175, 176 Organometallic polymers, 254-258 Osmium complexes, xenon fluoride reactions, 90 Osteocakin (SPARC), 484 Oxygen, noble gas combinations, 51 Oxygen-binding proteins, 495-496 Oxygen-evolving complex, 306, 309, 327-343 calcium requirement, 328, 338-339 catalase-like reaction, 411-419 chloride requirement, 328, 329, 338339,393 EPR spectroscopy, 388-389 manganese requirement, strict, 425 modeling, 344 oxidation states, 330-339 photosynthesis, 327-330 spectroscopy, 331-339, 342 structure, 327-330, 337-342 Tyrosine Z residue, 330, 338 water oxidation, 339, 342-343 water oxidation model system, 419-424 x-ray spectroscopy, 391-393 Ozone depletion, 109-110
550
INDEX
P Parasponia spp., leghemoglobin in, 496 Parkinson’s disease, peroxynitrite implicated, 402 Parvalbumin, 443, 445 Peribacteroid membranes, 533-537 Peroxotitanium, cerium(1V) reaction, 160-161 Peroxynitrite diseases implicated in, 310-311, 402 pulse studies, 138 Perxenate ion, 65 Perxenic acid, 67 Phenazine-bridged supramolecular complexes, 214-219 Phosphohydrolases, 306 Phosphoinositide-specific phospholipase C, 460, 464-465 Phospholipase A,, 442 Phosphorus complexes, xenon fluoride reactions, 86 Photosynthesis, 327-330, 495; see also Oxygen-evolving complex Photosystem I, 329 Photosystem 11, 328 8-8 Interaction framework supramolecular complexes, 228-239 Plant globin gene family, 498-499 Platinum, xenon fluoride reactions, 91 Polycatenane complexes, 271-273 Polynuclear manganese redox enzymes, see Manganese redox enzymes, polynuclear Polyrotaxene complexes, 271-273 Porphyrin-based manganese complexes, 400-402 Potassium pseudohalide clusters, 42 Products, of reaction, 101 Protein C, calcium binding, 466, 467, 473, 479-481 Protein-film voltammetry, 163 Protein kinase C, calcium2+-dependent, 460, 462-463 Protein kinase I, calmodulin-dependent, 448 Protein kinase 11, calmodulin-dependent, 447 Proteins calcium-binding, see Calcium-binding proteins
electrospray ionization mass spectrometry, 162-163 function observation, 155-156 manganese structural support, 305-306 oxygen-binding, 495-496 residue radicals, enzymes, 340 Protein S, calcium binding, 466, 474475, 479 Prothrombin, calcium binding, 466, 469 Prussian Blue, 41 Pulsed laser photolysis, 141 Pulsed-nozzle Fourier-transform microwave spectroscopy, 114-115 Pulse radiolysis, 137-138 Pump-and-probe techniques, 137 Pyrazine-bridged supramolecular complexes, 206-213, 240-243 Pyridine-bridged supramolecular complexes, 240-243 Pyruvate carboxylase, 306
Q QR1, 484
R Radon chemistry, 52-53, 54.91-93 Radon fluorides, 52, 91-93 Radon oxide, 91 Raman spectroscopy, time-resolved resonance, 138, 139, 143 Rare earth halide clusters, 2-3 Reactants, 101 Reaction rate, 101-102 Reactive intermediates, 101-107, 164 flow characterization methods, 156-164 gas-phase studies, 107-121 lifetimes, 106 retardation characterization methods, 107 gas-phase studies, low pressure, 107-113 gas-phase studies, supersonic jets, 113-121 matrix isolation, 122-131 solution studies, low temperature, 131-136 trapping, 121-122
INDEX
solid state studies, 154-156 solution-phase studies, 131-136, 159-160 time-resolved characterization methods, 136-156 Recoverin, calcium binding in, 457 Redox enzymes, see Manganese redox enzymes Resonance fluorescence, 156 Resonantly enhanced multiphoton ionization (REMPI), 147 Retardation reactive intermediate characterization, see Reactive intermediates, retardation characterization methods Retroviral reverse transcriptases, 307-308 Reverse transcriptases, 307-308 Rhenium complexes reactive intermediates, 136 xenon fluoride reactions, 90 Rhczobiurn spp., leghemoglobin in, 496, 499-500; see also Leghemoglobin Rhodium carbonyl complexes, reactive intermediates, 165-166 Ribonuclease enzymes, 307-308 Ribonucleotide reductases, 319; see also Manganese ribonucleotide reductase RNase H, HW-1, 307 Root nodules, leghemoglobin in, 495-501, 529-530 Rubidium pseudohalide clusters, 42 Ruthenium complexes matrix isolation, 128 xenon fluoride reactions, 90 Ruthenium dihydrides, reactive intermediates. 143-146
S So state, oxygen-evolving complex, 330, 333-334, 336-338 x-ray spectroscopy, 393 S, state, oxygen-evolving complex, 330, 331-334, 336-338 EPR spectroscopy, 389 S2state, oxygen-evolving complex, 330, 332-333,336-338 EPR spectroscopy, 388-389 S, state, oxygen-evolving complex, 330, 336-338
55 1
S, state, oxygen-evolving complex, 330, 336-338 S100A6,454-456 SlOOB, 453-454,455 SlOOD, 444,452-453 sloop, 453-454 S l O O proteins, calcium binding, 451-456 Salen-based manganese complexes, 394-400 SC1, 484 S cycle, oxygen-evolving complex, 330-339 Secreted Protein Acidic and Rich in Cysteine (SPARC), 484-485 Selenium complexes, xenon fluoride reactions, 87 Self-assembling metal complexes, 174175; see also Supramolecular copper( I)/silver(I)complexes Serine proteases, calcium binding in, 479-481 Serum albumin, 470 Silane discharges, 119-121 Silicon complexes, xenon fluoride reactions, 85 Silicon dichloride, 159 Silicon oxide, reactive intermediates, 108 Silver(I), 174-175 SilverU) supramolecular complexes, 174-175 channels, 222-227, 283-289,291-292 diamondoid frameworks, 240-251 graphite frameworks, 204-219 helical frameworks, 176-192 hexagonal frameworks, 204-219 hydrogen bond-assembled frameworks, 219-228 infinite-chain structures, 251-271 n-n interaction-assembled frameworks, 228-239 self-assembly pathway, 175 stereochemistry, 174-175 stoichiometries, 176 sulfur-sulfur contact-assembled frameworks, 192-204 three-dimensional structures, 283-292 two-dimensional structures, 271-283 Silver titanium sulfide clusters, intercalation, 155 Silver trifluoride, 59-60
552
INDEX
T
Smooth muscle light-chain kinase, 447-448
Solid-state reactive intermediates, 154-156
Solid-state supramolecular complexes, see Supramolecular copper(I)/silver(I) complexes Solution-phase reactive intermediates flow systems, 159-160 low temperature, 131-136 Solvent effects, reactive mechanisms and, 102 Soybeans, leghemoglobin in, 495-501; see also Leghemoglobin SPARC (Secreted Protein Acidic and Rich in Cysteine), 484-485 S l O O proteins, calcium binding, 451-456 Subtilisins, calcium binding, 481 Sulfur scavenging, 121 xenon fluoride reactions, 86-87 Sulfur-sulfur contact framework supramolecular complexes, 192-204 Supercritical fluids, reactive intermediate characterization, 135, 140 Superoxide, 310 destructive effects, 310-311 leghemoglobin reaction, 508-509 Superoxide dismutase enzymes, 310; see also Manganese superoxide dismutase Supramolecular copper(I)/silver(I) complexes, 174-176, 292-293 channels, 222-227,283-289,291-292 diamondoid frameworks, 240-251 graphite frameworks, 204-219 helical frameworks, 176-192 hexagonal frameworks, 204-219 hydrogen bond-assembled frameworks, 219-228
infinite-chain structures, 251-271 a-8interaction-assembled frameworks, 228-239
self-assembly pathway, 175 stereochemistry, 174-175 stoichiometries, 176 sulfur-sulfur contact-assembled frameworks, 192-204 three-dimensional structures, 283-292 two-dimensional structures, 271-283 Synaptotagmin, 460, 464-465
Tantalum bromide clusters, 26 Tantalum chloride clusters, 24-33 Tantalum halide clusters, 1-3, 24 chemically modified surfaces, 44 electronic structure, 3 1 extended solids, 43 ligand substitution, 27-31 molecular structure, 31-33 redox chemistry, 25-27 synthesis, 25 Tellurium complexes, xenon fluoride reactions, 87-88 Tetrakis(a1kylthio)-tetrathiafulvanebased supramolecular complexes, 196-200
Tetranuclear manganese redox enzymes, 376-379
Tetrathiafulvane derivatives, 175, 192 Thiols, leghemoglobin reaction, Fe(IV)=O, 522-524 Time-resolved characterization methods, reactive intermediates, 136-156 Time-resolved infrared spectroscopy, 138, 139, 146
Time-resolved resonance Raman spectroscopy, 138, 139, 143 Titanium, peroxotitanium-cerium(1V) reaction, 160-161 Transition metals halide clusters, see Group 5 metal halide clusters; Group 6 metal halide clusters root nodule damage initiator, 534 xenon fluoride reactions, 89-91 Transition states, 102 Trapping, reactive intermediates, 121-122
Trimethyleamine, thermal decomposition, 110-112 Trinuclear manganese redox enzymes, 373-375
Trolox C, 530 Troponin C, 450-451 Troponin I, 450 Troponin T, 450 Trypsin, calcium binding, 479 Trypsinogen, calcium binding, 479 TTC,-TTF-based supramolecular complexes, 196-200
553
INDEX
Tungsten carbonyl complexes, 4, 152-154 Tungsten chloride, 4 Tungsten complexes matrix isolation, 128 reactive intermediates, 152-154 Tungsten halide clusters, 1-4 electronic structure, 19-20 halide-chalcogenide clusters, 15 ligand chemistry, axial, 5-13 ligand chemistry, inner, 13-15 mixed halo clusters, 10 molecular structure, 20-24 photophysics, 17-19 redox chemistry, 15-17 synthesis, 4-5 Tungsten hexacarbonyl, 4 Tungstenocene, 125, 126
U Ultrafast time-resolved infrared spectroscopy, 146 Ultraviolet-visible absorption, 138-140 Ultraviolet-visible emission, 138
V Vanadium carbon hydrogen radical, 118
W Water oxidation complex, 328, 339, 342-343 model system, 419-424 White rot fungi, 315-316
X Xenate ion, 65 Xenic acid, 65, 67 Xenon-boron bonds, 54 Xenon-carbon bonds, 54, 81-84 Xenon cations, 68-72 Xenon chemistry, 52 advances in, 61-64 literature review, 53
Xenon difluoride, 58-59, 61-64 ammonolysis, 78 fluoride-ion donor chemistry, 69-70 iodide reactions, 88-89 metathesis reactions, anhydrous acids, 75-76, 77 physical properties, 64 Xenon fluorides, 52, 61-64 anions from, 72-75 cations from, 68-72 fluoride-ion donor/acceptor reactions, 68-72 fluorinating agents, 84-91 hydrolysis reactions, 65-67 metathesis reactions, anhydrous acids, 75-84 oxidizing/fluorinating strengths, 85 Xenon hexafluoride, 61, 62-64 ammonolysis, 78 fluoride-ion acceptor chemistry, 72-74 fluoride-ion donor chemistry, 70-71 physical properties, 64 Xenon-nitrogen bonds, 54, 78-81 Xenon oxide fluorides, 66-67 fluoride-ion acceptor chemistry, 74-75 fluoride-ion donor chemistry, 71-72 Xenon oxides, 65-66 Xenon(I1) pentafluoroorthotellurates, 76-78 Xenon-platinum hexafluoride, 5 1-52 Xenon tetrafluoride, 61-62 ammonolysis, 78 fluoride-ion acceptor chemistry, 74 fluoride-ion donor chemistry, 70 physical properties, 64 Xenon tetraoxide, 66 Xenon trioxide, 65-66 X-ray absorption near edge spectroscopy (XANES), 390 2
Zeolite-like supramolecular complexes, 274-277 Zeolites krypton difluoride decomposition, 61 noble gases in, 51 Zirconium halide clusters. 3
This Page Intention ally Left Blank
CONTENTS OF PREVIOUS VOLUMES
VOLUME 36 Inorganic Chemistry and Drug Design Peter J. Sadler Lithium and Medicine: Inorganic Pharmacology N . J . Birch and J. D. Phillips The Mo-, V-, and Fe-Based Nitrogenase Systems of Azobacter Robert R. Eady The Extraction of Metals from Ores Using Bacteria D. Keith Ewart and Martin N . Hughes Solid-state Bioinorganic Chemistry: Mechanisms and Models of Biomineralization Stephen Mann and Carole C. Perry Magnetic Circular Dichroism of Hemoproteins M. R. Cheesman, C. Greenwood, and A. J. Thomson
The Uptake, Storage, and Mobilization of Iron and Aluminum in Biology S. Jemil, A. Fatemi, Fahrni H. A. Kadir, David J. Williamson, and Geoffrey R. Moore Probing Structure-Function Relations in Ferritin and Bacteriofenitin P. M. Harrison, S. C. Andres, P. J. Artymiuk, G. C. Ford, J . R. Guest, J . Hirzmann, D. M. Lawson, J . C. Livingstone, J. M. A. Smith, A. Treffrey, and S. J. Yewdall INDEX
VOLUME 37 On the Coordination Number of the Metal Crystalline HalogenocupratedI) and Hal ogenoargentates(1) Susan Jagner and Garan Helgesson
Flavocytochrome b, Stephen K. Chapman, Scott A. White, and Graeme A. Reid
Structures of Organonitrogen-Lithium Compounds: Recent Patterns and Perspectives in Organolithium Chemistry Karina Gregory*Paul von Rague Schlqyer, and Ronald Snaith
X-Ray Absorption Spectroscopy and the Structures of Transition Metal Centers in Proteins C. David Garner
Cubane and Incomplete Cubane-Type Molybdenum and Tungsten 0x01 Sulfido Clusters Takashi Shibahara
Direct Electrochemistry of Proteins and Enzymes Liang-Hong Guo and H. Allen 0. Hill
Interactions of Platinum Amine Compounds with Sulfur-Containing Biomolecules and DNA Fragments Edwin L. M. Lempers and J a n Reedijk
Active-Site Properties of the Blue Copper Proteins A. G. Sykes
Recent Advances in Osmium Chemistry Peter A. Lay and W. Dean Harnian 555
556
CONTENTS OF PREVIOUS VOLUMES
Oxidation of Coordinated Diimine Ligands in Basic Solutions of Tris(diimine)iron(III),ruthenium(III), and -osmium(III) 0. M0nsted and G. Nord INDEX
VOLUME 38 Trinuclear Cuboidal and Heterometallic Cubane-Type Iron-Sulfur Clusters: New Structural and Reacticity Themes in Chemistry and Biology R. H. Holm Replacement of Sulfur by Selenium in Iron Sulfur Proteins Jacques Meyer, Jean-Marc Moulis, Jacques Gaillard, and Marc Lutz Dynamic Electrochemistry of Iron-Sulfur Proteins Fraser A. Armstrong EPR Spectroscopy of Iron-Sulfur Proteins Wilfred R. Hagen
VOLUME 39 Synthetic Approach to the Structure and Function of Copper Proteins Nobumasa Kitcy'ima Transition Metal and Organic RedoxActive Macrocycles Designed to Electrochemically Recognize Charged and Neutral Guest Species Paul D. Beer Structure of Complexes in Solution Derived from X-Ray Diffraction Measurements Georg Johansson High-Valent Complexes of Ruthenium and Osmium Chi-Ming Che and Vivian Wing-Wah Y a m Heteronuclear Gold Cluster Compounds D. Michael P. Mingos and Michael J. Watson
Structural and Functional Diversity of Ferredoxins and Related Proteins Hiroshi Matsubara and Kazuhiko Saeki
Molecular Aspects on the Dissolution and Nucleation of Ionic Crystals in Water Hitoshi Ohtaki
Iron-Sulfur Clusters in Enzymes: Themes and Variations Richard Cammack
INDEX
Aconitase: An Iron-Sulfur Enzyme Mary Claire Kennedy and C. David Stout Novel Iron-Sulfur Centers in Metalloenzymes and Redox Proteins from Extremely Thermophilic Bacteria Michael W. W. Adams Evolution of Hydrogenase Genes Gerrit Voordoww Density-Functional Theory of Spin Polarization and Spin Coupling in Iron-Sulfur Clusters Louis Noodleman and David A. Case INDEX
VOLUME 40 Bioinorganic Chemistry of PterinContaining Molybdenum and Tungsten Enzymes John H. Enemark and Charles G. Young Structure and Function of Nitrogenase Douglas C. Rees, Michael K. Chan, and Jongsun Kim Blue Copper Oxidases A. Messerschmidt Quadruply Bridged Dinuclear Complexes of Platinum, Palladium, and Nickel Keisuke Umakoshi and Yoichi Sasaki
CONTENTS OF PREVIOUS VOLUMES
Octacyano and 0x0- and Nitridotetracyano Complexes of Second and Third Series Early Transition Metals Johann G. Leipoldt, Stephen S. Basson, and Andreas Roodt Macrocyclic Complexes as Models for Nonporphine Metalloproteins Vickie McKee Complexes of Sterically Hindered Thiolate Ligands J. R. Dilworth and J . H u INDEX
VOLUME 41 The Coordination Chemistry of Technetium John Baldas Chemistry of Pentafluorosulfanyl Compounds R. D. Verma, Robert L. Kirchmeier, and Jean’ne M. Shreeve The Hunting of the Gallium Hydrides Anthony J. Downs and Colin R. Pulham The Structures of the Group 15 Element(II1) Halides and Halogenoanions George A. Fisher and Nicholas C. Norman Intervalence Charge Transfer and Electron Exchange Studies of Dinuclear Ruthenium Complexes Robert J. Crutchley Recent Synthetic, Structral, Spectroscopic, and Theoretical Studies on Molecular Phosphorus Oxides and Oxide Sulfides J . Clade, F. Frick, and M. Jansen Structure and Reactivity of Transferrins E. N . Baker INDEX
557
VOLUME 42 Substitution Reactions of Solvated Metal Ions Stephen F. Lincoln and Andre E. Merbach Lewis Acid-Base Behavior in Aqueous Solution: Some Implications for Metal Ions in Biology Robert D. Hancock and Arthur E. Martell The Synthesis and Structure of Organosilanols Paul D. Lickiss Studies of the Soluble Methane Monooxygenase Protein System: Structure, Component Interactions, and Hydroxylation Mechanism Katherine E. Liu and Stephen J. Lippard Alkyl, Hydride, and Hydroxide Derivatives in the s- and p-Block Elements Supported by Poly(pyrazoly1)borato Ligation: Models for Carbonic Anhydrase, Receptors for Anions, and the Study of Controlled Crystallographic Disorder Gerard Parkin INDEX
VOLUME 43 Advances in Thallium Aqueous Solution Chemistry Julius Glaser Catalytic Structure-Function Relationships in Heme Peroxidases A n n M. English and George Tsaprailis Electron-, Energy-, and Atom-Transfer Reactions between Metal Complexes and DNA H. Holden Thorp Magnetism of Heterobimetallics: Toward Molecular-Based Magnets Olivier Kahn
558
CONTENTS OF PREVIOUS VOLUMES
The Magnetochemistry of Homo- and Hetero-Tetranuclear First-Row d-Block Complexes Keith S. Murray Diiron-Oxygen Proteins K. Kristoffer Andersson and Astrid Graslund Carbon Dioxide Fixation Catalyzed by Metal Complexes Koji Tanaka INDEX
VOLUME 44
Recent Developments in Chromium Chemistry Donald A. House INDEX
VOLUME 45 Syntheses, Structures, and Reactions of Binary and Tertiary Thiomolybdate Complexes Containing the (O)Mo(S,) and (S)Mo(S,) Functional Groups (x = 1, 2, 4) Dimitri Coucouvanis
Organometallic Complexes of Fullerenes Adam H . H. Stephens and Malcolm L. H. Green
The Transition Metal Ion Chemistry of Linked Macrocyclic Ligands Leonard F. Lindoy
Group 6 Metal Chalcogenide Cluster Complexes and Their Relationships to Solid-state Cluster Compounds Taro Saito
Structure and Properties of Copper-Zinc Superoxide Dismutases Ivano Bertini, Stefan0 Mangani, and Maria Silvia Viezzoli
Macrocyclic Chemistry of Nickel Myunghyun Paik Suh Arsenic and Marine Organisms Kevin A. Francesconi and John S. Edmonds
DNA and RNA Cleavage by Metal Complexes Geneufve Pratviel, Jean Bernadou, and Bernard Meunier
The Biochemical Action of Arsonic Acids Especially as Phosphate Analogues Henry B. F. Dixon
Structure-Function Correlations in High Potential Iron Problems J . A. Cowan and Siu Man Lui
Intrinsic Properties of Zinc(I1) Ion Pertinent to Zinc Enzymes Eiichi Kimura and Tohru Koike
The Methylamine Dehydrogenase Electron Transfer Chain C. Dennison, G. W. Canters, S. de Vries, E. Vijgenboom, and R. J. van Spanning
Activation of Dioxygen by Cobalt Group Metal Complexes Claudio Bianchini and Robert W. Zoellner
INDEX
This Page Intention ally Left Blank
I S B N O-L2-023646-X