No title

INTERNATIONAL REVIEW OF

Neurobiology VOLUME 20

Associate Editors W. R. ADEY

H . J. EYSENCK

D. BOVET

C. HEBB

JOSE

DELGADO

S. KETY

SIR JOHN ECCLES

A. LAJTHA

0. ZANGWILL

Consultant Editors V. AMASSIAN

K. KILLAM

R. BALDESSARINI

C. KORNETSKY

F. BLOOM

B. A . LEBEDEV

P. B. BRADLEY

V. LONGO

0. CREUTZFELDT

P. MANDELL

J. ELKES

H . OSMOND

K. FUXE

S. H . SNYDER

B. HOLMSTEDT

S. SZARA

P. JANSSEN

W. GREYWALTER

INTERNATIONAL REVIEW OF

Neurobiology Edited by JOHN R. SMYTHIES Department of Psychiatry and the Neurosciences Program University of Alabama Medical Center Birmingham, Alabama

RONALD J. BRADLEY The Neurosciences Program University of Alabama Medical Center Birmingham, Alabama

VOLUME 20

1977

ACADEMIC PRESS

New York

San Francisco London

A Subsidiary of Harcourt Brace Jovanovich, Publishers

COPYRIGHT @ 1977, BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED. NO PART O F 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.

ACADEMIC PRESS, INC.

11 1 Fifth Avenue, New York, New York 10003

United Kiirgdom Edition piiblislied by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Koad, London NWI

LIBRARY OF CONGRESS CATALOG CARD NUMBER:59-13822 ISBN 0-12-366820-4 PRINTED IN THE UNITED STATES OF AMERICA

CONTENTS CONTRIBVTORS ..............................................................

ix

Functional Metabolism of Brain Phospholipids

G. BRIAN ANSELLA N D SHEILA SPANNER I . Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

........ ..............

11. Structure and Intracellular Distribution of Brain l'hospholipids

111. Origin and iMetabolism of' Brain Phospholipids . . . . IV. Relationship between Phospholipid Composition of Membranes .............. and Brain Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Phospholipids and Synaptic Transmission ...................... .... References . . . . . . . . . . . ......................................

1 2 6

]I 17 26

Isolation a n d Purification of the Nicotine Acetylcholine Receptor a n d Its Functional Reconstitution into a M e m b r a n e Environment

MICHAELs. BRILEYAND JEAN-PIERRE CHANGELIX I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............ 11. Isolation and I'iirification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 111. Reconstitution . . . ..................................................

I\'.

Conclusion. .......................... ..................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31 32 .+I) .iX

3)

Biochemical Aspects of Neurotransmission in the Developing Brain JOSEPH

T. COYLE

I . Introduction.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6.5

11. Prenatal Development of Central Catecholaminergic Neurons . . . . . . . . . . . . . I I I . Postnatal Development of the Nigrostriatal Circuit .......................

67

IV. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

97 99

X2

The Formation, Degradation, a n d Function of Cyclic Nucleotides i n the Nervous System JOHN

W. DALY

................................... . . . . . . . . . . . . . . . . 105

I. Introduction

................................... III. Cyclic AMP ........................................................... 111. CyclicGMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Cyclic Nucleotides and the Function 01' the Central and Peripheral Nervous System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. S u m m a r y . . . . . . . . . ......................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Y

109 13X

1.44 1.x 156

CONTENTS

vi

Fluctuation Analysis in Neurobiology

LOLllSJ. DEFELlCE

.................. ................... ........................ .................. .................. ............... I l l . Results . . . . . . . . . . . . . . ..................... ................... IV. Summary .. . . . . . ........................................ References ............................. I. Introduction . .

11. Methods

169 175 183 206 206

Lipotropin and the Central Nervous System

w. H . G 1 S P E N . J . M. V A N REE,A N D D. DE W l E D I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . 2 09

11. ACTH 4-10 . ............................ 211 . . . . . . . . 230 111. P-MSH . . . . . . .................................. ............................ 232 IV. P-Lipotropin 61–91 ..................................

V. Concluding Remarks .......................... References .............................

. . . . . . . . . . 239

...............

242

Tissue Fractionation in Neurobiochernistry: An Analytical Tool or a Source of Artifacts

PIERRE LADURON I. 11. Ill. IV.

Introduction .......................................................... Analytical Approach to Tissue Fractionation in the Brain . . . . . . . . . . . . . . . . . Interpretation of Tissue Fractionation Studies ............................ Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

25 1 253 269 280 281

Choline Acetyltransferase: A Review with Special Reference to Its Cellular and Subcellular Localization JEAN

ROSSIER

............. ..................................... .................... .............................

I. Introduction 11. History

III. Assay IV. Distribution of ChAc in Nonneuronal Tissue'. . . . . . . . . . . . . . . . . . V. Distribution of ChAc in Neuronal Tissue . . . . . ...................... V I . Purification of ChAc . . . . . VII. Biophysical Studies . . . . . . . . . . . . . . . . . ......................... VIII. Mechanism of Action . . . . . . . . . . . . . . . . . . . .......................... ............................. IX. Axonal Transport of ChAc . X. Immunology.. ......................... ...................... XI. Localization at the Cellular Level ..................... ............ XII. Subcellular Localization of ChAc ...............................

284 284 287 29 1 294 296 303 304 312 314 318 324

CONTENTS

vii

XI11 . Choline Transport and ChAc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIV . The Role of CI- in the Regulation of Ach Synthesis by ChAc . . . . . . . . . . . . . . X V . Pleiotropic Effect of Nerve Impulses on Ach Synthesis .................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

327 329 330 331

SCBJECT INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CONTENTS O F PREVIOLIS VOLUMES.............................................

339 345

This Page Intentionally Left Blank

C O NTRl B UTO RS Numbers in parentheses indicate the pages on which

the iiiitliors'

contributions begin

G. BRIANANSELL, Department of Pharmacology, Unirvrsity of Birmingliam, Medicul School, Birmingham, England ( 1 )

MICHAELS. BRILEY,Neurobiologie Moliculuire, Institut Pasteur, Puris, France (31) JEAN-PIERRE C H A N G E U X , Neurobiologze M o l h l a i r e , Insthit Pasteur, Paris, Frtim-e ( 3 1 ) T. COYLE, Departments of Pharmacology and Expm'mental Ttrera/writics and Psychiatry and Belimiorul Sciences, The Johns Hopkins Unizwrsity School of Medicine, Baltimore, Maryland (65)

JOSEPH

JOHN W. DALY,Laboratory of Chemistry, National Institute of Arthritis, Metabolism, and Digestive Disetises, National Institutes of Health, Bethesda, Mnryland ( 105) LOUISJ. DEFELICE, Department of Aniitomy, Emory Uniuersity, Atlanta, Georgza (169)

D. DE WIED,Rud0y'Mugnu.s Institute fM- Phamiacology, Medical Facul(y, Uniriersity of Utrecht, Utrecht, Tlie Netherlaiids (209) W. H . G I S P E N , Rudolf Magnus Institute ,for Pharmacology, Medical Faculty, Uniuersity of Utreckt, Utrecht, TIE Netherlands (209)

PIERRELADURON, Departvnmt of Biochemical Plinrmacology, Jan.s.sen Phcrrmuceutica, Beerse, BelgEu,m ( 2 51) JEANROSSIER,Tlie Salk Institute for Biological Studies, Ln Jolla, California (283) SHE1 LA SPANNER, Department of Pharmacology, I/ nii~ersity of Birmingham, Medical School, Birmingham, E n g l m d ( 1

J. M.

VAN REE, Rudolf Magnus h t i t u t e for Phnrmacology, Medical Faulty, Uni-r)ersityof Utrecht, Utrecht, The Netiierlands (209)

iX

This Page Intentionally Left Blank

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS By G. Brian Ansell and Sheila Spanner

Department of Pharmacology Univeniiy of Birmingham, Medical School Eirming ham, England

I. Introduction I I . Structure and Intracellular Distribution o f Brain Phosphdipids A. Structure . B. Distribution 111. Origin and Metabolism of Brain Phospholipids A. Transport to the Brain and the Origin of Precursors B. Dr Norm Synthesis-Neurones and Glia . C. Dc Nouo Synthesh-Subcellular Fractions 1V. Relationship between Phospholipid Camposition of Membranes and Brain Function A. Naturally Occurring Modifications of Phospholipid Composition B. Experimental Modifications of Phospholipid Composition C. Effects of Anesthetics V. Phospholipids and Synaptic Transmission A. Relationship Between Lipid-Bound Choline and Acetylcholine Formation in the Brain B, Effects of Neurotransmitters on Phospholipid Metabolism References

11 11 14 16 17

17 20 26

1. Introduction

This brief chapter is intended to cover recent developments in which metabolic processes involving phospholipids may be related to mechanisms important in the functioning of the brain. In general, basic information about the structure and metabolism of brain phospholipids can be found in the relevant chapters in the treatise edited by Ansell et al. (1973) and the review by Ramsey and Nicholas (1972). Such information will not be repeated here in detaiI. More details on the relationship between phospholipid metabolism and function in the nervous system can be found in the proceedings of a recent meeting (Porcellati et al., 1976). 1

2

G . BRIAN ANSELL A N D SHEILA SPANNER

It is unlikely that any new phospholipid structures will be found in the nervous system as a whole, and modern analytical techniques have made possible the identification of the molecular species of all those present. Currently, interest is focusing on, the phospholipid composition and metabolism of relatively homogeneous structural components of the nervous system, for example, myelin and the plasma membranes of nerve endings which are more likely to yield useful information and have, in fact, done so. It is by no means clear, however, why different membranous components of the nervous system have a different phospholipid composition. This comment applies, of course, to other tissues and the rationale for different phospholipid “species” is often totally obscure. The one certainty is that some membrane-bound enzymes need specific phospholipids in order to function normally (Coleman, 1973). What is clear is that when the metabolism of a phospholipid is abnormal then function can be seriously disturbed. In this way an abnormal metabolism may give a clue to normal function, as it has done for other tissue components.

II. Structure and lntracellular Distribution of Brain Phospholipids

A. STRUCTURE In the brain there are four classes of phospholipids; the first three have a glycerol backbone with a fatty acid, usually unsaturated, in the 2-position and a phosphorylated base (choline, ethanolamine, serine) or inositol in the %position. In the l-position is a fatty acid (diacylphospholipid) ( l ) , a long-chain aliphatic O-alk- l-enyl (unsaturated ether) moiety (plasmalogen) (2), or a long-chain aliphatic O-afkyl (saturated ether) moiety (3). Only choline and ethanolamine plasmalogens have been found and only ethanolamine-containing phospholipids with a saturated ether. The fourth class, of which the only representative is ceramide phosphorylcholine (sphingomyelin), contains a long-chain base, sphinganine, linked to phosphorylcholine through its primary hydroxyl group and with its amino group acylated by a long-chain fatty acid (4). There are also small but significant amounts of cardiolipin (phosphatidylglycerol phosphoglyceride) in which phosphatidylglycerol is linked to the basic phosphatidyl unit making it diphosphatidylglycerol (1). The chemistry of phospholipids has been clearly described by Strickland (1973).

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

3

0

I

X = H , choline, ethanolamine, serine, inositol, inositol monophosphate, inositol diphosphate, or phosphatidylglycerol. RCOO--, CI,-C2, acyl group (mainly saturated; some may be odd numbered or branched) R'COO--, CI4-C*2acyl group (unsaturated predominate)

H,C-OCR'

P I

B

RiCO H2C--O-P-O-X TH II

OH

1,2-diacyl-m-glycero-3-phosphoryl-X H H H2C-O<=CR'

(1)

X = choline, ethanolamine R'COO--, CI4-C2*acyl group (unsaturated predominate) R:'CH=CHO--, 1-alk-1-enyl ( C I or ~ Cln)

7 1

R*co7H i; H2C--O-P4-X

I

OH

I-alk- 1 '-enyl-2-acyl-\n-glycero-3-phosphorylcholine (ethanolamine) (2) HZC4CHZR' R'COO--, CI4-C22acyl group (unsaturated predominate) R'CH20-, long-chain alkyl (CI6,CIS) 0 1I

P I

R2c07H

H2C-O-P-O-ethanolamine

I

OH 1-alkyl-2-acyI-.~n-glycel-o-3-phosphorylethanolan~ine

(3)

CH3(CH2)&H
I

I I C=O I. R.'

OH NH

R"C0--, acyl group

JV-acyl-prythro-sphingosine1 -phosphorylcholine

(4)

B. DISTRIBUTION These phospholipids are found throughout the nervous system. The only one showing any specificity is cardiolipin which appears to be unique to mitochondria and structures which contain them, e.g., neurones (Table I). Apart from cardiolipin, the distribution of the various classes of phospholipids in neurones and glia seems to be very similar

4

G. BRIAN ANSELL AND SHEILA SPANNER

(Table I). If one compares the data in Tables I1 and 111it can be seen that the plasmalogens are of highest concentration in membranes, particularly in myelin, while phosphatidylcholine is the chief phospholipid of the synaptic vesicles. There are, however, differences in the fatty acid content of the various phospholipids particularly between the synaptosomal phospholipids and those of the myelin (Kishimotoet al., 1969).Thus, in both the synaptosomal plasma membranes (Breckenridge et al., 1972) and in the synaptic vesicles (Breckenridge et al., 1973), 33-48% of the fatty acids of phosphatidylserine, phosphatidylethanolamine, and phosphatidylinosito1 are polyunsaturated, mainly C20:4(n-6)and c22:6 (n-3). In myelin, the unsaturated fatty acids in these phospholipids are largely Another interesting difference is seen in sphingomyelin composition. I n synaptosomes the phospholipid contains almost exclusively C,,:, fatty acid while that of myelin contains 48% as and some C24:o. The composition of the phosphatidylcholine is similar in both structures with 80% of the fatty acids C16:O or CIgtl.

TABLE I PHOSPHOLIPID COMPOSITION OF NEURONES AND GLIA

Neurones

Rat"." 55'0 -

Phosphatidylcholine Choline plasmalogen

37'0 1.6

Ethanolamine Phosphatidylethanolamine plasmalogen Phosphatidylinositol Phosphatidylserine Sphingomyelin Cardiolipin

13.9 ""}25.2 2.9 6.8 11.7 5.4 5.6 4.4 5.7 -

Glia (astruglia)

Mouse" Rabbit"

} 58.9 } 46.8 } 23.1 } 32.2 11.9 1.6 3.4 -

5.6 7.5 7.9

-

Rat"." 3::,0}51.0 ::::}28.4 3.0 11.6 5.2

-

4.9 7.3 5.2

-

N o t e Measurements in percent of total phopholipid P in each fraction. " Freysz et al. (1968). " Norton and Poduslo (1971). ' Kohlschutter and Herschkowitz (1973). " Hamberger and Svennerholm (1971). Poduslo and Norton (1972). Fewster and Mead (1968). I'

'

Oligodendroglia (white matter)

Ox'.'

Rabbit"

} 42.6) } 37.8} 3.2 7.8 8.8 -

48

29

24

} 36

5.0 8.0 9.0

-

2.0 21.0 11.0 -

5

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

TABLE I 1 PHOSPHOLIPID COMPOSITION OF MYELIN A N D AXONS Myelin

Mouse" Phosphatidylchqline Choline plasmalogen Phosphatidylethanolaniine Ethanolamine plasmalogen Phosphatidylinositol Phosphatidylserine Sphingom yelin Phosphatidic acid

37 -

19 26

11.6 6.3

-

Rat"

Guinea pig","

32.4 0.5 16.1 32.4 3.0 X.5 7.1 I .5

29.0 0.5 11.5 25.0 3.0 13.5 9.7 1.5

}

Axon Ox"

Squirrel monkey'

Ox'

27,9

24.0

27.5 3 .o 10.7

25.5

14.6 16.6 1.9

2.0 14.0 16.0

5.7 9.3 15.5

-

-

Note: Measurements in percent of total phospholipid P in each fraction. " Singh et al. (197 1). '' Ansell ef al. (1973). ' Whittaker (1966). " Fleischer and Rouser (1965). 'Sun and Sun (1972). Poduslo and Norton (1972). deVries and Norton (1974). I'

TABLE 111 PHOSPHOLIPID COMPOSITION OF NERVE ENDINGS A N D THEIR MEMBRANESA N D VESICLES Sy naptosomal membranes

Intact nerve endings

Rat" Phosphatidylcholine Choline plasmalogen Phos p h atidylet h anolamine Ethanolamine plasmalogen Phosphatidylinssitol Phosphatidylserine Sph in gom yelin Phosphatidic acid

39-43 14.0 16.6 6.0 14.0 4.0-6.0 -

Guinea pig"." 39 0

1 7.6- 1 8 .0

16.0 4.0 13.0

5.1 0.7

Rat''

Guinea pig"

}41.6

24'0 ()

} 34'2

10.0 20.0 1 .o 3.0 6.0 6.0

3.5 13.2 5.1 0.6

Synaptic vesicles

Squirrel monkey"

}

Rat''

} }

Guinea pig"."

40.7-41 .0 0.6 13.1-15.0 19.8 36.3 12.0-1 5.0 16.3 5.0-5.9 2.9 14.5 1 1.8 10.2-12.0 11.7(incl. PI) 4.9 11.0-12.3 0-2.3 2.5 1.8 35.1

42.2

Note: Measurements in percent of total phospholipid P in each fraction. PI, phosphatidylinositol. " Ansell el rtl. (1973). " Whittaker (1966). ' Fleischer and Rouser (1965). Breckenridge el (11. (1972). " Sun and Sun (1972).

'

6

G. BRIAN ANSELL AND SHEILA SPANNER

111. Origin and Metabolism of Brain Phospholipids

A. TRANSPORT TO THE BRAINA N D

THE

ORIGINOF PRECURSORS

The synthetic pathways to the phospholipids of nervous tissue have been recently reviewed by Ramsey and Nicholas (1972) and Ansell (1973). One of the basic mechanisms by which the major ethanolamine and choline glycerophospholipids are synthesized is the cytidine pathway. It has been clearly demonstrated that when choline or ethanolamine are injected into the lateral ventricle of the brain, they are assimilated into the tissue, converted to phosphorylcholine (ethanolamine), cytidine diphosphate (C'DP)-choline (ethanolamine), and finally to the lipidbound forms, phosphatidylcholine, phosphatidylethanolamine and, via the saturated ether analog, to the ethanolamine plasmalogen (Ansell and Spanner, 1967, 1968). Further evidence for the de now0 synthesis of choline and ethanolamine glycerophospholipids was obtained by following the incorporation of [2-3H]glycerolafter intracerebral injection (Benjamins and McKhann, 1973; O'Brien and Geison, 1974). The brain is therefore capable of carrying out each of the anabolic steps of the cytidine pathway to major glycerophospholipids. However, although the mammalian brain can synthesize major phospholipids from component moieties, there is some evidence that phosphatidylcholine, or more probably lysophosphatidylcholine, can be transported to the brain by the blood. This was originally suggested by Ansell and Spanner ( 197 l), who deduced from experiments with intraperitoneally injected [ 1,2-14C]ethanolamine and [ 1,2-**C]dimethylaminoethanol that the brain received choline in a lipid-bound form. These bases yielded lipid-bound [ 1,2-'4C]choline in the liver, the blood, and the brain but there was no labeled free choline or phosphorylcholine in the blood. In these experiments no distinction was made between phosphatidylcholine and lysophosphatidylcholine (sn-l-acylglycero-3-phosphorylcholine)and sphingomyeliri or between blood cells and plasma. Subsequent studies, however (Spanner et al., 1976) clearly showed that there was heavy labeling in the lysophosphatidylcholine and phosphatidylcholine of the plasma. Moreover, Illingworth and Portman (1972) showed that when doubly labeled lysophosphatidylcholine,bound as a lipoprotein, was injected intravenously into squirrel monkeys, it was rapidly taken up and metabolized by the brain. The phosphatidylcholine formed had the same ratio of [Me-3H]cholineto [ l-'4C]palmitoyl as that originally injected. From this it was concluded that choline could be transported to the brain in a lipid-bound form. For a further discussion

FU N CTI ONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

7

of the fate of lysophosphatidylcholine, see Section V,A. To what extent the blood is a necessary donor of the phosphatidyl moiety as opposed to choline is unknown. In other words we do not know that lysophosphatidylcholine per se is anything other than a choline transporter as far as brain is concerned. Long-chain fatty acids, palmitic, oleic, linoleic, and linolenic are rapidly taken u p by the brain from the blood, though there is a preference for linolenic over linoleic acid (Dhopeshwarkar and Mead, 1973). They are rapidly incorporated into complex lipids. Nevertheless, fatty acid synthesis is also a rapid process in brain tissue as was shown quite clearly by Dhopeshwarkar et al. (197 1) by injecting [ l-"C]acetate into the carotid artery of the adult rat. Significant amounts of palmitic acid were synthesized de novo and found in many complex lipids within 15 seconds of the injection. The relative roles of transported fatty acids and those synthesized in situ in the synthesis of phospholipids in the brain are therefore unknown. Since fatty acids taken up from the blood are incorporated intact into phospholipids, it is not surprising that intracerebrally or intracisternally injected fatty acids can be rapidly assimilated into phospholipids also, as was shown by Sun and Horrocks (1969) and Bernsohn et al. (197 1). The origin of the long-chain alk- l-enyl moieties of plasmalogen has also been established as deriving from long-chain alcohols, e.g., hexadecanol, which are present in small amounts in brain (Takahashi and Schmid, 1970; Schmid and Takahashi, 1970; Bell et al., 1971; Horrocks, 197 1). By a series of reactions initially involving dihydroxyacetone phosphate, 1-alkyl-2-acyl-glycerolis formed (for details, see Ansell et al., 1973) which can then react with CDP-ethanolamine to yield 1alkyl-2-acyl-glycerol-3-phosphorylethanolamine (Radominska-Pyrek and Horrocks, 1972). The saturated ether analog is then presumably oxidized to the alk-l-enyl lipid by a reaction demonstrated in vitro (Horrocks and Radominska-Pyrek, 1972).

B. De Nouo SYNTHESIS-NEURONES A N D GLIA For the phospholipid composition of neurones and glia, see Table I. The enzymes responsible for the synthesis of major phospholipids appear to be found either in the soluble fraction or in the endoplasmic reticulum, normally isolated as the high-speed membrane fraction, the microsomes. In a conventional subcellular fractionation of brain tissue these fractions derive from both neurones and glia. Thus to distinguish between the synthetic capacities of the subcellular components of neurones and those of glia, it is necessary to separate the cell types. Even

8

G . BRIAN ANSELL A N D SHEILA SPANNER

so, only the neuronal perikarya are normally obtained so that the contribution of axonal components, and to a lesser extent dendritic components, are not included. In 1969, Freysz et al. (1969) followed the incorporiltion of 32P043into the phospholipids of neurones and glia by isolating them from rat brain cortex after an intraperitoneal injection. They showed that, in general, neuronal phospholipids had a faster turnover rate than those of glial cells but that, in both cell types, phosphatidylinositol and phosphatidylcholine had the shortest turnover times. There was an interesting finding that the turnover of the plasmalogen in the neurones was five times and that of the glia three times, as rapid as that in whole brain. Porcellati and his co-workers (Porcellati et al., 197 1) have studied in great detail the so-called calcium-dependent, energy-independent exchange mechanism by which the base moiety of the phospholipid is exchanged with a water-soluble pool. They demonstrated that choline, ethanolamine, and serine will exchange with a base in the intact phospholipid and this exchange has been studied in both the cell types and the subcellular fractions of brain. Raghaven et al. (1973) have also studied this pathway and have shown that, for brain tissue from young rats 13-20 days old, choline and ethanolamine exchange is greater in the isolated glia than in the neurones while serine exchange is the same in both neurones and glia. On the other hand in rabbits Goracci et al. ( 1973) found that the neuronal fraction had a much higher rate of exchange for serine and ethanolamine than the glial cell-enriched fraction. These workers have also studied the cytidine pathway in vivo and in uitro in neurones and glia. It was found in vitro that the incorporation of phosphorylcholine and phosphorylethanolamine into phospholipids in the neurones was greater than in the glia (Binaglia et al., 1973). For example the transfer of phosphorylcholine from CDP-choline to phosphatidylcholine was 21.8 nmoleslmg protein per 30 minutes in the neurones compared with 4.2 nmoleslmg protein per 30 minutes in the glia. Phosphorylethanolamine was transferred from CDP-ethanolamine into the three brain ethanolamine-containing phospholipids, phosphatidyl-ethanolam ine, 1 alk- I -enyl-2-acyl-glycerophosphorylethanolam ine (GPE), and l-alkyl-2-acyl-GPE. Here the direrence between the incorporation into the neurones and glia was less marked than with CDPcholine as the precursor but the neurones were at least three times as active as the glia. This difference was increased when alkylacylglycerol was added to the system when there was a stimulation of uptake which was greater in the neurones than the glia (Roberti et al., 1975). This greater activity of the cytidine pathway in neurones was confirmed in Z J ~ V O(Goracci et d.,1975).

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

9

C. De NOZJO SYNTHESIS-SUBCELLULAR FRACTIONS

For the phospholipid composition of some of the subcellular fractions, see Tables I1 and 111. It has already been stated that the enzymes responsible for the formation of phospholipids in the brain are found mainly in the cell sap or in the microsomal fraction. In 1972, Miller and Dawson (1972a) set out to discover if mitochondria and synaptosomes prepared from guinea pig brain cortex could synthesize phospholipids (Miller and Dawson, 1972a) and also whether or not intact phospholipids could be exchanged between membranes in vitro (Miller and Dawson, 1972b). In the first series of experiments it was found that mitochondria could not synthesize phospholipids from either CDP-[Me-’4C]choline or phosphoryl-[Me‘‘C]choline but they could synthesize phosphatidic acid and diphosphatidylglycerol from 32P043-. In the synaptosomes, the outer membrane and the intraterminal mitochondria did not appear to synthesize phosphatidylcholine from CDP-choline, though the vesicles and endoplasmic reticulum seemed to be active in this respect (see also Baker et al., 1976). Synaptosomes and microsomes could incorporate choline into phosphatidylcholine by a Ca”-mediated exchange and this has been demonstrated by Gaiti et al. (1974) and Kanfer (1972). Pasquini et al. (1973) studied the incorporation in vivo of [Me-’4C]cholineinjected intracranially into the subcellular fractions of adult rat brain. They found the peak of incorporation into phosphatidylcholine occurred at 5 hours in all fractions but with the highest specific radioactivity in the mitochondria. This would indicate that the transfer of phospholipid from microsomes to mitochondria was unlikely. However, Miller and Dawson (1972b) showed that such an exchange was possible in z&o. This work has been extended by various groups (Helmkamp et al., 1974; Possmayer, 1974; Brammer and Sheltawy, 1975; Wirtz et al., 1976), and it has been shown that the phospholipid exchange between membranes is catalyzed by a soluble protein in brain. Possmayer (1974) used rat brain microsomes labeled with [“Clphosphatidic acid, [”C]phosphatidylcholine, or [5H]phosphatidylinositol as donors and mitochondria isolated from the same rat brain for the “receiver” membranes. The protein was supplied by the cytosol, again from the same brain tissue. He found that 30-40%, of the available phosphatidylinositol was transferred to the mitochondria in 1 hour, while there was little or no transfer of either phosphatidic acid or phosphatidylcholine. The inability to transfer phosphatidylcholine could have been due to myelin contamination of the microsomes which was shown by Miller and Dawson (1972b) to inhibit phospholipid exchange between organelles. Further

10

G. BRIAN ANSELL A N D SHEILA SPANNER

evidence that inhibition was occurring can be found in the work of Helmkamp et al. (1974) and Harvey et al. (1973). They demonstrated the presence of a phospholipid exchange protein in brain which they subsequently purified. They used liver microsomes and mitochondria in their assay, so no inhibition by myelin would have been detected. They showed a higher exchange of phosphatidylinositol, and in contrast to Possmayer (1974) found the exchange of phosphatidylcholine was significant. Brammer and Sheltawy (1975) also demonstrated the exchange but their donor lipid was in the form of liposomes and not bound to microsomes. Very recently Wirtzet al. (1976) have extended the study of this protein and have shown that both synaptosomes and myelin preparations contain a soluble protein catalyzing the transfer of phosphatidylcholine and phosphatidylinositol from rat liver microsomes to phospholipid liposomes. So although myelin membrane contamination of microsomes appears to inhibit the transfer of phosphatidylcholine to mitochondria, myelin itself, though more probably the axoplasm associated within it, contains an extractable phosphatidylcholine exchange protein. In studies in vzvo using [ 1,2-14C]ethanolamineas the precursor, Ansell and Spanner (1967) showed that after an intracerebral injection, the peak of incorporation into phosphatidylethanolamine varied from fraction to fraction indicating a possible lipid exchange. At all times studied up to 24 hours the specific radioactivity of the microsomal fraction was greater than the mitochondria for phosphatidylethanolamine. I n the case of the ethanolamine plasmalogen, however, the reverse was true. When [Me-’4C]cholinewas used as the precursor in similar experiments the specific radioactivity of the microsomal phosphatidylcholine was higher than that of the mitochondria (Ansell and Spanner, unpublished results) up to 2 hours. Using the base as the labeled precursor makes it difficult to assess how much of the phospholipid molecule other than the base itself is turning over. Although experiments with radioactive choline or ethanolamine in z,zvo do not distinguish per se between incorporation by base exchange or transfer via the phosphate ester, the time sequence of radioactivities of base, phosphate ester, CDP-ester, and phospholipid indicated that the cytidine pathway is a major one. Such experiments do not, of course, indicate the turnover of the fatty acid moieties. For many years it was thought that central myelin, once laid down, became a metabolically stable entity (for reference, see Ansell, 1973). However, Ansell and Spanner (1965, 1967, 1968) demonstrated that when [ 1,P-’4C]ethanolaminewas injected intracerebrally into adult rats, myelin, particularly the so-called “small myelin,” contained labeled

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

11

phospholipids. This showed that at least the base moiety of the ethanolamine phospholipid molecule was being replaced. Jungalwala and Dawson (197 1) repeated these experiments with [U-’4C]glycerol, a marker for “total turnover,” and showed that the glycerol, too, was readily incorporated into the glycerol moiety of myelin phospholipids. Horrocks (1969) also demonstrated the fairly rapid incorporation of [1,2‘‘C]ethanolamine into the three ethanolamine-containing phospholipids of myelin in adult mice. How this incorporation takes place is still somewhat of an enigma. Miller and Dawson (1972b) were unable to demonstrate a labeling of myelin phospholipids when myelin fragments were incubated with microsomes containing labeled phospholipids, i.e., there was no exchange. However, Jungalwala (1974) failed to demonstrate the presence of cholinephosphotransferase (EC 2.7.8.2) in myelin which would be essential for a synthesis of phosphatidylcholine by the cytidine pathway. He did, however, demonstrate an uptake of [Me-’4C]choline into myelin phospholipid. Very recently, Gould and Dawson (1976) have demonstrated that, in sciatic nerve, the phosphatidylcholine of only the outer layers of myelin appeared to be labeled when tritiated choline was injected into the nerve. They believe that the choline was incorporated into the cytoplasm in the Schwann cell and transferred as a phospholipid to myelin. This may in part explain the experiments of Ansell and Spanner (1967) when the ethanolamine lipids of “small myelin,” in all probability the looser, outer lamellae, were so much more readily labeled after an intracerebral injection of [ l,2-’4C]ethanolaminethan the denser, “heavy myelin.” An excellent review of the exchange of phospholipids between cell membranes has been written by Dawson (1973).

IV. Relationship between Phospholipid Composition of Membranes and Brain Function

A. NATURALLY OCCURRING MODIFICATIONS OF PHOSPHOLIPID COMPOSITION The significance of the integrity of membranes for the maintenance of the mammalian central and peripheral nervous systems is in same ways self-evident. The method by which the composition is maintained is unknown, but the phospholipid components of the membranes appear to be replaced at constant, if dissimilar, rates (Section 111, B and C). Since, with the exception of phosphatidic acid, phospholipids are end products rather than metabolic intermediates, the meaning of their

12

G. BRIAN ANSELL A N D SHEILA SPANNER

metabolic turnover in an organ which, unlike the liver, is not exporting them, is obscure. Furthermore, in the central nervous system (CNS) there is a minimal requirement for cell division. It will be recalled that the turnover of brain phospholipids, as measured by radioisotopesinvivo, does not mean that all membrane organelles synthesize their own phospholipids. The likelihood is that they are transferred there from the phospholipid-generating system in the endoplasmic reticulum (Section 111, C). Modification of the phospholipid composition of membranes could therefore well be a result of some alteration in the system which transfers phospholipids, e.g., phospholipid exchange protein (Section 111, C), as well as a change in the synthetic capacity of the endoplasmic reticulum. The commonest naturally occurring changes in the phospholipid composition occur with myelin, and it is with myelin that most of the work on membrane phospholipids in the brain has been carried out. This is because myelin was the first membranous component of mammalian brain to be prepared in large, relatively pure amounts. I t has been apparent from morphological investigations over a period of decades that, if the myelin sheath is absent as in demyelinating diseases, then dysfunction occurs. In a “classical” central demyelinating condition such as multiple sclerosis demyelination is related to physical incapacity. Furthermore, there are numerous genetically determined disorders in which the chemical composition of myelin, particularly that of its lipid components, is significantly changed, and the effects of such a change can be profound. Nevertheless, the origin of the metabolic defect as far as phospholipids are concerned lies outside the myelin membrane since as has been discussed on Section 111, C, all the available evidence suggests that, although the phospholipid components of adult myelin “turn over” to a significant degree, they are transferred to the myelin structure after synthesis elsewhere. Exactly how this occurs and the degree of “molecular distortion” that the myelin membrane can accept before it malfunctions is unknown. Most studies have been carried out on relatively gross discrepancies and in recent years the known dysmyelinations of laboratory animals have received some attention. T h e fact that at a gross level myelin is essential for the proper functioning of the CNS is clearly seen in the myelin-deficient mutant, the “Quaking” mouse (autosomal recessive) which behaves normally until the onset of myelination at 12 days of age. Thereafter the animals show an abnormal gait, tremor, and seizures (Sidman et al., 1964). Analysis of myelin obtained from such animals showed that it was deficient in ethanolamine plasmalogen, sphingomyelin, and cerebroside (Dawson and Clarke, 1971; Singhet al., 1971).

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

13

The amounts of mono-unsaturated alk-1 -enyl and acyl groups in the plasmalogen were found to be reduced in the myelin from the deficient animals (Singh et al., 1971). This dysmyelination, i.e., a failure to produce myelin normally during development, is not due to a single missing enzyme since too many molecular constituents of the membrane are involved, but it may well be caused by the malfunction of some control mechanism the nature of which is obscure. A similar but more severe dysmyelination occurs in the ‘3impy” mouse (X-linked recessive) in which such myelin as is found contains a sphingomyelin resembling that found in the brain prior to myelination, being deficient in C22:o,C,,:, and particularly CZ4:fatty acids (Nussbaum et al., 1971). However, there is no evidence that this is related to the dysmyelination, and such observations indicate only that longer chain fatty acids are important for the correct functioning of a phospholipid in the myelin membrane. Since no specific function for sphingomyelin has ever been put forward, such a statement, of course, merely begs the question. A dysmyelination which is also a lipidosis is the rate Niemann-Pick disease (for references, see Ansell, 1973) in which sphingomyelin accumulates in the brain and other tissues, but this sphingomyelin, unlike that from normal myelin (Section I I ) , is rich in Clxand poor in C,?,, fatty acids, and although its accumulation can be accounted for by the deficiency in sphingomyelin phosphodiesterase (EC 3.1.4.13), its abnormal fatty acid pattern cannot, unless a fatty acid-specific sphingomyelinase is in fact absent. T h e essential role of longer chain fatty acids in myelin is demonstrated in Refsum’s disease (hereditary ataxic polyneuropathy) in which phytanic acid (3,7, 1 1, 15-tetramethyl hexanoic acid), a constituent of the diet which is normally metabolized, accumulates (Klenk and Kahlke, 1963: MacBrinn and O’Brien, 1968). Its accumulation leads to deficiencies in ethanolamine phospholipids in gray matter and myelin (though peripheral neuropathy is usually more marked than effects on the CNS). O’Brien (1967) suggested that the presence of the branchedchain phytanic acid could afFect the “packing” of the bimolecular lipid leaflet because of the branched methyl groups. The myelin lipids, including phospholipids, normally rich in longer chain fatty acids which enhance packing, might be more susceptible to the presence of a branched fatty acid likely to discourage packing. What is largely lacking is information about the nature of the malfunction when myelin is distorted or even absent. T h e relationship between the integrity of the myelin and its function in the CNS is by no means clearly established. Bunge (1968), in his excellent review on central myelin, noted that evidence for its role in saltatory conduction was limited to a paper published by Tasaki (1952) who worked with frog

14

G. BRIAN ANSELL A N D SHEILA SPANNER

spinal cord (see also BeMent and Ranck, 1969). More recent work on the central demyelination induced by diptheria toxin (McDonald and Sears, 1970a,b; Rasminsky and Sears, 1972) in spinal ventral roots indicates that, though axonal conduction is impaired in demyelinated 'fibers, conduction remains saltatory even to the point of conduction block, rather than becoming continuous as in unmyelinated fibers. Thus myelin alone is not responsible for saltatory conduction from node to node. As far as phospholipids are concerned one approach would be an investigation of the function of the plasmalogens. In a large recent volume on the subject (Snyder, 1972) very little could be said about the function of these phospholipids, which basically differ from the diacyl analogs only in the presence of an alk-l-enyl linkage instead of an acyl linkage in the I-position (see Section 11, A). They have a similar molecular shape to the diacyl phospholipids and are widely distributed in membranes. In mammalian tissues, however, they are most prominent in skeletal muscle, cardiac muscle, and nervous tissue, particularly myelin. This indicates some special function of plasmalogens in excitable tissues. It is a fact that, when demyelination takes place, these are the phospholipids lost to the greatest extent and there is evidence that plasmalogenase (alkenylglycerophosphorylethanolaminehydrolase, EC 3.3.2.-),an enzyme which cleaves the alk-l-enyl linkage, is more active (Ansell and Spanner, 1968; McMartin et ad., 1972). Moreover, Segall and Wood (1974) have shown that the neurotoxin methyl-mercury, not a particularly reactive molecule in aqueous systems, can promote the hydration and hydrolytic cleavage of the alk-l-enyl linkage in vitro (they did not examine myelin). This could be responsible for the lysis in neuroglia and granule cells seen in the victims of the Minimata disaster (Takeuchi, 1968). Thus there is indirect evidence that plasmalogens play some special function in the membranes of the nervous system, particularly myelin. B. EXPERIMENTAL MODIFICATIONS OF PHOSPHOLIPID COMPOSITION

'

The part played by phospholipids in axonal conduction has been studied by various groups of workers but as yet no clear conclusions have been drawn. Studies so far have of necessity been confined to peripheral nerve fibers or to model membranes. of known phospholipid content. The nerve studied has been, almost exclusively, the squid giant axon. Condrea and Rosenberg (1968) showed that when phospholipase A from snake venom was applied to the outside of the nerve this led to an increased permeability of the membrane, a block in conduction along the axon, and a desensitization to curare and acetylcholine (Rosenberg and

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

15

Hoskin, 1963; Rosenberg and Ng, 1963). There was, it is true, a considerable breakdown of the membrane phospholipids, but these workers concluded that it was the detergent action of the lysophospholipids produced within the axoplasm and not the disrupted nerve membrane which led to the inactivation of the nerve. Abbott et al. (1972), using phospholipase A and lysophospholipase from mouse intestine, failed to demonstrate any effect on nerve activity when the enzymes were applied to the outside of the nerve but these workers, unlike Rosenberg and his co-workers, made no. measure of phospholipids in the membrane or the axoplasm; nor did they check the production of lysophospholipids. Abbott et al. (1972) did find, however, that the phospholipases from mouse suppressed the action potential when they were applied to the inside of the nerve. The production of lysophosphatid ylcholine and its lytic effects in Rosenberg’s experiments is important in view of the experiments of Hall and Gregson (1971) and Gregson and Hall (1973). These workers using the peripheral, myelinated nerve (sciatic) of the mouse showed that when lysophosphatidylcholine was injected into the nerve there was rapid demyelination but the axon itself was undamaged and remained impermeable to exogenous ferritin. T h e role of the polar head groups of the phospholipids in nerve conduction was also studied by Rosenberg (1970). He applied phospholipase C (EC 3.1.4.3) from C. welchii externally to the squid giant axon and found that, though up to 60% of the membrane phospholipids were hydrolyzed and virtually all of the phosphorylated bases produced were released to the external bathing fluid, no functional inhibition was detected. The diglyceride moieties remained in the membrane. It would appear from these results that the polar head groups of the phospholipids are not necessary for nerve conduction. Several workers have studied the possible role of phosphoinositides in nerve function (see Michell, 1975, for a review of this work). Both diand triphosphoinositide have a very high binding a h i t y for Ca” ions and, since triphosphoinositide has a higher affinity than diphosphoinositide, the interconversion of the two lipids could change the Ca2+ binding at membrane surfaces (Hendrickson and Reinertsen, 1969). Buckley and Hawthorne (1972) have since demonstrated, using erythrocyte membranes, that there was an increase in the high-affinity binding of Ca2+in the membrane when phosphatidylinositol was converted to diand triphosphoinositide. However, it is not at all clear what effect such changes might have on nerve conduction. Very recently Tyson et al. (1976) have been exploring the possibility that some phospholipids can act as ionophores, transporting monovalent and divalent cations across an artificial membrane in a Pressman cell.

16

G. BRIAN ANSELL AND SHEILA SPANNER

They found that cardiolipin can translocate both mono- and divalent cations at similar rates over a very wide pH range. Other phospholipids appeared to inhibit the movement. Thus, although a great deal of work has been done on the structure of phospholipids vis 2 uis nerve conduction, it is impossible, as yet, to come to any firm conclusions.

C. EFFECTSOF ANESTHETICS “It is generally assumed that the primary actions of anaesthetics are on the cell membrane rather than on intracellular processes” Seeman ( 1972). The Overton-Meyer theory predicts that “anaesthesia occurs when a lipid-soluble anaesthetic reaches sufficient concentration in an excitable membrane to cause disorder” (Ansell, 1973). Studies on the effect of local anesthetics on lipid monolayers and on erythrocyte ghosts demonstrated that there was an increased fluidity and expansion of the membrane (Clements and Wilson, 1962), a phenomenon also described by Seeman (1972). Johnson and Bangham (1969) concluded from this that anesthetics increase the freedom of movement of groups in the lipid near the interface, sterically impeding the rearrangement which produces the sudden increase in permeability to Naf ions on which the generation of the action potential depends. The excitability of nervous tissue is, to a certain extent, controlled by Ca2+ions and some local anesthetics can displace Ca2+ from the membrane. Phosphatidylinositol can bind Ca2+ (Hauser and Dawson, 1968) and this ion can be released under varying conditions, not exclusively by local anesthetics. Work on the effect of local anesthetics on phospholipids and their metabolism in the nervous system has been minimal, but work on other tissues may be treated with cautious relevance. Scherphof and Westenberg (1975) studied the effect of the local anesthetics dibucaine and butacaine on pancreatic phospholipase A2 (EC 3.1.1.4).The effect was complex. In both cases there was an inhibition of the enzyme when the substrate was membrane-bound, but a high concentration of anesthetic was necessary. At low concentrations of dibucaine there was a stimulation of phospholipase Az activity toward the free substrate. The authors concluded that these two local anesthetics interfere with the phospholipase activity by interacting with the substrate rather than with the enzyme. Kunze et al. (1976) also studied the effect of local anesthetics on phospholipase Az from C. adamanteus and from seminal plasma and on lysophospholipase (EC 3.1.1.5) from rat liver cytosol. With the phospholipase A, from seminal plasma and with phosphatidylethanolamine

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

17

as a substrate they found the biphasic effect of local anesthetic described by Scherphof and Westenberg (1975), i.e., a stimulation at low and an inhibition at high anesthetic concentrations. Anesthetics had no effect on the lysophospholipase. These authors concluded from their studies that the substrate does not compete with the anesthetic for the enzyme. Neither group of workers found Ca" to play a significant part in the effect of anesthetics except, to a small extent, in the inhibition of phospholipase activity (Kunze et al., 1976). It seems possible that the stimulation of the production of lysophospholipids by the action of anesthetics on phospholipase Az may be a relevant factor in the action of anesthetics. In normal mammalian brain, the activity of the phospholipases is relatively low and they d o not require Ca'+. In view of the work with pancreas it would be interesting to study the effect of local anesthetics on the activity of phospholipase A:! in nerve cell membranes. Very recently Wu et al. (1976) have shown that triphosphoinositide, isolated from brain, will bind narcotics and narcotic agonists over a wide range of concentrations, but they did not study the subsequent catabolism of the lipid-drug complex. In these experiments it was shown that sodium ions inhibited the binding of levorphenol, a narcotic agonist but not of naloxone, a specific antagonist.

V. Phospholipids and Synaptic Transmission

A. RELATIONSHIPBETWEEN LIPID-BOUND CHOLINE ACETYLCHOLINE FORMATION I N THE BRAIN

AND

Recently there has been considerable interest in the metabolism of choline in the brain and the possible relationship between lipid-bound choline, free choline, and acetylcholine (Ansell and Spanner, 1975a). The following facts about free choline in the brain are noteworthy. A number of recent observations (e.g., Eade et al., 1973) have demonstrated that the amount of free choline in the brain is less than 30 nmoledgm wet weight with some variation in different brain areas. This is lower than the acetylcholine level, but there is a rapid release of choline postmortem from brain tissue (Stavinoha and Weintraub, 1974) which is presumably derived from a lipid-bound source, though neither the mechanism nor the anatomical and subcellular location of the release are known. In the rat there is an arteriovenous difference of free choline across the brain (Dross and Kewitz, 1972; Choi et al., 1975; Spanner et al., 1976) which represents a considerable efflux from the tissue. In the rabbit a higher level in the vencus return was not seen, though exper-

18

C. BRIAN ANSELL A N D SHEILA SPANNER

iments with ['HI choline showed that the specific radioactivity of the choline in the venous efflux was always lower than that of the arterial supply (Spanner et at., 1976). This also indicates an output of unlabeled choline from brain tissue in this species. An arteriovenous difference of choline across the brain is also present in man (Aquilonius et al., 1975). Although there is an apparent continuous release of free choline from brain tissue, this is not replaced by choline synthesized in situ since brain tissue cannot synthesize this base (for a summary of the evidence, see Ansell and Spanner, 1975a). It therefore receives it from an external source and it is highly likely that a major source is lysophosphatidylcholine as discussed in Section 111, A. Illingworth and Portman (1972) showed not only that doubly labeled lysophosphatidylcholine can readily enter the brain to form phosphatidylcholine, but that it can also donate choline to pools of free choline and acetylcholine. The experiments of Illingworth and Portman (1972) which were carried out on the squirrel monkey require confirmation for other laboratory species but seem clear enough. These observations raise some interesting questions. If there is a net efRux of free choline from the brain in the venous return and the input to the brain is lipid-bound, what is the source of the free choline in the arterial supply? Does lipid-bound choline liberate free choline as soon as it passes from the capillary beds to the neuropil, or is the lysophosphatidylcholine acylated to phosphatidylcholine which then moves as a lipoprotein from the nerve cell perikaryon down the axon? Evidence exists for an axoplasmic flow of a phosphatidylcholine-containinglipoprotein in sciatic nerve (Abe et al., 1973). On the other hand, if free choline is liberated at an earlier stage there is evidence that it can move rapidly throughout the brain tissue (Ansell and Spanner, 1975b). There are several mechanisms by which choline could be liberated from lipid-bound choline, e.g., phosphatidylcholine, and these are summarized in the following reactions. Choline can be released by a reaction in which a base such as serine can displace choline from phosphatidylcholine by a Ca2+-mediated exchange reaction, localized primarily in the endoplasmic reticulum in neurones and glia (see Section 11, B and C), though this reaction [Eq. ( l ) ] has largely been studied in the reverse direction. Ca"

Phosphatidylcholine

+ base + phosphatidyl - base + choline

(1)

Another mechanism is the release of choline by a hydrolytic pathway involving first the removal of fatty acids by the combined actions of phospholipase A, (EC 3.1.1.32) and A Band lyosphospholipase, Eq. (2)

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

19

(Gatt, 1968; Leibowitz and Gatt, 1968; Cooper and Webster, 1970), then the hydrolysis of glycero-3-phosphorylcholine(glycerophosphocholine) to choline and glycerol-3-phosphate (Webster et al., 1957), Eq. (3). Phosphatidylcholine+ glycero-3-phosphorylcholine + 2 fatty acids

(2)

<;a’-

Glycero-3-phosphorylcholine+ glycerol-3-phosphate+ choline

(3)

By yet another reaction recently discovered by Abra and Quinn ( 1975), glycero-3-phosphorylcholine can be hydrolyzed by an enzyme present in brain, but not in liver, to yield phosphorylcholine, Eq. (4). This

could then be hydrolyzed by an alkaline phosphatase (Strickland et al., 1956) to yield choline and inorganic phosphate, Eq. (5). Glycero-3-phosphorylcholine+ glycerol Phosphorylcholine + P, + choline.

+ phosphorylcholine

(4) (5)

It is possible that the reaction mechanism is located in cholinergic terminals but there is no hard evidence at present. Phospholipase AS is present in synaptosomal membranes (Woelk and Porcellati, 1973) and glycerophosphocholine phosphodiesterase (EC 3.1.4.2) is also enriched in synaptosomes (Mann, 1975). Eqs. (1) and (3) both require Ca2+ions which are also necessary for the release of acetylcholine from cholinergic terminals, so conceivably some production of intraterminal choline and the release of transmitter are coupled. The subcellular distribution of alkaline phosphatase has not been examined in detail but it is certainly present in synaptosomal membranes (Spanner, unpublished observations). Since specific inhibitors of the enzymes involved in Eqs. (1)-(4) are not available, the elucidation of the major route by which choline is released from phosphatidylcholine presents some difficulties. The source of choline for acetylcholine synthesis is perhaps of more than academic interest. In Huntington’s Chorea and the tardive dyskinesias it is believed that cholinergic activity in the striatum is impaired (McGeer et al., 1973) with a “predominance” of dopaminergic over cholinergic systems (Klawans and Rubovits, 1970). M.assive doses of free choline can control the dyskinesias to some extent (Davis et al., 1975), which suggests that the reduced cholinergic activity might be due to a reduction in acetylcholine production caused by a lack of free choline at the site of synthesis. Certainly large doses of intraperitoneally injected choline can raise acetylcholine levels in the brains of experimental animals (Cohen and Wurtman, 1975). Thus a failure of a system for the generation of free choline may be a relevant biochemical lesion in the dyskinesias.

20

G . BRIAN ANSELL A N D SHEILA SPANNER

A completely different and very interesting observation on the relationship of phospholipids to acetylcholine has &en made recently by Mantovani et al. (1976). They have shown that a phospholipid mixture from bovine cerebral cortex, when injected intraperitoneally or intravenously into rats, increased the output of acetylcholine from the cerebral cortex. The phospholipid largely responsible was found to be phosphatidylserine which produced a dose-dependent output of acetylcholine when injected intravenously. A dose of 150 mg/kg body weight gave a 3-fold increase in the spontaneous output of the transmitter. This release is presumably unrelated to choline production since phosphatidylcholine was ineffective. It could, however, be related to the fact that acetylcholine output from the cortex is Ca2+-dependent. Phosphatidylserine has an affinity for this cation and is known, for example, to enhance the Ca2+-dependent release of histamine from mast cells (Grosman and Diamant, 1975). Nevertheless, the relevance of a massive dose of a phospholipid already present in large amounts in brain tissue may be questioned. B. EFFECTSOF NEUROTRANSMITTERS ON PHOSPHOLIPIDMETABOLISM

There have been numerous attempts to demonstrate whether or not the metabolism of phospholipids is concerned with either the release of neurotransmitters or the action of these transmitters postsynaptically. Most of the work has been concerned with phosphatidic acid and phosphatidylinositol, since early experiments with szP04s-.in vivo and in vitro had indicated that these phospholipids or more specifically the phosphate (or inositol phosphate) moiety turned over much more rapidly than other phospholipids. They are of course present in much lower concentrations than other phospholipids in nerve endings (Table 111). In the 1950s Hokin and Hokin (1955, 1958) demonstrated that acetylcholine stimulated the uptake of labled orthophosphate into phosphatidylinositol and phosphatidic acid in brain slices. It has since &come clear that this effect on two anionic lipids occurs in a wide variety of tissues and in response to a wide variety of stimuli (Michell, 1975). It is also clear that only the inositol phosphate moiety of phosphatidylinositol and the phosphate moiety of phosphatidic acid are involved. Michell and his co-workers (for reviews, see Lapetina and Michell, 1973a; Michell, 1975) have made an extensive study of the phenomenon, and it is now certain that the primary response of phosphatidylinositol is its breakdown to inositol-l-phosphate plus some inositol-1,2-cyclicphosphate and a diglyceride. The latter is then phosphorylared to phosphatidic acid (Fig. 1) which then acts as a substrate for the formation of CDP-

21

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

4 CYTOSOL

To other mrm branrr

phosphate

myOlnOSltol

I - phosphate

Exchange Protein (PA)

myoinositol glycerol

4/

glucOse

3 -glycerolphosphate

FIG. 1 . The model suggested b y Michell and his colleagues for the events in lipid metabolism occurring in a cell exposed to a stimulus which produces an enhanced phosphatidylinositol (PI) turnover. PA, phosphatidic acid. (Reproduced from Michell, 1975, with the kind permission of the author and Elsevier/North Holland Biomedical Press.)

diglyceride, the precursor of phosphatidylinositol. Such a cycle explains the high labeling of phosphatidylinositol and phosphatidic acid in the systems which have been studied. Figure 1 summarizes the ideas expounded by Lapetina and Michell (1973a) and Michell (1975) and explains why it is that the phosphatidylinositol is labeled throughout the cell membranes though the response can be at the surface of the cell. Surface membranes of cells, including nerve cells, contain significant amounts of phosphatidylinositol, though synthesis de norm occurs in the endoplasmic reticulum as does that of most phospholipids. The enzyme catalyzing the hydrolysis of the glycerol-phosphate bond, though present in the cytosol, is certainly present in the plasma membranes of nerve cells (Michell and Lapetina, 1972; Lapetina and Michell, 1973b). 1,2Diacylglycerol kinase (EC 2.7.1.-) is also present in the cytosol and plasma membranes (Lapetina and Hawthorne, 1971). Metabolism of the polar head group of phosphatidylinositol and phosphatidic acid at the cell surface (or very near it) is a corollary of the primary action of the receptor, being the hydrolysis of the former, and there is selective reutilization of l-stearoyl-2-arachidonylglycerol for the resynthesis of the phosphatid ylinositol.

22

G. BRIAN ANSELL A N D SHEILA SPANNER

It seems likely that the acetylcholine receptor which when stimulated gives rise to the “phosphatidylinositol effect” is muscarinic rather than nicotinic, though the evidence is not conclusive (Michell, 1975). It also seems likely that the response to adrenergic stimuli is modulated via arather than @-receptors. T h e action of neurotransmitters is often modulated via Ca2+ ions (Rasmussen et al., 1972),but the phosphatidylinositol response per se does not require the presence of extracellular CaZf ions [for references, see Michell (1975), but for a contrary view, see Lennon and Steinberg (1973)], and Michell (1975) is of the opinion that the breakdown of phosphatidylinositol precedes the involvement of Ca‘+ but that the intracellular responses which follow are the consequence of the opening of some calcium “gate” (Fig. 2). If the breakdown of phosphatidylinositol is a very early response to a stimulus it would seem reasonable that the enzyme responsible is associated with plasma (surface) membranes and this is, to some extent, true for plasma membranes from the cerebral cortex (Lapetina and Michell, 1973b). However, it is not true for other cells, e.g., lymphocytes (Allan and Michell, 1974) in which most of the capacity to cleave phosphatidylinositol is clearly associated with the soluble fraction of the cell. In any event, breakdown of endogenous membrane-bound phosphatidylinositol in brain tissue does not occur until the preparation is treated with deoxycholate, subjected to sonication, or repeatedly frozen and thawed (Lapetina and Michell, 1973b, 1974). But the breakdown of phosphatidylinositol in isolated membrane preparations is not stimulated by acetylcholine. In the experiments of Abdel-Latif et al. (1974) even isolated neuronal perikarya failed to respond to acetylcholine, but this may have been due to the fact that the isolation of the perikarya was subsequent to the prior treatment of the tissue with trypsin. This could not, however, account for the failure of cloned neuroblastoma cells in a stationary phase to respond to either acetylcholine or noradrenaline (Eichberg et al., 1975), though here the abnormality of the cells may have been a factor. Thus the intact cell as found in a slice of tissue is apparently necessary for a response (but see below), and it is possible that the breakdown of endogenous phosphatidylinositol in cell surface membranes is carried out by the enzyme in the cytosol. However, as Michell(l975) points out, this rather assumes that the “received message” at the receptor must be transmitted across the thickness of the plasma membrane, and the phosphatidylinositol molecules which are broken down would have to be on the inner face of the bilayer, i.e., those in contact with the cytosol. So far we have considered only postsynaptic responses, but some investigations have suggested that the effects of acetylcholine on phos-

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

23

Phosphatidylinositol Metabolism and Receptor Mechanisms An agonist (e.g., a-adrenergic or muscarinic cholinergic)

interacts with its

1

Receptor which, in a manner independent of' extracellular Ca'. , triggers

Phosphatidylinositol

+

1,2-Diacylglycerol

1)

Cell surlace Ca" ates are then opened or Ca'- is released from membrane-bound sites) and

Intracellular Ca'+ rises

leading to

1

Responses such as secretion, contraction, K + efflux, elevation of cCMP concentration FIG. 2. An interpretation of receptor mechanisms in which the breakdown of phosphatidylinositol is seen as an early reaction which precedes the opening of Cay+gates. cCMP, cyclic guanidine monophosphate. (Reproduced from Michell el nl., 1976, with the kind permission of the authors and publishers.)

pholipid metabolism in the CNS are largely presynaptic and involve the presynaptic terminal and its contents rather than the postsynaptic receptor. (The arguments for pre-versus postsynaptic effects are given by Schacht et al., 1974.) Lunt and Pickard (1975) looked at the effect of carbamylcholine (an acetylcholine agonist which is not hydrolyzed by acetylcholinesterase) on phosphatidylinositol turnover in the cerebral cortex in vivo. They found that there was a breakdown of phos-

24

G. BRIAN ANSELL A N D SHEILA SPANNER

phatidylinositol, largely confined to synaptic vesicles and presynaptic membrane fragments. However, studies on isolated nerve endings in vitro have so far failed to show convincingly a significant response in the form of phosphatidylinositol breakdown. Yagihara and Hawthorne (1972) showed that low concentrations of acetylcholine stimulated the labeling of phosphatidic acid in the presence of s2P04:3- but had no significant effect on phosphatidylinositol. However, an effect on the latter was observed by Schacht and Agranoff (1972) using a rather different synaptosomal fraction. Subsequently, Yagihara et al. ( 1973) found that the increased labeling of phosphatidic acid was largely confined to the synaptic vesicle fraction when these were isolated following rupture of the nerve endings; the vesicles also showed a significant loss of phosphatidic acid previously noted for whole synaptosomes by Schacht and Agranoff (1973). This indicated significant turnover of the phosphatidic acid of the vesicle fraction and presumably that fraction of the vesicles derived from cholinergic terminals. Nevertheless, Schacht et al. ( 1974) have disputed that the stimulation of phosphatidic acid labeling by acetylcholine is necessarily associated with the vesicles, and in a more recent paper Bleasedale and Hawthorne (1975) note that a more rigorous identification of their vesicle fraction is necessary. Hawthorne and Bleasedale (1975) have extended their experiments on synaptosomes by stimulating them in bulk electrically, commenting that stimulation of presynaptic terminals with acetylcholine is not physiological, especially since a high proportion of the terminals are not, in any case, cholinergic. Electrical stimulation, simulating depolarization, increased the labeling of phosphatidic acid and the increase was largely confined to a fraction rich in synaptic vesicles (but see above). The authors conclude that this effect was not mediated by acetylcholine released presynaptically but was a direct effect, possibly on an enzyme in the synaptosome which attacked a substrate in the vesicle. Since the increase in phosphatidate labeling requires Ca2+ ions (Hawthorne and Bleasedale, 1975) (contrast the effect on phosphatidylinositol labeling), it may be concerned with transmitter release. If the transmitter is acetylcholine then some problems arise over interpretation because there is considerable evidence (Marchbanks, 1975) that this transmitter is released preferentially from the cytosol, not from vesicles. Furthermore, recent morphological investigation suggests that vesicles themselves may not exist in viuo but derive from tubular structures postmortem. Although most effects on brain phospholipids of pharmacological agents which stimulate or block neurotransmission seem to be largely confined to phosphatidylinositol and phosphatidic acid, a few observations suggest that other phospholipids “respond.” Hokin (1969) showed

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

25

that, while lo-‘ M noradrenaline stimulated the incorporation of ‘I”P0J:’into phosphatidic acid and phosphatidylinositol in slices of several different areas of guinea pig brain, higher concentrations (lo-“ M ) inhibited the incorporation into phosphatidylcholine and phosphatidylethanolamine + phosphatidylserine. High concentrations of’ noradrenaline in this study were justified by the author on the grounds that high concentrations of noradrenaline probably d o occur in the synaptic cleft during synaptic transmission. Sneddon and Keen (1970) found that 1.6 x lo-“ M noradrenaline first depressed and then stimulated the incorporation of :’2€‘0,:’into the total phospholipids of brain homogenates; phosphatidylcholine and phosphatidylethanolamine phosphatidylserine were affected though to a lesser extent than phosphatidylinositol and phosphatidic acid. These effects seemed to be modulated by a-receptors. The depressant effect seemed to be confined to synaptosomal membranes. The inhibitory effects could not be confirmed by Friedel et al. (1973) in their studies in 7 i z m in which noradrenaline and a- and /3-adrenergic agents were injected with the labeled phosphate intracisternally. They showed that noradrenaline (which is both an a- and P-agonist) stimulated the labeling of phosphatidylinositol and phosphatidic acid as did phenylephrine (aagonist), the effects of which were blocked by phenoxybenzamine (aantagonist); isoproterenol, a P-agonist, had no effect. Dopamine also stimulated phosphatidylinositol and phosphatidic acid synthesis in 7 ~ i 7 1 0 though probably through different receptors (Friedel et nl., 1974), a result which disagrees with the findings of Hokin (1970) in zlitro. There was no effect on phosphatidylclioline in contrast to the earlier studies of the same group (Friedel and Schanberg, 1972) with carbamylcholine which first stimulated the labeling of phosphatidylinositol and phosphatidic acid (5 minutes) and then stimulated the incorporation of the labeled phosphate into phosphatidylcholine (30 minutes). These effects were presumably muscarinic because they were blocked by atropine, a muscarinic antagonist. The effect of histamine on posphatidylcholine was similar to that of carbamylcholine (Friedel and Schanberg, 1975) except that histamine stimulated only phosphatidic acid and not phosphatidylinositol at short times (5 minutes). It was shown by the use of atropine and tripelennamine that the histamine efyect was mediated via receptors other than muscarinic ones. In this section numerous interactions between phospholipids and naturally occurring transmitters as well as pharmacological agents have been described. Many of the metabolic effects have been obtained with concentrations of the agonist comparable with the low concentrations producing pharmacological effects in ziivo. Thus, although some of the

+

26

G. BRIAN ANSELL A N D SHEILA SPANNER

observations are descriptive in that they are observed phenomena without functional correlates, certain patterns are emerging, particularly with respect to the phosphatidylinositol effect. I n this short chapter some recent findings on phospholipid metabolism in the brain and their possible relation to certain aspects of function, notably neurotransmission, have been discussed. It will be readily apparent that, at the present time, there are far too many questions and too few answers. To no single phospholipid can a specific role be ascribed whether this be a passive role or one related to its metabolism. The role of a given phospholipid will, of course, depend on its cellular and molecular environment, particularly that of the proteins. I t is interesting, therefore, that for phosphatidylinositol, for which it can reasonably be stated that function and metabolism are related, the available information has largely derived from studies on isolated cell systems, though these have generally been simpler than brain cell systems. This could be a pointer for the future. ACKNOWLEDGMENTS The work of the authorsdescribed in thischapter has been generously supported by the Medical Faculty Research Fund of the University of Birmingham, the Multiple Sclerosis Society of Great Britain, and the Medical Research Council. The interest of Professor P. B. Bradley is appreciated. REFERENCES Abbott, N . J., Degiichi, T., Frazier, D. T., Murayama, K., Narahashi, T., Ottolenghi, A., and Wang, C. M. (1972).J. Physiol. (London) 220, 73-86. Abdel-Latif, A. A., Yau, S.-J.,and Smith, J. P. (1974).J. Neurochem. 22, 383-393. Abe, T., Haga, T., and Kurokawa, M. (1973). Biochem. J. 136, 731-740. Abra, R. M., and Quinn, P. J. (1975). Biochim. Biophys. Acta 380, 436-441. Allan, D., and Michell, R. H. (1974). Biochem. J . 142,591-597. Ansell, G . B. (1973). In “Form and Function of‘ Phospholipids” (G. B. Ansell, R. M. C. Dawson, and J. N. Hawthorne, eds.), pp. 377-422. Elsevier, Amsterdam. Ansell, G. B., and Spanner, S. (1965). Bzorhem. J . 96, 64P. Ansell, G. B., and Spanner, S. (1967).J. Neurochem. 14, 873-885. Ansell, G . B., and Spanner, S. (1968).J . Neurochem. 15, 1371-1373. Ansell, G. B., and Spanner, S. (1971). B i o c h a . J . 122, 741-750. Ansell, C. B., and Spanner, S. (1975a).In “Cholinergic Mechanisms” (P. G. Waser, ed.), pp. 117-129. Raven, New York. Ansell, G. B., and Spanner, S. (1975b). Biochem. Pharmucol. 24, 1719-1723. Ansell, G . B., Dawson, R. M. C., and Hawthorne, J. N. (1973). “Form and Function of Phospholipids.” Elsevier, Amsterdam. Aqiiilonius, S.-M., Ceder, G . . Lying-Tiinell, U., Malmlund, H. O., and Schiiberth, J. (1975). Brain R a . 99, 430-433. Baker, R. R., Dowdall, M . J., and Whittaker, V. P. (1976). B i o c h a . J . 154,65-75. Bell, 0. E., Jr., Blank, M. L., and Snyder, F. (1971). Biochim. Biophys. Acta 231, 579-583. BeMent, S. L., and Ranck, J. B., Jr. (1969). E@. Neurol. 24, 147-170. Reiijamins, J. A,, and McKhann, G. M. (1973).J. Neurochem. 20, 11 11-1 120.

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

27

Bernsohn, J., Stephanides, L. M., and Morgello, H. (197 1). Brain Res. 28,327-337. Binaglia, L., Goracci, G., Porcellati, G., Roberti, R., and Woelk, H. (1973).J. Neurochem. 21, 1067-1082. Bleasdale, J. E., and Hawthorne, J. N. (1975).J. Neurochem. 24, 373-379. Brammer, M. J., and Sheltawy, A. (1975).J. Neurochem. 25, 699-705. Breckenridge, W. C., Combos, G., and Morgan, I. G. (1972). Biochim. Biophys. Acla 266, 695-707. Breckenridge, W. C., Morgan, 1. G., Zanetta, J. P., and Vincendon, G. (1973). Bwchim. Biophys. Acta 320, 68 1-686. Buckley, J. T., and Hawthorne, J. N. (1972).J. Biol. C h m . 247, 7218-7223. Bunge, R. P. (1968). Physiol. Rev. 48, 197-251. Choi, R. L., Freeman, J. J.. and Jenden, D. J. (1975).J. Neurochem. 24, 735-741. Clements, J. A., and Wilson, K. M. (1962). Proc. Natl. Acud. Sci. U.S.A. 48, 1008-1014. Cohen, E. L., and Wurtman, R. J. (1975). L f e Sci. 16, 1095-1 102. Coleman, R. (1973). Biochim. Biophys. Acta 300, 1-30. Condrea, E., and Rosenberg, P. (1968). Biochim. Biophys. Actu 150, 271-284. Cooper, M. F., and Webster, G. R. (1970).J. Neurochem. 17, 1543-1554. Davis, K. L., Berger, P. A., and Hollister, L. E. (1975). New Engl. J. Med. 293, 152. Dawson, R. M. C. (1973). Sub-cell. Biochem. 2, 69-89. Dawson, R. M. C., and Clarke, N. (1971).J. Neurochem. 18, 1313-1316. deVries, G. H., and Norton, W. T. (1974).J. Neurochem. 22, 259-264. Dhopeshwarkar, G. A., and Mead, J. F. (1973). Adz). Lipid Res. 11, 109-142. Dhopeshwarkar, G. A., Subrarnanian, C., and Mead, J. F. (1971).Biochim. l3iophy.r. Actu 248, 41-47. Dross, K., and Kewitz, H. (1972). Naunyn-Schmiedeberg's Arch. Phurmacol. 274, 91 -106. Eade, I., Hebb, C., and Mann, S. P. (1973).J. Neurochem. 20, 1499-1502. Eichberg, J., Shein, H. M., and Hauser, G. (1975).J. Neurochem. 24, 67-70. Fewster, M. E., and Mead, J. F. (1968).]. Neurochem. 15, 1041-1052. Fleischer, S., and Rouser, G. (1965).J. Am. Oil Chem. SOC. 42, 588-607. Freysz, L., Bieth, R., Judes, C., Sensenbremmer, M., Jacob, M., and Mandel, P. (1968).J. Neurochm. 15, 307-313. Freysz, L., Bieth, R., and Mandel, P. (1969).J. Neurochem. 16, 1417-1424. Friedel, R. O., and Schanberg, S. M. (1972).J. Pharmacol. Exp. Ther. 183, 326-332. Friedel, R. 0..and Schanberg, S. M. (1975).J. Neurochem. 24, 819-820. Friedel, R. O., Johnson, J. R., and Schanberg, S. M. (1973).J. Phurmacol. Exp. Ther. 184, 583-589. Friedel, R. O., Berry, D. E., and Schanberg, S. M. (1974).J. Neurochem. 22, 873-875. Gaiti, A., de Medio, G. E., Brunetti, M., Amaducci, L., and Porcellati, G. (1974). J. Neurochem. 23, 1153-1 159. Gatt, S. (1968).Biochim. Biophys. Acta 159, 304-316. Goracci, G., Blomstrand, C., Arienti, G., Hamberger, A., and Porcellati, G. (1973). J. Neurochem. 20, 1167-1 180. Goracci, G., Francescangeli, E., Piccinin, G. L., Binaglia, L., Woelk, H., and Porcellati, G. (1975).J. Neurochm. 24, 1181-1186. Could, R. M., and Dawson, R. M. C. (1976).J. Cell Bwl. 68,480-496. Gregson, N. A., and Hall, S. M. (1973).J. Cell Sci. 13, 257-277. Grosman, N., and Diamant, B. (1975). Agents Actions 5,296-301. Hall, S. M>,,and Gregson, N. A. (197 I).J. Cell Sci. 9, 769-789. Hamberger, A,, and Svennerholm, L. (1971).J. Neurochem. 18, 1821-1829. Harvey, M. S., Wirtz, K. W. A., Kamp, H. H., Zegers, B. J. H., and van Deenen, L. L. M. (1973). Biochim. Biophys. Actu 323, 234-239.

28

G. BRIAN ANSELL AND SHEILA SPANNER

Hauser, H., and Dawson, R. M. C. (1968).Bi0chem.J. 109,909-916. Hawthorne, J. N., and Bleasdale, J. E. (1975).Mol. Cell. Biochem. 8 , 83-87. Helmkamp, G . M. Jr., Harvey, M. S., Wirtz, K. W. A., and van Deenen, L. L. M. (1974).J. Biol. Chem. 249, 6382-6389. Hendrickson, H. S., and Reinertsen, J. L. (1969). Biochernktry 8, 48554858. Hokin, L. E.. and Hokin, M. R. (1955).Biochim. Biophys. Acta 18, 102-110. Hokin, L. E., and Hokin, M. R. (1958).J . Biol. C h m . 233, 818-821. Hokin, M. R. (1969).]. Neurochem. 16, 127-134. Hokin, M. R. (1970).J. Neurochmn. 17, 357-364. Horrocks, L. A. (1969).J. Neurochtm. 16, 13-18. Horrocks, L. A. (1971). Int. SOC. Neurochem., 3rd Int. Meet., Budapest p. 312. (Abstr.) Horrocks, L. A., and Radominska-Pyrek, A. (1972). FEBS Lett. 22, 190-194. lllingworth, D. R., and Portman, 0. W. (1972). Biochem. J. 130,557-567. Johnson, S. M., and Bangham, A. D. (1969). Biocbim. Biophyr. Actn 193,92-104. Jungalwala, F. B. (1974). Brain Res. 78,99-108. Jungalwala, F. B., and Dawson, R. M. C. (1971). Biochem. J. 123, 683-693. Kanfer, J. N. (1972).J. Lipid Res. 13, 468-476. Kishimoto, Y., Agranoff, B. W., Radin, N. S., and Burton, R. M. (1969).J. Netirochem. 16, 397404. KLawans. H. L., and Rubovits, R. (1970),In “L-Dopa and Parkinsonism” (A. Barbeau and R. H. McDowall, eds.), pp. 107-1 16. Davis, Philadelphia, Pennsylvania. Klenk, E., and Kahlke, W. (1963). Hoppe-Seyler’s Z. Physiol. Chem. 333, 133-139. Kohlschutter, A., and Herschkowitz, N. N. (1973). Brain Res. 50,379-385. Kunze, H., Nahas, N., Traynor, J. R., and Wurl, M. (1976). Biochim. Ewphys. Acta 441, 93-102. Lapetina, E. G., and Hawthorne, J. N. (1971). Biochem. J. 122, 171-179. Lapetina, E. G., and Michell, R. H. (1973a). FEBS Lett. 31, 1-10. Lapetina, E. G., and Michell, R. H. (1973b). Bi0chem.J. 131,433-442. Lapetina, E. G., and Michell, R. H. (1974).J. Neurocha. 23, 283-287. Leibowitz, Z., and Gatt, S. (1968). Biochim. Biophys. Acta 164,439-441. Lennon, A. M., and Steinberg, H. R. (1973).J . Neurochem. 20,337-345. Lunt, G. G . , and Pickard, M. R. (1975).J. Neurochmn. 24, 1203-1208. MacBrinn, M. C., and OBrien, J. S. (1968).J. Lipid Res. 9, 552-561. McDonald, W. I., and Sears, T. A. (1970a). Brain 93,575-582. McDonald, W. I., and Sears, T. A. (1970b). Brain 93,583-598. McGeer, P. L., McGeer, E. G., and Fibiger, H. C. (1973). Lani-et ii, 623-624. McMartin, D. N., Horrocks, L. A., and Koestner,A. (1972).Acta Neuropathd. 22,288-294. Mann, S. P. (1975). Experientia 31, 1256-1258. Mantovani, P., Pepeu, G., and Amaducci, L. (1976). In “Function and Metabolism of Phospholipids in the Central and Peripheral Nervous Systems” (G. Porcellati, L. Amaducci, and C. Galli, eds.), pp. 285-292. Plenum, New York. Marchbanks, R. M. (1975). Int. J. Bwchem. 6, 303-312. Michell, R. H. (1975). Bwchim. Biophys. Acta 415, 81-147. Michell, R. H., and Lapetina, E. G . (1972). Nature (London) New Biol. 240,258-259. Michell, R. H., Jones, L. M., and Jafferji, S. S. (1976).1n “Stimulus-Secretion Coupling in the Gastrointestinal Tract” (R. M. Case and H. Goebell, eds.), pp. 89-103. MTP Press, Lancaster, England. Miller, E. K., and Dawson, R. M. C. (1972a).Eiochem.J. 126,805-821. Miller, E. K., and Dawson, R. M. C. (1972b). Biochem.J. 126,825-835. Norton, W. T., and Poduslo, S. E. (1971).J. Lipid Res. 12, 84-90. Nussbaum, J. L., Neskovic, N., and Mandel, P. (1971).J. Neurochem. 18, 1529-1543. OBrien, J. S. (1967).J. Theor. Biol. 15, 307-324.

FUNCTIONAL METABOLISM OF BRAIN PHOSPHOLIPIDS

29

OBrien, J. F., and Geison, R. L. ( 1 974). J. Lipid Res. 15,44-49. Pasquini, J. M., Krawiec, L., and Soto, E. F. (1973).J. Neurochem. 21, 6 4 7 4 5 3 . Poduslo, S. E., and Norton, W. T. (1972).J. Neurochem. 19, 727-736. Porcellati, G., Arienti, G., Pirotta, M., and Giorgini, D. (1971).]. Neurochem. 18,13951417. Porcellati, G., Amaducci, L., and Galli, C., eds. (1976). “Function and Metabolism of Phospholipids in the Central and Peripheral Nervous Systems.” Plenum, New York. Possmayer, F. (1974). Brain Res. 74, 167-174. Radominska-Pyrek, A., and Horrocks, L. A. (1972).J. Lipzd Res. 13, 580-587. Raghaven, S..Rhoads, D., and Kanfer, J. (1973). J . B i d . Chem. 247, 7153-7156. Ramsey, R. B., and Nicholas, H. J. (1972). Adv. Lipid Res. 10, 144-232. Rasminsky, M., and Sears, T. A. (1972).J. Physiol. (London) 227, 323-350. Rasmussen, H., Goodman. D. B. P., and Tenenhouse, A. (1972).CRC Crir. Rn).Hiodirtn. 1, 95-148. Roberti, R., Binaglia, L., Francescangeli, E., Goracci, G., and Porcellati, G. (I975).Lipids 10, 121-127. Rosenberg, P. (1970). Toxicon 8,235-243. Rosenberg, P., and Hoskin, F. C. G. (1963).J. Gen. Physiol. 46, 1065-1073. Rosenberg, P., and Ng, K. Y. (1963). Biochim. Biophys. Acta 75, 116-128. Schacht, J., and Agranoff, B. W. (1972).J. Bwl. Chem. 247,771-777. Schacht, J., and Agranoff, B. W. (1973). Biochem. Biophys. Res. Commun. 50, 934-941. Schacht, J., Neale, E. A., and Agranoff, B. W. (1974).J. Neurochem. 23, 211-218. Scherphof, G., and Westenberg, H. (1975). Biochim. Biophp. Acta 398,442451. Schmid, H. H. 0..and Takahashi, T. (1970).J. LipidRes. 11,412-419. Seeman, P. (1972). Pharmacol. Rev. 24, 583-655. Segall, H. J., and Wood, J. M. (1974). Nature (London) 248, 456-458. Sidman, R. L., Dickie, M. M., and Appel, S. H . (1964). Science 144, 309-31 1 . Singh, H., Spritz, N., and Geyer, B. (1971).J. Lzpzd Res. 1 2 , 4 7 3 4 8 1 . Sneddon, J. M., and Keen, P. (1970). Biochem. Phannacol. 19, 1297-1306. Snyder, F., ed. (1972). “Ether Lipids: Chemistry and Biology.” Academic Press, New York. Spanner, S., Hall, R. C., and Ansell, G. B. (1976). Biochem. J . 154, 133-140. Stavinoha, W. B., and Weintraub, S. T. (1974). Science 183, 964-965. Strickland, K. P. (1973). In “Form and Function of Phospholipids” (G. B. Ansell, R. M. C. Dawson, and J. N. Hawthorne, eds.), pp. 9-42. Elsevier, Amsterdam. Strickland, K. P., Thompson, R. H. S., and Webster, G. R. (1956).J. Neurol., Neurosurg. Psychiatty 19, 12-16. Sun, G. Y., and Horrocks, L. A. (1969).J. Neurochem. 16, 181-189. Sun, G. Y., and Sun, A. Y. (1972). Biochim. Biophys. Actn 280, 306-315. Takahashi, T., and Schmid, H. H. 0. (1970). C h m . Phys. Lip& 4, 243-246. Takeuchi, T. (196X). In “Minamata Disease” ( M . Kirksnna, ed.), p. 141. Kumamoto Univ. Press, Kumamoto, Japan. Tasaki, I. (1952).Jnp.J. Physiol. 3, 73-74. Tyson, C. A., Zande, H. V., and Green, D. E. (1976).J. Biol. Chem. 251, 1326-1332. Webster, G. R., Marples, E. A., and Thompson, R. H. S. (1957).Biochem. J. 65, 374-377. Whittaker, V. P. (1966). Ann. N . Y. h a d . Sci. 137, 982-998. Wirtz, K. W. A., Jolles, J., Westerman, J., and Neys, F. (1976). Nature (London) 260, 354355. Woelk, H., and Porcellati, G. (1973). Hoppe-Seyler’s 2. Physiol. Chem. 354, 90-100. Wu, Y.-C., Cho, T. M., Lok, H. H., and Way, E. L. (1976). Biochem. Pharmucol. 25, 15511553. Yagihara, Y., and Hawthorne, J. N. (1972).J. Neurochem. 19, 355-367. Yagihara, Y., Bleasdale, J. E., and Hawthorne, J. N. (1973).J. Neurochem. 21, 173-190.

This Page Intentionally Left Blank

ISOLATION AND PURIFICATION OF THE NICOTINIC ACETYLCHOLINE RECEPTOR AND ITS FUNCTIONAL RECONSTITUTION INTO A MEMBRANE ENVIRONMENT By Michael S. Briley' a n d Jean-Pierre Chongeux Neurobiologie Moleculaire lnstitut Pasteur, Paris, France

I. Introduction ........................................................... [ I . Isolation and Purification ... A. The Problem of Identification . . . . ............................ B. Purification of the Subsynaptic Membrane Fragments . . . . . . . . . . . . . . . . . . . C. Solubilization of the Receptor Protein ................................. D. Purification of the Solubilized Receptor Protein . E. Chemical and Structural Properties of the Purifie Receptor Protein ........ 111. Reconstitution ... ...... ...

31 32 32 34 39

44

A. Reconstitution Measured by Na+ Flux.................. 49 B. Reconstitution Measured by Bilayer Membrane Conductance . . . . . . . . . . . . 52 IV. Conclusion ............................................................. 58 R e f e r e n c e s . . . . . . . . . . . . . . . . . . . . ....................................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 59 9

1. Introduction

Synaptic transmission at the neuromuscular junction is mediated through nicotinic cholinergic receptors situated in the postsynaptic membrane. The binding of acetylcholine results in an increased cation permeability leading to a partial membrane depolarization. The nicotinic receptor has been studied in greatest detail using the electric organs of two electric fish, E l e c t r o p h m , a freshwater electric eel and Torpedo, a marine electric ray. The electric organs of these fish represent rich sources of large quantities of a single type of cholinergic synapse amenable to both electroph ysiological and biochemical analysis. The morphology and electrophysiology of these systems have received considerable attention and several recent reviews are available (Bennett, 1970; Rang, 1974; Changeux, 1975; Magazanik, 1976). In this chapter we will concentrate on the isolation and purification of the receptor pro-

' Present address: Synthelabo, 5X Rue de la Glacikre, Paris, France 31

32

MICHAEL S. BRILEY A N D JEAN-PIERRE CHANGEUX

tein, its physical and chemical properties, and attempts at the reconstitution of a functional receptor-membrane complex. Various functional aspects of the receptor such as changes in the binding affinity for agonists and the relationship between these changes and the structural and functional states of the receptor have been reviewed recently in considerable detail (Changeux et al., 1976) and will not be covered here. 11. Isolation and Purification

A. THEPROBLEM OF IDENTIFICATION Important to the purification of any biological material is its easy and unambiguous identification at all stages of purification. In the case of the cholinergic receptor this problem has been largely overcome by the use of covalent affinity labels (Singer et al., 1973; Karlin et al., 1973) and specific snake venom a-toxins (Lee, 1972).

1 . A@nity Labels The first reagent used to covalently label the receptor was the diazo derivative of phenyltrimethylammonium (PTA) (Changeux et al., 1967). This reagent acted as an irreversible antagonist and its action was delayed by the presence of reversible antagonists. After prior reduction with dithiothreitol (DTT)the receptor may be labeled with 4-(N-maleimido)-benzyltrimethylammoniumiodide (MBTA), a quaternary maleimido derivative (Karlin and Winnick, 1968). This irreversibly blocks the response to agonists. Its high degree of specificity has enabled the labeling of the receptor protein in vivo, on membrane fragments and in solubilized preparations (Reiter et al., 1972). A variety of other alkylating and acylating reagents have been tested (for review, see Karlin et al., 1973) but only MBTA and 4(N-maleimido)-phenyl-trimethylammonium iodide (MPTA) have been widely used. Recently Hucho et al. (1976) have used the photoaffinity reagent, 4-azido-2-nitrobenzyltrimethylammoniumfluoroborate, to label the receptor. Unlike the maleimido derivatives, this product labels all of the subunits. Labeling is inhibited by cholinergic agonists and antagonists. The specificity of this reagent has, however, been questioned (Ruoho et al., 1973) and thus its value as a specific receptor label has to be investigated further. 2. Snuke Venom a-Toxins Small basic proteins (molecular weight about 7000)*isolatedfrom the venom of certain -snakes block neuromuscular transmission and the

ISOLATION AND PURIFICATION OF NICOTINIC RECEPTOR

33

synapse of the electroplaque by acting at the level of the acetylcholine receptor (Lee, 1972). Although lacking any structural resemblance to cholinergic effectors, they bind with very high specificity and high affinity and act as very slowly reversible antagonists (Changeux et al., 1970, 1971). They probably bind to a site overlapping the cholinergic ligand binding site (Prives et al., 1972). These a-toxins have been radioactively labeled by a variety of methods (Lee and Tseng, 1966; Menezet al., 1971; Barnard et al., 1971; Cooper and Reich, 1972) and are now used in most laboratories for routine identification and assay of the receptor. The details of the assays used vary considerably, but in all cases involve separation of the receptor-toxin complex from the free toxin. With membrane fragments this is usually achieved by ultracentrifugation or the use of Millipore filters (Weber and Changeux, 1974). In the case of the detergent-solubilized receptor, the receptor-toxin complex has been separated from the free toxin by ammonium sulfate precipitation (Meunier et al., 1972a), filtration through Sephadex (Biesecker, 1973),or Bio-gel (McNamee et al., 1975b) columns. Since the toxin is a small positively charged molecule it can be readily separated from the receptor-toxin complex by passage through diethylaminoethyl (DEAE) resin or filter disk (Fulpius et aZ., 1972; Schmidt and Raftery, 1973). Although too small to be retained on a Millipore filter (0.45 p m diameter pore), the detergent-solubilized receptor forms large aggregates when the detergent is diluted to a concentration below the critical micelle concentration (cmc) in the presence of other detergent-soluble proteins and phospholipids. The aggregates may then be separated from the free toxin by Millipore filtration (Olsen et al., 1972; Meunieret al., 1974). All of these methods allow detection of picomole amounts of bound toxin and in some cases as low as 50 femtomoles. There are several contradictory reports concerning the relationship between the number of toxin binding sites and those binding small ligands or affinity labels. Kasai and Changeux (1971) found that in membrane fragments from Electrophorus the number of toxin binding sites were somewhat greater than those binding decamethonium. The purified receptor from the same source had a similar ratio of 1.6: 1 (Meunier and Changeux, 1973). Both membranes and purified receptor from Torpedo were found to have twice as many toxin binding sites as sites for acetylcholine, d-tubocurarine, or decamethonium (Moody et al., 1973). Chang (1974) found a ratio of 2 : 1, toxin : acetylcholine binding sites with the purified receptor. McNamee et al. (1975a) have found that purified receptor from Torpedo has twice the specific activity when measured by toxin binding than by affinity labeling with MBTA. In a detailed investigation, however, Weber and Changeux (1974) found the relationship between

34

MICHAEL S . BRILEY A N D JEAN-PIERRE

CHANGEUX

the number of toxin and decamethonium binding sites to be very close to unity. This has been confirmed by Sugiyama and Changeux (1975) who found it to be true for the purified receptor as well. Although the use of toxins and affinity labels has largely overcome the problem of identification of the ligand binding site of the receptor, the problem remains of determining whether, after detergent solubilization, the protein carrying the ligand binding site includes all of the apparatus necessary for receptor function (see Section 111). B. PURIFICATION

OF T H E

SUBSYNAPTIC MEMBRANEFRAGMENTS

The isolation of the receptor in the form of receptor-rich subsynaptic membrane fragments many of which exist as closed vesicles or “microsacs” represents an important stage in the total purification of the receptor. This material is “functional” in that the membranes are sensitive to permeability control by cholinergic effectors in a manner closely analogous to that found in vivo (Hazelbauer and Changeux, 1974; Popot et al., 1974, 1976). It is, at the same time, a greatly simplified system, comprising in its purest form only four protein bands by sodium dodecylsulfate (SDS)-gel electrophoresis (Sobel and Changeux, 1977; Hucho et al., 1976). Subcellular fractionation of electric tissue from either Electrophurus or Torpedo results in the separation of membrane fragments derived from the innervated face of the cells (rich in the acetylcholine receptor and acetylcholinesterase) from those derived from the noninnervated face (rich in NafKt ATPase) (Bauman et al., 1969; Duguid and Raftery, 1973a,b). The membrane fragments containing the receptor may then be separated from those containing the esterase (Cohen et al., 1972). Recent modifications of the original procedure (Sobel and Changeux, 1977) have enabled the separation of receptor-rich membrane fragments from the electric organ of Torpedo containing more than 4 nmoles of toxin binding sites per mg/protein with the acetylcholinesterase at least 100 times less. On the basis of the molecular weight of the receptor, this suggests that more than 50% of the protein of these membranes is present as the receptor. In addition to centrifugation, the method of affinity partition has been applied to the purification of these membrane fragments (Flanagan et al., 1976).This method is based on the principles of affinity chromatography (Cuatrecasas, 1970) and phase partition (Albertsson, 1960). It involves partitioning membrane fragments between two immiscible aqueous polymer solutions, one of which carries a bound affinity ligand. The results achieved are comparable to those described above but poten-

ISOLATION A N D PURIFICATION OF NICOTINIC RECEPTOR

35

tially suffer from the same disadvantage as those from affinity chromatography (see Section 11, D). 1. Functional Aspects The receptor-rich membranes form closed vesicles or "microsacs" which are capable of retaining ions such as Na+ and K+. Kasai and Changeux (197 1) using membranes from Electrophorus showed that these microsacs were chemically excitable, i.e., they responded to cholinergic agonists by an increase in their permeability to "Na'. This effect was blocked reversibly by d-rubocurarine or flaxedil (gallamine) and irreversibly by a-bungarotoxin. TheK,,, for agonists and antagonists agree well with those determined electrophysiologically on the intact electroplaque (Table I ) . Recently using a method which maximizes the proportion of agonist-inducible Na' flux, Hess et al. (1975, 1976) have analyzed the kinetics of this process. Similar results have recently been obtained using purified membrane fragments from Torpedo (Hazelbauer and Changeux, 1974; Popot et al., 1976). Here a set of seven ligands were tested and found to have the same relative order of affinities in ziztro as in z h o although the absolute I

TABLE I EFFECTORSWITH T H E RECEPTORFROM ELECTROPHORLT in Vir1o A N D in Vitro

INTERACTIONS OF C H O L I N E R C I C

K,,,,, (in 7 J 7 7 w ) " fM)

K,,, (in vitro)"

K , (a-toxin)"

(M)

(M)

-

1x IOP 8 x lor' 4 x lo-,>

Ag"n b h

2 x lo-" 2 x 10-1;

Acetylcholine Decamethonium Carbam ylcholine

3 x lo-.!

Antagonists d-Tirbocurarine Flaxedil Hexamethonium

3 x lo-' 3 x 10-5

2x

10-7

1 x lo-" 4 x lo-%

2 x lo-' 3 x 10-7 6 x lo-;'

2x 4x

10-7 10-7

6 X lo-"

KallP(in rlirro) apparent dissociation constant determined from depolarisation response of monocellular electroplaqiie (Higman et a / . , 1963: Changeux and Podleski, 1968: Weber and Changeux. 1974: Bartels and Nachmansohn, 1965: Mautner et al., 1966: Karlin and Winnik, 1968). " K,,,,, (in 7tilro) apparent dissociation constant determined from the N a + Hux respor.se of membrane vesicles isolated from El~c/rophorus(Kasai and Changeux, 197 1). ' K,, (a-toxin): protection constant determined from the ability of drugs to decrease by 50% the initial rate of binding 01. ['HI a-toxin o t N . nigrirollis to isolated membranes (Weber and Changeux. 1974). 'I

36

MICHAEL S . BRILEY A N D JEAN-PIERRE

CHANGEUX

values of the apparent dissociation constants appeared higher in vitro (Popot et al., 1976). Conductance measurements of the Torpedo electroplaquein vivo have shown that prolonged application of an agonist causes a reduction in the conductance response (Lester et al., 1975).This pharmacological “desensitization,” which resembles that seen at the neuromuscular junction (Katz and Thesleff, 1957), may also be observed in vitro using microsacs from Torpedo (Sugiyama et al., 1976).Preincubation of the microsacs for several minutes with agonists such as acetylcholine or carbamylcholine caused a subsequent decrease in the magnitude of the response to the same (or a different) agonist (Popot et al., 1974).This effect was blocked by antagonists and reversed by dilution. Both local anaesthetics and Ca2+ions, which enhance desensitization at the neuromuscular junction (see reference in Magazanik and Vyskocil, 1973), show a similar effect in vitro with Torpedo microsacs. A similar effect has been observed on the agonist binding affinity (Weber et al., 1975;for review of desensitization in vitro, see Changeux et al., 1976).

2. Physical Aspects Electron microscopy of the rece tor-rich membranes after negative diameter (Fig. 1) (Cartaud et al., staining shows particles of 80-90 1973; Raftery et al., 1975) or 60-70 A diameter (Nickel and Potter, 1973).These particles have been interpreted as rosettes of 5-6 subunits surrounding a central core. The rosettes are densely packed and sometimes seen arranged as a hexagonal lattice with a center-to-center distance of about 90 A. The lattice is seen more clearly with freeze etching (Cartaud et al., 1973). The purified receptors from Electrophms also appears as rosettes of 80-90 A (Meunier et al., 1974) (Fig. 2). X-ray diffraction studies of centrifugally aligned hydrated membrane specimens show a definite repeating unit in the plane of the membrane (Dupont et al., 1974) which, assuming a hexagonal lattice, has a center-to-center distance of about 90 A. Raftery et al. (1975)have confirmed the presence of an array but have calculated a center-to-center distance of 173 A. I n addition they find a repeating unit perpendicular to the plane of the membrane every 74 A with a secondary repeat every

1

37 A.

Brisson et al. (1975a)have recently introduced a potentially valuable probe for the physical environment of the receptor. The spin-labeled amphipathic cholinergic analog, 8(4’,4’-dimethyloxazolidine-N-oxyl) palmitoylcholine, appears to act as a reversible antagonist of the receptor. The choline end is thus anchored to the receptor binding site allowing the spin-labeled lipid chain to report on the immediate lipid envi-

ISOLATION A N D PURIFICATION OF NICOTINIC RECEPTOR

FIG. 1 . Electron micrograph o f a purified fragment of subsynaptic membrane from electric organ o f Torpedo mrmorata. The preparation according to Sobel and Changeux (1977) was negatively stained with uranyl formate (1 R). (Micrograph kindly provided by Dr. J . Cartaud, Universite de Paris VII.)

38

MICHAEL S. BRILEY A N D JEAN-PIERRE CHANGEUX

FIG. 2 . Electron micrographs of' highly purified receptor protein solubilized from electric organ of E/~trop/zonrselectricits by negative staining. (Reproduced from Meunier ei nl., 1974.)

I S O L A T I O N AND PURIFICATION OF N I C O T I N I C RECEPTOR

39

ronment of the receptor. Preliminary results suggest that this immediate lipid environment is completely immobilized. OF THE RECEPTOR PROTEIN C. SOLUBILIZATION

1. Detergent Sokrbilimtion The cholinergic receptor protein is not solubilized by aqueous salt solutions of high or low ionic strength (Olsen et al., 1972; Potter, 1973) or by prolonged sonication (Olsen et ul., 1972). Nondenaturing, nonionic detergents and bile salt solutions that grossly disrupt the membrane structure do, however, solubilize the receptor protein (i.e., it remains in solution after centrifugation at 100,000 g for 1 hour). Of the various detergents used, Triton X-100 (Miledi et al., 1971) and cholate or deoxycholate (Changeuxet al., 197 1 ) are those most commonly reported. The receptor protein may therefore be classified as an integral membrane protein (Singer and Nicolson, 1972). 2. Extraction with Organic Solwnts A novel approach to the isolation of the cholinergic receptor was introduced by De Robertis and his co-workers (for review, see De Robertis, 1971). This method involves the isolation of the hydrophobic receptor protein as a proteolipid by extraction with chloroform-methanol. The extract is incubated (still in chloroform-methanol) with a radioactive cholinergic ligand and then the mixture passed through a column of lipophilic Sephadex LH-20. Several protein peaks are eluted with chloroform-methanol mixtures of increasing polarity but only one, the “cholinergic proteolipid,” is associated with the radioactive ligand. A criticism of this technique (Levinson and Keynes, 1972), that the coelution of protein and ligand was artifactual, has since been refuted (Donellan & Cattell, 1975). There remains, however, several serious objections to this work. Potter (1973) found that the toxin-receptor complex is not extracted with chloroform-methanol, although apparently the cholinergic proteolipid, after being transferred to an aqueous medium in Triton X- 100, can then bind the a-toxin (De Plazas and De Robertis, 1972). Similar attempts in other laboratories, however, have found that treatment with chloroform-methanol irreversibly denatures the toxin binding site (Potter, 1973; Barrantes et al., 1975; Heilbronn, 1975). Similarly the covalent receptor-affinity label complex is not extracted with chloroformmethanol (Karlin, 1973: Barrantes et al., 1975). The fact that treatment with DTT (a preliminary to affinity labeling) greatly reduces the amount

40

MICHAEL S. BRILEY A N D JEAN-PIERRE CHANGEUX

of total proteolipid extracted (De Robertis et nl., 1976) has been offered as an explanation for this. However, if after affinity labeling the membranes are reoxidized with 5,5’-dithiobis(2-nitrobenzoicacid) (DTNB), the amount of proteolipid extracted is increased but there is no increase in the extraction of affinity label (Barrantes et al., 1976). Using a Torpedo-like electric fish,Narkejaponica, Kametari et al. (1975) found that most chloroform-methanol soluble acetylcholine binding material could be extracted from the membrane fraction rich in acetylcholinesterase and that very little was extracted from the receptor-rich membrane fraction. The values they obtained for the overall binding of acetylcholine and the dissociation constants were similar to those found by De Plazas and De Robertis (1972).Kametari et ul. (1975)were unable to demonstrate characteristic protein properties for the acetylcholine binding material, but it appears that this may have arisen from their lack of experience with proteolipids and the modifications necessary for their protein assay. Rabbit antiserum raised against purified detergent-extracted receptor from Electrophorus blocks the physiological response in vivo by an immune reaction, confirming its identity as the cholinergic receptor (Patrick and Lindstrom, 1973; Sugiyama et aL, 1973; Heilbronn and Mattson, 1974).Such antisera did not react against the cholinergic proteolipid transferred to detergent solution nor did antisera raised against the cholinergic proteolipid react against the purified receptor (Barrantes et al., 1975;Heilbronn, 1975).Finally a comparison of the protein composition of receptor-rich membranes from Torpedo by SDSpolyacrylamide electrophoresis shows that the band at 40,000 (the only one to be labeled with [ 3 HIMPTA) is not significantly diminished by extraction with chloroform-methanol. The cholinergic proteolipid, therefore, appears to differ in many respects from the now wellcharacterized, detergent-solubilized cholinergic receptor protein. D. PURIFICATION OF

THE

SOLUBILIZED RECEPTOR PROTEIN

Once the receptor had been solubilized in bile salts or nonionic detergents the way was clear for its purification. Difficulties arising from the presence of detergents and the hydrophobic nature of the receptor protein have meant that conventional methods of purification have not, in general, been very successful. However, starting from‘ a preparation of Torpedo membranes ( 1000 nmoles a-toxin boundgm protein), Potter (1973)used a purification procedure involving a sequence of sucrosedensity gradient centrifugation, ammonium sulfate precipitation, gel filtration, anion-exchange chromatography, and ultracentrifugation to

ISOLATION AND PURIFICATION OF NICOTINIC RECEPTOR

41

achieve a preparation close to purity (9500 nmoles a-toxidgm protein) but with a very low yield. Using Electrophorus, which has a much lower specific activity in the membrane fraction (about 50 nmoles a-toxin/gm protein), this approach has not been successful. In the case of mammalian muscle (about l nmoles/gm protein in the membrane fraction) the problems are even greater. The technique of affinity chromatography, introduced by Cuatrecasas (1970), was soon used in the purification of the cholinergic receptor. From the beginning, two basic types of affinity columns were developed. In one group various cholinergic toxins have been coupled directly to activated Sepharose beads. The receptor which binds to the immobilized toxin is eventually displaced by high concentrations of cholinergic ligands. In the other group a cholinergic ligand is attached to Sepharose via a spacer “arm.” Again the bound receptor protein is usually eluted with high concentrations of cholinergic ligand. Results obtained with both types of affinity columns are summarized in Table 11. Toxin columns have the advantage of high specificity. The first unambiguous demonstration that acetylcholinesterase and the acetylcholine receptor are different proteins came from the binding of the receptor to Naja nigricollk toxin coupled to Sepharose beads while acetylcholinesterase remained in solution (Meunieret al., 1971).Desorption from toxin columns is slow and requires very high concentrations of cholinergic ligands. With columns using coupled cholinergic ligands separation of the esterase and receptor is usually achieved by selective desorption using ligands with greater affinity for the receptor than for the esterase. Thus carbamylcholine (Karlin and Cowburn, 1973), flaxedil (Olsen et al., 1972; Meunier et al., 1974), and decamethonium (Biesecker, 1973) have all been used successfully. Similarly a salt gradient also gives selective elution (Schmidt and Raftery, 1972). Both methods usually give yields of 20-50%. While this is adequate from a preparative point of view, it leaves open the question of whether a specific subpopulation of receptor molecules is being selected and that the receptor thus purified is not typical of that existing in vivo. As yet no evidence has been published to support or deny this disturbing possibility. Another criticism of the purification of the receptor by affinity chromatography is the possibility of densitization of the receptor (see Changeux et al., 1976) by the cholinergic effectors which are used in high concentrations both coupled to the gel and as eluants. Sugiyama and Changeux (1975), using a crude Triton X-100-solubilized preparation from Torpedo, found that the affinity for acetylcholine of most of the

TABLE I1 SUMMARY OF AFFINITY CHROMATOGRAPHY METHODSUSED FOR PURIFICATION OF THE NICOTINIC ACETYLCHOLINE RECEPTOR AND COMPARISON WITH CONVENTIONAL METHODS

(fi )

Purity achieved specific activity pmole toxin bound/ gm protein"

E. electricus E . electricus

Hexa-50 Deca-10 or Benzo Q-0.1

6.6 4.1

T. munorata

Carb 0-500 gradient

2.2

T. manorata T. califonica T. nubiliana T. ocelhta E . electricus E . electricus E . electricus E. electricus T. mannoruta T. californica T. califmica

Carb- 1000 Carb- 1000 C a r b 100 Carb-1000 Deca-1 Carb-50 Bis Q-0.003 Flax-2.5 Flax-2.5 NaCl gradient Carb-50

Elution ligand and concentration Affinity ligand

Source material

Reference

ELECTRIC ORGAN P

N

a-Toxin, Najn nnjn

PTA

Flaxedil analog Quaternary ammonium PTA

7.8" 10 12.2" 10 4.5 4" 54.5 5.9 6.9 6.9 6

Klett et al. (1973) Lindstrom and Patrick (1974) Patrick et al. (1975) Karlsson et al. (1972) Heilbronn and Mattson (1974) Eldefrawi and Eldefrawi (1973) Eldefrawi et al. (1975a) Ong and Brady (1974) Rubsamen et al. (1976b) Biesecker (1973) Karlin and Cowburn (1973) Chang (1974) Meunier el al. (1974) Sugiyama and Changeux (1975) Schmidt and Raftery (1972,1973) Weill et a/.( 1 974)

MAMMALIAN SOURCES

a-Toxin. Nnjo nrda

PTA NOA'AFFINI TY METHODS

Conventional methods Conventional methods starting from extensively purified receptor-rich membranes

Mammalian sources Rat diaphragm muscle

Rabbit hindlimb muscle Mouse brain Cat denervated leg muscle

T. mnnornta T. mcinnorntcc

Carb-1000 Carb- 1000 Carh-200 or Hexa-200 Flax-2.1

0.19' 0.53' 0.2'

Not reported 3.5-6

9.3 7 .0

Brockes and Hall (1975) Bradley

~t 01.

(1976)

Romine et nl. (1 974) Dolly and Barnard (1975)

Potter (1973) Sobel and C;liange~x( 1977); Clrangeux P / rrl. ( I 977)

Ahh-Pi1icitions: Hexa, hexamethonium: Deca, decanrethoniiim: Benzo Q. benzoqitinacrium: Carh. carhamykho\ine; Flax, flaxedil (gallamine): bromide: PTA, phenyltrimethylanrmonium. Bis Q, 3,3'-his(n-(triniethylammoni~1m)metl~yl)-azohenzene " These values are final purities L I S L involving ~ ~ ~ sucrose-gradient centrifugation and/or ion-exchange chromatography after the affinity chromatography. " Acetylcholine binding sites. ' Protein by amino acid analysis (all others by the Lowry method). " MBTA binding sites which equal approximately 50% of toxin binding sites (McNaniee et 01.. 1975a). "Junctional receptors. ' Extrajunctional receptors.

44

MICHAEL S . BRILEY AND JEAN-PIERRE CHANCEUX

sites was decreased by about two orders of magnitude after passage through an affinity column (a flaxedil derivative coupled to Sepharose 4B with elution by 2.5 mM flaxedil [Meunier et al., 19741). In the light of these possible problems a purification procedure has recently been developed to avoid the use of affinity chromatography. Using a modification of the method of Cohen et al. (1972), receptor-rich membranes from Torpedo have been very highly purified so that their solubilization in Triton X-100, followed by centrifugation in a sucrose gradient, results in a purification at least equivalent to those achieved with affinity chromatography (Sobel and Changeux, 1977; Changeux et al., 1977). After affinity chromatography one or more other steps are required to purify the receptor to homogeneity. Centrifugation in a sucrose gradient (Meunier et al., 1974; Lindstrom and Patrick, 1974), passage through an ion-exchange column (Klett et al., 1973), and electrophoresis (Eldefrawi and Eldefrawi, 1973) are the most commonly used. Many of the preparations of the receptor protein are homogeneous by various criteria. Ultracentrifugation in a sucrose gradient gives a symmetrical peak which coincides with toxin binding (Meunier et al., 1974; Raftery et al., 1975; McNamee et al., 1975a). Polyacrylamide-gel electrophoresis in nondenaturing detergents at different pH gives one band (Dolly and Barnard, 1975; Raftery et d.,1975; Eldefrawi et al., 1975b) which may be labeled with toxin (Klett et al., 1973; Meunier et al., 1974). Cross-linking of the receptor and toxin with suberimidate (Hucho and Changeux, 1973) or glutaraldehyde (Biesecker, 1973) gives a single band on SDS-gel electrophoresis. Isoelectric focusing (Eldefrawi and Eldefrawi, 1973) and column chromatography (Eldefrawi et al., 1975a) have also been used. Electron microscopy of the purified receptor shows a homogeneous distribution of identical particles (Fig. 2) (Meunier et al., 1974). AND STRUCTURAL PROPERTIES OF E. CHEMICAL RECEPTORPROTEIN

THE

PURIFIED

1. Composition The amino acid composition has now been established for the purified receptor from Electrophmus and three species of Torpedo (Table 111). With the exception of one analysis (Klett et aZ., 1973) which failed to detect tryptophan, the receptor has been found to contain all of the commonly occurring amino acids. According to the classification of Capaldi and Vanderkooi (19721, the receptor contains about 46% polar residues, a value typical of globular water-soluble proteins. This would

45

ISOLATION AND PURIFICATION OF NICOTINIC RECEPTOR

TABLE I11 AMINOACIDC O M P O ~ I T I(mole/l00 ON mole$) O F T H F RELEITOR PROTEIN PURIFIED FROM T H E ELECTRIC O R ~ ~ oAt NElectrophom\ A N D Torpedo Amino acid Lysine Histidine Arginine Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Cysteic acid Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Tryptophan Glucosamine Reference

E. electricus 6.3 2.5 4.2 9.8 6.0 8.2 9.0 6.7 4.8 5.4 1.7 6.9 3.4 8.1 10.7 3.8 5.1 2.4

-

4.6 2.2 4.2 11.4 5.6 6.2 10.2 5.7 .5.9 5.8 2.0 8.6 2.0 6.4 10.5 4.0 5.7 0 -

n

b

T.mannorata

T.californzca

5.0 2.5 3.3 12.4 6.2 8.1 8.7 5.6 5.0 5.0

6.1 2.1 3.5 11.8 6.3 7.1 10.7 6.2 6.4 6.0 2.0 5.5 1.7 5.2 9.3 3.6 4.4 2.1 -

6.1 2.7 4.1 11.9 6.3 6.6 10.2 5.9 4.9 5.1 0.9 7.0 1.8 7.5 9.7 3.8 4.6 0.9

d

e

-

7.3 2.5 7.4 10.1 3.5 4.5 C

5.4 2.4 3.9 11.6 6.4 7.9 10.0 5.9 4.6 5.0 1.2 7.1 2.0 8.2 9.5 3.7 4.5 2.4 -

f

T.nobdinna 4.5 2.5 3.7 12.2 6.8 6.4 9.7 7.1 5.0 4.5 2.8 6.2 1.6 6.2 10.2 4.2 4.2 1.5 2.0

R

Kqr to Refwenus: (a) Meunier et al. (1974); (b) Klett et al. (1973); (c) Heilbronn and Mattson (1974); (d) Eldefrawi and Eldefrawi (1973); (e) Michaelson el al. (1974): (f) Eldef rawi ct NI. ( 1 975a): (g) Moore y t NI. ( 1 974).

suggest that the obvious hydrophobic nature of the receptor is derived from an asymetric distribution of the polar amino acids rather than its overall amino acid composition. On the other hand, using the same data, analyses of hydrophobicity by the methods of Barrantes (1975) or Bigelow (1967) suggest a similarity between the receptor and known integral membrane proteins (Raftery el al., 1976). The receptor appears not to possess any covalently bound phospholipid. No lipid phosphorous was detected (down to a limit of 1 mole P/mole toxin sites) in receptor purified from Electrophorus (Klett et al., 1973). The receptor does, however, contain carbohydrates. By its reaction with concavalin A and various other plant lectins, Meunier et al. ( 1974) detected the presence of D-mannose and N-acetyl-D-galactosamine. N-acetyl-D-glucosamine has been detected in receptor preparations from Torpedo (Michaelson et al., 1974; Moore et al., 1974). Preparations from this source also contain about 5% by weight neutral

46

MICHAEL S . BRILEY AND JEAN-PIERRE CHANGEUX

sugars : mannose, galactose, and glucose in the ratio 8 :2 : 1 (Raftery et al., 1975). In a‘similar preparation, Heilbronn (1975) reported mannose, galactose, and glucose in a similar ratio 8 : 1.8 :0.2. Tests for sialic acid were negative. On a cautionary note, contamination from agarose columns used during preparation exists as a strong possibility and may substantially alter the very small amounts of sugars detected. Analysis by atomic absorption revealed that the receptor contained 4.7% by weight of bound Ca”+ which was not removed by extensive dialysis (Eldefrawi et al., 1975~).Receptor prepared with Ca2+-freesolutions containing l mM ethylenediametetraacetate (EDTA) still bound 0.7% (by weight) Ca2+ (15 moles Ca2+/moleacetylcholine binding site). The significance of this bound divalent cation has not yet been established. Using the fluorescent lanthanide, terbium, as a fluorescent probe, Riibsamen et al. (1976a) demonstrated two types of terbium binding sites both withK,,, of 1.8 x M . About 60% of these sites bind Ca2+with a K,,, of about 1 x 10-3M. These sites which are located on the 40,000 dalton subunit (Riibsamen et al., 197613) also interact with cholinergic agonists but not antagonists such as the a-toxin. These results provide some support for a possible mechanism of changes in membrane ion permeability initiated by activator-induced displacement of Caz+ as has been suggested by Nachmansohn and Neumann (Nachmansohn, 1974; Nachmansohn and Neumann, 1975). 2. Size and Subunit Structure Sucrose-density gradients, in detergent, of crude receptor preparations from Torpedo show two distinct bands of receptor, a predominant one of 9 s and the other of 12s (Raftery et al., 1972; Potter, 1973). The corresponding Stokes radii were found to be 7.0 nm and 8.5 nm, respectively. Triton X-100 solubilization in the presence of 10 mM P-mercaptoethanol gives a single band of receptor of 9 s when centrifuged in a gradient containing 10 mM P-mercaptoethanol (Changeux et al., 1977; Sobel and Changeux, 1977). Centrifugation of the purified receptor from Electrophwus in a detergent-containing sucrose gradient gives a single symmetrical peak with a sedimentation coefficient of 9 S (Meunier et al., 1972a,b; Raftery et al., 1971). Gel filtration through Sepharose 6B in the presence of detergent gives a Stokes radius of 7.3 nm (Meunier et al., 1972b; Raftery et al., 1971). This value is considerably larger than that expected from sedimentation data and was explained by the presence of a large amount of bound detergent (greater than 0.1 mg Triton X-lOO/mg protein) increasing the size and bouyancy of the receptor (Meunier et al., 1972a). This data, after correction for the detergent present, gave estimates of

ISOLATION A N D PURIFICATION OF NICOTINIC RECEPTOR

47

320,000-350,000 daltons (Hucho and Changeux, 1973). Sedimentation equilibrium centrifugation of the purified receptor from Torpedo gave two species of 330,000 and 660,000 daltons, the latter reducing to 330,000 daltons when the Triton X- 100 concentration was increased to 0.1% (Edelstein et al., 1975). Centrifugation of crude purified preparations of receptor from three different species of Torpedo each gave t w o major bands of toxin binding material of molecular weights 190,000 and 330,000 (Gibson et al., 1976). Both oligomers showed similar binding properties. Gel electrophoresis in SDS of the purified receptor from Electrophorus after cross-linking with gluteraldehyde (Biesecker, 1973) or suberimidate (Hucho and Changeux, 1973) gave apparent molecular weights of 260,000 and 230,000, respectively, using globular, nonmembrane proteins as standards. After correction for the underestimation often found when using globular proteins as standards for membrane proteins (Spatz & Strittmatter, 1973), a value of 275,000 was determined (Hucho and Changeux, 1973). SDS-gel electrophoresis of homogeneous preparations from Electrophorus under denaturing conditions has, in general, resulted in t w o bands of about 40,000 and 50,000 daltons (Hucho and Changeux, 1973: Biesecker, 1973; Lindstrom and Patrick, 1974). There are also reports of three bands (40,000, 47,000, and 53,000 daltons) (Karlin et al., 1976), four bands (48,000, 54,000, 60,000, and 110,000 daltons) (Patrick et al., 1975), and one band (size not reported) (Klett et al., 1973). I t is well established that only the subunit of 40,000 daltons binds the affinity label MBTA (Karlin and Cowburn, 1973; Meunier et al., 1974) and therefore carries the cholinergic binding site. This is further demonstrated by the finding that the a-toxin cross-linked to the receptor by suberimidate is found associated with only one subunit of about 40,000 daltons (Gordon et al., 1974). Similarly, the affinity label has been found only associated with the band of 40,000 daltons after labeling the receptor in intact electroplaques (Reiter et al., 1972) and in membrane fragments (Karlin and Cowburn, 1973). Patrick et al. (1975) have suggested that components of less than 48,000 molecular weight arise from proteolytic digestion. This, however, would appear to be an oversimplification since the band at 40,000 daltons occurs equally in receptor purified from Twpedo in the presence of- the protease inhibitor phenylmethyl-sulphonylfluoride (PMSF) (Sobel and Changeux, 1977: Hucho et nl., 1976). Partial cross-linking of the purified receptor gives at least six bands on subsequent SDS-gel electrophoresis (Hucho and Changeux, 1973), suggesting that the receptor is composed of at least five subunits. This is compatible with the rosettes of 5-6 subunits observed by electron microscopy both in membrane fragments (Cartaud et al., 1973; Nickel and

48

MICHAEL S. BRILEY A N D JEAN-PIERRE

CHANCEUX

Potter, 1973) and the purified receptor in detergents (Meunier et al., 1974; Eldefrawi et al., 1975b). The situation with the receptor from Torpedo appears to be more complex. Although a single band on SDS-gel electrophoresis has been reported (Potter, 1973), most studies show multiple bands. Gordon et al. (1974) showed the existence of five bands of 37,000, 49,000, 74,000, 93,000, and 148,000 daltons. T h e higher molecular weight bands, however, are almost certainly aggregates since their abundance increased with the age of preparation. Four bands of 40,000, 50,000, 60,000, and 64,000 daltons in a constant ratio of 4 : 2 : 1 : 1 have been reported (Raftery et al., 1976). The amino acid compositions of the 4 subunits did not differ greatly from each other and each had a similar mean hydrophobicity. Similar subunit patterns have been obtained by Karlin et al. (1976) who reported four subunits of 39,000,48,000,58,000, and 64,000 daltons in the ratio of 5 : 1.5 : 1 : 1.5 and by Rubsamen et al. (1976b); 4 subunits of 40,000, 50,000, 61,000, and 81,000 daltons in the ratio of 3.1: 1.4: 1.0: 1.1. Highly purified Torpedo membrane fragments showed four bands on SDS-gel electrophoresis of 40,000, 43,000, 50,000, and 66,000 daltons (Sobel and Changeux, 1977; Changeux et al., 1977) or 40,000, 48,000, 62,000, and 66,000 daltons (Hucho et nl., 1976). After solubilization and further purification only three bands (40,000, 50,000, and 60,000) were detected. Furthermore these bands varied in their relative proportions from one preparation to another suggesting that the two minor bands may be tenacious impurities (Sobel and Changeux, 1977; Hucho et nl., 1976). In the purest preparation these bands were found in a ratio of 6.3 : 1.4 : 1 (Sobel and Changeux, 1977). As with the receptor from Electrophorus, the only protein band from Torpedo to be affinity labeled is that of 40,000 daltons (Karlin et al., 1976; Changeux et nl., 1977; Sobel and Changeux, 1977). All four of the peptides found by Karlin et al. (1975) and by Raftery et al. (1976) react positively with periodic acid-Schiffs reagent (PAS) indicating the presence of carbohydrate moieties. In each band, mannose, glucose, galactose, and a trace of N-acetylglucosamine have been identified (Raftery et al., 1976). T h e purest protein preparations of receptor from fish electric organ have specific activities near or approaching 10 pmoles of a-toxin binding sites/gm protein (Table I I) or 1 mole of binding sitell 00,000 gm protein. In combination with the molecular weight estimations by physicochemical methods (in the region of 200,000-300,000 daltons), this suggests that there are at least two and possibly three cholinergic binding sites per receptor molecule. In the case ofElectrophorus this might be three of each

ISOLATION A N D PURIFICATION

OF N I C O T I N I C RECEPTOR

49

of the two types of subunit (i.e., 3a 3b) giving a hexamer. With Torpedo the picture is less clear and it is too early to suggest any molecular structure.

IV. Reconstitution

The advances made in our understanding of the nature of the receptor binding site, its various affinity states, its chemical and physical structure have not been matched by our knowledge of how agonist binding controls permeability and the nature of the associated ionophore.‘ At a more specific level one would like to know if all of the components of the functional receptor complex exist in the purified receptor protein isolated on the criterion of the toxin binding site. Equally, do the different types of subunits carry different functional components and if so how do they relate to one another? Permeability changes in artificial biological membranes can be measured directly as the flux of radioactive ions through the membrane or as changes in membrane resistance using the “black lipid” membrane or bilayer technique.

A. RECONSTITUTION MEASUREDBY NA+ FLUX Physical reintegration of the receptor into a membrane structure has been achieved by several groups. Receptor-rich membrane vesicles solubilized in ionic detergents such as sodium cholate or deoxycholate have been reconstituted by removal of the detergent by extensive dialysis. This results in the recovery of particulate material containing the toxin binding activity. Under the electron microscope this material appears as closed membrane vesicles similar to the native “microsacs” (Changeux et al., 1972). A similar preparation was reported to retain **Na+and to show a carbamylcholine-sensitive increase in Na+ efflux which was blocked by toxin, in other words a functional reconstitution. Details of this work, however, have never been published (unpublished observations, cited in Potter, 1973). Later Hazelbauer and Changeux (1974) provided the first demonstration that the receptor could, under certain conditions, be reintegrated into a membrane in a functional form. Receptor-rich membrane vesicles from Torpedo solubilized in sodium cholate were dialyzed for 48 The term ionophore is used here in a general sense and refers to any selective ion pathway, channel, or pore.

50

MICHAEL S. BRILEY A N D JEAN-PIERRE CHANGEUX

hours at 4OC. During this time most of the protein and toxin binding capacity was retained. There was, however, a significant loss of phospholipid. To compensate for this a sonicated aqueous dispersion of phospholipids extracted from Torpedo membrane fragments was added after dialysis. To reproducibly obtain closed vesicles it was necessary to add divalent cations, Mpf+ and Ca2+.Centrifugation in a sucrose gradient then gave two fractions, the heaviest of which retained 22Na' and under the electron microscope appeared as vesicles 300-1000 A in diameter bordered by a membrane 70 A thick. T h e rate of Na' efflux was accelerated by carbamylcholine; this increase was blocked by the a-toxin (Fig. 3A). TheK,,, for carbamylcholine was found to be in the region of 5 x lo-" M (Fig. 3B) similar to that obtained for native Torpedo microsacs (Popot et al., 1976). Functional reintegration of the purified receptor into phospholipid vesicles has proved to be difficult. Some partial success has been reported by Michaelson and Raftery (1974).The purified receptor from Torpedo in Triton X-100 was retained on a column of DEAE-cellulose and washed extensively and eventually eluted with solutions of sodium cholate in order to exchange the virtually nondialyzable Triton X-100 for the more freely dializable sodium cholate. A mixture, in 2% sodium cholate, of phospholipids and neutral lipids from Torpedo membranes and receptor protein in the ratio of lipid to protein 10: 1 (w/w) was dialyzed against detergent-free buffers at room temperature for 48 hours (dialysis at 4°C gave a physical but nonfunctional reintegration). After dialysis physical

50 A

0

I

10

I

I

20

30

minutes

10-6

10-7 10-6 1 0 - 3

I O - ~ 10-3

[carbamylcholine].M

FIG. 3. Na' emlux from reconstituted receptor vesicles. A. Reconstituted vesicles equilibrated with 2L'NaC were diluted into Turpedo Ringer's solution ( O ) ,Ringer's solution containing carbamylcholine lo-' M (A),or Ringer's solution containing carbamylcholine lo-' M after preincubation with a-toxin 10P M (m). B . Concentration dependence of' excitability in membrane vesicles (0)and reconstituted receptor vesicles ( A ) . Excitability is calculated a$ ( / t / t + )- 1 . (Reproduced from HaLelbauer and Changeux. 1974.)

ISOLATION A N D PURIFICATION OF NICOTlNIC RECEPTOR

I

7-

---i&c

51

10

TIME (rnin)

FIG. 4. Na+ efflux from vesicles reconstituted from purified receptor from Torpedo. Vesicles were equilibrated with "Na' and diluted into 200 mM NaCI, 10 mM Tris/HCI. p H 7.4 (01the , same solution containing carbamylcholine lo-' M ( O ) , or the same solution containing carbamylcholine lo-' M after preincubation with excess toxin ( W j . (Reproduced from Michaelson and Raftery, 1974.)

reassociation of the receptor and lipids was demonstrated by cosedimentation in a sucrose gradient. Intact reconstituted vesicles and those resolubilized with Triton X- 100 showed similar toxin binding capacities, indicating that the toxin binding sites were asymetrically arranged in an "all-facing out" orientation. The vesicles, which showed osmotic sensitivity, retained "Na+ although they were rather leaky (ti efflux varied between 5 and 15 minutes at room temperature). The rate of efflux was increased in the presence of carbamylcholine at 10-.' M . This excitability was blocked by preincubation with a-toxin (Fig. 4). No evidence was presented on the concentration dependence of the excitability and therefore its similarity to the native system. Furthermore many preparations yielded nonexcitable vesicles and there was no apparent correlation between membrane excitability and the binding capacity for the a-toxin. As yet these results have not been confirmed by other workers. McNamee et al. (1975a,b)have used two methods to incorporate receptor protein purified from Electrophomis and from Torpedo into phospholipid vesicles. In one method detergent-depleted receptor was cosonicated with a mixture of egg lecithin and phospholipids from Electrophwzis receptor-rich membranes. Physical reassociation only was reported (McNamee et al., 1975a). The other method was essentially the same as that used by Michaelson and Raftery (1974) but in this case no excitability could be achieved.

52

MICHAEL S . BRILEY AND JEAN-PIERRE CHANGEUX

Receptor in Triton X-100 purified from Torpedo was detergentexchanged for sodium cholate by centrifuging into a sucrose gradient containing cholate or by repeated dilution with cholate and reconcentration by ultrafiltration. T h e receptor in cholate was mixed with egg lecithin and Twpedo phospholipids and neutral lipids in the ratio of lipid to protein 10: 1 (w/w) and dialyzed at 4°C or 25°C for 72 hours against detergent-free buffers. This treatment resulted in no loss of MBTA labeling capacity. Physical reassociation was again demonstrated by cosedimentation of the receptor and lipids. T h e vesicles formed had a low permeability to .“a+ (t+ influx = 2-3 hours at 25°C) but as found by Michaelson and Raftery (1974) showed no difference between those containing the receptor and those corn posed solely of lipids. The rate of sodium influx could be accelerated by the addition of such artificial ionophores as gramicidin D and valinomycin. Carbamylcholine, however, had no effect on the rate of sodium influx (McNamee et al., 1975b). The uncertain “state of the art” is exemplified by various contradictions between different reports. Hazelbauer and Changeux (1974), for example, stress the importance of divalent cations for the formation of closed vesicles. Others, however, have found no such requirement (Michaelson and Raftery, 1974; McNamee et al., 1975b). Michaelson and Raftery (1974) find that the dialysis must be carried out at room temperature to obtain functional reintegration, whereas Hazelbauer and Changeux (1974) achieved a functional reintegration after dialysis at 4°C. Michaelson and Raftery (1974) have found that 100% of the toxin sites were on the outside of their reconstituted vesicles. Using similar reconstitution methods of cholate dialysis, others have found a symmetrical 50% inside, 50% outside arrangement (Potter, 1973; Briley and Changeux, 1976, 1977; Changeux et al., 1977). Thus the results obtained to date suggest that the functional reconstitution of the purified receptor into membrane vesicles is possible but that all of the parameters involved are not yet identified or controlled.

B. RECONSTITUTION MEASUREDBY BILAYERMEMBRANE CONDUCTANCE 1. “Reconstitution” of Nonionophoric Proteins

Very soon after Mueller and Rudin introduced the technique of bilayer membrane (black lipid films) (Mueller et al., 1962) as a model for biological membranes, attempts were made to modify the membrane properties by inserting specific functional proteins. Del Castillo et a!.

ISOLATION A N D PURIFICATION OF NICOTINIC RECEPTOR

53

(1966) found that trypsin (and other enzymes such as chymotrypsin, lactate dehydrogenase, glutamate dehydrogenase, urease, and acetylcholinesterase) added to the bath became associated with the membrane. Subsequent addition of a substrate caused a drop in membrane resistance of about three orders of magnitude in a manner compatible with the activity of the enzyme. Similarly the incorporation of antigens caused a resistance decrease on subsequent addition of the corresponding antibodies. Although the membranes had a relatively high conductivity (- lov6 mho/cm2 compared with lO-'O mho/cm2 for good membranes made currently), the enzymes appeared to be acting only at the membrane surface since the addition of the substrate and enzyme to opposite sides of the membrane caused no resistance change. Leuzinger and Schneider (1972), working under the misapprehension that acetylcholinesterase and the acetylcholine receptor were identical, incorporated purified preparations of the esterase into bilayers of high resistance (mho/cm2)and found resistance decreases of two orders of magnitude on addition of acetylcholine to the opposite side of the membrane. This effect was blocked by eserine (an inhibitor of acetylcholinesterase which has no action on the acetylcholine receptor at the concentration used). Further work with purified esterase (again on the assumption that it was identical with the receptor) incorporated into high resistance (- lo-' mho/cm2) bilayers gain et al., 1973) showed enzyme-mediated conductance changes on addition. of acetylcholine or carbamylcholine3 to the opposite side of the membrane. These effects were blocked by neostigmine,' atropine," a-toxin;j and d-tubocurarine." In the light of current knowledge that acetylcholinesterase and the acetylcholine receptor are different proteins with different binding properties, these results are difficult to interpret. In a later paper Jain (1974) suggested that the receptor was an impurity of the esterase preparation used. Using a commercial preparation of acetylcholinesterase, he went on to show that dose-response curves for various nicotinic agonists were similar to those with the eel electroplaque. In addition a cation selectivity similar to that occurring in the electroplaque (Na : K : C 1, 3 : 3 : 1) was demonstrated. The previous results with neostigmine and atropine, however, were not explained.

-

Carbamylcholine is a cholinergic agonist which is not a substrate for acetylcholinesterase. ' Neostigmine is an inhibitor of the esterase but does not block the nicotinic receptor. Atropine blocks the muscarinic receptor but not acetylcholinesterase or the nicotinic receptor. 0-Toxin and d-tubocurarine block the nicotinic receptor but are without effect on the esterase.

54

MICHAEL S. BRILEY AND JEAN-PIERRE

CHANGEUX

The incorporation of enzymes into bilayer membranes which results in enzyme activity-induced conductivity changes clearly demonstrates the ambiguity of this method and the difficulty of distinguishing between artifacts inherent in the technique and any conductivity changes due to the acetylcholine receptor. Other experiments which, in the light of more recent findings, appear more confusing than enlightening are those of De Robertis and co-workers. Using a system which gave control bilayers of relatively low resistance ( mhokm’), they incorporated the cholinergic proteolipid (see Section 11, C, 2) into the “membrane-forming solution” prior to bilayer formation (Parisi et al., 1971, 1972). Subsequent local addition of acetylcholine to an unstirred bath gave a small ( 5 - to 10-fold) increase in membrane conductance. This increase was transient, presumably due to the dispersion of the concentrated agonist added. The addition of d-tubocurarine also gave an increase in conductance but less than that of acetylcholine. The effect was not seen with a “noncholinergic” proteolipid. Later the work was extended to electron microscopy of the membrane in the presence and absence of acetylcholine (Vasquez et al., 1971). A change in the appearance of the membrane was interpreted as an opening of the ionophoric channels. As already discussed (see Section 11, C) the cholinergic proteolipid probably differs considerably from the acetylcholine receptor and thus these results are more a demonstration of bilayer artifacts than a recovery of receptor function. A possible explanation for these artifacts has come from Parisi et al. (1975). They have shown that the presence of negatively charged lipids in a bilayer membrane can produce an apparent cholinergic excitability which can be prevented by d-tubocurarine! Whether negatively charged lipids could also explain the binding properties of the cholinergic proteolipid (De Robertis, 197 1) was not investigated.

2. Reconstitution of the Acetylcholine Receptor into Bilayers The first attempt to reintegrate the receptor itself into a bilayer used a partially purified preparation of receptor from rat diaphragm muscle (Kemp et al., 1973). The receptor in 0.6% Triton X-100 was added (at a final dilution of 1000-fold) to the bath after the membrane had thinned. Conductance measurements showed a linear increase with time (Fig. 5). In the presence of acetylcholine, at 5 x M, the rate of increase was accelerated about 4-fold. This effect was blocked by both a-toxin and d-tubocurarine. The sensitivity to acetylcholine was lost on storage although the toxin binding capacity was not altered. No dose-response curves were reported but the authors did note an ion specificity of 3 : 1, K’ : Na’ both in the presence and absence of acetylcholine.

ISOLATION A N D PURIFICATION OF NICOTINIC RECEPTOR

55

'Oi

9 1 N

'E

-I

E

l

6 1 u C c

U U

5

0

4'

TIME (Min)

FIG. 5 . Increase of bilayer membrane conductance by incorporation of receptor. The receptor preparation was added to one side at zero time. Bathing solutions (both chambers) contained 0.1 M KCI (lower curve) or 0.1 M KCI plus acetylcholine (upper curve). (Reproduced from Kemp et al., 1973.)

Shamoo and Eldefrawi (1975) recently described experiments using a purified receptor preparation. A purified and well-characterized receptor preparation from Torpedo (Eldefrawi et al., 1975b) was incorporated into high-resistance bilayer (lo-' mhokm'). With Ca'+ in the bath the conductance increased with time, this increase being proportional to the amount of receptor added. In the presence of Na+ instead of Ca2+the conductance increase was 20- to 100-fold less for the same protein concentration. In either case there was no reproducible effect of carbamylcholine at M. In an attempt to facilitate receptor incorporation into the membrane the receptor was partially digested with trypsin and then treated with Sephadex CM 50 to remove small positively charged peptides. This resulted in a 5% loss of protein and a decrease in the molecular weight of most of the protein bands on SDS-gel electrophoresis. This trypsin treatment, however, had no effect on the acetylcholine binding activity of the receptor. In the presence of Ca2+ the trypsin-treated receptor showed little difference to the native receptor. In the presence of Na+, however, at low protein concentrations no conductance increase with M caused a large time was seen. The addition of carbamylcholine at increase in conductance which could be blocked by curare (Fig. 6). The addition of curare after carbamylcholine could not reverse the conductance increase. The ion specificity for the carbamylcholine-induced con-

56

MICHAEL S. BRILEY A N D JEAN-PIERRE CHANGEUX

1

50

TIME(MIN)

FIG.6. Excitability of trypsin-treated purified receptor incorporated into a lipid bilayer membrane. Trypsin-treated receptor (0.8 pglml) was present (in 0.1 M NaCI, 5 mM histidine pH 7.3) in both compartments before formation o f the bilayer at zero time. The concentration of carbamylcholine and curare were 10P M . Arrow indicates membrane rupture. (Reproduced from Shamoo and Eldefrawi, 1975.)

ductance was Na' : K+ : C1-, 4.4 : 4.4 : 1. The increased conductance level remained constant in the presence of constant concentration of carbamylcholine. In other words it did not show desensitization as seen in uivo and in vitro (see Section 11, B). No dose-response curves were reported. While these experiments show some of the properties expected from in uivo and in vitro studies, the necessity of tryptic digestion and the lack of reversibility and desensitization make their interpretation less straightforward. The conductance of a single-receptor channel has recently been estimated from statistical analysis of random "noise" fluctuations in muscle end-plates during iontophoretic application of acetylcholine (Katz and Miledi, 1970, 1972a,b) and found to be about lO-'O mho with an average lifetime of about 1 msec. Further studies with Fourier Transform spectra and end-plate current "noise" (Anderson and Stevens, 1973) gave estimates of unit channel conductance of 3.2 X 10"' mho with an openchannel lifetime of 6-1 1 msec. Some workers have recently attempted to detect such unit-channel conductance in lipid bilayers containing the receptor. Goodall et al. (1974) purified the nicotinic receptor from mouse brain by applying a Triton X-100 extract to a toxin affinity column. The fraction eluted with a carbamylcholine gradient was taken to be the receptor. No biochemical determination of its identity or purity was made. The putative receptor was incorporated into bilayers in the presence of carbamylcholine at 5 x lop4M. Two series of discrete conductance steps or quanta were observed, 1.5 X lo-'' mho (Na'), 2.4 x lo-"' mho (Cl-), and 3.7 x lo-'' mho (Na+),5.9 x lo-'' mho (Cl-).

ISOLATION A N D PURIFICATION OF NICOTINIC RECEPTOR

57

The larger quanta, which appear to be four times the smaller, were diminished in the presence of d-tubocurarine or if the carbamylcholine was removed. This study was extended by Romine et al. (1974) who used receptor purified from both hog brain and mouse brain by toxin affinity columns. Again no biochemical analyses of the receptor were made. The receptor from hog brain gave minimal quanta of 3.8 X lo-” mho (Naf) and 4.7 x lo-’’ mho (C1-) (Fig. 7). Again larger steps were also observed which could be diminished by the addition of d-tubocurarine or atropine (sic). The lifetime of these large steps (the order of several seconds) is too great to result from single channel openings which are more probably part of the membrane “noise.” The aggregated steps were suggested to be due either to groups of receptor opening in a cooperative manner or to some phenomenon associated with desensitization. Equally they may reflect an irreversible channel opening. It is also possible that they may result from some experimental artifact such as membrane instability or receptor exchange with the bath medium. Further work with the nicotinic receptor from mammalian skeletal muscle (Bradley et al., 1976) again showed similar quantal conductance changes again of long duration (tens of seconds). Their frequency was enhanced by the addition of carbamylcholine and reduced by curare and a-toxin but not by atropine. Concanavalin A also showed an antagonistic effect suggesting the involvement of glycoproteins in the phenomenon. Preincubation of the receptor with the reducing reagent, DTT, also reduced the frequency of quantal events. No new explanations were offered to account for the long lifetime of the quanta. The authors did,

10-l0

L 10s

24

FIG. 7 . Bilayer response after addition of receptor extract from hog brain tissue following elution with hexamethonium. Note the increase in “noise” as more quantal jumps in conductance occur. Smallest size conductance increase to Na+ is 3.8 X 10-l’mho. (Reproduced from Romine et al., 1974.)

58

MICHAEL S. BRILEY AND JEAN-PIERRE

CHANGEUX

however, note that in similar reconstitution experiments with receptor isolated from fish electric organ they were unable to observe any quantal changes. The study of membrane “noise” seems to be a potentially sensitive and informative method for determining the properties of receptorcontaining bilayers especially if the temporal resolution in the millisecond range is available. As yet, however, the properties of these quantal events have only some of the properties to be expected for the reconstituted receptor system.

IV. Conclusion

The isolation and purification of the receptor is now well advanced largely due to an abundant source of receptor, the electric organ, and the highly specific, high-affinity ligand, the snake a-toxin. Soon one can expect the quaternary structure to be unequivocally assigned and probably the determination of the primary sequence, at least in the region of the binding site. At the membrane level the various functional states of the receptor and their transitions are becoming increasingly available to study through the use of fluorescent probes (see Changeux et al., 1976, 1977). In spite of considerable attention and some partial successes the functional reconstitution of the receptor and hence our understanding of its mode of action has not progressed very far. Physical association of the receptor with both bilayers and vesicles has been inferred by various methods but the integration of the receptor into the lipid bilayer and its orientation across the membrane have yet to be demonstrated. The functional reintegration of the receptor into vesicles has not been reproducibly demonstrated. With bilayers the distinction between artifactual and receptor-mediated conductance changes has complicated the interpretation of this data. Furthermore, in many cases, the so-called receptor is ill defined in biochemical terms. One may pose the question: Why does the reconstitution of the receptor appear to be so difficult in comparison with other apparently analogous systems such as the adenosine triphosphatase (ATPase) Ca2+ pump (MacLennan, 1975; MacLennan and Holland, 1975)? In order to achieve a functionally active reconstituted receptor-membrane complex it is necessary, on the basis of our current ideas, that the receptor span the membrane orientated with its binding site to the source of agonist. (In a random arrangement this would of course be true for 50% of the sites.) T h e receptor must be in the resting state but capable of activation. The ionophore may be present as a pore in the tertiary

ISOLATION AND PURIFICATION OF NICOTINJC RECEPTOR

59

structure of each subunit, as a specific ionophoric subunit, as a pore formed by the quaternary striictiire of the s~tbunitsof’a single receptor oligomer, or as a pore formed by the quinternary str~ctiireof‘ protein and lipids in the membrane. T h e latter may require a specific “annulus” of lipids in the immediate environment of the receptor as postulated for the ATPase (Warren ut d.,1975). Each of these possible striictiire~ implies its own conditions to be fiilfilled for activity. In addition one may suggest that the lateral stabililty of‘ the receptor in the suhsynaptic membrane (ie., the fact that the receptor remains subsynaptic in spite of lateral diffusion which usually occurs in the plane of the membrane) was also required probably involving a rather inu usual local rigidity in the membrane. T h e possibility that residual detergent may act as a local anesthetic also exists. T h e local anesthetic action of Triton X-100 dt micromolar concentrations (>1000 times lower than the concentration used in receptor purification) has been demonstrated (Bi-isson et al., 197517). This exhausting but not exhaustive list shows that it is not so surprising that the relatively crude methods of reconstitution employed so far have not been routinely successful. I t would thus appear that a more analytical approach to reconstitution is required. I t is not sufficient simply to look for the final product, the excitable membrane, but to study and compare as many properties as possible throughout the process of solubilization, purification, and reconstitution. In this way one may hope to be able to gradually fulfill the requirements for functional reconstitution and in the process probably learn a great deal about the functioning of the receptor! ACKNOWLEDGMENTS Michael S. Briley is the recipient of a long-term EMBO (European Molecular Biology Organization) fellowship. We are indebted to Drs. A . Sobel and J. Cartaud for the preparation and electron microscopy, respectively, of the pitrified membranes shown in Fig. 1. T h e original research was supported by grants from the National Institutes of Health, United States Public Health Service, the Centre National d e la Recherche Scientifique, the Delegation Generale a la Recherche Scientifique et Technique, the Fondation pour la Recherche Midicale FranCake, the College d e France, and the Commissariat a I’Energie Atorniqne. REFERENCES

Albertsson, P. A. (1960). “Partition of Cell Particles a n d Macromolecules.” Almqvist & Wiksell, Stockholm. Anderson, C. R., and Stevens, C. F. (1973).J . Physiol. (London) 235,665-691. Barnard, E. A., Wieckowski, J., and Chiu, T. H. (197 1). Nature ( L o d o n ) 234, 207-209. Barrantes, F. J. (1975).Biochem. Biophy.7. Res. Cornmiin. 62, 4 0 7 4 1 4 . Barrantes. F. J., Changeux, J. P., Lunt, G. G., and Sobel, A. (197.5).Nature (London) 256, 325-327.

60

MICHAEL S. BRILEY A N D JEAN-PIERRE

CHANGEUX

Barrantes, F. J., Changeux, J. P., Lunt, G. G., and Sobel,A. (1976).Nature (London) 259, 605-606. Bartels, E., and Nachmansohn, D. (1965). Biochem. Z. 342, 359-374. Bauman, A., Benda, P., and Changeux, J. P. (1969). FEES Lett. 8, 145-148. Bennett, M. V. C. (1970). Annu. Rev. Physiol. 32, 471-528. Biesecker, G. (1973). Biochemistry 12,4403-4409. Bigelow, C. C. (1967).J. Theor. Biol. 16, 187-211. Carl, G. F., and Kemp, G. E. (1976). Biochem. Bradley, R. J., Howell, J. H., Romine, W. 0.. Biophys. Res. Commun. 68, 577-584. Briley, M. S., and Changeux, J . P. (1976). Meet. Eur. Sor. N~urochem.,1 s t . Bafh No. 23p. (Abstr .) Briley, M. S., and Changeux, J. I? (1977). Submitted for publication. Brisson, A. D., Scandella, C. J., Bienvenue, A., Devaux, P. F., Cohen, J. B., and Changeux, J. P. (1975a). Proc. Natl. Acad. Sci. U.S.A. 72, 1087-1091. Brisson, A. D., Devaux, P. F., and Changeux, J, P. (397513). C.R. Acad. Sn’., Ser. D 280, 2 153-2 156. Brockes, J. P., and Hall, Z. W. (1975). Biochemistry 14, 2092-2106. Capaldi, R. A., and Vanderkooi, G. (1972). Proc. Natl. Acad. Sci. U.S.A. 69, 930-932. Cartaud, J., Benedetti, E. L., Cohen, J. B., Meunier, J . C., and Changeux, J. P. (1973). FEES Lett. 33, 109-1 13. Chang, H. W. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 21 13-21 17. Changeux, J. P. (1975).In “Handbook of Psychopharmacology” (L. L. Iversen, S. D. Iversen, and S. H. Snyder, eds.), Vol. 6, pp. 211-230. Plenum, New York. Changeux, J. P., and Podleski, T. R. (1968). Proc. Natl. Acad. Sci. U.S.A. 59, 944-950. Changeux, J. P., Podleski, T. R., and Wofsy, L. (1967). Proc. Natl. Acad. Sci. U.S.A. 58, 2063-2070. Changeux, J. P., Kasai, M., and Lee, C. V. (1970). Proc. Natl. Acad. Sci. U.S.A. 67, 12411247. Changeux, J. P., Meunier, J. C., and Huchet, M. (1971). Mol. Pharmacol. 7, 538-553. Changeux, J. P., Huchet, M., and Cartaud, J. (1972). C.R. Acad. Sci. 274, 122-125. Changeux, J. P., Benedetti, L., Bourgeois, J. P., Brisson, A., Cartaud, J., Devaux, P., Criinhagen, H. H., Moreau, M., Popot, J. L., Sobel,A , , and Weber, M. (1976). Cold Spring Harbor Symp. & a n t . Biol. 40, 2 11-230. Changeux, J. P., Bon, C., Briley, M . S., Grunhagen, H. H., Iwatsubo, M., Sobel, A., and Teichberg, V. I. (1977). Proc. Int. Sri. Conf. Muscular Dystrophy Asor., 5lh, Durangv, Calo. in press. Cohen, J. B., Weber, M., Huchet, M., and Changeux, J. P. (1972). FEES Lett. 26,43-47. Cooper, D., and Reich, E. (1972).J. Bzof. Chem. 247, 3008-3013. Cuatrecasas, P. (1970).J. Biol. Chem. 245, 3059-3065. Del Castillo, J., Rodriguez, A., Romero, A., and Sanchez, V. (1966).Science 153,185-188. De Plazas, S. F., and De Robertis, E. (19723. Biochim. Biophys. Acta 474, 258-265. De Robertis, E. (1971). Science 171, 963-971. De Robertis, E., Fiszer de Plazas, S., and De Carlin, M. C. L. (1976). Nature (London) 259, 605. Dolly, J. O., and Barnard, E. A. (1975). FEES Lett. 57, 267-271. Donellan, J. F., and Cattell, K. J. (1975). Biochem. SOC. Trans. 3, 106-109. Duguid, J. R., and Raftery, M. A. (I973a). Arch. Biochem. Biophys. 159,512-516. Duguid, J. R., and Raftery, M. A. (1973b). Biochemistry 12, 3593-3597. Dupont, Y., Cohen, J., and Changeux, J. P. (1974). FEES Lett. 40, 130-133. Edelstein, S . J., Beyer, W. B., Eldefrawi, A. T., and Eldefrawi, M. E. (1975).J. Bzol. Chem. 250,6l01L6106.

ISOLATION A N D PURIFICATION OF NICOTINIC RECEPTOR

61

Eldefrawi, M. E., and Eldefrawi, A. T. (1973). Arch. Biochem. Biophys. 159, 362-373. Eldefrawi, M. E., Eldefrawi, A. T., and Wilson, D. B. (1975a). Eiochemisby 14, 430443 10. Eldefrawi, M. E., Eldefrawi, A. T., and Sharnoo, A. E. (1975b). Ann. N.Y. Acad. Sci. 265, 183-202. Eldefrawi, M. E., Eldefrawi, A. T., Penfield, L. A., OBrien, R. D., and Van Campen, D. (1975~). Life Sri. 16, 925-936. Flanagan, S. D., Barondes, S. H., and Taylor, P. (1976).J. Eiol. Chem. 251, 858-865. Fulpius, B., Cha, S., Klett, R., and Reich, E. (1972). FEBS Lett. 24, 323-326. Gibson, R. E., O’Brien, R. D., Edelstein, S. J., and Thompson, W. R. (1976).Biochemirtry15, 2377-2383. Goodall, M. C., Bradley, R. J., Saccomani, G., and Rornine, W. 0. (1974).Nature (London) 250968-70. Gordon, A,, Bandini, G., and Hucho, F. (1974). FEES Lett. 47,204-208. Hazelbauer, G. L., andChangeux,J. P. (1974).Proc. Natl. Acad. Sci. U.S.A. 71,1479-1483. Heilbronn, E. (1975). Croat. Chem. Acta 47, 395-408. Heilbronn, E., and Mattson, C. (1974).J. Neurochem. 22, 315-317. Hess, G. P., Andrews, J. P., Struve, C . A., and Coornbs, S. E. (1975).Proc. Natl. Acad. Sci. V.S.A. 72, 437 1-4375. Hess, G. P., Andrews, J. P., and Struve, G. A. (1976).E w c h m . Biophys. Res. Commnn. 69, 830-837. Higman, H., Podleski, T. k.,and Bartels, E. (1963).Eiochim. Ezophys. Acta 75, 187-193. Hucho, F., and Changeux, J. P. (1973).FEES Lett. 38, 11-15. Hucho, F., Layer, P., Kiefer, H. R., and Bandini, G. (1976).Proc. Nad. Acad. S r i . U.S.A. 73, 2624-2628. Jain, M. K. (1974). Arch. Bwchem. Ewphys. 164, 20-29. Jain, M. K., Mehl, L. J., and Cordes, E. H. (1973). Bwchem. Biophys. Res. Commun. 51, 192-1 93. Karnetari, T., Ikeda, Y., and Kasai, M. (1975). Eiochim. Eiophys. Acta 413, 415-424. Karlin, A. (1973).Proc. Int. Congr. Pharmacol., 5th 5,86-97. Karlin, A., and Cowbnrn, D. (1973).Proc. Natl. Acad. Sci. U.S.A. 70, 3636-3640. Karlin, A,, and Winnik, M. (1968). Proc. Natl. Acad. Sci. U.S.A. 60,668-674. Karlin, A,, Cowburn, D. A., and Reiter, M. J. (1973). In “Drug Receptors” (H. Rang, ed.), pp. 193-210. Macrnillan, London. Karlin, A., Weill, C., McNamee, M., and Valderrarna, R. (1976).Cold Spring Harbor Symp, Quant. Bwl. 40,203-210. Karlsson, E.. Heilbronn, E., and Widlund, L. (1972). FEBS Lett. 28, 107-1 11. Kasai, M., and Changeux, J. P. (1971).J. Membr. Biol. 6 , 1-23, 24-57, 58-80. Katz, B., and Miledi, R. (1970). Nature (London) 226, 124-126. Katz, B., and Miledi, R. (1972a). Nature (London),New Eiol. 232, 124-126. Katz, B., and Miledi, R. (1972b).J. Physiol. (London) 224,665-700. Katz, B., and Thesleff, S. (1957).J. Physiol. (Lundon) 138, 63-80. Kemp, G., Dolly, J. O., Barnard, E. A,, and Wenner, C. E. (1973). Bwchem. Biophys. Res. Commun. 54, 607-613. Klett, R. P., Fulpius, B. W., Cooper, D., Smith, M., Reich, E., and Possani, L. D. (1973).J. Biol. Chem. 248,68416853. Lee, C. Y. (1972). Annu. Rev. Pharmacol. 12, 265-28 1. Lee, C. Y., and Tseng, L. F. (1966). Toxicon 3, 281-290. Lester, H., Changeux, J. P., and Sheridan, R. (1975).J. Gen. Physiol. 65, 797-816. Leuzinger, W., and Schneider, M. (1972).ExpPrientia 28, 256-257. Levinson, S. R., and Keynes, R. D. (1972). Biochim. Eiophys. Actu 288, 241-247.

62

MICHAEL S. BRILEY A N D JEAN-PIERRE CHANGEUX

Linstrom, J.. and Patrick, J. (1974). In “Synaptic Transmission and Neuronal Interaction” (M. V. L. Bennett, ed.), pp. 191-216. Raven, New York. MacLennan, D. H. (1975). Can.]. Biochm. 53,257-261. MacLennan, D. H., and Holland, P. C. (1975). Annu. Rev. Biophys. Bioeng. 4,377. McNamee, M. G., Weill, C. L., and Karlin, A. (I975a).Ann. N.Y. Acud. Sci. 265, 175-182. McNamee, M. G., Weill, C. L., and Karlin, A. (1975b).In “Protein-Ligand Interactions” (H. Sund and G. Blane, eds.), pp. 316-327. d e Gruyter, Berlin. Magazanik, G. G. (1976). Annu. Ren. Pharmucol. 16, 161-175. Magazanik, G . G., and Vyskocil, F. (1973). In “Drug Receptors” (H. P. Rang, ed.), pp. 105-1 19. Macmillan, London. Mautner, H. G.. Bartels, E., and Webb, G. D. (1966). Bzochem. Phurmurol. 15, 187-193. Menez, A., Morgat, J. L., Fromageot, P., Ronseray, A. M., Boquet, P., and Changeux, J. P. (I97 1). FEES Lett. 17, 333-335. Meunier, J. C., and Changeux, J. P. (1973). FEES Lett. 32, 143-148. Meunier, J. C., Huchet, M., Boquet, P., and Changeux, J. P. (1971). C.R. Acud. Sci., Ser. D 272, 117-120. Meunier, J. C., Olsen, R. W., and Changeux, J . P. (1972a). FEES Lett. 24,63-68. Meunier, J. C., Olsen, R. W., Menez, A., Fromageot, P., Boquet, P., and Changeux, J . P. (1972b). Biochemistry 11, 1200-1210. Meunier, J. C., Sealock, R., Olsen, R., and Changeux, J. P. (1974). Eur. J , Biachem. 45, 37 1-394. Michaelson, D. M., and Raftery, M. A. (1974).Proc. Nutl. Acad. Sci. U.S.A. 71,4768-4772. Michaelson, D. M., Vandlen, R., Brode, J., Moody. T., Schmidt, J., Raftery, M. A. (1974). Arch. Biochem. Biophys. 165, 796-804. Miledi, R., Molinoff, P., and Potter, L. T. (1971).Nature (London) 299, 554-557. Moody, T., Schmidt, J., and Raftery, M. A. (1973). Biochem. Biophys. Res. Commun. 53, 761-772. Moore, W. M., Holladay, L. A., Puett, D., and Brady, R. N. (1974). FEESLett. 45, 145-149. Mueller, P., Rudin, D. O., Tien, H. T., and Westcott, W. C. (1962). Nature (London) 194, 979-980. Nachmansohn, D. (1974). In “Biochemistry of Sensory Functions” (L. Jaenicke ed.), pp. 431 4 6 4 . Springer-Verlag, Berlin and New York. Nachmansohn, D., and Neumann, E. (1975). “Chemical and Molecular Basis of Nerve Activity.” Academic Press, New York. Nickel, E., and Potter, L. T. (1973).Brazn Res. 57, 508-517. Olsen, R.. Meunier, J. C., and Changeux, J. P. (1972).FEBS Lett. 28,96-100. Ong, D. E., and Brady, R. N. (1974). Biochemistry 13,2822-2827. Parisi, M., Rivas, E., and De Robertis, E. (1971). Science 172, 56-57. Parisi, M., Reader, T. M., and De Robertis, E. (1972).]. Gen. Physiol. 60, 454-470. Parisi, M., Adragna, N. C., and Salas, P. J. 1. (1975). Nature (London) 258, 245-247. Patrick, J., and Lindstrom, J. (1973). Science 180, 871-872. Patrick, J., Boulter, J., and O’Brien, J. C. (1975). Bwchem. Biophys. Res. Commun. 64, 219225. Popot, J. L., Sugiyama, H., and Changeux, J. I? (1974i.C.R. Acad. Sci., So.D 279, 17211724. Popot, J. L., Sugiyama, H., and Changeux, J. P. (1976).J. Mol. B i d . 106, 469-484. Potter, L. (1973). In “Drug Receptors” (H. Rang, ed.), pp. 295-312. Macmillan, London. Prives, J. M., Reiter, M. J., Cowburn, D. A., and Karlin, A. (1972). Mol. Phurmurol. 8, 786-789. Raftery, M. A., Schmidt, J., Clark, D. G., and Wolcott, R. G. (1971). Biochm. Biophys. Res. Commun. 45, 1622-1629.

I SO LA TI ON A N D P URIF ICAT ION O F N I C O T I N I C R EC EPTO R

63

Raftery, M. A., Schmidt, J., and Clark, D. G. (1972).Arch. Biochem. Biopliy.~.152,882-886. Raftery, M. A., Bode, J.. Vandlen, R., Michaelson, D., Deutsch, J., Moody, T., ROSS,M. J., and Stroud, R. M. (1975).1n “Protein-Ligand Interactions” ( H . Sund & G. Blane, eds.), pp. 328-355. de Grityter, Berlin. Raftery, M . A . , Vandler, R. L., Reed, K. L., and Lee, T. (1976). Cold Spring Harbor Sjmp. Quant. Biol. 40, 193-202. Rang, H. P. (1974). Qzrnrt. Rex!. Uio,bhp. 7, 283-400. Reiter, M. J., Cowburn, D. A., Prives, J. M . , and Karlin, A. (1972).Proc. Nntl. ,4eod. Sci. L‘.S.A. 69, 1168-1 172. Romine, W. O., Goodall, M. C., Peterson, J.. and Bradley, R. J . (1974).Hiochirn.Biop/ty,v.Actn 367, 316-32.5. Rubsamen. H.. Hess. G. P., Eldefrawi, A. T., and Eldefrawi, M. E. (1976a).Rioc//~m. Uioply~. Res. C m m i c n . 68, 56-63. Riibsamen, H., Montgomery, M., Hess, G. P., Eldefrawi, A. T., and Eldefrawi, M. E. (1976b).Bjochtm. Uiophys. Res. Commtcn. 70, 1020-1027, Ruoho, A . E., Kiefer, H.. Roeder, P. H.. and Singer, S. J . (1973). Prcx. Nntl. A r d . S r t . L‘.S..4. 70,2567-257 1. Schmidt. J.. and Rattery, X I . A . (1072).Hiorlrrm. H w / J / ~Rut. ~ \ . Conintitn. 49, 572-57X. Schmidt, J.. and Raftery, M. .4. (1973). Biochemistty 12, 852-856. Shamoo, A. E., and Eldefrawi, M. E. (1975).J. Memhr. Biol. 25, 47-63. Singer, R. J., and Nicolson, G. L. (3972).Science 175, 720-73 I . Singer, S. J., Ruoho, A., Kiefer, A . , Lindstrom. J.. and Lennox, E. S. (1973). 1tr “Drug Receptors” (H. Rang, ed.), pp. 183-192. Macmillan, London. . Trum. 5, .51 1-514. Sobel. A , , and Changeitx. J . 1’. ( I O i ’ i ) . H i o c h / ~ mSor. Spatr, L., and Strittmatter, P. (1973).J. Uiool. CIiettz. 248, 793-799. Sugiyama, H., and Changeus, J. P, (1975).Etct-.J. Uioc/tm. 55, 50.5-515. Sugiyama. H., Benda, P., Mennier, J. C., and Changeus, J. P. (1973). FEHS Lett. 35, 124-128. Sugiyama, H., Popot, J. L., and Changeux, J. P. (1976).J. Mol. Bid. 106, 485-496. Vasquez, C., Parisi, M., and De Robertis, E. (1971).J. Membr. Biol. 6, 353-367. Warren, G. B.. Hoiislay, M. D., Metcalf, J . C., and Birdsall. N. J . M. (1975).Nntrrre (London) 255,684-687. Weber, M., and Changeux, J. P. (1974).Mol. Phnrmncol. 10, 1-14, 1.5-34. 35-40, Weber. M., David-Pfeuty, T., and Changeus, J. 1.’ (1975).Proc. N n k Acnd. Sci. Ii.S.,4. 72, 3443-3447. Weill, C. I.., McNamee, M. G., and Karlin, A. (1974). Uioclvm. Biojdry.~Re.5. Commirn. 61, 997-1003.

This Page Intentionally Left Blank

B IOCH EM1CAL ASPECTS 0F NE UROTRANSMISSlON IN THE DEVELOPING BRAIN By Joseph T. Coyle

Departments of Pharmacology and Experimental Therapeutics and Psychiatry and Behavioral Sciences The Johns Hopkins University School of Medicine Baltimore, Maryland

1. Introduction 11. Prenatal Development of Central Catechdaminergic Neurons

A. Background B. Appearance of Catecholaminergic Neurons in Fetal Brain C. THIThymidine Autoradiography D. Octopamine and 8-Phenylethanolamine in Fetal Brain E. Regional Innervation by Catecholaminergic Processes F. Effects of Drugs on Fetal Brain Catecholamines G. Synaptogenesis on Catecholaminergic Neurons H. Catecholaminergic lnriervation in the Immature Neocortex I . Summary and Speculation 111. Postnatal Development of the Nigmstriatal Circuit A. Background B. Dopaminergic lnnervation to Striatum C. Striatal Cholinergic Neurons D. Striatal GABAergic Neurons E. Ontogenesis of the Function of the Nigrostriatal Circuit 1V. Conclusion References

1. Introduction

T w o related issues have long provoked the interest of both basic and clinical investigators whose focus of research is the development of the central nervous system. First, when are specific neuronal pathways formed in the developing brain; and second, when does effective neurotransmission among the components of such neuronal pathways commence during brain maturation? Biochemical analysis of the processes that mediate chemical synaptic transmission can provide information relevant to both these questions. By measuring the biochemical parame65

66

JOSEPH T. COYLE

ters that are specific for a neuronal type, it should be possible to “map” in a quantitative fashion the development of the neuronal pathway in the brain. Since neuronal communication occurs by means of chemical transmitters (Krijevic, 1974), the relative development of the processes that mediate neurotransmission should reflect the functional influence of a neuronal pathway at a particular stage of brain development. Four parameters play primary and specific roles in mediating neurotransmission for many identified classes of neurons as demonstrated by the fact that interference with any one of these parameters can profoundly affect the efficacy of neurotransmission (Cooper et al., 1974) (Fig. 1 ) . The neuron possesses the enzyme or enzymes necessary to synthesize its neurotransmitter; in most cases, these enzymes are exclusively localized in or, at least, highly concentrated in the particular neuronal population. Effective neurotransmission requires the presence of adequate stores of neurotransmitters available for release at the synapse; and the vesicles that populate terminal boutons concentrate the neurotransmitter for release as “quanta.” On the neuronal membrane, there are high-affinity transport sites specific for the neurotransmitter utilized by the neuron (Kuhar, 1973). The uptake process terminates the action of released neurotransmitter and may keep the synaptic cleft free of neurotransmitter during periods of presynaptic neuronal quiescence. For interneuronal communication to occur, there must be appropriate receptors for the neurotransmitter on the postsynaptic neuron to decode

REUPTAKE

FIG. 1 . Schematic model of a typical synapse indicating major steps involved in regulating chemical synaptic neurotransmission. These include the specific enzymes that synthesize the neurotransmitter (enzymes), the stores o f neurotransmitter in terminal vesicles (transmitter), the interaction of released neurotransmitter with appropriate postsynaptic receptors (receptors), and the inactivation of the neurotransmitter through removal from the synaptic cleft by high-affinity uptake specific for the neurotransmitter (reuptake).

NEUROTRANSMISSION

I N T H E DEVELOPING BRAIN

67

the chemical message. On one hand, the presynaptic parameters can be used as specific biochemical markers for quantifying neuronal differentiation; and, on the other hand, the relative development of all four parameters in a brain region provides insight into the potential influence of a neuronal pathway. During the last 25 years numerous investigators have examined the ontogenetic development of one or more of these parameters in the brains of a variety of species, and many comprehensive reviews of these studies have appeared (Baker and Quay, 1969; Filogamo and Marchisio, 1971; Coyle, 1973; Haber and Kuriyama, 1973; Lanier et al., 1976). Accordingly, the purpose of the present chapter is not to review again the diverse aspects of neurotransmitters in the developing brain but rather to critically evaluate the possible correlation between the development of the processes that regulate neurotransmission and the morphologic and functional aspects of neuronal differentiation in rat brain. In the first part of the chapter, studies concerning the early development of central catecholaminergic neurons will be examined because the correlation between morphologic and neurochemical differentiation can be assessed. In the second part, the nigrostriatal circuit shall be the focus of discussion because of the opportunity to compare and contrast neurochemical differentiation of three types of neurons involved in defined synaptic relationships. II. Prenatal Development of Central Catecholaminergic Neurons

A. BACKGROUND The central catecholaminergic neurons provide a unique opportunity for studying the neurochemical aspects of neuronal development because of the possibility of correlating the biochemical results with the neuroanatomic characteristics of the neurons. The cell bodies for the catcholaminergic neurons are limited to the brainstem regions whereas they provide innervation to cortical and subcortical regions quite distant from their perikarya (Ungerstedt, 1971); hence, for neurochemical studies brain regions containing exclusively axons and terminals of the catecholaminergic neurons can be anatomically separated from those possessing the cell bodies of these neurons. In addition, the catecholaminergic neurons can be visualized for histologic studies by histofluorescent (Dahlstrom and Fuxe, 1964) and, more recently, immunochemical techniques (Hartman et nl., 1972; Pickel et d.,1976); thus, one can correlate the neurochemical and morphologic aspects of their differentiation,

68

JOSEPH T. COYLE

B. APPEARANCE OF CATECHOJAMINERGIC NEURONSIN FETALBRAIN

Information about the stage of brain development at which a neuronal pathway appears is important for assessing possible influences of this pathway on brain function during subsequent differentiation. For the rat, an animal that is born at a relatively early stage of brain maturation, it has been generally accepted that the catecholaminergic neurons develop primarily after birth (Agrawal et al., 1966). However, it has become apparent with application of sensitive radiometric-enzymatic assay techniques that the presynaptic markers for the catecholaminergic neurons are present in the fetal rat brain well before birth. The enzymes in the synthesis pathway for catecholamines, tyrosine hydroxylase, Dopa decarboxylase, and dopamine-P-hydroxylase, as well as the neurotransmitters themselves, dopamine and norepinephrine, are detectable in the fetal rat brain as early as 15 days of gestation when the brain weighs 1.5% of that of the adult (Coyle and Axelrod, 1972a,b; Coyle and Henry, 1973;Lamprecht and Coyle, 1972).The specific activities of the enzymes are approximately 5-10% of that of the adult brain, whereas the concentrations of the catecholamines are only 3% of adult levels. During the last week of fetal development, there is a relatively linear and coordinated increase in the levels of the presynaptic markers for the catecholaminergic neurons to achieve approximately 30% of adult levels by birth. Based upon the linear rate of their rise during the late fetal stages, it has been predicted that these catecholaminergic neuronal markers appear in the brain between 13 and 14 days of gestation (Coyle, 1974). The developmental stage of appearance of the major catecholaminergic nuclei including the noradrenergic cell bodies in the locus coeruleus and the dopaminergic cell bodies in the substantia nigra has been examined with the histofluorescent microscopic technique. With this technique, thin sections of tissues are exposed to paraformaldehyde vapor under rigorously defined conditions; the condensation product formed by paraformaldehyde and catecholamine within neurons can be visualized with the fluorescent microscope (Falck et al., 1962). Nascent noradrenergic perikarya in the medulla pons and dopaminergic perikarya in the midbrain region become apparent at approximately 13 to 14 days of gestation (Maeda and Dresse, 1969; Olson and Seiger, 1972).At this stage, the catecholaminergic cells are still in the process of migration to their ultimate site of localization. Thus, there is an excellent correlation between the neurochemical and histofluorescent microscopic techniques with regard to the time of appearance of the major catecholaminergic nuclei in the rat brain.

NEUROTRANSMISSION I N T H E DEVELOPING BRAIN

c.

69

[’H]‘rHYMlDlNE AUTORADIOGRAPHY

Histofluorescent microscopy and the neurochemical assays are, in fact, measuring a similar parameter, i.e., the presence of catecholamines within the neurons. It is possible that the catecholaminergic neurons are formed well before they develop the capacity to synthesize and store catecholamines, and thus, the “birth” of these neurons may antedate their neurochemical differentiation. Therefore, it is essential to know at what stage in development the catecholaminergic neurons cease dividing or are, in effect, “born.” Such information can be obtained by [3H]thymidine authoradiography (Sidman, 1970). With this method, fetuses of various stages are administered a single dose of rH]thymidine; the animals are then allowed to mature to adulthood for histologic examination. Neuroblasts that are undergoing mitosis during the brief exposure to the [3H]thymidine incorporate it into their deoxyribonucleic acid (DNA), which remains “labeled” unless diluted by subsequent cell division or removed by cell death. Neurons that have ceased dividing no longer incorporate the [3H]thymidine into their DNA. The c3H]thymidine incorporated into the DNA of the neuronal nuclei can be demonstrated in adulthood by autoradiographic techniques: the identity of the radiolabeled neurons is determined by light and histofluorescent microscopy of adjacent sections. With this technique, the neurons in the locus coeruleus exhibit heavy incorporation of [3H]thymidine into their nuclear DNA when injected on days 11 and 13 of gestation but no incorporation when injected on day 14 of gestation or thereafter (Lauder and Bloom, 1974). The nuclei of the dopaminergic cell bodies in the substantia nigra exhibit heavy labeling after exposure to rHIthymidine on days 13-15 of gestation, but no incorporation occurs on subsequent days. Thus, cell division of the noradrenergic neurons in the locus coeruleus ceases by 14 days of gestation and for the dopaminergic neurons in the substantia nigra ceases by 16 days of gestation. I n brief, the full complement of catecholaminergic neurons that form these two major nuclei is attained fully a week before birth in the rat. The developmental stage when cell division of the catecholaminergic neurons ceases coincides with the time when they can first be demonstrated by the histofluorescent technique and the time when their neurotransmitters and biosynthetic enzymes appear. The striking agreement among these three different methodologic approaches provides compelling evidence that the catecholaminergic neurons acquire those specialized processes essential for neurotransmission during

70

JOSEPH

T. COYLE

or immediately after the terminal (phase) mitoses of their precursor neuroblasts. Since the total complement of catecholaminergic neurons in the brain is achieved by 15 days of gestation in the rat, increases in the specific neurochemical markers for these neurons after this date represent neuronal differentiation and not neuronal multiplication. Since the central nervous system is a relatively closed system with respect to the catecholaminergic neurons (i.e., their cell bodies and axonal processes are limited to the cerebrum except for a small portion that innervate the spinal cord), an approximate quantitative assessment of the total increase in these parameters on a per neuron basis can be made by multiplying specific activities of the enzymes or concentrations of the neurotransmitters by brain weight (Fig. 2). When the developmental changes in the neurochemical parameters for the catecholaminergic neurons are expressed in this fashion, there is a 500- to 1000-fold increase between 15 days of gestation and adulthood (Coyle, 1974). This value, of course, I

i 6

Y-

O

Conceptual Age ( Days)

FIG. 2. Development of the biosynthetic enzymes for catecholamines and endogenous dopamine and norepinephrine in fetal rat brain. The activities of tyrosine hydroxylase ( X - . - x ) and dopamine-p-hydroxylase ( x ...X ) and levels of dopamine (0---0) and are expressed in terms of percent of whole brain activity or norepinephrine (0-0) content. The rectangles indicate the period of mitosis, and the arrows indicate the date of appearance by histofluorescent microscopy of the noradrenergic neurons in the locus coeruleus (L.C.) and of the dopaminergic neurons in the substantia nigra (S.N.).

NEUROTRANSMISSION

71

IN THE DEVELOPING BRAIN

provides only a hint of the enormous changes that occur during the process of differentiation from a discrete cell at one week before birth to a neuron with axonal extensions that make several thousand synaptic contacts. D.

OCTOPAMINE AND P-PHENYLETHANOLAMINE IN

FETALBRAIN

P-Phenylethanolamine and octopamine result from the direct decarboxylation of phenylalanine or tyrosine, respectively, with subsequent P-hydroxylation of the amine by dopamine-/3-hydroxylase (MolinofFand Axelrod, 1972). In mammals, it has generally been thought that these two biogenic amines occur as biochemical “mistakes” due to a side reaction in the normal synthetic pathway for norepinephrine in adrenergic neurons (Kopin et al., 1964). Accordingly, under usual conditions, these two amines are found in relatively low concentrations in tissues with noradrenergic innervation; however, the fact that neither entirely disappears after sympathetic denervation suggests that they may have an additional localization outside of noradrenergic neurons (Saavedra and Axelrod, 1973; Coyle et al., 1974). Notably, in certain invertebrates, octopamine and P-phenylethanolamine appear to be neurotransmitters in their own right for which there are specific receptors (Nathanson and Greengard, 1975; Saavedra et al., 1976). In fetal rat brain at 15 or 16 days of gestation, the concentrations of P-phenylethanolamine and octopamine are 5- to 6-fold higher than those that occur in adult brain (Saavedra et al., 1974) (Table I). More importantly, the concentrations of P-phenylethanolamine and ocTABLE I LEVELS OF’

P-PHENYLETHANOLAMINE, OCTOI’AMINE, A N D N O R E P I N E P H R I N E IN

Conceptual age (days) 15

16 17 18 20 22 (Birth) Adult

FETAL RAT BRAIN“

P-I’henylethanolamine Octopamine ( p m o l e h g tissue) 0. I47 0.154 0.19 1 0.044 0.029

0.037 0.044

Norepinephrine

0.085 0.176 0.065 0.020 0.035 0.039 0.033

” Values are obtained from Coyle and Henry ( 1 973) and Saavedra PI are the mean o f five or more separate determinations.

0.062 0.097 0 . I66 0.297 0.645 0.923 2..556

NI.

( 1 974) and

72

JOSEPH T. COYLE

topamine exceed that of endogenous norepinephrine by 2.5-fold at this stage of brain development. Treatments of pregnant mothers at 16 days of gestation with a monoamine oxidase inhibitor, parachlorophenylalanine, or the combination of these drugs with phenylalanine considerably increase the ratio of the amines to norepinephrine in the fetal brain. At 18 days of gestation, there is dramatic decrease in the concentration of these two amines to adult levels, and their concentrations remain relatively constant thereafter. The reasons for this transient high concentration of P-phenylethanolamine and octopamine remain unclear. Although in the adult rat inhibition of monoamine oxidase results in a considerable increase in the levels of these two amines, their levels do not surpass that of endogenous norepinephrine (Molinoffand Axelrod, 1972).Thus, the low activity of monoamine oxidase (10% of adult specific activity) in fetal rat brain does not adequately explain the high levels of these amines (Saavedra et al., 1974; Gripois, 1975). In peripheral sympathetic neurons, the levels of tyrosine hydroxylase are regulated by the cholinergic input to the ganglion and corticosteroids (Mueller et al., 1969; Otten and Thoenen, 1976). It is noteworthy that the pituitaryadrenal axis becomes functional and synapses appear on the locus coeruleus (see below) at 18 days of gestation when the levels of P-phenylethanolamine and octopamine decrease precipitously (Roffi, 1968; Lauder and Bloom, 1975). Thus, it is possible that their elevated levels reflect a disparity between the activity of tyrosine hydroxylase and dopamine-P-hydroxylase within the immature central noradrenergic neurons. The possibility that octopaminergic neurons appear transiently during the early stages of brain development cannot be ruled out. Regardless of the explanation, phenomenologically this represents an example of neurotransmitter ontogeny recapitulating phylogeny.

E. REGIONAL INNERVATION BY CATECHOLAMINERGIC PROCESSES During the last week of gestation, these is a marked increase in the specific activities of the biosynthetic enzymes for catecholamines in fetal rat brain with tyrosine hydroxylase, dopa decarboxylase, and dopamine-P-hydroxylase increasing 4- to 6-fold. The concentrations of endogenous dopamine and norepinephrine increase 15-fold during this time frame, dramatizing the marked differentiation that is occurring in this neuronal population (Coyle and Henry, 1973). The specific activities of the biosynthetic enzymes in the sheared off nerve terminals or synaptosomal fractions increase 10-fold between 15 days of gestation and birth (Coyle and Axelrod, 1972a,b). In accordance with this evidence that

NEUROTRANSMISSION IN THE DEVELOPING BRAIN

73

terminals are being formed at this early stage of development, the highaffinity uptake process for norepinephrine in synaptosomes can be demonstrated by 18 days of gestation (Coyle and Axelrod, 1971). Finally, during the late stages of fetal development, there is a translocation in the distribution of the biosynthetic enzymes with a progressive increase in their activity in the forebrain regions which receive innervation but do not possess intrinsic catecholaminergic cell bodies. T h e extensive histofluorescent microscopic study of Olsen and Seiger on the prenatal development of the central monoaminergic neurons in the rat provide fine-structural correlates for the neurochemical observations (Olsen and Seiger, 1972; Seiger and Olsen, 1973). In the rat embryo with a crown-rump length of 13 mm, which corresponds to 15 days 13rnrn embryo

0

CA cell groups

(3 S-HT re11 groups (.% ’

axon bundles

R-HTaxon bundles (‘.%+h-HT axon bundles

FIG.3. Schematic representation of the distribution of monoaminergic cell groups and axonal projections in the 15day gestational rat brain (13 mm embryo). CA, catecholamine: S-HT, 5-hydroxytq ptarnine. Midsagittal section through fetal b n i n demonstrating cell groups and axon bundles: in, mesencephalic flexure: 1). pontine flexure: c, cervical flexure. (Reproduced from Olson and Seiger, 1972, with kind permission of the authors and publisher, Springer-Verlag.)

74

JOSEPH T. COYLE

of gestation, a prominent fluorescent axon bundle arising from the rostral noradrenergic cell complex transverses the lateral part of the diencephalon and reaches the preoptic area (Fig. 3). In the prosencephalon, the catecholaminergic processes have a distinct beaded appearance, a structural characteristic that in adult brain is associated with synaptic specialization. Although it has been generally assumed that synaptogenesis occurs later in rat brain development (Aghajanian and Bloom, 1967), especially in the telencephalon, Konig et al. (1975) have demonstrated by electron microscopy unequivocal synaptic contacts in the developing rat neocortex as early as 16 days of gestation. At approximately 18 days of gestation, widespread projections of the catecholaminergic neurons are evident throughout the fetal brain. Processes with terminals are observed in the medulla, tectum, discrete areas of the hypothalamus, and the tuberculum olfactorium. The medial forebrain bundle is well developed with a large number of catecholaminergic axons; the lateral part of the bundle, which contains dopaminergic fibers, projects to the primordial neostriatum. The striatum itself has a weakly fluorescent marginal zone consisting of small densely packed islands of catecholaminergic processes, and the interior aspects of the nucleus exhibit sparsely distributed varicosities. In the more medial parts of the medial forebrain bundle, which contain primarily noradrenergic fibers, there are axonal projections that can be traced to the hippocampus and the cerebral cortex. During the final 3 days of fetal development, the major morphologic change in the catecholaminergic neurons consists of a proliferation of terminal varicosities in all regions of the central nervous system.

F. EFFECTSOF DRUGSON FETAL BRAINCATECHOLAMINES The neurochernical and morphologic studies demonstrate that catecholaminergic axonal projections reach most regions of the fetal brain by a week before birth and commence the formation of terminal varicosities. During this period (15-16 days of gestation), neurons in the caudal region of the brain are in early stages of differentiation, whereas in the rostra1 regions neuronal cell division continues and the major aspects of neuronal organization are in their earliest stages (Schultze et al., 1974: Hine and Das, 1974; Hicks and D’Amato, 1968). The early, widespread projections of the catecholaminergic neurons may influence the neuronal differentiation or function in the areas receiving these fibers. Since such an influence would likely be communicated by the neurotransmitter itself, it is particularly important to assess the presynaptic processes regulating the disposition of catecholamines at this stage of development. Accordingly, the effects of administration of sev-

75

NELJROTRANSMISSION I N T H E DEVELOPING BRAIN

era1 drugs known to interfere with strategic processes involved in the intraneuronal disposition of catecholamines were examined in fetal rats at 18 days of gestation (Coyle and Henry, 1973; Coyle, 1974) (Table 11). Reserpine, a drug that interferes with the vesicular storage process for monoamines (Iversen, 1967), causes a profound depletion of endogenous dopamine and norepinephrine in the fetal brain. Amphetamine, a drug that displaces catecholamines from their storage sites as well as inhibits the presynaptic high-affinity transport process for them (Axelrod, 1970),causes a 70% decrement in the amine levels in the fetal brain. Thus, endogenous catecholamines are primarily sequestered in storage vesicles in the immature neurons. Inhibition of monoamine oxidase, the enzyme that plays a major role in the intraneuronal catabolism of catecholamines (Iversen, 1967), causes a 37% increase in the levels of both dopamine and norepinephrine. The efFects of reserpine, amphetamine, and monoamine oxidase inhibition are qualitatively similar to that which occur in the adult brain. Administration of a large dose of L-dopa, the amino acid precursor to catecholamines, produces a 60% increase in the concentration of norepinephrine and a massive 70-fold increase in the concentration of dopamine in the fetal brain. A particularly interesting aspect of these pharmacologic studies concerns the effects of inhibitors of catecholamine synthesis. Since inhibition of catecholamine synthesis prevents the restitution of catecholamines lost as a result of release, the time-dependent decrement in catecholamine levels after inhibition of their synthesis is proportional to neuronal TABLE I I EFFECTS01.' PHARMACOLOCIC TREATMENTS ON CATECHOLAMINE LEVELS I N ~ X - D A Y GESTATIONAL FETALRAT B R A I N "

Percent of control Treatment Control Reserpine (4 mg/kg) d-Amphetamine (20 nig/kg) a-Methyl-p-tyrosine (400 mg/kg) Diethyldithiocarhamate (500 mg/kgj Phenipra~ine(20 mglkg) L-dopa (300 mg/kg) + MK486 (150 mg/kg)

Duration

Dopamine

Norepinephrine

100 t 3 .5 t 1 31 ? 6 60 t 3 141 t 6 134 5 6

00 -c 4

5 hours 3 hollr$ 3 hours 3 hours 4 hours

4 hours

7460

230

X k 1

27 t 3

55 2 3 60 2 2 33 t 6 160 2 I 1

" Drugs were administered to pregnant rats at 18 days of gestation and fetuses were delivered by caeserian section between 3 to 5 hours after treatment. Results are expressed in terms of the mean percentage of the untreated control 5 S.E.M. Absolute \dues tor controls are 97 pg/mg for dopamine and 5 0 p g h g for norepinephrine (Coyle. 1973).

76

JOSEPH T. COYLE

activity (Anden el al., 1967). Inhibition of tyrosine hydroxylase, the initial enzyme in the synthesis pathway for catecholamines, with a-methyltyrosine results in a 40-45% decrement in the levels of dopamine and norepinephrine after 3 hours in the fetal brain. Inhibition of dopamine-P-hydroxylase with diethyldithiocarbamate causes a 40% fall in the levels of norepinephrine and a 40% increase in the levels of dopamine in the fetal brain. The fact that inhibition of the first and last enzymes in the synthesis pathway for norepinephrine results in the same decrements in the levels of the amine indicates that negligible amounts of the intermediates L-dopa and dopamine are present in the noradrenergic neurons; thus, the synthesis pathway for norepinephrine is tightly coupled as is the case in the mature neurons (Goldstein and Nakajima, 1967). More importantly, the decrease in the concentrations of dopamine and norepinephrine after inhibition of their synthesis indicates that they are in a dynamic state. The half-life for the two neurotransmitters is approximately 3.5 hours, which is 50-7’5% of the half-life of the neurotransmitters in the whole brain of adult rats under basal conditions (Korf el al., 1973; Neff et al., 1971). This half-life suggests that fetal catecholaminergic neurons are sustaining an impulse flow near that of the adult.

G. SYNAPTOGENESIS O N CATECHOLAMINERCIC NEURONS These pharmacologic studies which suggest that catecholaminergic neurons are spontaneously active as early as 18 days of gestation raise the question of when synapses are formed on the catecholaminergic soma and dendrites. This issue has been examined by a combination of histofluorescent and electron microscopy (Lauder and Bloom, 1975). For the dopaminergic neurons in the substantia nigra, both neuropil and somatic synapses are evident by 18 days of gestation, the earliest time point examined; however, the density of synapses is extremely low until well after birth when they markedly increase in number between 15 and 30 days postpartum. With regard to the noradrenergic neurons of the locus coeruleus, faintly stained synaptic profiles are first observed in the neuropil at 19 days of gestation; by 20 days of gestation, synaptic contacts are clearly apparent both in the neuropil and on the somata, I n the locus coeruleus, the most rapid phase of synaptogenesis occurs between 5 and 10 days after birth. The combination of the morphologic and pharmacologic results suggest that the noradrenergic and possibly dopaminergic neurons are spontaneously active prior to the development of presynaptic input. That neuronal activity appears spontane-

NEUROTRANSMISSION IN T H E DEVELOPING BRAIN

77

ously prior to development of presynaptic input has been well established in other neuronal systems (Woodward et al., 1971). The role that presynaptic input to the central catecholaminergic neurons may play in terms of regulating their differentiation remains unclear; however, it should be noted that 95% of their total differentiation occurs after the appearance of synapses on their somata and dendrites. In a series of elegant experiments, the cholinergic input to the superior cervical ganglion has been demonstrated to modulate the differentiation of the sympathetic neurons. In this system, a major increase in the activity of tyrosine hydroxylase occurs after the development of cholinergic synapses in the ganglia; decentralization of the ganglia to prevent its cholinergic innervation significantly reduces the developmental increases in tyrosine hydroxylase (Black et al., 1971, 1972). Treatment with drugs that block the acetylcholine nicotinic receptor at this critical time also reduces the subsequent increase in tyrosine hydroxylase activity, thus demonstrating that acetylcholine liberated by the presynaptic terminals is the chemical cue that promotes the differentiation of the postsynaptic neurons (Black and Green, 1973, 1974). Furthermore, this cholinergic input not only regulates enzyme levels in the perikarya but also the formation of target organ innervation (Black and Mytilineou, 1976). Based upon these observations, it is not unreasonable to speculate that the presynaptic input to the central noradrenergic and dopaminergic neurons may play an important but yet undefined role in modulating their differentiation. H. CATECHOLAMINERGIC INNERVATIONI N

THE

IMMATURE NEOCORTEX

Although these histofluorescent and biochemical studies indicate that axonal projections reach the primordial neocortex by a week before birth, the innervation of the cortex at birth appears to be surprisingly sparse. The specific activities of the biosynthetic enzymes for catecholamines are only 10-20% of adult (Porcher and Heller, 1972; Nomura et al., 1976; Coyle and Axelrod, 1972a,b). In agreement, Loizou ( 1972) reports that catecholaminergic terminal varicosities are extremely rare in the rat neocortex at birth. The disparity between the antenatal and postnatal observations suggest that either there is a long delay between the arrival of catecholaminergic axons to the fetal cortex and the subsequent formation of synapses, or that immature catecholaminergic axons innervating the cortex at birth are deficient in their ability to synthesize endogenous catecholamines. To circumvent the possible pitfall of relying upon endogenous catecholamines as the markers for the

78

JOSEPH T. COYLE

nerve terminals, neonatal rats have been pretreated with catecholamine congeners and precursors prior to electron microscopic and biochemical analysis in an attempt to uncover previously inapparent terminals. Monoaminergic terminals can be identified at the ultrastructural level by the presence of small (40-50 nm) granular vesicles which are storage sites for the amines (Bloom, 1973; Hokfelt and Ljungdahl, 1972). Although the small granular vesicles are rarely seen in the central nervous system in routine ultramicroscopic preparations, their demonstration can be markedly enhanced by exposure of brain tissue to the catecholamine congener 5-hydroxydopamine (Richards and Tranzer, 1970; Ajika and Hokfelt, 1973). This “false” neurotransmitter is selectively taken up and concentrated in the synaptic vesicles of the monoaminergic nerve terminals, whereupon it forms an electron-dense precipitate after aldehyde fixation. Thus, this indirect method permits the ultrastructural visualization of those synaptic terminals which have an uptake-storage mechanism specific for monoamines. During the first week after birth, the lateral cortex consists of three layers divided into the marginal zone, cortical plate, and subplate layer (Kristt and Molliver, 1975). Synapses are quite sparse throughout the layers of the neonatal cortex; however, there are regions of relatively greater synaptic density in the marginal zone, in the deep third of the cortical plate, and in the subplate layer. In untreated animals, all synaptic terminals contain clear or “empty” vesicles. In neonatal rats treated with 5-hydroxydopamine, approximately 30% of all synaptic terminals are filled with small round vesicles that contain a granular deposit (Molliver and Kristt, 1976) (Fig. 4). By 6 days after birth, the synapses containing small granular vesicles are mostly concentrated in the deep third of the cortical plate wherein 70% of the synaptic boutons contain these granular vesicles. Many of the axons form multiple synapses de passage: notably the same type of granular vesicles is found at every junctional site for a given axon. Pretreatment of the neonatal rat with reserpine, a drug that blocks the monoamine vesicular storage process prior to the administration of 5-hydroxydopamine, abolishes the development of vesicular precipitates. Hence, the vesicular accumulation of 5-hydroxydopamine occurs in a reserpine-sensitive site that is probably restricted to monoaminergic terminals. Since an apparent high density of monoaminergic terminals can be demonstrated in the neonatal cortex only after loading with the catecholamine congener, the effect of pretreatment with the catecholamine precursor L-dopa on the levels of endogenous catecholamines in the cortex of the neonatal rat has also been examined (Coyle and Molliver, 1977) (Table 111). Previous studies in the fetal and

NEUROTRANSMISSlON IN T HE DEVELOPING BRAIN

79

Flc. 4. Aminergic synapses in lateral newortex of the 6-day-old rat. Rats were pretreated with 5-hydroxydopamine: cortical sections were fixed in aldehyde-osmiiim. Llpper left: Synaptic Imuton with many sinall vesicles containing dense granules and two large dense core vesicles. LJpper right: .A termirial containing g n n n l a r vesicles t h i t edi ibits three areas of synaptic contact with asymmetric membrane speckalizations. Middle: Ax011 forms two ,ytup.s<.\ dr / K I . \ . S O ~ S Pwith granular vesicles at both contacts. Lowei left: T w o adjacent synaptic boiitons with one containing granular vesicles and the other containing clear vesicles. (Reproduced from Molliver and Kristt, 1976, with kind permission of the authors and publisher, Elsevier/Nortli-Holland Biomedical Press.)

neonatal rat have indicated that pretreatment with L-dopa results in marked increases in the catecholamine levels in the neocortex (Coyle, 1974; Kellogg and Lundborg, 1972): however, the specificity of such increases remains unclear since L-dopa can be decarboxylated to dopamine at many other sites besides aminergic neurons (Romero et nl., 1973). To reveal specific (i.e., reserpine-sensitive) storage sites for catecholamines, the amount of catecholamines that accumulate in the neocortex of animals pretreated with reserpine have been compared with those receiving L-dopa alone. Administration of a dose of L-dopa that has negligible effects on catecholamine levels in adult rat cortex results in a 15-fold increase in the catecholamine levels of the lateral cortex of the newborn rat. In neonates pretreated with the reserpine prior to the administration of L-dopa, the accumulation of

80

JOSEPH T. COYLE

TABLE 111 EFFECTSOF PHARMACOLOCIC TREATMENTS ON THE LEVELS OF CATECHOLAMINES IN LATERAL CORTEXOF RAT"

Age Newborn

6 Days after birth

Adult

(A-B) ReserpineTotal sensitive catecholamine catecholamine (pg/mg tissue)

Treatment None A. L-dopa (100 mg/kg) B. L-dopa (100 mg/kg) + Reserpine (2.5 mg/kg) None A. L-dopa (100 mg/kg) B. L-dopa ( I 00 mg/kg) + Reserpine (2.5 mg/kg) None A. L-dopa (100 mg/kg) B. L-dopa (100 mg/kg) Reserpine (2.5 mg/kg) A. RO-4063 (50 mglkg) + t-dopa (250 mg/kg) B. RO-4063 (50 mg/kg) L-dopa (250 mg/kg) + Reserpine (2.5 mg/kg)

+

Adult

142 f 42 2267 f 280" 1362 905 f 81 115 f 20 1205 f 145" 796 409 f 42 313 f 35 314 f 39" 253 61 f 8 1141 f 124 301 840 k 79

' I Rats were administered reserpine (2.5 mg/kg) or an equivalent volume of vehicle 6-8 hours prior to sacrifice. Two hours before killing, they received L-dopa (100 mg/kg) by subcutaneous injection. In one experiment, adult rats were pretreated with a peripheral decarboxylase, RO-4063 (50 mg/kg), 30 minutes before the administration of L-dopa (250 mg/kg) to increase the amount of L-dopa in the brain. Results are expressed in terms of picograms of catecholamine per milligram tissue with S.E.M. indicated: each value is the mean of at least five experiments. (Derived from Coyle and Molliver, 1977.) I' A vs. B: P < 0.01 by student's t-test.

catecholamines in the cortex is markedly attenuated, indicating that nearly 1400 pg/mg of the accumulated catecholamine is localized in the reserpine-sensitive storage site. Although pretreatment of adults with a peripheral decarbosylase inhibitor and 2.5-fold higher doses of L-dopa significantly increase the catecholamine levels in their neocortex, this is largely in a reserpine-resistent compartment. These precursor loading studies indicate that the immature cortex can take up and store catecholamines in a reserpine-sensitive compartment with a capacity per milligram of tissue that exceeds the endogenous stores (142 pg/mg) by 8to 10-fold (1362 pglmg) and even exceeds the adult storage capacity (253 pg/mg) by 5-fold.

NEUROTRANSMISSION I N T H E DEVELOPING BRAIN

81

The number of synapses in neonatal cortex is extremely low as compared with the adult: nevertheless, these morphologic and biochemical studies suggest that the catecholaminergic neurons make a major contribution toward the innervation of the immature rat neocortex. The magnitude of the innervation indicates that this is one of the earliest major cortical inputs to be established. The density of catecholaminergic synapses in the adult lateral cortex is estimated to be less than 1% of the total number of synapses (Bloom, 1970; Hokfelt, 1968); hence, this dense monoaminergic input to the immature cortex is a transient feature of early postnatal development. Since the absolute density of synapses in the neonatal cortex is quite low as compared with the adult cortex, it is likely that the density of these monoaminergic terminals is subsequently diminished by the formation of more numerous synapses of other types from proliferating extrinsic and intrinsic neurons. The possibility that immature neurons which are not monoaminergic may take u p 5-hydroxydopamine and synthesize and store catecholamines in a reserpine-sensitive site cannot be excluded, and the recent demonstration that certain endocrine type cells exhibit such capabilities does offer a precedent (Pearse, 1975). I. SUMMARY AND SPECULATION

Taken together, these studies suggest that the central catecholaminergic neuronal system is one of earliest long-tract neuronal pathways to develop in the rat brain. These neurons provide widespread projections, especially to the rostra1 portions of the brain, at a time when the neurons in these regions are in the process of mitosis, migration, and the early stages of differentiation. Furthermore, the endogenous catecholamines are in a dynamic state compatible with neuronal activity. Histofluorescent investigation of fetal human brain indicates a similar remarkably early and generalized development of the catecholaminergic neuronal innervation (Nobin and Bjorklund, 1973; Olson et al., 1973). The precocious development of the central catecholaminergic neurons suggests that they may play a disproportionately important role in fetal brain function and possibly in regulating neuronal development. The functional role played by the catecholaminergic neurons in immature brain remains speculative. Important receptors for dopamine and norepinephrine in the central nervous system are specific adenylate cyclases (Kakiuchi and Rall, 1968; Shimizu et al., 1969; Kebabian et al., 1972). Adenosine-3',5'-cyclic nucleotide has been shown to regulate several fundamental aspects of neuronal differentiation including morphologic specialization, axogenesis, and differentia! expression of

82

JOSEPH T. COYLX

neurotransmitter synthesizing enzymes (Prasad et al., 1973 ; Simantov and Sachs, 1972; Roisen et al., 1972; Schrier and Shapiro, 1973). Conceivably, activation of specific adenylate cyclases by released dopamine and/or norepinephrine could regulate such processes during brain differentiation; surprisingly, however, norepinephrine-stimulated adenylate cyclase is not detectable until several days after birth (Schmidt et al., 1970; Perkins and Moore, 1973). Nevertheless, recent studies have demonstrated that the destruction of the cortical noradrenergic innervation during the perinatal period results in distortion of the pyramidal cell dendrites (Maeda et al., 1974) and alterations in the activity of choline acetyltransferase in the neocortex of the adult rat (Jaim-Etcheverry et al., 1975). Moreover, in several studies in which drugs that interfere with catecholaminergic neuronal transmission at pre- and postsynaptic sites were administered to fetal or neonatal rats, long-term abnormalities in neurochemical and behavioral parameters have been observed (Ahlenius et al., 1973, 1975; Engel and Lundborg, 1974; Nyakas et al., 1973; Middaugh et al., 1974). Because of the complex actions of these drugs, the causal relationship between interference with catecholaminergic neuronal transmission and the resulting abnormalities is highly suggestive but remains to be established.

111. Postnatal Development of the Nigrortriatol Circuit

A. BACKGROUND During postnatal development when complex neuronal circuits are being established, a better understanding of the functional importance of the differentiation of the processes mediating neurotransmission may be obtained by examining the developmental relationships among the components of a circuit as opposed to studying a single component in isolation. Because of the obvious complexities of the central nervous system, a neuronal circuit with relatively well-defined synaptic relationships and behavioral manifestations is particularly suitable for such studies. The nigrostriatal circuit satisfies these requirements. Dopaminergic, cholinergic, and y-aminobutyric acid (GABA)ergic neurons form synaptic relationships among themselves (McGeer and McGeer, 1976), and alterations in the functional balance among these three neuronal components, produce characteristic behavioral responses (Creese and Iversen, 1975; Price and Fibiger, 1974). The dopaminergic terminals that innervate the striatum arise from cell bodies localized in the substantia nigra in the base of the midbrain

NEUROTRANSMISSION I N T H E DEVELOPING BRAIN

83

(Ungerstedt, 1971). The innervation to the striatum by these neurons is extremely dense with an estimated 10-15% of all synapticcontacts in the adult striatum being dopaminergic (Hokfelt et al., 1970). The cholinergic innervation to the striatum is derived almost exclusively from neurons intrinsic to the region; hence, striatal cholinergic neurons are primarily local circuit neurons (McCeer et al., 1971; Bak et al., 1975). The striatal GABAergic neurons consist of two populations. One group, like the cholinergics, are local circuit neurons, whereas others have cell bodies within the striatum but send axonal processes that innervate the substantia nigra (Hattori et al., 1973; Fonnum et al., 1974). Although the dopaminergic input to the striatum has been thought to exert primarily inhibitory postsynaptic effects (York, 1972; Siggins ~t al., 1975), recent electrophysiologic studies provide compelling evidence that postsynaptic excitation also occurs (Kitai et al., 1976). As with other neurotransmitters, the neuronal response to dopamine reflects the characteristics of the postsynaptic receptor, of which there are probably more than one type (Cools and Van Rossum, 1976). Nevertheless, the dopaminergic influence on striatal cholinergics is inhibitory since stimulation of dopamine receptors results in decreased turnover of acetylcholine, whereas blockade of these receptors accelerates the firing rate of the striatal cholinergic neurons (Sethy and Van Woert, 1974; Racagni et al., 1976; Trabucchi et al., 1975). The postsynaptic effects of acetylcholine in the striatum are complex and dependent upon the characteristics of the cholinergic receptors: however, muscarinic cholinergic receptors are found in high concentration in the striatum and play a major role in extrapyramidal function (Yamamura and Snyder, 1974a). GABA has nearly universal inhibitory effects at pre- and postsynaptic sites (Davidson, 1976); hence the striatonigral GABAergic pathway may serve as an inhibitory feedback loop regulating the firing rate of the nigra! dopaminergic cell bodies (McCeer and McCeer, 1976). B. DOPAMINERCIC INNERVATION TO STRIATUM The synthesis pathway for dopamine involves ring hydroxylation of L-tyrosine to form L-dopa by tyrosine hydroxylase, which is highly concentrated within the dopaminergic processes. The L-dopa formed is rapidly converted to dopamine by dopa decarboxylase, an enzyme which unlike tyrosine hydroxylase is found in other neuronal and nonneuronal constituents of the striatum (Hokfelt et al., 1973). At birth, the specific activity of tyrosine hydroxylase in the striatum is 10% of adult specific activity when assayed at saturating levels of' substrate and cofactor (Porcher and Heller, 1972; Coyle and Campochiaro, 1976) (Fig. 5 ) .

84

1. L U I L D

I DA Sensitive

.-

Dopamine

cu

I;./

n

0

0

-

5

4-

a

,

,

,

I

, . .

I

400- Tyroaine

/ / Dopomine

Hydroxyloae

Uptake

.-a

*I 0

0

7

14

28

Adul?

14

28

Adul?

Age (Days After Birth)

FIG. 5. Development of the neurochemical indices of dopaminergic innervation in the striatum. The activity of tyrosine hydroxylase (Coyle, 1972), the concentration of dopamine (Coyle and Henry, 1973), the activity of dopamine-sensitive (DA-sensitive) adenylate cyclase (ad. cyclase) (Kebabian et al., 1972), and the synaptosomal uptake of [:'H]dopamine (Holz and Coyle, 1974) were measured in rat striatum. Values are presented in terms of milligrams of wet weight of tissue and are the mean kS.E.M. of five or more preparations. These results are summarized from Coyle and Campochiaro (1976).

The specific activity doubles during the first 7 days after birth, nearly trebles between 1 week and 2 weeks after birth, to achieve 75% of adult specific activity by 4 weeks after birth. The activity of dopa decarboyxlase exhibits a different developmental profile in the striatum (Lamprecht and Coyle, 1972). Its specific activity remains relatively constant for the first week after birth, doubles between the first and second week, but achieves only 55% of adult specific activity by 4 weeks after birth. The concentration of dopamine in the neonatal striatum is 10% of that of the adult. It exhibits a linear increase with respect to time to achieve 75% of adult levels by 4 weeks after birth. Thus, there is a striking correlation between the developmental profile of tyrosine hydroxylase, the initial and rate-limiting enzyme in the synthesis pathway, and the levels of

NEUROTRANSMISSION IN THE DEVELOPING BRAIN

a5

endogenous dopamine in the striatum. Since tyrosine hydroxylase and dopamine are almost exclusively localized to dopaminergic terminals in the striatu m, their developmental increases presumably reflect i ngrowt h of dopaminergic axons and terminals. Dopamine is taken up by a specific, high-affinity transport process that is localized on dopaminergic terminals innervating the striatum (Coyle and Snyder, 1969). This transport process exhibits pharmacologic and kinetic characteristics unique to the dopaminergic terminals and is driven by the sodium gradient across the neuronal membrane (Holz and Coyle, 1974). Since the high-affinity uptake of dopamine is proportional to the density of dopaminergic innervation, and since most of the uptake can be recovered in synaptosome-enriched particulate fractions (Coyle and Snyder, 1969), the developmental changes in the synaptosomal uptake of dopamine can be used as a means for quantifying the formation of dopaminergic terminals. The high-affinity uptake of dopamine by particulate fractions prepared from striatum of the newborn rat is 10% of that of the adult. During subsequent development, there is a linear increase in the uptake with age such that 70% of the adult activity is attained by 4 weeks after birth. Examination of the kinetics of uptake into particulate fractions at different ages reveals that the developmental changes reflect an increase in the maximum velocity of transport (V,,,), whereas the affinity of the carrier for dopamine (KO remains constant throughout development. Hence, the number of transport sites for dopamine increases during striatal maturation: this presumably corresponds with the formation of dopaminergic terminals that can be converted to synaptosomes by the homogenization process. The striking similarity in the developmental profiles of the three presynaptic biochemical markers for the dopaminergic neurons, tyrosine hydroxylase, dopamine levels, and activity of synaptosomal uptake of dopamine merits closer scrutiny. It must be recalled that tyrosine hydroxylase activity measured in vitro is done under optimal conditions for pH, cofactor, and substrate and therefore may be an unreliable index of enzyme activity in vim (Richelson, 1976). The levels of dopamine in the neuron even at the earliest stages of differentiation are primarily dependent upon the vesicular storage capacity and are only indirectly related to tyrosine hydroxylase activity (Coyle and Henry, 1973). The synaptosomal uptake studies utilize a highly derivative preparation; although synaptosomes have been demonstrated in homogenates prepared from immature brain, it is unclear how growth cones, immature boutons, and nascent synapses en passage characteristic of the catecholaminergic neurons behave during the process of homogenization and subcellular

86

JOSEPH T. COYLE

fractionation (Gonatas et al., 1971; Jones and Revell, 1970). In spite of these caveats, the close correlation in development among the three presynaptic parameters suggests that dopaminergic neuronal differentiation is a tightly coordinated process in which the biosynthetic enzymes, their cofactors and regenerating systems, the storage vesicles, and the presynaptic high-affinity transport sites develop in close temporal association with the formation of terminals. The postnatal development of the dopaminergic innervation to the striatum as characterized by histofluorescent microscopy reveals important regional and microanatomic heterogeneity of this process. Significant dopaminergic innervation to the caudate putamen has occurred prior to birth of the rat. In the neonate, the catecholamine histofluorescence of the neostriatum is concentrated along a marginal zone of the nucleus. Within the nucleus there are patches or islands of densely packed terminals which are often associated with the fibers of the internal capsule: between these islands, the characteristic fine fluorescent processes of dopaminergic terminals are sparse and faint (Seiger and Olson, 1973; Olson et al., 1972). During the next 3 weeks of postnatal development, the fluorescent fibers between the islands become progressively more prominent and confluent, thus obscuring the islands. By 4 weeks after birth, the striatum exhibits the diffuse green fluorescence of the adult preparations. The fluorescent islands observed in the immature striatum represent an anatomically and physiologically unique class of dopaminergic terminals since they can be revealed in adult striatum after inhibition of dopamine synthesis (Olson et al., 1972). In a detailed study of the development of the rabbit caudate putamen, Tennyson et al. ( 1972) have described a correlation between the ontogenetic changes in the density of fluorescent processes and the maturation of growth cones into more differentiated axons with synapses en passage characteristic of catecholaminergic neurons. The nonlinear relationship between fluorescence and catecholamine concentration at high levels of the amines confounds interpretations of terminal density even in the neonatal period when innervation is restricted primarily to dense patches of terminals (Jonsson, 1969). Nevertheless, accepting such restraints, there appears to be a reasonable correlation between the development of dopaminergic innervation as measured by histofluorescent microscopy and as measured by the biochemical markers for these neurons. The process of neurotransmission requires not only the availability of releasable neurotransmitter at the presynaptic site but also the presence of appropriate receptors on the postsynaptic neuron. There is compelling evidence that a specific dopamine-sensitive adenylate cyclase mediates certain effects of dopamine on neurons innervated by

NEUROTRANSMISSION IN T H E DEVEL.OPING BRAlN

87

dopaminergic terminals (Kebabian et al., 1972; Siggins et a!., 1975). In the striatum, dopamine-sensitive adenylate cyclase has a postsynaptic localization primarily on neurons intrinsic to the region (Schwarcz and Coyle, 1977). The development of dopamine-sensitive adenylate cyclase has been assessed in cell-free homogenates from rat striatum. Dopamine-sensitive adenylate cyclase activity is present in the newborn striatum with a specific activity of 20% of that of the adult; there is a linear increase in the activity of the enzyme to adult levels at 3 weeks after birth (Coyle and Campochiaro, 1976). Whereas the specific activity of the dopamine-stimulated adenylate cyclase increases !&fold between birth and adulthood, the basal cyclase increases 10-fold during this same period. Utilizing a different assay technique, Von Hungen et al. (1974) has similarly shown that a relatively high activity of dopamine-stimulated adenylate cyclase is present in the forebrain regions of the neonatal rat. When the development of the pre- and postsynaptic markers are expressed in terms of percent of adult levels, it is evident that the development of the dopamine-sensitive adenylate cyclase parallels but precedes that of the presynaptic markers. In another system, the chick neural retina, the developmental appearance of dopamine-sensitive adenylate cyclase unequivocally precedes that of the presynaptic markers for the dopaminergic interneurons (Coyle, unpublished). Although these studies suggest that the apparent dopamine receptor develops in advance of its presynaptic input, such an interpretation must be made cautiously because of the complex nature of the adenylate cyclase system. Receptors may be quantified by measuring the specific binding of radioactive ligands that have a high affinity for the receptor site. Such binding studies can provide a more accurate assessment of receptor density because they do not require receptor interaction with effector enzymes such as the adenylate cyclase system. An apparent dopamine receptor has been identified by means of measuring the specific binding of the dopamine receptor antagonist, [3H]haloperidol, to neuronal membranes; it is currently unclear whether the haloperidol binding site is separate from the dopamine receptor as measured by the adenylate cyclase technique (Burt et nl., 1976a). The specific haloperidol binding site is present in the striatum at birth but at a concentration of only 10% of that of the adult: it remains at this level until 1 week after birth, whereupon receptor density increases linearly to achieve adult levels by 25 days after birth (Burt et al., 1976b). Although the ['Hlhaloperidol binding exhibits a more gradual development in the striatum than dopamine-sensitive adenylate cyclase, its developmental profile is nearly identical with that of the presynaptic markers for stnatal cholinergic neurons, the neurons thought to bear the receptor (see below).

88

JOSEPH T. COYLE

C. STRIATAL CHOLINERGIC NEURONS The synthesis of acetylcholine is catalyzed by choline acetyltransferase, an enzyme that is highly localized to cholinergic neurons in the nervous system (McGeer et al., 1974). The factors that regulate choline acetyltransferase activity in vim are poorly understood, although end product inhibition and the availability of the substrate choline may play important roles (Simon el al., 1976; Kaita and Goldberg, 1969). In the newborn rat striatum, the specific activity of choline acetyltransferase is only 2% of that of the adult, and during the first week after birth increases only an additional 2% (Coyle and Yamamura, 1976) (Fig. 6). After the first week postpartum, there is a dramatic inflection in the rate of increase in choline acetyltransferase with a linear 20-fold increase in

Murcarinic

E I? 3 0

Choline Uptake

c

0.10

5

0.05

t 9

t0 0

E 0

E 0

7

14

28

Adult

0

7

14

28

Adult

Age (Days After Birth 1 FIG. 6. Development of the neurochemical indices of cholinergic innervation in the striatum. The activity of choline acetyltransferase (Bull and Oderfeld-Nowak, 197 1). the concentration of acetylcholine (Goldberg and McCaman, 1973), the concentration of specific binding sites of ['H]quinuclidinylbenzilate (Yamamura and Snyder, 1974b), and the synaptosomal uptake of choline (Simon et al., 1976) were measured in rat striatum. Val~iesare presented in terms of milligrams of wet weight of tissue and are the mean *S.E.M. of five or more preparations. These results are summarized from Coyle and Yamamrira (1976).

NEUROTRANSMISSION IN T H E DEVELOPING BRAIN

89

specific activity of the enzyme such that 70% of the adult level is attained by 4 weeks after birth. Because of the extremely rapid turnover of endogenous acetylcholine that can be markedly affected by decapitation, it is essential to arrest neurotransmitter metabolism as rapidly as possible to obtain reliable steady-state levels of acetylcholine (Cheney et d.,1976). Focused microwave radiation nearly instantaneously stops metabolism by heatdenaturation of the enzymes; and it has been shown to provide the most reliable preparation for acetylcholine determinations. With this technique, the concentration of acetylcholine in the neonatal striatum is remarkably high, being 23% of that of the adult. This represents a relative disparity of 10-fold between the level of acetylcholine and the activity of its biosynthetic enzyme, choline acetyltransferase. Like the transferase, there is only a modest 2% increase in the concentration of acetylcholine during the first week after birth; subsequent to this, there is a linear increase in the concentration of the neurotransmitter, which attains adult levels by 4 weeks after birth. Although the cholinergic neurons do not have an uptake mechanism for acetylcholine, they do possess a sodium-dependent high-affinity transport for its precursor, choline (Simon etal., 1976).Suszkiw and Pilar (1976) have shown that the high-affinity transport mechanism for choline is a specialized process restricted to cholinergic terminals and is not found on cholinergic soma or dendrites. Hence, the development of this transport process in synaptosomal preparations could provide an index of cholinergic terminal formation. The specific activity of the high-affinity uptake process for choline in striatal synaptosomes is quite low until 10-15 days after birth (Coyle and Yamamura, 1976; Sorimachi and Kataoka, 1976). Then, there is a rapid, linear increase in specific activity of the transport process such that 60-75% of adult activity is achieved by 4 weeks after birth. The low-affinity transport process for choline, which is revealed by its lack of dependence on sodium ion, remains relatively constant in activity throughout the postnatal maturation; thus, during early stages of striatal differentiation, most of the choline is transported by this nonspecific mechanism. The developmental increase in the high-affinity choline uptake process is due to an increase in the maximal velocity of transport (V,,,), an index of the number of transport sites, whereas no significant change in the affinity ( K J of the carrier for choline is noted. Although acetylcholine has a number of different postsynaptic effects in the central nervous system, a subset of acetylcholine receptors with “muscarinic” characteristics can be labeled and quantified with potent radioactive muscarinic receptor antagonists (Yamamura and Snyder,

90

JOSEPH T. COYLE

1974a; Birdsall and Hulme, 1976). This muscarinic receptor, which has a particularly high concentration in the striatum, is localized primarily on neurons postsynaptic to the cholinergic neurons (Yamamura and Snyder, 1974b). The muscarinic receptor is present in the newborn striatum with a concentration of 10%of that of the adult; specific binding of the radioactive ligand shows a negligible change during the first week after birth and then increases in a linear fashion to achieve 75% of adult concentration by 4 weeks after birth (Coyle and Yamamura, 1976). Although the developmental rise in the muscarinic receptor precedes that of the activity of choline acetyltransferase, it is nearly superimposable on the ontogenic changes in the concentration of endogenous acetylcholine. The neurons intrinsic to the striatum, which include the cholinergic interneurons, exhibit a period of intense cell division during the final days of gestation in the rat that continues in a much more attenuated form for a few days after birth (Das and Altman, 1970). Hence, the low levels of choline acetyltransferase activity and negligible activity of synaptosomal high-affinity uptake process for choline in the neonatal striatum are compatible with the relatively late developmental appearance of these neurons. Nadler et al. (1975) and Burt (1968) have shown that the activity of choline acetyltransferase increases acutely at the time of formation of cholinergic synapses in regions of the central nervous system (CNS) that receive these terminals from extrinsic cholinergic cell bodies. The marked and coordinated increase in the activity of choline acetyltransferase and the synaptosomal high-affinity uptake process for choline beginning during the second week of age are suggestive of cholinergic terminal formation in the striatum. In support of this hypothesis, electron microscopic studies reveal that the synaptic density in the striatum increases progressively after birth with the greatest increment between 13 and 17 days of age (Hattori and McGeer, 1973). The high concentration of acetylcholine in the neonatal striatum is at variance with the relative immaturity of these neurons at this stage. Certain factors may contribute to this disparity. Acetylcholinesterase has been shown to play an important role in regulating the neurotransmitter levels in the adult brain, since inhibition of its activity results in considerable increases in the acetylcholine content (Cheney et al., 1976). The activity of acetylcholinesterase is low in the neonatal rat striatum, but more importantly, is deficient in the small interneurons thought to be cholinergic (Butcher and Hodge, 1976; McGeer et al., 1971). In adult striatum, acetylcholine levels are inversely related to the firing rates of the neurons with an estimated half-life of the neurotransmitter of only several minutes (Schuberth et al., 1970). Since it is known that spontane-

91

NEUROTRANSMISSION IN THE DEVELOPING BRAIN

ous firing rates of neurons in the immature nervous system are considerably slower than in the adult brain, the resulting lower rate of utilization of acetylcholine could contribute to higher steady-state levels of the neurotransmitter in the immature neuron. It is noteworthy that the acetylcholine in the neonatal brain is not concentrated in synaptosomes but rather localized primarily in a soluble pool that may not be accessible for release (Abdel-Latif et al., 1970).

D. STRIATAL GABAERGICNEURONS The GABA synthesis in brain is catalyzed by glutamic acid decarboxylase, an enzyme that is highly concentrated in but not exclusively localized to neurons (Barber and Saito, 1976). The specific activity of glutamic acid decarboxylase in the newborn rat striatum is 7% of that of the adult (Fig. 7). The specific activity of the enzyme remains constant during the first week after birth: then there is a linear increase in the activity between 1 and 4 weeks after birth, at which time 75% ofthe adult activity is reached (Coyle and Enna, 1976). In contrast to the relatively low activity of glutamic acid decarboxylase, the concentration of GABA in the striatum of the newborn rat is quite high, being 40% of the adult level. Thus, there is a 6-fold disparity between the levels of the neurotransmitter and the activity of its biosynthetic enzyme. The concentration of GABA is lower at 1 week after birth but then subsequently rises in

Glutornote

n 100

U

0

7

14

28

Adull

0

7

14

28

Adult

T

T

7

T

T

r

Z

U

l

Age (Days After Birth)

FIG. 7 . Development of the neurochemical indices of GABAergic innervation in the striatum. The activity ofglutamic acid decarboxylase (Wilson etal., 1972), the concentration of GABA (Graham and Aprison, 1966), and the synaptosomal uptake of ["HIGABA were measured in rat striaturn. Values are presented in terms of milligrams of wet weight of tissue and are the mean k S.E.M. of five or more preparations. These results are summarized from Coyle and Enna (3976).

l

92

JOSEPH T. COYLE

close association with the activity of glutamic acid decarboxylase. These observations concerning the developmental lag between GABA and its synthesizing enzyme are in general agreement with the pioneering studies of Roberts et al. (1951) and of subsequent investigators in several species (Van den Berg et al., 1965; Vernadakis and Woodbury, 1962). The GABAergic neurons possess a high-affinity uptake process for their neurotransmitter. Although glia also accumulate GABA by a highaffinity transport process, the neuronal uptake mechanism can be distinguished from the glial on the basis of its kinetics, pharmacologic characteristics, and subcellular localization (Levi and Raiteri, 1973: Schon and Kelley, 1975; Sellstrom and Hamberger, 1975); furthermore, specific lesions of the neurons intrinsic to the striatum result in 7 0 4 0 % decrements in the apparent neuronal uptake of GABA in the region (Coyle and Schwarcz, 1976). Although there has been disagreement about whether the accumulation of [BHIGABA measured in synaptosomal fractions represents a net uptake process or homoexchange, it has recently been demonstrated that net uptake of GABA does occur into the synaptosoma1 fractions. The development of the uptake process for GABA has been examined in washed, crude synaptosomal fractions prepared from striatum; the washing procedure is essential because during the homogenization process large amounts of GABA are released into the soluble phase (Mangan and Whittaker, 1966). The uptake of GABA by the synaptosomal fractions from newborn striatum is nearly identical to that of the adult. During the first 2 weeks after birth, there is a marked increase in the activity of the uptake process which peaks at 300% of adult levels by 14 days postpartum. During the subsequent 2 weeks, the activity of the uptake process decreases progressively to a level only 50% greater than that of the adult. The high-activity of the uptake process at 2 weeks after birth is not an artifact of the preparation since comparably elevated activities have been noted in cortical slices from immature rat and chick (Schousboe et al., 1976: Levi, 1972). Because of the unusual changes in the development of the GABA transport process, the kinetics of GABA uptake have been examined in partially purified synaptosomes. The K t or apparent affinity of the carrier mechanism for GABA does not change significantly between birth and adulthood; however, the Vmaxof uptake increases to 225% of adult by 15 days after birth and then decreases toward adult activity subsequently. The characteristics of the GABA transport process at the peak of uptake at 15 days after birth as compared with adult are quite similar; the uptake is a sodium-dependent process occurring in an osmotically sensitive site and is 20-fold more sensitive to inhibition by diaminobutyric

NEUROTRANSMISSION IN THE DEVELOPING BRAIN

93

acid, an inhibitor of neuronal CABA transport, than P-alanine, an inhibitor of the glial transport process (Coyle and Enna, 1976). T h e kinetic studies indicate that the number of transport sites for GABA is markedly elevated in the striatum at 15 days after birth. It is possible that the immature GABAergic neurons have a particularly high density of transport sites on their neurolemma at 2 weeks postpartum, which decreases as the neuron undergoes subsequent differentiation. Alternatively, since the V,,, of transport appears to be proportional to the density of nerve terminals, GABAergic neurons may elaborate a high density of terminals or processes capable of forming synaptosomelike particles by 2 weeks after birth that atrophy or are diluted by the subsequent growth of other neuronal or non-neuronal constituents of the striatum. Regardless of the reason for this elevated transport capacity of the 15-day-old striatum, there is a considerable disproportionality between the levels of endogenous GABA and the activity of the process that terminates its action at the synapse. E. ONTOGENESIS OF

THE

FUNCTION OF

THE

NIGROSTRIATAL CIRCUIT

The presence of the neuronal pathway as documented by biochemical markers is suggestive but not conclusive evidence that the neurons may exert a physiologically significant influence. It is necessary, in addition, to demonstrate that the neurons in the pathway are active, i.e., releasing a neurotransmitter. Neurochemically, such a determination can be made by examining the turnover of the neurotransmitter. Unfortunately, because of the complex nature of the precursor pools of choline and glutamic acid for acetylcholine and GABA respectively, and because of the lack of specific inhibitors for their biosynthetic enzymes, there is currently no information on the turnover of these important neurotransmitters at early stages of development. With respect to the nigrostriatal dopaminergic pathway, several investigators have utilized different approaches to examine the ontogenesis of dopamine turnover. Using the semiquantitative technique of histofluorescent microscopy to measure catecholamine concentration in the striaturn, Loizou (197 1) has found that dopamine depletion in the striatum after treatment with a-methylparatyrosine, an inhibitor for tyrosine hydroxylase, is comparable in rats from birth through to adulthood; this suggests that the halflife of neuronal catecholamines is qualitatively similar at all ages. In a study on the effects of inhibition of tyrosine hydroxylase on the levels of dopamine in whole brain of postnatal rats (nearly 80% of whole brain dopamine is limited to the striatum), Kellogg and Lundborg (1973) observe only a 25% decrease in dopamine 4 hours after inhibition of its

94

JOSEPH T. COYLE

synthesis in l-day-old rats, whereas in 4-day, 10-day, 2-week, and 3-week-old rats the concentration falls approximately 50% by 2 hours after injection of the inhibitor. The turnover plots have biphasic characteristics for the 4-day-old and 10-day-old rats with an inflection at 2 hours; in contrast, at subsequent stages of development, there is linear decrease in dopamine concentration throughout the 4-hour period. Later studies limited to the striatum suggest that dopamine turnover at 14 and 28 days after birth are similar, whereas at 4 days it is somewhat slower (Kellogg and Wennerstrom, 19’74). Keller et al. (1973) have assessed the turnover of dopamine by examining the developmental changes in the primary metabolite of released dopamine, homovanillic acid (HVA). The concentration of HVA increases 4-fold between 4 days and 12 days after birth and then decreases about 30% to adult levels by 4 weeks after birth; during the same period, the concentration of dopamine in the striatum exhibits a linear 8-fold increase. Hence, there is a progressive decrease with maturation in the ratio of HVA to dopamine in whole brain and in the striatum. Since the ratio of HVA/DA may be taken as an indicator for the turnover of dopamine which is due to neuronal activity, the maturational decrease in this ratio suggests that dopamine turnover is higher at earlier than at later periods of postnatal development. Although the turnover studies suggest that the dopaminergic neurons in the nigrostriatal pathway are active by at least 4 days after birth, if not before, these studies do not indicate whether dopaminergic neuronal firing occurs in an unregulated fashion or is controlled by neuronal feedback in the nigrostriatal circuit. The establishment of the functional relationship among the components of this circuit can be assessed by examining the effects of drugs that alter their neurotransmission. Accordingly, treatment with drugs that directly increase o r decrease postsynaptic dopamine receptor stimulation result in compensatory alterations in dopaminergic neuronal activity in the adult brain, presumably through changes in the activity of a striatonigral feedback loop. Administration of apomorphine, a direct dopamine receptor agonist, causes a compensatory decrease in the firing rate and turnover of dopamine in the adult rat striatum; in contrast, apomorphine minimally affects the dopamine turnover in the 4-day-old rat (Kellogg and Wennerstrom, 1974). Administration of apomorphine causes a 68 % decrease in the concentration of HVA of whole brains of adult rats; however, the reduction at 4 days is only 25% but increases progressively with age to a 47% reduction in the 30-day-old rat. Blockade of the postsynaptic dopamine receptors causes a compensatory increase in the firing rate of

NEUROTRANSMISSION IN T H E DEVELOPING BRAIN

95

dopaminergic neurons. Administration of the potent dopamine receptor antagonist, haloperidol, causes a significant increase in the brain HVA levels at 4 days after birth through to adulthood. The increase at 4 days after birth is only 30%, by 12 days after birth is loo%, and in the adult brain is 300% (Keller et al., 1973). These studies suggest that the feedback regulation of dopaminergic neuronal function is probably absent at birth but appears by 4 days after birth. The ability of the immature dopaminergic neurons to alter firing rates to compensate for pharmacologic perturbations of their postsynaptic receptors is, however, limited. To assess the development of the inhibitory dopaminergic input on cholinergic neurons intrinsic to the striatum, the effects of administration of dopamine receptor antagonists on acetylcholine levels in the region have been examined. For the adult striatum, treatment with dopamine receptor blockers results in significant decreases in the levels of endogenous acetylcholine as well as increases in the turnover of the neurotransmitter. It should be noted that such postsynaptic effects of a receptor antagonist should occur only when there is a tonic presynaptic input, the effects of which the drug can block. The administration of dopamine receptor blockers to rats from birth to 1 week of age has no effects on steady-state levels of acetylcholine in their striata (Guyenet et al., 1975; Coyle and Campochiaro, 1976). However, at 8 days after birth, such treatments cause a significant reduction in the levels of acetylcholine; there is a progressive increase in this response at subsequent ages until 17 days after birth, at which time the adult response is attained. These results suggest that dopaminergic neurons assume a functional inhibitory control over the cholinergic striatal neurons at 8 days after birth. Notably, at this point, the marked developmental increases in activity of choline acetyltransferase, activity of the high-affinity uptake of choline, and the density of muscarinic receptors occurs in the striatum (Coyle and Yamamura, 1976). Disturbances in the function of the dopaminergic, cholinergic, and GABAergic neurons in the nigrostriatal circuit have been implicated in the presentation of abnormal behaviors including stereotypy, catalepsy, and possibly hyperactivity (Cools and Van Rossum, 1970; Creese and Iversen, 1973; Costal1 and Olley, 1971; Drayet al., 1975). The postnatal development of the inducability of such behaviors by pharmacologic perturbations of the components of the circuit can provide insight into the functional maturation of the nigrostriatal circuit. It must be emphasized, however, that the behavioral manifestations are limited by the motor repertoire of the animal at a given stage of development; hence, the lack of a behavioral response may indicate that either the primary

96

JOSEPH T. COYLE

neuronal circuit or the efferents from the circuit which manifest the behavior have not yet developed. As noted in the biochemical studies in the preceding section, a positive response to an agonist demonstrates that the circuit distal to the neuron bearing the receptor for the agonist is functional, whereas a positive response to an antagonist provides evidence that the presynaptic neurons impinging upon the receptor blocked by the antagonist are functioning. Because of the limited behavioral repertoire of neonatal rats, there is disagreement concerning the exact age of onset of stereotypic behavior induced by directly acting apomorphine or indirectly acting amphetamine and L-dopa. McGeer et al. (1971) report that 80% of newborn rats treated with 2.5 mglkg of &hetamine exhibit stereotypic gnawing behavior. La1 and Sourkes (1973) date the onset of stereotypic behavior in rats at 2 days postpartum, although in the 1-day-old rats increased locomotion and head-rearing are observed; these latter two behaviors are considered by some investigators as components of dopaminergic receptor overstimulation (Roberts et al., 1975; Creese and Iversen, 1975). Similarly, these authors report the onset of stereotypic behavior in response to injection of 10 mg/kg of amphetamine at 2 days of age. In contrast, Kellogg and Lundborg (1972) indicate that the stereotypic behavior after administration of directly acting apomorphine or indirectly acting L-dopa appears much later in development. Nevertheless, they observe marked head-raising behavior and hyperactivity after administration of L-dopa as early as 1 day of age; and this behavior can be abolished by a high dose (20 mg/kg) of the dopamine receptor blocker, haloperidol. It is noteworthy that after treatment with L-dopa, there are massive increases in the dopamine concentration of the rat brain (25- to 50-fold; up to 5 pglgm), whereas norepinephrine levels change little if at all. Because the behavioral effects of L-dopa are similar to that of the a-receptor agonist clonidine, at early stages of development, these authors ascribe the hyperactivity to excessive stimulation of norepinephrine a-receptors; however, recent studies suggest that clonidine may interact with dopaminergic systems (Creese and Iversen, 1973, 1975). Cataleptic behavior occurs with marked dopamine receptor blockade; the ontogenic appearance of such behavior presumes a functional dopaminergic input. In a study on the development of catalepsy induced by dopaminergic receptor antagonist spiperone, Baez et al. (1976) observe a cataleptic effect in rats at all ages between birth and adulthood. In brief, these studies indicate that the behavioral responses to extremes of dopaminergic neuronal activity are demonstrable in the rat within a day or two after birth. Thus, the behavioral studies are in agreement with the neurochemical observations that dopaminergic neurons are functional in the early neonatal period.

NEUROTRANSMISSION IN THE DEVELOPING BRAIN

97

Phenomenologically, there appears to be a balance between dopamine receptor and muscarinic receptor stimulation such that behaviors produced by excessive dopaminergic activity can be attenuated by muscarinic receptor stimulation or potentiated by muscarinic receptor blockade (Scheel-Kruger, 1970; Zettler, 1968). This may, in part, reflect the inhibitory dopaminergic input on striatal cholinergic neurons but also suggests that these two neuronal types may exert opposing influences on a common effector. Hence, pharmacologic perturbations of the cholinergic system may uncover the ontogenesis of its functional role in the nigrostriatal circuit. Fibiger et al. (1970) demonstrated that the cholinergic agonist, pilocarpine does not antagonize the excitatory effects of amphetamine at 15 days but does at 20 days after birth and thereafter; furthermore, the muscarinic antagonist, scopolamine, does not potentiate the effects of amphetamine at 15 days but does at 25 days postpartum in the rat. I n a similar vein, McGeer et al. (1971) have observed that the muscarinic antagonist, scopolamine, does not potentiate stereotypic behavior induced by amphetamine in 10-day-old rats but does in 30day-old animals. I n their study on the ontogenesis of catalepsy, Baez et al. (1976) have demonstrated that the cholinergic agonist, pilocarpine, does not produce catalepsy in 10-day-old rats but does produce the behavioral syndrome in 15-day-old rats and at ages thereafter. These studies suggest that the cholinergic contribution to these behaviors develops considerably later than the dopaminergic. These behavioral studies correlate remarkably well with the neurochemical findings that the pre- and postsynaptic components of the cholinergic neurons in the striatum d o not differentiate until 1 to 2 weeks after birth and that the dopaminergic neurons do not exert an inhibitory influence on the cholinergic neurons in the striatum until at least 8 days after birth. IV. Conclusion

It was proposed in Section I that the processes that mediate synaptic transmission may provide biochemical markers for monitoring differentiation of specific neuronal pathways in the brain. In the case of the central catecholaminergic neurons, which can be visualized by histofluorescent microscopy, there is an impressive correlation between their biochemical and morphologic differentiation. T h e date of “birth” of these neurons as assessed by rH]thymidine autoradiography, histofluorescent microscopy, and neurochemical measures is nearly identical. Furthermore, the regional development of innervation by terminal projections of the catecholaminergic neurons corresponds with the ontogenetic changes in their presynaptic biochemical markers. Although specific histologic methods have not yet been applied to characterize

98

JOSEPH T. COYLE

fine-structural differentiation of the GABAergic and cholinergic neurons, the developmental changes in the presynaptic markers for these neurons in the striatum are compatible with less specific indices of neuronal differentiation in this region. It was proposed, in addition, that the relative development of the processes mediating neurotransmission would provide an insight into the functional influence of a neuronal pathway at a particular stage of brain development. T h e neurochemical studies in the striatum indicate that this region possesses postsynaptic receptors for dopamine and receives innervation by dopaminergic neurons by birth, whereas the preand postsynaptic markers for the cholinergic neurons exhibit a more delayed rate of development, commencing between I and 2 weeks after birth. In striking agreement with these neurochemical observations, behaviors that reflect imbalances between the dopaminergic and cholinergic neuronal function in the striatum can be produced within a day or two of birth by drugs that alter dopaminergic neurotransmission, whereas the cholinergic component of such behaviors does not appear until approximately 2 weeks after birth. It must be emphasized, however, that the various presynaptic parameters that regulate neurotransmission for a particular class of neurons do not invariably develop in a unitary or coordinate fashion. The GABAergic neurons are a most dramatic example of this, for there are marked disparities in the relative rate of development of glutamic acid decarboxylase activity, the levels of endogenous GABA, and the activity of the high-affinity uptake process for GABA. Such a developmental disparity may also be operative with regard to the catecholaminergic terminals innervating the cortex of the newborn rat, which appear to be deficient in the ability to synthesize endogenous neurotransmitters. Although the reasons for such differential rates of development for the presynaptic parameters remain speculative, the occurrence of such developmental disparities for catecholaminergic, cholinergic, and GABAergic neurons underlines the necessity of examining several parameters to assess neuronal differentiation and the hazards of relying on only one. In summary, these neurochemical studies provide an added dimension to the complex and heterogenous nature of brain development that has long been appreciated by neuroanatomists. T h e presynaptic processes that regulate the availability of neurotransmitters at the synaptic cleft exhibit different rates of development within a particular neuronal group; the shift in balance among these processes may significantly affect the physiologic influence of the neuronal pathway. Furthermore, among neuronal pathways, there exist major differences in the time of their appearance and the rate of their differentiation. This ontogenetic heterogeneity may explain, in part, the anomalous responses of imma-

NEUROTRANSMISSION IN T H E DEVELOPING BRAIN

99

ture animals to neuropharmacologic agents as well as the developmental changes in the symptoms of pediatric neurologic and psychiatric disorders. Finally, the precocious development of the processes that mediate neurotransmission for certain neuronal groups, such as the catecholaminergic neurons, suggest that these neurons may play an important role in modulating brain differentiation, ACKNOWLEDGMENTS T h e author gratefully acknowledges the excellent secretarial assistance of Carol Kenyon and Vickie Rhodes. Research support was received from USPHS Grants DA-00266 and MH-26654 and a March of Dimes Basil OConnor Award. REFERENCES Abdel-Latif, A. A., Smith, J. P., and Ellington. E. P. (1970). Erain Reg. 18, 441-450. Aghajanian, G. K., and Bloom, F. E. (1967).Brain Res. 6, 716-727. Agrawal, H. C.. Glisson. S. N.. and Himwich. W.A. (1966). Eiodtirrt. Bic@j>. A r l o 130, 51 1-513. Ahlenius, S., Brown, R., Engel, J., and Lundborg, P. (1973). Nnunyn-Schmie&herg’.s Arch. Pharmacol. 279, 3 1-37. Ahlenius, S., Engel, J., and Lundborg, P. (1975). Nniri?yn-Srhmipdeherg’s Arch. Pitarmacol. 288, 185-193. Ajika, K., and Hokfelt, T. (1973). Brain Re.$.57, 97-1 17. Anden, N.-E., Corrodi, H., Fuxe, K., and Hokfelt, T. (1967).Errr.]. Pharmmol. 2,59-64. Axelrod, J . (1970). In “Amphetamine and Related Compounds” (E. Costa and S. Garattini, eds.), pp. 207-216. Raven, New York. Baez, L. A., Eskridge, N . K., and Schein, R. (1976).Ezrr. J . Phnrmarol. 36, 155-1 62. Bak, 1. J., Choi, W. B., Hassler, R., Usunoff, K. G., and Wagner, A. (1975).Adz). NPtrrol. 9, 25-41. Baker, P. C., and Quay, W. E. (1969).Brain Res. 12, 273-295. Barker, R., and Saito, K. (1976).1n “GABA in Nervous System Function” (E. Roberts, T. N. Chase, and D. B. Tower, eds.), pp. 113-132. Raven, New York. Birdsall, N. J. M., and Hulme, E. C. (l976).J. Netrrochnn. 27, 7-16. Black, 1. B., and Geen. S. C . (1973).E m i N R n . 63, 291-302. Black, 1. B., and Geen. S. C. ( 1 9 7 4 ) , j .Nut,rochum. 22, 301-:306. Black, I. B., and Mytilineou, C. (1976). Brnin Re.\. 101, 503-521. Black, I. B., Hendry, I. A., and Iversen, L. L. (1971).Brain Re.7. 34, 229-240. Black, I . B., Hendry, I. A,, and Iversen, L. L. (1972).j. Neitrochtm. 19, 156771377. Bloom, F. E. (1970). Int. Reir. Neiirohiol. 13, 27-53. Bloom, F. E. (1973).J. Historhem. Cytoclrcm. 21, 333-348. Bull, G., a n d Olderfeld-Nowak, B. (1971).J. Nezrrorh,um. 19,935-947. Burt, A. M. (1968).J. Exp.Zoo1. 169, 107-112. Burt, D. R., Creese, I., and Snyder, S. H . (1976a). Mo/. Plinnnarol. 12, 800-812. Burt, D. R., Creese, I., Pardo, J.. Coyle, J. T., and Snyder, S. H . (1976b). Neirmrri. Ahstr. 2, 775. Butcher, L. L., arid Hodge, G. K. (1976). Brain Res. 106, 223-240. Cheney, D. L., Racagni, G . ,Zsilla, G . , and Costa, E. ( I 976).j. Phann. Phormacol. 28,75-78. Cools, A. R., and Van Rossum, J . M. ( 1 9 7 0 ) , ~ 4 w I tI .n / . Phromtrco+. Ther. 187, 165-172. Cools, A. R., and Van Rossum. J . M. (1976).Ysyrlio~itannacologia45, 243-251. Cooper, J. R., Bloom, F. E., and Roth, R. H . (1974). “The Biochemical Basis of’ Neuropharmacology.” Oxford IJniv. Press, London a n d New York.

100

JOSEPH T. COYLE

Costall, B., and Olley, J. (1971).Neuropharmacology 10, 297-306. Coyle, J. T. (1972). Biochm. Phannacol. 21, 1935-1944. Coyle, J. T. (1973). Neurosci. Res. 5, 35-52. Coyle, J. T. (1974). In “Dynamics of Degeneration and Growth in Neurons” (K. Fuxe, L. Olson, and Y. Zotterman, eds.), pp. 425-434. Pergamon, New York. Coyle, J. T., and Axelrod, J. (1971).J. Neurochm. 18, 2061-2075. Coyle, J. T., and Axelrod, J. (1972a).J. h‘eurochem. 19,449-459. Coyle, J . T., and Axelrod, J. (1972b).J. Neurochem. 19, 1117-1 123. Coyle, J. T., and Campochiaro, P. (1976).J. Neurochem. 27, 673-678. Coyle. J. T., and Enna, S. J. (1976).Brain Res. 111, 119-133. Coyle, J. T., and Henry, D. (1973).J. Neurochem. 21, 61-67. Coyle, J . T., and Molliver, M. E. (1977). SciPnrP 196,444447, Coyle, J. T., and Schwarcz, R. (1976).Nature (London) 263, 244-246. Coyle, J. T., and Snyder, S. H. (1969).J. Pharmacol. Exp. Ther. 170,221-231. Coyle, J. T., and Yamamtrra, H. I. (1976).Brain RP.L118, 4 2 9 4 4 0 . Coyle, J. T., Wooten, G. F., and Axelrod, 1 . (1974).,/. Neurochem. 22, 923-929. Creese, I., and Iversen, S. D. (1973). Brain Res. 55, 369-382. Creese, I., and Iversen, S. D. (1975).Brain Res. 83, 419-436. Dahlstrom, A., and Fuxe, K. (1964).Acta Physiol. Scand., Suppl. 232. Das, G. P., and Altman, J. (1970).Brain Res. 21, 122-127. Davidson, N. (1976). “Neurotransmitter Amino Acids.” Academic Press, New York. Dray, A., Oakley, N. R., and Simmonds, M. A. (1975). J . Phann. Pharmacol. 27, 627-629. Engel, J., and Lundborg, P. ( 1 974). Naunyn-Schmiedeberg‘sArch. Pharmacol. 282, 327-334. Falck, B., Hillarp, N.-A., Threine, G., and Torp, A. (1962).J. Hislochem. Cytochem. 10, 348-354. Fibiger, H. C., Lytle, L. D., and Campbell, B. A. (1970).J. Comp. Physiol. Psychol. 72, 384-389. Filogama, G., and Marchisio, P. C. (1971).Neurosci. Res. 4, 29-54. Fonnum, F., Grofara, I . , Rinvik, E., Storm-Mathisen, J., and Walberg, F. (1974).Brain Res. 71, 77-92. Goldberg, A. M., and McCaman, R. E. (1973).J. Neurochm. 20, 1-8. Goldstein, M., and Nakajima, K. (1967).J. Phannacol. Exp. The/. 157, 96-102. Gonatas, N. K., Antilio-Gambetti, L., Gambetti, P., and Shafer, P. (1971).J. Cell Biol. 51, 484-498. Graham, L. T., and Aprison, M. H. (1966). Anal. BiOchem. 15, 587-497. Gripois, D. (1975). Comp. Biochem. Physiol. C 51, 143-151. Guyenet, P. G., Beaujouan, J. C., and Glowinski, J. (1975). Nuunyn-Schmiedeberg’sArch. Plunacol. 288, 329-334. Haber, B., and Kuriyama, K. (1973).In “Biochemistry of the Developing Brain” (W. A. Himwich, ed.), pp. 37-54, Dekker, New York. Hartman, B. K., Zide, D., and Udenfriend, S. (1972). Proc. Natl. Acud. Sci. U.S.A. 69, 2722-2726. Hattori, T., and McGeer, P. (1973). Exp. Neurol. 38, 70-81. Hattori, T., McGeer, P. L., Fibiger, H. C., and McGeer, E. G. (1973). Brain Res. 54, 103114. Hicks, S . P., and D’Amato, C. J. (1968). Anat. Rev. 160, 6 1 9 6 3 4 . Hine, R. J., and Das, G. P. (1974). 2. Anat. Entwkklzingsgesch. 144, 173-186. Hokfelt, T. (1968). Z. Zellforsch. Mikrosk. Anat. 91, 1-13. Hokfelt, T., and Ljungdahl, A. (1972). Ad.cr. Biochem. Psychopharamacol. 6, 1-36. Hokfelt. T.,Jonsson, G., and Lidbrink, P. (1970).Brain Res. 22, 147-151.

NEUROTRANSMISSION IN THE DEVELOPING BRAIN

101

Hokfelt, T., Fuxe, K., and Goldstein, M. (1973). Brain Res. 62, 461-469. Holz, R. W., and Coyle, J. T. (1974). Mol. Phamucol. 10, 746-758. Iversen, L. L. (1967). “The Uptake and Storage of Noradrenalin.” Cambridge University Press, London and New York. Jairn-Etcheverry, G., Teitelrnan, G., and Zieher, L. (1975). Brain Res. 100, 699-704. Jones, D. G., and Revell, E. (1970). Z. Zellforsch. Mikrosk. Anat. 111, 179-208. Jonsson, G. (1969).J. Hirtochem. Cytochem. 17,714-723. Katia, A. A., and Goldberg, A. M. (1969).J. Neurocha. 16, 1185-1 19 I. Kakiuchi, S., and Rall, T. W. (1968). Mol. Phannacol. 4, 367-378. Kebabian, J. W., Petzold, G. L., and Greengard, P. (1972). Proc. Nutl. Acud. Sci. U.S.A. 69, 2145-2149. Keller, H. H., Bartholini, G., and Pletscher, A. (1973).Brain Res. 64, 371-378. Kellogg, C., and Lundborg, P. (1972). Psychophamacologia 23, 187-200. Kellogg, C., and Lundborg, P. (1973).Bruin Rex. 61, 321-329. Kellogg, C., and Wennerstrorn, G. (1974). Brain Res. 79, 451-464. Kitai, S. T., Sugurnori, M., and Kocsis, J. D. (1976). Exp. Bruin Res. 24, 351-363. Konig, N., Roch, G., and Marty, R. (1975). Anat., Histol., Eminyol. 148, 73-87. Kopin, I. J., Fischer, J. E., Musacchio,J. M., and Horst, W. D. (1964). Proc. Nutl. Acad. Sci. U.S.A. 52, 716-721. Korf, J., Aghajanian, G. K., and Roth, R. H. (1973). Neurophurmucology 12, 933-938. Kristt, D. A., and Molliver, M. E. (1975). Neurosci. Abstr. 1, 795. Krnjevic, K. (1974). Physial. Rev. 54, 418-540. Kuhar, M. J, (1973). L f e Sci. 13, 1623-1634. Lal, S., and Sourkes, T. L. (1973). Arch. Int. P h a m c o d y n . Ther. 202, 171-182. Larnprecht, F., and Coyle, J. T. (1972).Bruin Res. 41, 503-506. Lanier, L. P., Dunn, A. J., and Van Hartesveldt, C. (1976). Rev. Neurosci. 2, 195-256. Lauder, J. M., and Bloom, F. E. (1974).J. C m p . Neurol. 155,4694182. Lauder, J. M., and Bloom, F. E. (1975).J. C m p . Neural. 163,251-264. Levi, G . (1972). Arch. Biochem. Bwphys. 151,8-2 1 . Levi, G., and Raiteri, M . (1973). Life Sci. 12, 81-88. Loizou, L. A. (1971). 8 r . J . Phannaol. 41, 41-48. Loizou, L. A. (1972).Bruin Res. 40, 395418. McGeer, P. L., and McGeer, E. G. (1976). In “GABA in Nervous System Function” (E. Roberts, T. N. Chase, and D. B. Tower, eds.), pp. 487-495. Raven, New York. McGeer, P. L., McCeer, E. G., Fibiger, H. C., and Wickson, V. (1971). Brain Res. 35, 308-3 14. McGeer, P. L., McGeer, E. G., Singh, V. K., and Chase, W. H. (1974). Bruin Res. 81, 3 73-3 79. Maeda, T., and Dresse, A. (1969). Arfa Neurol. B d g . 69, 5-10. Maeda, T., Tohyarna, M., and Shimizu, N. (1974). Bruin Res. 70, 515-520. Mangan, J. L., and Whittaker, V. P. (1966). Biochem. J. 98, 128-137. Middaugh, L. D., Blackwell, L. A., Santos, C., and Zernp, J. W. (1974). Dev. Psychobwl. 7 , 429-438. Molinoff, P. B., and Axelrod, J. (1972).]. N e u r o c h . 19, 157-163. Molliver, M. E., and Kristt, D. A. (1976). Neurosci. Lett. 1, 305-310. Mueller, R. A., Thoenen, H., and Axelrod, J. (1969). Mol. Phurmucal. 5,463-469. Nadler, J. V., Matthews, D. A., Cotrnan, C. W., and Lynch, G. (1975). Dev. Biol. 36, 142154. Nathanson, J. A., and Greengard, P. (1975). Science 180, 308-310. Neff, N. H., Spano, P. F., Groppetti, A., Wang, C. T., and Costa, E. (1971).J. Pharmucol. Exp. Ther. 176, 701-710.

102

JOSEPH T. COYLE

Nobin, A., and Bjorklund, A. (1973). Acta Physiol. Scand., Suppl. 388, 1 4 0 . Nomura, Y., Naitoh, F., and Segawa, T. (1976). Brain Res. 101, 305-315. Nyakas, C., van Delf, A. M. L., Kaplanski, J., and Smelik, P. G. (1973).J. Neural Traxsm. 34, 253-266. Olson, L., and Seiger, A. (1972). Z. Anat. Entwicklungsgesch. 137, 301-316. Olson, L., Seiger, A., and Fuxe, K. (1972).Brain Res. 44,283-288. Olson, L., Boreus, L. O., and Seiger, A. (1973). Z. Anat. Entvickhngsgesch. 139, 250-282. Otten, V., and Thoenen, H . (1976). Mol. Phrmacol. 12, 353-362. Pearse, A. G. E. (1975).Z. Krehfursch. 84, 1-18. Perkins, J. P., and Moore, M.M. (1973). Mol. Pharmazol. 9, 774-782. Pickel, V. M., Joh, T. H., and Reis, P. J. (1976).J. Histochem. Cytochem. 24,792-806. Porcher, W., and Heller, A. (1972).J. Neurockm. 19, 1917-1930. Prasad, K. N., Kamar, S., Gilmer, K., and Vernadakis, A. (1973). Biochem. Biophys. Res. Commun. 50,973-977. Price, M. T. C., and Fibiger, H. C. (1974). Eur. J. Phaaacol. 29, 249-252. Racagni, G . , Cheney, D. L., Trabucchi, M., and Costa, E. (1976).J. Pharmacol. Exp. Ther. 196,323-332. Richards, J. G., and Tranzer, J. P. (1970). Brain Res. 17,463-469. Richelson. E. (1976).J. Neurochem. 27, 1113-1118. Roberts, D. C., Zis, A. P., and Fibiger, H . C. (1975).Brain Res. 93,4411154. Roberts, E., Harman, P. J., and Frankel, S. (1951).Proc. SOC.Exp. Biol. Med. 78,799-803. Roffi, J. (196S).J. Physiol. (Paris) 60, 455-494. Roisen, F. J., Murphy, R. A., and Braden, W. G. (1972). Science 177,809-811. Romero, J. A., Lytle, L. D., Ordonez, L. A., and Wurtman, R. J . (1973).J. Pharmazol. Exp. Ther. 184,67-75. Saavedra, J. M., and Axelrod, J. (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 769-772. Saavedra, J. M., Coyle, J. T., and Axelrod, J. (1974).J. Neurochem. 23, 511-515. Saavedra, J. M., Ribas, J.. Swann, J., and Carpenter, D. (1976). Neurosci. Abstr. 2, 355. Scheel-Kruger, .I. (1970). Acta Phaaacol. ToxLol. 28, 1-16. Schmidt, MI J., Palmer, E. C., Dettbarn, W.-E., and Robison, G. A. (1970).Dez1.Psychobiol. 3, 53-67. Schon, F., and Kelley, J. S. (1975). Brain Res. 86, 243-257. Schousboe, A., Lisy, V., and Hertz, L. (1976).J. Neurochan. 26, 1023-1027. Schrier, B. K., and Shapiro, D. L. (1973). Exp. Cell Res. 80, 459-462. Schuberth, J., Sparf, B., and Sundwall, A. (197W.J. Neurochem. 17,461-468. Schultze, B., Nowak, B., and Maurer, W. (1974).J . Comp. Neurol. 158, 207-218. Schwarcz, R., and Coyle, J. T. (1977).Brain Re,<. 127, 235-249. Seiger, A., and Olson, L. (1973). Z. Anat. Enhuicklungsgesch. 140, 281-318. Sellstrom, A., and Hamberger, A . (1975).J. Neurochem. 24,847-852. Sethy, V. H., andVan Woert, M. H. (1974).Res. Commun. Chem. Pathol. Pharmacol. 8,13-18. Shimizu, H., Daly, J. W., and Crevling, C. R. (1969).J. Neurochem. 16, 1609-1619. Sidman, R. L. (1970). I n “Contemporary Research Methods in Neuroanatomy” (W. J. H. Nauta and S. 0. E. Ebbesson, eds.), pp. 252-274. Springer-Verlag, Berlin and New York. Siggins, G. R., Hoffer, B. J., and Ungerstedt, U. (1975).L f e Sci. 15, 779-792. Simantov, R., and Sachs, L. (1972). E u r . 1 . Biochem. 30, 123-129. Simon, J. R., Atweh, S. F., and Kuhar, M. J. (1976).J. Neurochem. 26, 909-921. Sorimachi, M., and Kataoka, K. (1976). Brain Res. 94, 325-336. Suszkiw, J. P., and Pilar, G. (1976).J. Neurochem. 26, 1133-1 138. Tennyson, V. M., Barrett, R. E., Cohen, G., C6te, L., Heikkila, R., and Mytilineou, C . (1972). Brain Res. 46, 251-285.

NEUROTRANSMISSION IN T H E DEVELOPING BKAlN

103

Trabucchi, M . , Cheney, D. L., Racagni, G.. and Costa, E. (1975).RminRe.\. 85, 130-134. Ungerstedt, L1. (197 1). Acta P l y i o / . Scand., S i i j ~ p / .367, 1-93. Van den Berg, C. J., Van Kempen, G. M . J., Schade, J. P., and Veldstra, H . ( 1 96Fi).J. Neuroclzni. 12, 863-869. Vernadakis, A,, and Woodbury, D. M . (1962). Am. J . P l y i o l . 203, 748-752. Von Hungen, K., Roberts, S., and Hill, D. F. ( 1974).5. Neirrochm. 22, 8 1 1-X 19. Wilson, S. H., Schrier, B . K., Farber,J . L., Thompson, E. J., Rosenberg, R. N., Blume, A. J., and Niremberg, M. W. (1972).J. Biol. C/wm. 247, 3159-3169. Woodward, D. J., HoR'er, B. J., Siggins, G . R., and Bloom, F. E. (1971). Brain Re.\. 34, 73-97. Yamamura, H. I., and Snyder, S. H. (l974a).Brain Re.\. 78, 320-326. Yamamura, H. I . , and Snyder, S. H. (1974b).P~.or. Natl. Acad. Sci. U.S.A. 7 1 , 1725-1729. York, D. H . (1972).Brain Res. 37, 91-99. Zettler, C. (1968). Int. J . Nezrro~)liarmacol.7, 325-335.

This Page Intentionally Left Blank

THE FORMATION, DEGRADATION, AND FUNCTION OF CYCLIC NUCLEOTIDES IN THE NERVOUS SYSTEM John W. Daly By John

Laborutory of Chemistry Chemistry Laboratory and Digestive Diseases Diseases National Institute of Arthritis, Metabolism, and Health National Institutes of Health National Bethesdu, Maryland Maryland Bethesda,

....... ........ ................................... .............................. A. Localization of Adenylate Cyclases . . . ............

II.. Introduction 11. Cyclic AMP

B. Regulation o f Adenylate Cyclases . C. Localization of Phosphodiesterases .................................... D. Regulation of Phosphodiesterases . . . . . . ...................... E. Inhibitors and Activators of Phosphodies ...................... F. Cyclic AMP-Dependent Protein Kin G. Phosphoprotein Phosphatases ........................................ ................. H. Cyclic AMP-Dependent Autophosphorylation of Proteins. . . . . . . . . . . . . . . . .......... I 11. Cyclic GMP . . . . . .

A.. Localization of Guanylate Cyclases ........................... B. Regulation of Guanylate Cyclases . . . . . . . . . . . . . . . ........... C. Cyclic GMP Phosphodiesterases ....................................... Kinasa ... .......... ................... D. Cyclic GMP-Dependent Protein Kinases

109

127 128 130 134 I34 135

138 139 142 143

1V. Cyclic Nucleotides and the Function o f the Central and Peripheral References N e r v o u s S y s t e m . . . . . . . . . . . . . . . . . ....................................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1144 44 A. Biochemistry .............................. B. Neurophysiology .................... C. Central Pharmacology, Behavior, and C yclic Nucleotides .............. 152 V. Summary .............................................................. 156 ............... ........ References ........... ........................................ 156

Introduction 1. Introduction

Cyclic nucleotides are formed by the action of adenylate and guanyto late cyclase, which respectively convert ATP to cyclic AMP and GTP to cyclic GMP. Adenylate cyclase is primarily located in the plasma membrane of cells, where its activity in the nervous system is controlled controlled by neurotransmitters and neuromodulators at exoreceptor sites and by GTP, calcium ions, macromolecular regulators, and availability of ATP 105

106

JOHN W. DALY

within the cell. Putative neurotransmitters and neuromodulators which appear to regulate activity of adenylate cyclase include (nor)epinephrine, dopamine, serotonin, histamine, adenosine, prostaglandins, certain peptides, and possibly glutamate-aspartate. GTP at intracellular sites activates the enzyme. Calcium at low concentrations activates adenylate cyclase, while being inhibitory at higher concentrations. Of the macromolecular factors which influence adenylate cyclase, the best defined at present is a calcium-dependent activator protein. Guanylate cyclase appears to be associated with both cell membranes and cytosol. The membrane enzymes are usually inhibited by calcium ions, while the cytosol enzymes are usually activated by calcium. Excitatory neurotransmitters such as acetylcholine and glutamate appear to activate guanylate cyclase in intact cells, but it is as yet unclear whether the activation is direct or due to a transmitter-elicited influx of calcium ions. After formation, the cyclic nucleotides interact with cyclic nucleotide-dependent protein kinases, resulting in a dissociation of an inhibitory regulator unit from the inactive holoenzyme to yield an active catalytic unit. This mechanism has been well established for cyclic AMP-dependent protein kinases and frobably also pertains for cyclic GMP-dependent protein kinases. Cyclic nucleotides have high affinities for binding sites on their respective regulatory units, and thus generation of only small quantities of cyclic AMP or cyclic GMP is presumably sufficient to fully activate the kinases. Phosphorylation of endogenous protein substrates by activated kinases results in alterations of the protein's properties and accounts for the physiological sequelae evoked by cyclic nucleotides. Many endogenous proteins serve as substrates for cyclic nucleotide-dependent kinases, and it has proven difficult to establish the physiological protein substrate or substrates for cyclic AMP or cyclic GMP systems. Metabolic enzymes such as glycogen synthetase and phosphorylase kinase, enzymes concerned with synthesis of neurotransmitters, plasma membrane, and intracellular membrane proteins concerned with control of ionic fluxes, microtubular proteins concerned with neurosecretion and axonal transport, ribosomal proteins, and nuclear proteins suc'h as histones concerned with control of transcription of DNA represent some of possible protein substrates of cyclic AMP-dependent kinases in intact cells of the nervous system. The protein substrates of cyclic GMP-dependent kinases are less well defined. Membrane proteins controlling ionic fluxes and nuclear proteins such as histones are two possibilities. Modulator proteins are present in nervous tissue which can alter the substrate specificities of cyclic nucleotidedependent protein kinases. Calcium can be inhibitory to kinases. Adenosine, AMP, and ADP are possible endogenous inhibitors of cyclic AMP-dependent kinases. Termination of the effect of cyclic nucleotides

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

107

is a dynamic process involving both inactivation of cyclic nucleotides through hydrolysis by phosphodiesterases to 5’-AMP or 5‘-GMP, respectively, and “deactivation” of phosphorylated proteins through hydrolysis by membrane and cytosol phosphoprotein phosphatases. A variety of isozymes of phosphodiesterases exist in neurons and it is as yet unclear to what extent these different isozymes are concerned with hydrolysis of cyclic AMP and cyclic GMP. There are intracellular macromolecular factors which either activate or inhibit phosphodiesterases. Phosphodiesterases and a calcium-dependent activator protein are associated in neurons both with membranes and with the cytosol. Calcium ions can be inhibitory to certain phosphodiesterases. At least in the case of cyclic AMP systems a complex feedback control system serves to regulate formation, degradation, and action of the cyclic nucleotide. Thus, formation of cyclic AMP can lead to deactivation or inhibition of adenylate cyclases, activation of phosphodiesterases, and alteration in the properties of the regulatory nit of protein kinases. Translocation of calcium-dependent activator protein from membrane to cytosol and of kinases from cytosol to cell nucleus appear dependent on cyclic AMP. Many of these feedback control systems appear to require the activation of cyclic AMP-dependent protein kinases. In addition, cyclic AMP appears to directly activate or inhibit certain phosphoprotein phosphatases. The complex regulation of cyclic AMP and cyclic GMP system: is !inked intimately to divalent ions and therefore to apparent cyclic nucleotideelicited translocations of ions. Adenylate cyclases require magnesium ions. Guanylate cyclase is activated by manganese ions, and to a lesser extent by mangesium ions at least in cell-free preparations. Phosphodiesterases and protein kinases require magnesium ions. Calcium ions have inhibitory and stimulatory effects on the various enzymes of the cyclic nucleotide systems (see above). Phosphoprotein phosphatases are the only metal ion-independent enzymes of the cyclic nucleotide systems. ATP subserves the entire system as a substrate for adenylate cyclase and protein kinases, while GTP is both a substrate for guanylate cyclase and a regulatory activator of adenylate cyclase. ATP-regenerating systems in mitochondria and cytosol are of course intimately linked to the function of the cyclic nucleotide systems. A model summarizing and illustrative of the complexity of functional cyclic AMP and cyclic CMP systems is depicted in Fig. 1. The mode is intended only to illustratepossibk pathways for formation, degradation, and action of cyclic nucleotides. During the past two decades, a plethora of data have been amassed concerning the formation, degradation, and actions of cyclic nucleotides, particularly cyclic AMP in the central and peripheral nervous system. This research, documented in over 1200 publications, has been sum-

108

JOHN W. DALY

FIG. 1. A stylized model illustrative of the complexity of functional cyclic nucleotide systems in intact cells of the nervous system. The broken line divides the figure into four arbitrary sections. A, Formation and degradation of cyclic AMP: B, Effects of cyclic AMP; C, Formation and degradation of cyclic GMP: and D, Effects of cyclic GMP. Abbreviations are as follows: cyclic AMP, CAMP;cyclic GMP, cGMP; adenylate cyclase, A. cyclase: guanylate cyclase, G. cyclase: phosphodiesterase, PDE; calcium-dependent phosphodiesterase, PDE I I; calcium-dependent activator protein, AP; cyclic AMP-dependent protein kinase, A. kinase; cyclic GMP-dependent protein kinase, C . kinase; regulatory units for protein kinases, R-unit; stimulatory and inhibitory modulators for protein kinases, stim. modulator and inhib. modulator: phosphoprotein phosphatases, PPP: p h o s p h o r y l a t i o n , m ; activation, --->;inhibition, . . . . ’ >.

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

109

marized in a recent monograph (Daly, 1977). The present chapter will, therefore, attempt only an outline of current research and advances in this field of cyclic nucleotide research. The following areas will be covered: ( 1) the properties of adenylate cyclases, guanylate cyclases, phosphodiesterases, cyclic nucleotide-dependent kinases, and phosphatases as delineated with cell-free preparations and as studied in intact cells of brain slices, ganglia, and cultured brain cells; (2) the possible relationships of cyclic nucleotides to the control of biochemistry and neurophysiology in the intact nervous system; (3) the possible relationships of cyclic nucleotides to the pharmacology of centrally active drugs to behavior and mental dysfunctions. Literature has been surveyed to September, 1976, but the coverage is intended to be selective rather than comprehensive. II. Cyclic AMP

A. LOCALIZATION OF ADENYLATE CYCLASES The high levels of adenylate cyclases in brain, particularly in gray matter of neocortex and cerebellum (Klainer el al., 1962; Sutherland et al., 1962; Weiss and Costa, 1968), and the association of a major portion of the enzyme with synaptosome fractions from brain (De Robertis et al., 1967) strongly suggested important roles for cyclic AMP in the nervous system, presumably related to synaptic transmission. The activation of adenylate cyclases from brain and ganglia by various putative neurotransmitters lent further support to this belief. Synaptosomes, however, consist of both a presynaptic vesicular moiety and an attached postsynaptic membrane, and definitive evidence for presynaptic versus postsynaptic localization of adenylate cyclase has not, as yet, been obtained with brain synaptosome preparations. I n this regard, it should be noted that adenylate cyclase, usually an enzyme of the plasma membrane, has been demonstrated in the membranes from epinephrine-secretory vesicles of the adrenal medulla (Nikodijevic el al., 1976). In contrast to adenylate cyclases from plasma membranes, the cyclase of these secretory vesicles was inhibited by P-adrenergic agonists. Whether adenylate cyclases are associated with presynaptic noradrenergic vesicles and other neurotransmitter vesicles of central neurons is unknown. Recent studies have employed sucrose gradients to assess to what extent adenylate cyclases are associated with the synaptosome fractions which exhibit active uptake mechanisms for catecholamines. Active uptake mechanisms are, of course, thought to serve as a marker for the

110

JOHN W. DALY

presynaptic entities. With synaptosomes from rat cerebral cortex, a significant portion of the adenylate cyclase activity was not associated with fractions exhibiting marked uptake of radioactive norepinephrine (Davis and Lefkowitz, 1976). High levels of binding sites for the /I-adrenergic antagonist dihydroalprenolol were associated almost entirely with the fractions which accumulated norepinephrine. It is interesting that these ‘‘&receptors,” normally thought to be associated with adenylate cyclase, were not associated with the major portion of adenylate cyclase in synaptosome fractions. An activation of adenylate cyclase by catecholamines was stated to be undetectable in these synaptosome preparations. With synaptosomes from rat striatum, the distribution of cyclase activity did not exactly correspond to the fractions which accumulated radioactive dopamine (Sieghart et al., 1976). Basal adenylate cyclase appeared on these sucrose gradients as one broad peak, while dopamine-stimulated cyclase activity appeared to represent two broad peaks. These studies, while not definitive, are at least consonant with significant postsynaptic rather than only presynaptic localization of adenylate cyclase in synaptosome preparations. Accumulations of cyclic AMP elicited by norepinephrine in synaptosomes appeared in one study to be derived from endogenous ATP within the synaptosome (Harris, 1976), a result consonant with a presynaptic generation of cyclic AMP. Intracellular ATP of brain slices was labeled by incubation with radioactive adenine, followed by homogenization, isolation of synaptosomes, and measurement of catecholamineelicited generation of cyclic AMP from “intrasynaptosomal” ATP. Norepinephrine appeared to be activating a /3-adrenergic receptorlinked cyclase. Other studies on adenylate cyclases associated with synaptosomes have been based on assays with exogenous ATP, a technique which will measure only the cyclases which have catalytic sites accessible to the medium. Synaptosomes are normally isolated after homogenization of brain tissue in isotonic sucrose. If, instead, brain tissue is homogenized in physiological medium, vesicular entities were formed whose adenylate cyclases appeared to accept as substrate not exogenous ATP, but instead the intravesicular ATP (Chasin et al., 1974; Shimizu et al., 1975a; Horn and Phillipson, 1976). The major morphological components of these homogenates were 100-800 nm vesicular membrane fragments, but in addition larger synaptosomelike entities were present whose postsynaptic elements when present had formed a vesicular structure. The responses of the cyclic AMP-generating systems in these cell-free preparations to neurotransmitters were remarkably similar to responses of cyclic AMP systems in brain slices. Responses of adenylate cyclases in synaptosomes isolated from sucrose homogenates have in contrast often

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

111

lost or greatly reduced responsiveness to neurotransmitters (cf. Daly, 1977). The morphological localization of adenylate cyclase and of cyclic AMP in tissue slices from brain has been studied to a limited extent. In rat cerebral cortex, a histochemical assay demonstrated aden ylate cyclase activity at a limited number of synapses on plasma membranes of astrocytes and associated with capillaries goo et al., 1975).The validity of such assays for adenylate cyclase has been questioned (Lemay and Jarett, 1975). Lead ions used in many such assays are potent inhibitors of adenylate cyclase (Nathanson and Bloom, 1975, 1976). Catecholaminesensitive adenylate cyclase was stated to have been detected histochemically at presynaptic sites in cerebral cortex and caudate nucleus (cited in Hervonen and Rechardt, 1976). In rat cerebellum, high levels of cyclic AMP were detected by immunofluorescent assay in Purkinje neurons and granule neurons (Bloom et al., 1972; Siggins et al., 1973). The weak fluorescence in the cerebellar molecular layer was associated with the Purkinje cell dendrites. Norepinephrine greatly increased the fluorescence of Purkinje cells. In view of these results, it is perhaps surprising that the basal levels of cyclic AMP in cerebellum were not lower in a mutant strain of mice, the nervous mouse, in which Purkinje cells are virtually absent (Ma0 et al., 1975). Levels of cyclic AMP were slightly higher in the granular layer than in the molecular layer of Swiss-Webster mice (Rubin and Ferrendelli, 1976). The generation of cyclic AMP has been studied extensively in incubated slices of brain tissue, but little has been learned from these studies as to the morphological sites at which cyclic AMP is formed (cf. Daly, 1977). Basal levels of cyclic AMP were similar in incubated cerebellar slices from control and X-irradiated rats (Hoffer et al., 1976). Thus, elimination of neurons of the granular layer by neonatal X-irradiation had little effect on basal levels of cyclic AMP. However, the accumulation of cyclic AMP elicited by norepinephrine was markedly reduced in slices from X-irradiated rats. The data were indicative of the presence of cyclic AMP systems controlled by P-adrenergic receptors both in Purkinje cells and in neurons of the granular layer of cerebellum. Another aspect of brain slice studies is relevant to the localization of cyclic AMP systems. Thus, adenine phosphoribosyltransferase, the enzyme responsible for incorporation of adenine into intracellular adenine nucleotides, was closely associated with cyclic AMP-generating systems in brain slices (cf. Daly, 1977). Adenosine incorporation into adenine nucleotides was much less closely associated with the cyclic AMP compartments. Labeling of synaptosomes, presumably the presynaptic elements, occurred more rapidly with adenosine than with adenine (Kuroda and Mcllwain, 1974). Such data provide further evidence for a postsynaptic

112

JOHN W. DALY

rather than presynaptic localization of cyclic AMP systems in brain tissue. Destruction of central noradrenergic presynaptic terminals with 6-hydroxydopamine had either no effect or resulted in an apparent adaptive increase in responses of norepinephrine-sensitive cyclic AMPgenerating systems in brain slices (Huang et al., 1973b; Kalisker et al., 1973; Dismukes and Daly, 1975b; Dismukes et al., 1976b; Skolnick and Daly, 1976a, 1977). Similar results pertain after destruction of ascending noradrenergic fibers by lesions of the medial forebrain bundle (Dismukes et al., 1975, 1976b). Levels of adenylate cyclase in homogenates of rat cerebral cortex were only marginally decreased by the B-hydroxydopamine treatment (Kalisker et al., 1973). Electrolytic or 6-hydroxydopamine lesions of the nigrostriatal dopaminergic pathway had no effect or apparently resulted in an increase in dopamine-sensitive adenylate cyclases in striatum (Krueger et al., 1976; Mishra et al., 1974; Von Voightlander et al., 1973). The lack of change or, in some instances, an apparent adaptive increase in catecholamine-sensitive cyclases after destruction of presynaptic terminals, is at least consonant with postsynaptic loci for the cyclic AMP systems. In a region of the substantia nigra containing dendrites of the dopaminergic nigrostriatal neurons, the dopamine-sensitive cyclase activity was unaffected by 6-hydroxydopamine-induced destruction of the dopaminergic neurons (Kebabian and Saavedra, 1976). Thus, the dopamine-sensitive cyclases must be associated with some other cell type in this brain area. One further aspect of sites of generation of cyclic AMP in the nervous system needs to be considered. Brain consists mainly of neurons and glia. The latter cells, the astrocytes and the oligodendroglia are more numerous than neurons and comprise about one-half of the total brain tissue. A variety of data with cultured glioma (astrocytoma) and neuroblastoma cells suggested that catecholamine-sensitive adenylate cyclases might, in brain, be associated primarily with astrocytes, not neurons. Indeed norepinephrine elicited a profound accumulation of cyclic AMP in cultures of fetal brain cells in which extensive proliferation of cells, presumably glial cells, had occurred (Gilman and Schrier, 1972). Norepinephrine elicited a much lower accumulation of cyclic AMP in reaggregation cultures in which proportions of neurons and glia are probably still similar to those of fetal brain (Seeds and Gilman, 1971). Studies with brain tissue have not resolved the neuron-glia question. It would appear likely that neurotransmitter-sensitive adenylate cyclase will be found to have roles in physiological regulation in both neurons and glia. Adenylate cyclase has been detected in homogenates of both neuron and gliaenriched fractions from rat brain (Palmer, 1973). Cyclase activity was lower in the homogenates from glia-enriched fractions. Norepinephrine,

CYCLIC NUCLEOTIDES I N THE NERVOUS SYSTEM

113

dopamine, and histamine-sensitive cyclases were reported from the neuron and glia-enriched preparations (Palmer, 1973, 1976; Palmer and Manian, 1976; Spiker et al., 1976). In cultured rat superior cervical ganglia, basal levels and responses of cyclic AMP-generating systems to isoproterenol were found to decrease markedly during culture (Cramer et al., 1973). In view of survival of postganglionic noradrenergic cell bodies and degeneration of interstitial cells and presynaptic terminals during culture, it was proposed that the P-adrenergic cyclases were associated in part with ganglionic glial cells. However, in other studies on superior cervical ganglion, the in vivo increase in cyclic AMP elicited by isoproterenol was found to be virtually lost when the noradrenergic cell bodies had been destroyed by prior administration of nerve growth factor or 6-hydroxydopamine (Otten et al., 1974). These treatments had little effect on basal levels of cyclic AMP. Dopamine-elicited accumulations of cyclic AMP in bovine superior cervical ganglion were subsesquently shown by immunofluorescent assay to occur primarily in the dendrites and cell bodies of the noradrenergic neurons (Kebabian et al., 1975a). Norepinephrine-elicited accumulations of cyclic AMP occurred in the postganglionic noradrenergic neurons, but also occurred in fibroblast and blood vessellike cells. In the superior cervical ganglion as in brain slices, radioactive adenine selectively labeled ATP compartments associated with cyclic AMP-generating systems (Lindl et al., 1975). In lumbar sympathetic ganglia from chick, norepinephrine (0.1 mM)- and dopamine (3 mM)-sensitive adenylate cyclase activity appeared by histochemical assay to be localized at both postsynaptic dendrites of the sympathetic neurons and at presynaptic “aminergic” nerve terminals (Hervonen and Rechardt, 1976). Isoproterenol has been reported to increase levels of cyclic AMP in desheathed frog sciatic nerves (Horn and McAfee, 1976). Accumulation of particulate adenylate cyclases, proximal to a constriction of chicken sciatic nerve, was indicative of significant axonal transport of the enzyme to distal cholinergic terminals (Bray et al., 1971). In summary, the majority of data on localization of cyclic AMPgenerating systems in nervous tissue is consonant with a postsynaptic localization, probably primarily in neurons but perhaps to some extent in glia. Presynaptic cyclic AMP-generating systems are probably present in certain types of neurons.

B. REGULATION OF ADENYLATE CYCLASES Cyclic AMP-generating systems in brain tissue appear important to central homeostatic mechanisms as evident in the adaptive changes in

114

JOHN W. DALY

responsiveness of these systems which occur as a result of alterations in transsynaptic input of specific neurotransmitters (cf. Dismukes and Daly, 1976b, 1977). Such adaptation could involve changes in adenylate cyclases, phosphodiesterases, and cyclic AMP-dependent protein kinases, but such alterations are studied with difficulty in heterogeneous brain tissue. In pineal gland, during the day when noradrenergic input is low, levels of /3-adrenergic-sensitive adenylate cyclases were found to have undergone a compensatory increase, while during the night when noradrenergic input is elevated, P-adrenergic-sensitive cyclases had undergone a compensatory decrease (cf. Romero and Axelrod, 1975; Kebabian et al., 1975~).Isoproterenol-elicited decreases in pineal gland adenylate cyclase did not appear to involve protein synthesis. Adaptive changes in phosphodiesterases and protein kinases also occurred in pineal gland. The adaptive changes in adenylate cyclase activity which occurs in cultured neuroma cells on exposure to neurotransmitters which elevate cyclic AMP levels appeared in certain cell lines to involve cyclic AMP-elicited synthesis of a protein which inhibits adenylate cyclase (De Vellis and Brooker, 1974). In another cell line cyclic AMP-dependent reductions of cyclase activity did not require protein synthesis (Browning et al., 1976). Stimulation of cyclic AMP-generating systems of glioma cells by norepinephrine or prostaglandin resulted in specific rather than general reductions in responsiveness of adenylate cyclases (Perkins et al., 1975). 1. Norepinephrine

T h e effect of norepinephrine on formation of cyclic AMP has been studied extensively with brain slice preparations. Unfortunately, adenylate cyclases assayed with exogenous ATP in homogenates of brain retains little of the norepinephrine-mediated regulation of activity seen in brain slice preparations (cf. Daly, 1977). Certain investigators, most notably Von Hungen, Roberts, and co-workers, have been able to obtain small but reproducible amine-elicited stimulations of adenylate cyclase in brain homogenates through the use of EGTA-inhibited preparations (cf. Von Hungen and Roberts, 1974). Recently, the stimulation of asolubilired adenylate cyclase from bovine brain by norepinephrine and dopamine, but not by epinephrine or serotonin, was reported (Stellwagen and Baker, 1976). Norepinephrine-sensitive cyclic AMP-generating systems have been studied in cell-free vesicular preparations from guinea pig cerebral cortex and cerebellum (Chasin et al., 1974; Shimizu et al., 1975a) and from rat limbic forebrain (Horn and Phillipson, 1976). In the cortical preparations, the response to epinephrine was blocked by an a-adrenergic antagonist, while in the cerebellar preparations it was

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

115

blocked by a P-adrenergic antagonist. In rat limbic forebrain, the response to norepinephrine was blocked by &antagonists and partially blocked by various phenothiazines and other antipsychotics such as clozapine and haloperidol. Thus, the nature of the adrenergic receptors modulating cyclic AMP generation in these cell-free preparations was quite consonant with data from brain slice preparations (see below). With slices from various brain regions, (nor) epinephrine reproducibly elicits a marked activation of cyclic AMP-generating systems. The nature of the noradrenergic receptors regulating formation of cyclic AMP differs greatly among brain regions and between species. Thus, in rat cerebral cortex, the response has been characterized as due to a mix of a- and P-adrenergic receptors (Perkins and Moore, 1973; Skolnick and Daly, 1976b), while in rat cerebellum only P-adrenergic receptors pertain (Skolnick et al., 1976; Schwabe and Daly, 1977). In guinea pig cerebral cortex, in contrast to rat, the activation of cyclic AMPgenerating systems by norepinephrine involved virtually only a-adrenergic receptors (Chasin et al., 1971, 1973; Sattin et al., 1975). In guinea pig cerebellum only P-adrenergic receptors pertained. In mice the nature of responses again differed, involving primarily P-adrenergic receptors in cerebral cortex (Schultz and Daly, 1973d) and a mix of aand /3-adrenergic receptors in cerebellum (Ferrendelli et al., 1975). Thus, in different brain regions and different species, the interaction of adrenergic agonists and antagonists with the function of norepinephrineresponsive cyclic AMP systems will differ markedly. I t follows that in z~ivo differences in the effects of such drugs might provide insights into the functional role of norepinephrine-sensitive cyclic AMP systems. The marked synergism between norepinephrine and adenosine with respect to activation of cyclic AMP-generating systems in brain slices appeared to involve mainly a-adrenergic receptors (Perkins and Moore, 1973; Schultz and Daly, 1973c,d; Perkins et al., 1975; Sattin et al., 1975; Skolnick and Daly, 1975b). The synergism was particularly striking in guinea pig cerebral cortical slices where a-adrenergic receptors predominate and where norepinephrine has virtually no effect on cyclic AMP levels except in the presence of adenosine. In rat cortical slices, a-adrenergic receptors appeared involved to a greater extent than were P-adrenergic receptors in the synergism between norepinephrine and adenosine, while in mouse cerebral cortical slices where P-adrenergic receptors predominate, there was no clear synergism between norepinephrine and adenosine. In rat caudate slices, 2-chloroadenosine and isoproterenol had greater than additive effects on cyclic AMP levels (Wilkening and Makman, 1975). Synergisms between norepinephrine and histamine were quite pronounced in guinea pig cerebral cortical

116

JOHN W. DALY

slices (Huang et al., 1971, 1973a; Chasin et al., 1973; Schultz and Daly, 1973a) but were minimal in rabbit (Kakiuchi and Rall, 1968a) and rat (Huang et al., 1971; Palmer et al., 1973; Schultz and Daly, 1973d; French et al., 1975) cerebral cortical slices. It appeared that a histamine-elicited release of adenosine might be partially responsible for the synergism in guinea pig cortical slices. The physiological significance of the synergisms between adenosine and biogenic amines and between norepinephrine and histamine is unknown. Norepinephrine, based on studies with brain slices, can be expected to elicit accumulations of cyclic AMP in brain via interaction with both aand P-adrenergic receptors, methoxamine via interaction with a-receptors (Skolnick and Daly, 1975b), and isoproterenol via interaction with P-receptors. The a-component of norepinephrine responses is blocked by a-antagonists such as phentolamine, phenoxybenzamine, ergot alkaloids, and clonidine (Skolnick and Daly, 1975c, 1976a), while the ,&component is blocked by P-antagonists such as alprenolol, propranolol, dichlorisoproterenol, and sotalol (Skolnick and Daly, 197613). Fluphenazine appeared capable of blocking the P-adrenergic receptors associated with cerebellar Purkinje cells (Skolnick et al., 1976; Hoffer et al., 1976). Antipsychotic drufs such as clozapine, chlorpromazine, and haloper idol partially blocked the response of P-adrenergic-controlled cyclic AMP systems to norepinephrine in slices from limbic forebrain (Blumberg et al., 1976). In brain slices, dopamine-elicited accumulations of cyclic AMP were minimal so that it appeared unlikely that a significant portion of the norepinephrine response was due to activation of dopaminergic receptors. In bovine superior cervical ganglia, norepinephrine elicited accumulations of cyclic AMP via interaction with both a P-adrenergic receptor and via interaction with what appeared to be a dopaminergic receptor (Kebabian and Greengard, 1971; Kalix et al., 1974). In rat superior cervical ganglia, dopamine was relatively ineffective and norepinephrine appeared to elicit accumulations of cyclic AMP via interaction with aand P-adrenergic receptors (Lindl and Cramer, 1975).

2 . Dopamine In contrast to norepinephrine, the effects of dopamine on formation of cyclic AMP are studied most satisfactorily with brain homogenates rather than in brain slices. In homogenates or with synaptosomes, dopamine elicited a small but reproducible activation of adenylate cyclases which was unaffected by P-adrenergic antagonists and was blocked by antipsychotics such as ffuphenazine and haloperidol (cf, Iversen, 1975; Kebabian et al., 1975d). Dopamine-sensitive adenylate cyclases

CYCLIC NUCLEOTIDES I N THE NERVOUS SYSTEM

117

have been studied primarily in homogenates of striatum (caudate nucleus), but have also been studied in homogenates from substantia nigra (Kebabian and Saavedra, 1976; Phillipson and Horn, 1976), amygdala (Racagni and Carenzi, 1976; Weinryb and Michel, 1976), olfactory tubercle, and nucleus accumbens (Kebabian et al., 1972, 1975d; Clement-Cormier et al., 1974; Horn et al., 1974; Miller et al., 1974; Carenzi et al., 1975; Mishra et al., 1975; Weinryb and Michel, 1976). No pronounced differences in the potencies of dopamine antagonists in different brain regions have emerged from these studies. Thus, these studies have not clarified the factors involved in antipsychotic, extrapyramidal, and endocrinological effects of dopamine antagonists. Antipsychotic effects have been proposed to relate to antagonism of dopaminergic receptors in the mesolimbic system (olfactory tubercle, nucleus accumbens), extrapyramidal side effects to antagonism of receptors in striatum (caudate nucleus), and endocrinological side effects to antagonism of receptors in the median eminence of the hypothalamus. The relatively low potency of haloperidol and other butyrophenones as antagonists of dopamine-sensitive cyclases is also puzzling in view of their high potency as antipsychotics (cf. Laduron, 1976). In preparations from brain regions other than the limbic system, dopamine-elicited activations of adenylate cyclases have not been well characterized. Based on studies with brain homogenates dopamine, apomorphine, 2-amino-l,2,3,4-tetrahydronaphthalene, and probably lysergic acid diethylamide (Spano et al., 1975a; Von Hungen et al., 1975) can be expected to elicit accumulations of cyclic AMP in brain via interaction with dopaminergic receptors. Phenothiazines such as fluphenazine, trifluoperazine, chlorpromazine, and thioridazine, thioxanthenes such as flupenthixol and chlorprothixene, butyrophenones such as droperidol and haloperidol, and certain other antipsychotic agents such as butaclamol, pimozide, and clozapine block these dopaminergic receptors. The reason underlying the lack of responsiveness of dopaminesensitive cyclic AMP-generating systems in brain slices is not apparent. Dopamine did elicit small accumulations of cyclic AMP in slices of rat caudate nucleus (Fornetal., 1974; Kruegeretal., 1976) and in slices from cerebral cortex (Dismukes and Daly, 1974; Harris, 1976), but the presence of high concentrations of a phosphodiesterase inhibitor was required. In caudate slices, the dopamine response was blocked by fluphenazine, while the response in cortical slices has not been evaluated and might represent a partial activation of adrenergic receptors by dopamine. Adenosine did not appear to potentiate dopamine responses in rat (Schwabe, unpublished results) or guinea pig (Shimizu et al., 1970; Sattin et al., 1975) cerebral cortical slices. Dopamine in recent studies has

118

J OHN W. DALY

been reported to elicit small accumulations of cyclic AMP in rat caudate (Wilkening and Makman, 1975) and mouse and rat cortical (Martres et al., 1975; Schwabe and Daly, 1977) slices in the absence of a phosphodiesterase inhibitor. I n adenine-labeled striatal synaptosomes where endogenous rather than exogenous ATP is serving as substrate for adenylate cyclase, dopamine was ineffective and formation of cyclic AMP appeared activated by interaction of catecholamines with a P-adrenergic receptor (Harris, 1976). In a preliminary report on dopamine and p-adrenergic-sensitive cyclic AMP-generating systems in homogenates from rat caudate nucleus, it was proposed that the dopamine-sensitive compartment lacked endogenous ATP and required lysis by hypotonic media to permit exogenous ATP access to catalytic sites of adenylate cyclase (Sheppard and Burghardt, 1976). The p-adrenergic-sensitive compartment was proposed to contain sufficient endogenous ATP which was lost on lysis and could not be effectively replaced by exogenous ATP. Whether similar lack of ATP in intact cells is responsible for minimal responses of dopamine-sensitive cyclic AMP-generating systems in caudate slices appears somewhat unlikely. However, in an intracellular cyclic AMP system with high levels of phosphodiesterases and low ATPgenerating capacity, high turnover of cyclic AMP after activation of cyclases by dopamine might result in depletion of ATP. The lack of responses to dopamine in terms of generation of radioactive cyclic AMP in adenine-labeled slices of rat striatum suggests that adenine did not significantly label the dopamine-sensitive compartment (Harris, i976). P-Adrenergic agonists did elicit generation of radioactive cyclic AMP in striatal slices. In bovine and rabbit superior cervical ganglion, dopamine activated cyclic AMP-generating systems via interaction with receptors antagonized both by a-adrenergic antagonists, and less effectively by central dopaminergic antagonists such as chlorpromazine and haloperidol (Kebabian and Greengard, 1971; Kalix et al., 1974; Roch and Kalix, 1975a,b),while in rat superior cervical ganglion, convincing evidence for the presence of dopamine-sensitive cyclases has not been obtained (cf. Lindl and Cramer, 1975). 3. Serotonin

Stimulations of cyclic AMP-generating systems by serotonin in homogenates from brain tissue and in brain slices have usually been marginal. Responses of adenylate cyclases to serotonin were reported in EGTA-inhibited homogenate preparations from rat hippocampus, anterior and posterior colliculi, midbrain, and hypothalamus (Von Hungen

CYCLIC NUCLEOTIDES I N T H E N E R V O U S SYSTEM

I19

and Koberts, 1974). Methysergide antagonized the serotonin response. Serotonin has been reported to elicit small, often marginal accumulations of cyclic AMP in brain slices from rabbit cerebral cortex and cerebellum (Kakiuchi and Rall, 1968a,b), guinea pig cerebral cortex (Chasin et al., 197l), hippocampus, amygdala, diencephalon, and brain stem (Chasin et al., 1973), mouse cerebral cortex (Martres et al., 1975), monkey polysensory cortex (Skolnick et al., 1973), and human cerebral cortex (Shimizu el al., 1971; Kodama et al., 1973). In rat cerebral cortical slices serotonin had only marginal effects on cyclic AMP levels even in the presence of phosphodiesterase inhibitors (Dismukes and Daly, 1974: Dismukes et al., 1975; French et ad., 1975). Many of these studies have been with adenine-labeled slices, and it is possible that adenine does not effectively label serotonin-sensitive compartments. It is also possible that high phosphodiesterase activity or low activity of ATP-regenerating systems associated with serotonin-sensitive compartments make difficult the detection of serotonin responses in brain slices. Serotonin, although having no effect in cerebral cortical slices from control rats, did elicit a small response in slices from rats during withdrawal from chronic ethanol treatment (French et al., 1975). The response to serotonin was, however, blocked by methysergide, phenoxybenzamine, or propranolol. The apparent nonspecific nature of the response casts doubts on whether it really represents a specific activation of a serotonin-sensitive cyclic AMP system. Other marginal responses to serotonin may also represent nonspecific or indirect activation of cyclic AMP systems. Further studies on serotonin responses in brain tissue are clearly required before any speculations as to central function of serotonin-sensitive cyclic AMP systems are warranted. In guinea pig cerebral cortical slices, serotonin had, at best, marginal effects on cyclic AMP-generating systems except in the presence of exogenous adenosine (Shimizu et al., 1970; Huang et al., 1971; Huang and Daly, 1972; Schultz and Daly, 1973a,c; Dismukes et al., 197613). The synergism between adenosine and serotonin was antagonized in guinea pig cortical slices by methysergide. No synergism between serotonin and low concentrations of adenosine was reported in one study with guinea pig cerebral cortical slices (Sattin et al., 1975). In rat and mouse cortical slices, combinations of serotonin and adenosine had effects on accumulations of cyclic AMP no greater than that elicited by adenosine alone (Huang et al., 1973b: Schultz and Daly, 1973d; Skolnick and Daly, 1974b). Serotonin had no effect on cyclic AMP levels in cultured cells from fetal rat brain (Gilman and Schrier, 1972). Effects of serotonin on cyclic AMP levels in sympathetic ganglia of mammals have not apparently been investigated. Serotonin did increase cyclic AMP levels

120

JOHN

W. DALY

in the abdominal ganglion of the mollusc Aplysia califomica (Cedar and Schwartz, 1972; Cedar et al., 1972; Levitan et al., 1974) and in cockroach thoracic ganglion (Nathanson and Greengard, 1973, 1974). 4. Histamine T h e stimulation of cyclic AMP generation in brain tissue by histamine has been studied extensively in brain slices and more recently in homogenates. Histamine activated adenylate cyclases in homogenates from guinea pig neocortex, hippocampus, and striatum (Hegstrand et al., 1976). The 2-fold stimulation in hippocampal preparations was antagonized by an Hz-antagonist, metiamide, but not by an HI-antagonist, pyrilamine. Histamine elicited a small increase in adenylate cyclase activity in preparations from rabbit cortex and hippocampus (Palmer, 1973; Spiker et al., 1976). Histamine has no, or only marginal effects, on cyclases from rat brain (Burkard and Gey, 1968; Von Hungen and Roberts, 1973a,b; De Belleroche et al., 1974; Izumi et al., 1975a; Hegstrand et al., 1976) or on cyclases from monkey hippocampus (Weinryb and Michel, 1976). I n brain slices, the magnitude of responses to histamine had been largely predictive of the recent results with brain homogenates. Thus, histamine elicited large accumulations of cyclic AMP in slices from rabbit cerebral cortex (Kakiuchi and Rall, 1968a) and in slices from guinea pig cerebral cortex and hippocampus (Chasin et al., 1973; Rogers et al., 1975), while in slices from rat and mouse cerebral cortex or hippocampus minimal responses pertained (Krishna et al., 1970; Schultz and Daly, 1973d; Skolnick and Daly, 1974b, 1975a; Dismukeset al., 1975).Characterization of the nature of the histaminergic receptors regulating cyclic AMP generation in brain tissue is far from complete. However, studies with H1-agonists such as 2-methylhistamine and 2-aminoethylthiazole, with the He-agonist 4-methylhistamine, with HI-antagonists such as brompheniramine, pyrilamine, and diphenhydramine, and with H2antagonists such as metiamide provided evidence that both HI- and H2histaminergic receptors are involved in activation of cyclic AMP systems in brain tissue. In guinea pig cerebral cortical and hippocampal slices a mix of HI- and H2-receptors appeared involved in activation of cyclic AMP systems (Baudry et al., 1975; Rogers et al., 1975; Dismukes et al., 1976a). In rat cortical slices, the results were less conclusive because of the magnitude of the small response. H2-receptors, however, appeared to be primarily involved (Dismukes et al., 1975). Histamine responses in chick cortical slices involved mainly Hz-receptors (Nahorski, et al., 1974). Responses to histamine in other species or brain regions have not been well characterized. Histamine had no effect on cyclic AMP levels in cultured cells from fetal rat brain (Gilman and Schrier, 1972). The syner-

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

121

gism between histamine and adenosine in guinea pig cortical slices appeared to involve mainly HI-receptors (Dismukes et al., 1976a). The apparent proportion of HI- and Hz-receptor contributions to histamine responses will, therefore, be strongly dependent on endogenous levels of adenosine in the brain slice preparations. Recently, clonidine (EC,,, 50 1.1M ) has been proposed to elicit accumulations of cyclic AMP in guinea pig hippocampal slices via interaction with an Hz-receptor (Audiger et al., 1976). In chick cortical slices histamine elicits a large accumulation of cyclic AMP via interaction with an Hz-receptor, but clonidine had no effect (Nahorski et al., 1975b). In rat superior cervical ganglion, histamine activated cyclic AMPgenerating systems via interaction with a receptor that was antagonized by both HI- and Hz-antagonists (Lindl and Cramer, 1974). In bovine superior cervical ganglion, histamine-elicited accumulations of cyclic were antagonized by an Hz-antagonist (Kebabian et al., 1975b; Roch and Kalix, 1975a). 5. Adenosine

Activation of cyclic AMP-generating systems by adenosine occurs, apparently at an extracellular site, in brain slices from a variety of species and brain regions (cf. Mah and Daly, 1976; Daly, 1977). Theophylline, isobutylmethylxanthine, caffeine, and certain adenosine analogs such as 2'-deoxyadenosine, act as adenosine antagonists in brain slices. Other adenosine analogs such as N"-phenylisopropyladenosineand 2-chloroadenosine are agonists. In addition to stimulatory effects on cyclic AMP-generating systems, adenosine has striking effects on the responsiveness of amine-sensitive systems. Synergistic interactions of adenosine with norepinephrine, serotonin, histamine (see above), and glutamate (Shimizu et al., 1974, 1975b) have been reported. In addition, adenosine has been reported to prevent refractoriness of amine-sensitive cyclic AMP-generating systems in guinea pig cerebral cortical slices (Schultz and Daly, 1973b; Schultz, 1975a,b). Adenosine appeared to be involved as an intermediary in the activation of cyclic AMP-generating systems in brain slices by depolarizing agents and certain metabolic inhibitors (cf. Daly, 1977). Although firmly established as an important factor in the regulation of cyclic AMP generation in brain slices, activation of adenylate cyclases by adenosine has as yet not been demonstrated in homogenates of brain tissue prepared in isotonic sucrose (Sattin and Rall, 1970; McKenzie and Bar, 1973). It should be noted that exogenous ATP employed for assay of adenylate cyclase in such preparations will undergo enzymatic conversion to adenosine. Basal levels of adenylate cyclase in homogenates of rat

122

JOHN W. DALY

caudate nucleus were indeed lower when assayed in the presence of the adenosine antagonists theophylline and isobutylmethylxanthine, than when assayed in the presence of other phosphodiesterase inhibitors which are not adenosine antagonists such as papaverine or dipyridamole (Fredholm et al., 1976). Theophylline reduced levels of cyclic AMP in rat cerebral cortical synaptosornes (De Belleroche et al., 1974). It should be noted that adenosine is converted to an inactive metabolite inosine by the action of the enzyme adenosine deaminase. This enzyme occurs in various brain regions (Sun et al., 1976) and has been used for the study of adenosine-dependent activation of cyclic AMP systems in brain slices (Huang et al., 1973a; Schwabe et al., 1977). Adenosine has been demonstrated to activate adenylate cyclases in homogenates of blood platelets ( Haslarn and Lynham, 1972), cultured neuroblastoma (Blume and Foster, 1975, 1976a,b; Penit et al., 1976), and glioma (Clark et al., 1975; Perkins et al., 1975; Clark and Seney, 1976) cells, and undoubtedly conditions will be found under which activation of adenylate cyclases of synaptosome preparations by adenosine can be demonstrated. Indeed, in the vesicular preparations obtained after homogenization of guinea pig cerebral cortex in physiological medium, adenosine (ECoo 10 p M ) elicited a marked stimulation of cyclic AMP-generating systems (Chasin et al., 1974; Shimizu etal., 1975a). In cultured cells from fetal rat brain, adenosine elicited accumulations of cyclic AMP (Gilman and Schrier, 1972; Sturgill et al., 1975). Unlike results with brain slices, combinations of catecholamines and adenosine did not have synergistic effects on cyclic AMP generation in cultured cells. In rat superior cervical ganglia, effects of adenosine on cyclic AMP levels were not detected (Roch and Kalix, 1975b). Furthermore, unlike results with brain slices, theophylline did not antagonize, but was instead required in order for depolarizing agents to elicit accumulations of cyclic AMP (Kalix and Roch, 1975; Roch and Kalix, 1975b; Webbetal., 1975). 6. Acetylcholine Acetylcholine had no effect or inhibited the activity of adenylate cyclases from rat brain (Von Hungen and Roberts, 1973a,b; Duffy and Powell, 1975) and rat striatum (Walker and Walker, 1973). In brain slices, acetylcholine or cholinergic agonists had virtually no effect on levels of cyclic AMP (cf. Daly, 1977), although in some instances a slight reduction pertained in the presence of acetylcholine (Kuo et al., 1972). In bovine superior cervical ganglion, acetylcholine and cholinergic agonists enhanced the levels of cyclic AMP (McAfee et al., 1971; Greengard et al., 1972; Kalix et ad., 1974). The mechanism appeared to involve an acetylcholine-induced release of doparnine from ganglionic

CYCLIC NUCLEOTIDES IN THE NERVOUS SYSTEM

123

interneurons and a resultant stimulation of cyclic AMP-generating systems in postganglionic noradrenergic cell bodies by the released dopamine. Acetylcholine, however, slightly inhibited the accumulation of cyclic AMP elicited by exogenous dopamine (Kebabian et al., 1975b). In rat superior cervical ganglion acetylcholine had no significant effect on levels of cyclic AMP (Cramer and Lindl, 1974). Acetylcholine had no effect on levels of cyclic AMP in peripheral neurons (Kebabian et nf., 1975b). 7. Prostaglandins Prostaglandin-sensitive cyclic AMP-generating systems have been demonstrated in homogenates and slices of brain tissue, but quite high concentrations have been required to elicit significant responses. Early studies with the low concentrations of prostaglandins, which are effective in other tissues, had no effect in brain preparations (Schmidt et al., 1970; Robison et al., 1970; Zanella and Rall, 1973). In recent years stimulation of adenylate cyclases in rat brain homogenates by prostaglandins of the E series has been reported (Collier and Roy, 1974a,b; D ~ f f yand Powell, 1975). Concentrations of prostaglandin from 2.5 to 150 CLMwere employed. Morphineand other narcotic analgesics were effective antagonists of prostaglandin-elicited activation of brain adenylate cyclases. Other groups, however, have been unable under a variety of conditions to demonstrate significant activation of adenylate cyclases by prostaglandins in homogenates of rat brain, cerebral cortex, or caudate nucleus (Van Inwegen et al., 1975; Tell et nl., 1975). Similarly, in brain slices reports on prostaglandin-sensitive cyclic AMP systems have not been definitive. Very high concentrations of prostaglandins of the E series (EC,, 20 P M ) did elicit significant accumulations of cyclic AMP in slices from rat cerebral cortex (Berti et al., 1972; Kuehl et al., 1972; Dismukes and Daly, 1975a). Prostaglandin antagonists such as 7-oxa- 13-prostynoic acid and the dibenzooxazepine hydrazine, SC 19220, had no effect of responses to prostaglandin E, in rat cortical slices. Morphine slightly potatiated the prostaglandin response, in contrast to its inhibitory effects in brain homogenates (Collier and Roy, 1974a,b) and cultured cells (cf. Sharma et al., 1975). Tentative evidence for partial antagonism of’ norepinephrine and isoproterenol-elicited accumulations of cyclic AMP by prostaglandin El in rat cortical slices has been reported (Dismukes and Daly, 1975a). However, in guinea pig cerebellar slices, prostaglandin E, did not antagonize the norepinephrine response (Ohga and Daly, 1977a). In view of the lack of potency of prostaglandins in brain preparations, the relevance of these efTects to physiological roles of prostaglandin in the central nervous system must be subjected to further investigation. Even at high concentrations, prostaglandin El had no apparent

124

JOHN W. DALY

effect on cyclic AMP levels in cerebral cortical slices from species other than rat (Berti et al., 1972). Prostaglandin El ( E G O< 3 p M ) elicited large accumulations of cyclic AMP in cultured cells from fetal rat cerebral cortex (Gilman and Schrier, 1972). Effects of prostaglandin on cyclic AMP-generating systems in ganglia have apparently not been studied. 8. Amino Acids

The status of putative amino acid neurotransmitters, particularly glutamate and aspartate, with respect to function of cyclic AMPgenerating systems in brain tissue, is as yet not completely resolved. Neither glutamate (Shimizu et al., 1974) nor y-aminobutyrate (Von Hungen and Roberts, 1973a,b) had any effect on adenylate cyclase activity in brain homogenates. Both glutamate and aspartate at very high concentrations (EGO1.5 mM) stimulated cyclic AMP generation in slices from cerebral cortex and cerebellum of various species (Ferrendelli et al., 1974, 1975; Shimizu et al., 1974, 1975b,c; Mah and Daly, 1976; Schmidt et aE., 1976). y-Aminobutyrate either had no effect or slightly reduced levels of cyclic AMP in brain slices. y-Aminobutyrate has been reported to antagonize responses to norepinephrine in rat cortical slices (French et al., 1975) and mouse cerebellar slices (Ferrendelli et al., 1975). Glycine had only marginal effects on cyclic AMP levels in brain slices. It would appear possible that enhanced formation and release of adenosine are responsible in incubated brain slices for the responses to glutamate and aspartate. The excitatory amino acid, glutamate, might, through stimulation of neuronal activity o r through ATP-dependent uptake in brain slices, deplete ATP and enhance adenosine release (cf. Pull and McIlwain, 1975). The inhibitory amino acid y-aminobutyrate might, by reduction in neuronal activity, tend to reduce adenosine release. Theophylline, an adenosine antagonist, did block glutamate-elicited accumulations of cyclic AMP in brain slices. However, another adenosine antagonist, 2’-deoxyadenosine, had no effect on glutamate responses. In addition, the response to combinations of glutamate and adenosine was greater than additive. Glutamate (EC,, 20 p M ) elicited significant accumulations of cyclic AMP in vesicular preparations obtained after homogenization of guinea pig cerebral cortical tissue in physiological medium (Shimizu et al., 1957a). Clearly, further studies are needed to resolve the mechanism and significance of glutamate-elicited accumulations of cyclic AMP in brain slices. At least two interpretations of the results are possible: one, that glutamate directly activates cyclic AMP systems and that theophylline is not a specific adenosine antagonist. It would appear more likely that glutamate elicits “release” of adenosine in brain slices and vesicular brain preparations and that “released”

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

125

adenosine is responsible for stimulation of cyclic AMP-generating systems, hence the blockade by theophylline. I n order to explain the lack of antagonism of glutamate responses with 2'-deoxyadenosine and at least additive responses to combinations of glutamate and adenosine, it must be proposed that exogenous adenosine and 2'-deoxyadenosine d o not readily reach the sites at which glutamate elicits release of adenosine. It should be noted that if glutamate does directly stimulate cyclic AMPgenerating systems in brain slices then it, along with substance P, will represent the only excitatory putative neurotransmitters which appear to have such activity. All other substances-norepinephrine, dopamine, serotonin, histamine, adenosine, and prostaglandin-which stimulate cyclic AMP systems have pronounced inhibitory effects on electrical activity of certain central neurons. 2,3-Diaminopropionate has been reported as a specific antagonist of glutamate-elicited accumulations of cyclic AMP (Shimizu et al., 1975~).Cysteine sulfinate and kainic acid represent glutamate analogs which stimulate cyclic AMP accumulation in brain slices (Shimizu et al., 1974, 1975b,c; Schmidt et al., 1976). At least in the case of kainic acid the mechanism of stimulation of cyclic AMP-generating systems appeared different from that of glutamate. 9. Peptides

Substance P, an extremely potent excitatory peptide, has been reported to stimulate adenylate cyclase activity in homogenates from rat brain (Duffy and Powell, 1975). Enkephalins in the presence of a peptidase inhibitor caused a small reduction in cyclic AMP levels in slices from rat cerebral cortex (Minneman and Iversen, 1976b). In rat brain homogenates, enkephalins inhibited formation of cyclic AMP (H. 0. J. Collier and A. C. Roy, cited in Goldstein, 1976). The relationship of the central analgesic and behavioral activity of enkephalins and of other larger endorphins to cyclic AMP mechanisms is at present under active investigation. In rat superior cervical ganglia, nerve growth factor elicited a transient increase in levels of cyclic AMP (NikodiJevicet al., 1975). Angiotensin, a ganglionic depolarizing agent, had no effect on cyclic AMP levels in bovine superior cervical ganglia (Kebabian et al., 1975b). Inhibitory peptides from neurosecretory cells of the mollusc Aplysia cal$mica increased levels of cyclic AMP in the neuropil of Aplysia abdominal ganglia (Treistman and Levitan, 1976). 10. Macromolecular Factors Heat-stable factors from brain have been reported to activate adenylate cyclases (Kauffmanet al., 1972; Izumietal., 1975a, 1976). The heat-

126

JOHN W. DALY

stable calcium-binding protein which is required for activation of calcium-dependent phosphodiesterases has recently been established as an activator of adenylate cyclase in brain homogenates (Brostrom et al., 1975; Cheunget al., 1975a; Lynch et al., 1976). This activator protein(s) was present in brain homogenates nearly equally distributed between soluble and membrane fractions (Cheung et al., 1975a; Gnegy et al., 1976a). The activator protein was released from membrane fractions of rat brain or striatum after cyclic AMP-dependent phosphorylation of a membrane protein (Gnegy et al., 1976a,b). It is tempting to speculate that in the membrane the activator protein is associated with adenylate cyclases, thus rendering enzymatic activity sensitive to regulation by calcium, and that as a result of cyclic AMP generation, a feedback control mechanism elicits a release of the activator protein into the cytosol where it and calcium ion can activate hydrolysis of cyclic AMP by soluble phosphodiesterases. In a preliminary communication, the activation of adenylate cyclase in membrane fractions from caudate nucleus by dopamine was reported to be greatly reduced after cyclic AMP-dependent release of the activator protein from the membrane (Gnegy et al., 1 9 7 6 ~ )Levels . of activator protein in caudate membranes were increased after treatments of animals with agents such as a-methyltyrosine, reserpine, clozapine, and haloperidol, which cause in vivo supersensitivity to dopaminergic agonists. Further studies on the role of macromolecular activators of adenylate cyclase, for example ubiquitin, an ubiquitous activator of P-adrenergic receptor-controlled cyclases (Goldstein et al., 1975) and on inhibitory factors (Levey et al., 1975; Izumi et al., 1975b), should provide valuable insights into the complex intracellular control of cyclic AMP systems in the nervous system. 11. Calcium Ions The activity of adenylate cyclases in homogenates was inhibited by high concentrations of calcium ions, but was activated by low concentrations of calcium, apparently through interaction with a high-affinity calcium-binding site (cf. Johnson and Sutherland, 1973; MacDonald, 1975). EGTA, by chelation of calcium ions associated with adenylate cyclases, markedly reduced enzymatic activity. The relationship of this calcium dependency to the presence of calcium-dependent activator protein has not been established. In brain slices, responses of cyclic AMP-generating systems to biogenic amines and adenosine were significantly influenced by removal of extracellular calcium with EGTA (cf. Schwabe and Daly, 1977; Schwabe et al., 1977). Responses to amines were reduced in the absence of extracellular calcium, while responses to adenosine appeared

C Y C L I C NUCLEOTIDES IN T H E NERVOUS S Y S T E M

127

somewhat enhanced. The a-adrenergic component of responses to norepinephrine in rat brain slices appeared to be completely dependent on the presence of extracellular calcium ions. 12. G T P .

The regulation of activity and hormone responsiveness of adenylate cyclases by GTP has been extensively studied with cell-free preparations. In such preparations GTP or a stable analog, guanylylimidodiphosphate (Gpp(NH)p),activates adenylate cyclases and often potentiates responses to biogenic amines and other activators. In recent studies with adenosine-sensitive and prostaglandin-sensitive adenylate cyclases from neuroblastoma cells, a model was proposed in which activation of adenylate cyclase was dependent upon a relatively irreversible binding of GTP to a guanine nucleotide site (Blume and Foster, 1976a). Hydrolysis of GTP to GDP on the enzyme was proposed to yield an inactive cyclaseGDP complex. Dissociation of GDP was proposed to be facilitated by “hormones” (adenosine or prostaglandin) to yield adenylate cyclases with the guanine nucleotide binding site again available for activation by GTP. In this model GTP would be the physiological activator of adenylate cyclase with hormones merely facilitating dissociation of inhibitory GDP from the enzyme from the cyclase. In unstimulated systems the adenylate cyclase would be present mainly in the inactive GDPcomplexed form. Such a model is consonant with GTP control of adenylate cyclase even in intact cells where levels of GTP (100-200 p M ) are far greater than the concentrations required for activation of adenylate cyclase in cell-free preparations. It would appear possible that under conditions in which GTP levels are reduced in a morphological compartment by activation of guanylate cyclase, associated adcnylate cyclases would become refractory to hormonal activation. Such inhibitory interrelationships of cyclic AMP and cyclic GMP-generating systems have been proposed (Goldberg et al., 1973, 1975), but have not been clearly demonstrated in either brain slices (Ohga and Daly, 1977a) or in ganglia (Kebabian et al., 1975b). Methods for the study of regulation of cyclic AMP-generating systems by GTP in intact cells are, unfortunately, not available.

C. LOCALIZATION OF PHOSPHODIESTERASES High levels of phosphodiesterases were associated in homogenates of brain tissue with both soluble and membrane fractions, including soluble and membrane fractions from lysed synaptosomes (De Robertis et al., 1967; Weiss and Costa, 1968; Beavo et nl., 1970; Gaballah and Popoff,

128

JOHN W. DALY

1971a). The presence of variety of phosphodiesterase isozymes some of which hydrolyze both cyclic AMP and cyclic GMP and some of which are more or less specific for either cyclic AMP or cyclic GMP complicates studies of this enzyme in both brain (Uzunov and Weiss, 1972a; Fertel and Weiss, 1974; Weiss et al., 1974; Kakiuchiet al., 1975a,b; Pledgeret al., 1975) and ganglia (Boudreau and Drummond, 1975; Lindlet al., 1976). Calcium-dependent activator protein for calcium-dependent phosphodiesterases occurred nearly equally distributed between soluble and membrane fractions from brain (Cheung et al., 1975a,b; Gnegy et al., 1976a) and underwent a cyclic AMP-dependent release from synaptosome membranes (Gnegy et al., 1976a,b). The morphological localization of phosphodiesterase in tissue slices from brain has been studied to a limited extent. Cyclic AMP phosphodiesterases assayed histochemically with high concentrations of cyclic AMP appeared localized at postsynaptic dendritic sites of neurons in rat cerebral cortical slices (Florendo et al., 1971) and at postsynaptic dendritic sites in the molecular layer of developing mouse brain (Adinolfi and Schmidt,. 1974). In earlier histochemical studies with rabbit, phosphodiesterase activity was found associated with glial cells, synaptic areas of neurons, in the neuropile of the cerebral cortical plexiform layer, and in the cerebellar molecular layer which contains Purkinje cell dendrites (Shanta et al., 1966). Phosphodiesterase activity assayed in homogenates from gray matter of rabbit cerebral cortex or olfactory bulb did not appear uniquely associated with particular layers (Breckenridge and Johnston, 1969). Phosphodiesterase activity in cerebral cortex was not decreased after lesions of the ascending noradrenergic, serotoninergic, and histaminergic nerve fibers of the medial forebrain bundle or after denervation of the superior cervical ganglia (Breckenridge and Johnston, 1969). Phosphodiesterase activity in rat cerebral cortex was only marginally decreased after destruction of presynaptic noradrenergic terminals with 6-hydroxydopamine (Kalisker et al., 1973). A soluble phosphodiesterase increased proximal to a constriction of the chicken sciatic nerve, suggesting axonal transport of phosphodiesterases to distal cholinergic terminals (Bray et al., 1971). In summary, phosphodiesterases appear to be localized to a significant extent at postsynaptic neuronal sites. The enzyme is probably also associated with glia and presynaptic terminals. D. REGULATION OF PHOSPHODIESTERASES

A role for a calcium-dependent activator protein in the control of calcium-dependent cyclic AMP and cyclic GMP phosphodiesterases has

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

129

been apparent since the discovery of this activator in 1970 (Kakiuchi and Yamazaki, 1970; Cheung, 1970). The activator has been studied extensively (cf. Teshima and Kakiuchi, 1974; Wickson et al., 1975; Brostrom and Wolff, 1976; Liu and Cheung, 1976; Uzunov et al., 1976) and has been recently shown to be capable of activating both adenylate cyclases and phosphodiesterases (Brostrom et al., 1975; Cheung et al., 1975a). Translocation of the activator provides an attractive basis for feedback control of cyclic AMP formation. An activator isolated from porcine brain appeared to be a phosphoprotein (Wolff and Brostrom, 1974), while activator isolated by other groups from rat, porcine, or bovine brain did not contain phosphate (Liu and Cheung, 1976; Watterson et al., 1976). A factor from rat brain activated phosphodiesterases of synaptosomes but not soluble phosphodiesterases (Izumi et al., 1976). Certain phospholipids have been reported to activate the calcium-dependent phosphodiesterases (Wolff and Brostrom, 1976). A protein isolated from bovine rod outer segments inhibited phosphodiesterases from rat brain (Dumler and Etingof, 1976). Similar heat-stable inhibitory proteins were present in brain homogenates (T. Kanamori, C. R. Creveling, and J. W. Daly, unpublished results). Inhibition of phosphodiesterases by calcium ions has been reported (cf. Cheung, 1971; Boudreau and Drummond, 1975). Inhibitions of cyclic AMP phosphodiesterases by cyclic GMP and vice versa represent a possible interrelationship between metabolism of the two cyclic nucleotides (Goldberg et al., 1970; Roberts and Simonsen, 1970; Weiss et al., 1974; Weiss and Greenberg, 1975). Stimulations of cyclic AMP phosphodiesterases by cyclic GMP have been reported (Beavoetal., 1971; Boudreau and Drummond, 1975; Hidakaetal., 1975). Regulation of hydrolysis of cyclic AMP in intact cells is obviously quite complex. In addition, adaptive changes in apparent levels of phosphodiesterase appear responsible in part for alterations in responsiveness of cyclic AMP-generating systems to changes in neurotransmitter input. Such adaptation in responsiveness is difficult to study in heterogeneous tissue such as brain (cf. Dismukes and Daly, 1976b; Daly, 1977). I n pineal gland, refractoriness of postsynaptic cyclic AMP-generating systems to isoproterenol was due in part to an adaptive increase in phosphodiesterases (Oleshansky and Neff, 1975). During the night when noradrenergic input to the pineal gland is elevated, levels of phosphodiesterases have apparently undergone a compensatory adaptive increase (Minneman and Iversen, 1976a). The mechanisms involved in adaptive changes in phosphodiesterase activity appear to involve cyclic AMP-dependent mechanisms and, at least in cultured neuroma cells, to involve protein synthesis (Schwartz and Passonneau, 1974; Browning et al., 1976). In certain cell lines, exposure to stimulants of cyclic AMP-

130

JOHN W. DALY

dependent generating systems did not result in adaptive changes in phosphodiesterases (De Vellis and Brooker, 1974; Perkins et al., 1975).

E. INHIBITORS A N D ACTIVATORS OF PHOSPHODIESTERASES Drugs which specifically inhibit or active phosphodiesterases would provide extremely valuable tools for the study of functional roles of cyclic AMP and cyclic GMP in the central and peripheral nervous system. Unfortunately, most known inhibitors of phosphodiesterases have side effects and few appear to be specific with respect to inhibition of specific phosphodiesterase isozymes. The situation is even less satisfactory with respect to agents which activate phosphodiesterases. Thus, imidazoles, generally accepted as phosphodiesterase activators, are maximally effective with brain phosphodiesterases only at 10-20 mM concentrations (cf. Cheung, 1971). Inhibition of phosphodiesterases by imidazole has also been reported (Goldberg et al., 1970). 1 . MethylxanthineJ Theophylline and caffeine have low potencies as inhibitors of phosphodiesterases from brain (Weinryb et al., 1972; Fredholm et al., 1976; Levin and Weiss, 1976). Thus, it remains rather doubtful that the central pharmacological activities of these two compounds are primarily due to inhibition of phosphodiesterases. Isobutylmethylxanthine was manyfold more potent than theophylline or caffeine as a phosphodiesterase inhibitor in brain preparations (cf. Fredholm et aL, 1976). Isobutylmethylxanthine appeared to be more potent as an inhibitor of calciumdependent phosphodiesterases than of crude phosphodiesterases from rat cerebral cortex (DuMoulin and Schultz, 1975). Isobutylmethylxanthine had similar potency with regard to inhibition of hydrolysis of cyclic AMP and cyclic GMP (Fredholm et al., 1976). Methylxanthines, in addition to their activity as phosphodiesterase inhibitors, are active antagonists of adenosine-sensitive cyclic AMP systems in brain slices (cf. Mah and Daly, 1976).

2 . Benzodiazepines The centrally active benzodiazepines, which include diazepam, chlordiazepoxide, and medazepam, were potent inhibitors of phosphodiesterases from brain (Weinryb et al., 1972; Dalton et al., 1974; DuMoulin and Schultz, 1975; Levin and Weiss, 1976). N o striking selectivity, with respect to inhibition of phosphodiesterases from different brain regions, was noted for this class of inhibitor. Benzodiazpeines inhibited hydrolysis of cyclic AMP and cyclic GMP by brain phosphodies-

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

131

terases equally effectively. Pharmacological activities of benzodiazepines may not be, in all cases, linked to inhibition of phosphodiesterases. Diazepam and chlordiazepoxide have, for example, been demonstrated to be potent, presumably directly acting antagonists of y-aminobutyrate-elicited inhibition of Purkinje cells in explants of rat cerebellum (Gahwiler, 1976). 3. Phenotliiazines

Although relatively potent as phosphodiesterase inhibitors, compounds of this class such as fluphenazine, trifluoperazine, and chlorpromazine have too many other activities, for example as dopaminergic antagonists, inhibitors of adenylate cyclases (Uzunov and Weiss, 1972b; Palmer and Manian, 1974a,b),and inhibitors of norepinephrine uptake, to be useful as specific tools for in situ investigation of phosphodiesterases. The phenothiazines were very potent inhibitors of calciumdependent phosphodiesterases (Uzunov et al., 1974; Weiss et al., 1974; Weiss and Greenberg, 1975; Levin and Weiss, 1976). I t would appear that trifluoperazine, chlorpromazine, and other antipsychotics such as pimozide inhibited calcium-dependent phosphodiesterases primarily through competitive antagonism of the activation by the calciumdependent activator protein. Chlorpromazine was much more potent in inhibiting hydrolysis of cyclic AMP than in inhibiting hydrolysis of cyclic GMP by calcium-dependent phosphodiesterases. Trifluoperazine inhibited phosphodiesterases from rat cerebrum and brain stem more effectively than phosphodiesterases from cerebellum (Uzunov and Weiss, 1971).

4. Papaverine The alkaloid papaverine is a relatively potent inhibitor of brain, phosphodiesterases, but in view of other pharmacological activities, such as inhibition of ATP-generating systems and inhibition of uptake of adenosine, it is not a particularly selective tool for the in situ study of phosphodiesterases. Inhibition of cyclic AMP phosphodiesterases by papaverine was competitive at low concentrations of the alkaloid and noncompetitive at high concentrations (Weiss, 1975). Papaverine was more potent as an inhibitor of certain calcium-independent enzymes than as an inhibitor of the calcium-dependent enzymes. Particulate enzymes appeared to be inhibited more readily than soluble enzymes by papaverine (Furlanut et al., 1973; Fredholm et al., 1976). Papaverine was a potent inhibitor of hydrolysis of both cyclic AMP and cyclic GMP by calcium-dependent phosphodiesterase (Kakiuchi et al., 1975a; Weiss, 1975; Fredholm et al., 1976.)

132

JOHN

W. D A L Y

5. 1-H-Pyrazolo[3,4b]pyridines This class of compounds, exemplified by SQ 20009, represents a group of extremely potent inhibitors of brain phosphodiesterases (Weinryb et al., 1972; Hess et al., 1975; Kakiuchi et al., 1975a). S Q 20009 was a potent inhibitor of hydrolysis of cyclic GMP by calcium-dependent phosphodiesterases (Kakiuchi et al., 1975a). I n studies with brain slices SQ 20009 had, unlike other potent phosphodiesterase inhibitors, little effect on accumulations of cyclic AMP elicited by biogenic amines, adenosine, or glutamate (Schultz, 1974a,b; Mah and Daly, 1976). Potentiations of amine responses in brain slices by relatively high concentrations of SQ 20009 have been reported by other groups (Forn et al., 1974; Hess et al., 1975; Wilkening and Makman, 1975). The potency of SQ 20009 in intact brain cells would appear much lower than would have been predicted based on its potency with cell-free preparations. 6. Dipyridamole This polar compound probably penetrates intact cells to a limited extent, thus circumscribing its usefulness as a tool for the in situ study of phosphodiesterases. With brain phosphodiesterases, dipyridamole was a relatively potent inhibitor (Weinryb et al., 1972; Fredholm et al., 1976). Dipyridamole was ineffective with cyclic GMP phosphodiesterases. Potentiation of amine responses by dipyridamole in brain slices appeared mainly due to inhibition of adenosine uptake rather than inhibition of phosphodiesterases (Huang and Daly, 1974). 7. Dialkoxybenzyl-2-imidazolzdinones This widely used class of phosphodiesterase inhibitors is exemplified by RO 20-1724. In brain homogenates RO 20-1724 was not nearly as potent an inhibitor of phosphodiesterases as isobutylmethylxanthine, diazepam, papaverine, or SQ 20009 (Sheppard et al., 1972; DuMoulin and Schultz, 1975). RO 20-1724 had little or no effect on cyclic GMP phosphodiesterases (Sheppard et al., 1972; Schwabe et al., 1976). RO 20- 1724 was quite effective in potentiating amine and adenosine-elicited accumulations of cyclic AMP in brain slices (Schultz, 1974a,b; Mah and Daly, 1976; Schwabe el al., 1976). In part, potentiation of amine responses by RO 20-1724 in brain slices appeared due to inhibition of uptake of endogenous adenosine resulting in synergistic amineadenosine interactions (Mah and Daly, 1976; Schwabe et al., 1977). 8. Dialkoxyphenyl-2-pyrlidones

Recently, a dialkoxyphenyl-2-pyrrolidone,ZK 627 11, has been proposed as the phosphodiesterase inhibitor of choice for study of the enzyme in situ in the nervous system (Schwabe et al., 1976b). ZK 6271 1 was

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

133

many-fold more potent as an inhibitor of calcium-dependent phosphodiesterases from brain than RO 20-1724 and was much more potent than RO 20- 1724 in potentiating norepinephrine and adenosine-elicited accumulations of cyclic AMP in brain slices. Adenosine mechanisms appeared to have only a minor role in potentiation of amine responses by ZK 6271 1 in brain slices. ZK 6271 1 had little effect on cyclic GMP phosphodiesterases. F. CYCLICAMP-DEPENDENT PROTEINKINASES

Cyclic AMP-dependent protein kinases in homogenates from brain tissue were present both in soluble fractions, and associated with particulate fractions including synaptosomes (Gaballah et al., 1971; Maeno et al., 1971; Gaballah and Popoff, 1971b; Uno et al., 1976) and microtubules (cf. Sloboda et al., 1975; Rappaport et al., 1976). Synaptic membranes from synaptosomes contained high levels of cyclic AMP-dependent protein kinases. Only low levels of protein kinase activity were associated with nuclei, although in other tissues such as liver, protein kinases from nuclei appeared involved in phosphorylation of histone and nonhistone chromatin proteins (cf. Kish and Kleinsmith, 1974). Low- and highaffinity binding sites for cyclic AMP were present in synaptic membranes (Weller and Rodnight, 1975). Presumably, the high-affinity binding sites represent, at least in part, sites on the regulatory unit of cyclic AMPdependent protein kinases. A number of different protein kinases were present in brain (cf. Miyamoto et al., 197 1 ; Inoue et al., 1973; Kuo, 1974; Takahashi et al., 1975; Uno et al., 1976). Soluble cyclic AMP-dependent protein kinases obtained from membrane and cytosol fractions appeared distinct in terms of physical properties and substrate specificities (Uno et al., 1976). Cyclic AMPdependent protein kinases have been divided into two types (Corbin et al., 1975). Type I enzymes were only slowly dissociated into regulatory and catalytic subunits by histunes and reassociated quite rapidly after cyclic AMP-elicited dissociation. Type I1 enzymes were dissociated relatively rapidly by histone and reassociated relatively slowly after cyclic AMP. The major enzyme in rat brain was of type 11. The type I kinase and another kinase were minor constituents. The physiological substrates for cyclic AMP-dependent protein kinases in brain tissue will be dependent on both the substrate specificity of the kinase and the accessibility of various potential substrates. Thus, while histones are active substrates for most cyclic AMP-dependent protein kinases, their nuclear localization renders them inaccessible to, for example, protein kinases associated with synaptic membranes. Translocation of the catalytic unit of cyclic AMP-dependent protein kinase from cytosol to nucleus, however, occurred in glioma cells (Salem

134

JOHN W . DALY

and DeVellis, 1976) and in adrenal medulla (Costa et al., 1976) after activation of cyclic AMP-generating systems. Synaptic membranes (Johnson etal., 1971) and ribosomes (Schmidt and Sokoloff, 1973) represent two particulate fractions from brain which contain substrates for cyclic AMP-dependent protein kinases. An “inhibitory” modulator protein has been reported from various tissues including brain (cf. Kuo, 1975; Kuo et al., 1976a,b). This inhibitory modulator protein inhibited cyclic AMP-dependent protein kinase activity assayed with histone as substrate, but actually stimulated activity when assayed with protamine as substrate. In brain homogenates, the inhibitory modulator protein was associated with soluble proteins and with synaptosome fractions (Roskoski et al., 1976). Virtually all of the modulator in lysed synaptosomes was associated with soluble proteins. Other endogenous factors would appear capable of regulating cyclic AMP-dependent protein kinases. These include calcium ions, adenosine, AMP, and ADP which inhibit kinase activity (Kuo and Greengard, 1969; Miyamoto et al., 1969; Kuoet al., 1970). ADP was the most active inhibitor. Formation of cyclic AMP in various tissues including brain has apparently been accompanied by the formation of a compound similar in properties to cyclic AMP which inhibited activation of kinases by cyclic AMP (Murad et al., 1969; Wasner, 1975). Activation of cyclic AMP-dependent protein kinases by cyclic GMP does not appear, in view of the specificity of binding sites on the regulatory subunits of the kinase, to be of significance under physiological conditions. Cyclic GMP has, however, been reported to antagonize cyclic AMP-dependent phosphorylation of microtubular protein (Sandoval and Cuatrecasas, 1976b). Cyclic AMP-dependent protein kinases catalyzed phosphorylation of their own regulatory subunits (Maeno et al., 1974). The significance of phosphorylation of the regulatory subunit is unclear. In heart preparations phosphorylation facilitated cyclic AMP dissociation of regulatory and catalytic subunits (Erlichman et al., 1974; Rangel-Aldao and Rosen, 1976). In summary, it is clear that regulation of phosphorylation of endogenous substrates by cyclic AMP-dependent protein kinases is complex. Steady-state levels of cyclic AMP-dependent protein phosphorylation will reflect not only the activity of kinases, but also the rates of hydrolysis by phosphatases.

G. PHOSPHOPROTEIN PHOSPHATASES These enzymes which presumably serve to terminate the physiological responses to cyclic nucleotides in intact cells were present in homogenates of brain tissue in both soluble and membrane fractions (Weller and Rodnight, 1971; Maeno and Greengard, 1972; Maeno et al., 1975;

CYCLIC NUCLEOTIDES I N T H E NERVOUS SYSTEM

135

Miyamoto and Kakiuchi, 1975). A significant proportion of the phosphoprotein phosphatases was associated with synaptosome fractions. Lysed synaptosome fractions afforded both soluble and membrane phosphatase activity. Histochemically, phosphoprotein phosphatases have been cited as appearing to be associated with postsynaptic dendritic sites of neurons in rat cerebral cortex (Greengard ct al., 1972). At least three isozymes with differing substrate profiles have been detected in brain preparations (Maeno and Greengard, 1972). Magnesium and calcium ions had little effect on phosphatase activity; manganese was sornewhat stimulatory and zinc was inhibitory. A compound which had properties similar to cyclic AMP and which was apparently formed along with cyclic AMP in various tissues has been reported to stimulate a phosphoprotein phosphatase from beef muscle (Wasner, 1975). Cyclic AMP inhibited the phosphatase. Dephosphorylation of the regulatory unit of cyclic AMP-dependent protein kinase has been reported to be stimulated by cyclic AMP (Maeno ef nl., 1975). In membranes, the localization of kinases, phosphatases, and their protein substrates are probably interrelated, perhaps as functional complexes (cf. Ueda et d,, 1975). H. CYCLICAMP-DEPENDENT AUTOPHOSPHORYLATION OF PROTEINS

The activity of cyclic AMP-dependent protein kinases can be studied either with exogenous substrates or with endogenous substrates present in soluble, synaptosome, or microtubule preparations from brain tissue or in brain slices. The latter technique, which will be referred to as autophosphorylation, provides data relevant to the normal physiological substrates whose phosphorylation is regulated by cyclic AMP. Incorporation of radioactive phosphate from [:"P]ATP into such endogenous proteins with soluble or membrane preparations will be dependent on accessibility of ATP to the catalytic site of protein kinases and on the activity of protein kinases and phosphoprotein phosphatases. Rates of turnover of phosphorylated proteins are conveniently assessed by addition of EDTA to inhibit the kinases. The rate of dephosphorylation by phosphatases can then be measured. In tissue slices, steady-state incorporation of phosphate into protein will, in addition, be dependent on the rate of incorporation of radioactive phosphate into ATP and the degree to which alternate pathways for ATP utilization compete with the kinase for labeled ATP. 1. Soluble Fractions

Autophosphorylation of proteins in soluble fractions from brain might be expected to provide data on possible physiological substrates for cyclic AMP-dependent kinases in the cytosol of brain cells. However,

136

JOHN W. DALY

interpretation of the data must be tempered by a realization that many morphological relationships in the cytosol have been disrupted on homogenization, and that nonphysiological substrates can now compete with physiological substrates for the kinase. Nonetheless, in soluble fractions from brain homogenates, cyclic AMP clearly stimulated phosphorylation of one protein, apparently identical with the regulatory subunit of cyclic AMP-dependent protein kinase (Malkinson, 1975). Presumably, the phosphorylation of other cytosol proteins, such as phosphorylase 6 kinase, glycogen synthetase I, and tyrosine hydroxylase-activator protein was also stimulated by cyclic AMP but was undetectable in heterogeneous soluble fractions from brain.

2 , Synaptosomes Cyclic AMP stimulated phosphorylation of a t least three proteins in synaptic membranes from rat brain (Johnson et al., 1972; Ueda et al., 1973; Ehrlich and Routtenberg, 1974; Krueger et al., 1975; Maenoet al., 1975; Malkinson et al., 1975; Routtenberg and Ehrlich, 1975; Weller and Morgan, 1976). A protein with a molecular weight of about 49,000 appeared to correspond to the regulatory unit of cyclic AMP-dependent protein kinase and was phosphorylated in soluble fractions, synaptic membrane fractions, and other membrane fractions suggesting that it has no special o r unique role in synaptic events. However, two higher molecular weight proteins (80,000 and 86,000) appeared to be unique to synaptic membranes (Weller and Morgan, 1976).Levels of these proteins increased markedly in rat cerebrum 2 to 3 weeks after birth, a time at which synapses are rapidly being formed in the rat central nervous system (cited in Greengard, 1976). Calcium ions have been shown to stimulate phosphorylation of a number of proteins in synaptic membranes, in particular proteins with molecular weights of about 62,000 and 49,000 (DeLorenzo, 1976). The relationship of these proteins to those phosphorylated by cyclic AMP-dependent mechanisms has not been investigated.

3. Microtubules The autophosphorylation of protein constituents of microtubules has been studied extensively since an initial report on phosphorylation of microtubular protein in 1970 (Goodman et al., 1970). It now appears that a high molecular weight cyclic AMP-dependent protein kinase is associated with microtubular protein and phosphorylates certain other trace, high molecular weight proteins rather than tubulin itself (cf. Sloboda et al., 1975; Rappaport et al., 1976; Sandoval and Cuatrecasas, 1976a,b). The significance of the phosphorylation of microtubular pro-

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

137

teins is unclear, although it has been proposed that phosphorylation may facilitate polymerization of tubulin (Sandoval and Cuatrecasas, 1976a). However, neither the rate nor extent of polymerization of microtubular proteins was significantly affected by inhibition of protein phosphorylation or by cyclic AMP-elicited phosphorylation (Rappaport et al., 1976). Others have proposed the stabilization of microtubules by cyclic AMP (Gillespie, 1971; cf. review by Daly, 1977).

4. Brain Slices and Ganglia Phosphorylation of proteins in brain slices has been studied extensively during incubations with agents or under conditions expected to stimulate cyclic AMP accumulations (Reddington et al., 1973; Weller and Rodnight, 1973a,b; Williams et al., 1974a,b; Williams and Rodnight, 1975, 1976). Norepinephrine, histamine, serotonin, and electrical pulsation increased incorporation of radioactive phosphate into proteins of guinea pig cortical slices. Adenosine, another stimulant of cyclic AMPgenerating systems, but in addition an inhibitor of protein kinase, had little effect alone, and selectively blocked the increase in protein phosphorylation elicited by histamine. T h e increase in protein phosphorylation elicited by norepinephrine was blocked by P-adrenergic but not by a-adrenergic antagonists. The increase in protein phosphorylation elicited by norepinephrine and electrical pulsation appeared to have been associated with neuronal elements, while the increase elicited by histamine and serotonin appeared to have been associated with glial elements in guinea pig cortical slices. The stimulations of protein phosphorylation elicited by norepinephrine and electrical pulsation were not additive. The data are difficult to completely rationalize in terms of a-adrenergic receptor-elicited accumulations of cyclic AMP in guinea pig cortical slices, and synergistic responses of cyclic AMP-generating systems to the combination of electrical pulsation with norepinephrine and to combinations of adenosine with histamine, serotonin, or norepinephrine. In slices of rat striatum, all of the biogenic amines-norepinephrine, histamine, serotonin, and dopamine-nhanced protein phosphorylation (Williams, 1976). y-Aminobutyrate and acetylcholine also stimulated protein phosphorylation. Adenosine reduced protein phosphorylation. T h e stimulatory effect of dopamine was antagonized by fluphenazine and haloperidol. 8-Bromo cyclic AMP or the phosphodiesterase inhibitor, isobutylmethylxanthine, enhanced phosphorylation of three proteins in slices of rat caudate nucleus (Krueger et al., 1975). The two higher molecular weight proteins (80,000 and 85,000) appeared associated with synaptosome membranes, while the lower molecular weight protein (49,000) appeared to be the regulatory subunit of cyclic AMP-dependent protein kinase.

138

JOHN W. DALY

In the abdominal ganglion of the mollusc, Aplysiu culijimica, dibutyryl cyclic AMP enhanced the phosphorylation of a specific high molecular weight protein ( 1 18,000) probably associated with synaptic entities of the ganglia (Levitan and Barondes, 1974; Levitan et al., 1974). Octopamine and serotonin also enhanced phosphorylation of this protein. 111. Cyclic GMP

A. LOCALIZATION OF GUANYLATE CYCLASES Guanylate cyclase, unlike adenylate cyclase, was associated with both soluble and particulate fractions in brain homogenates (Goridis and Morgan, 1973; Bensinger et al., 1974; Nakazawa and Sano, 1974; Kimura and Murad, 1974, 1976a,b; Nakazawa et al., 1976; Troyer and Ferrendelli, 1976). High levels of guanylate cyclase were found in synaptosome fractions. Lysis of synaptosomes, however, afforded mainly soluble guanylate cyclase, a result strongly indicative of a presynaptic 10calization of this enzyme. Recent data indicated that soluble and particulate guanylate cyclases from cerebellum were distinct entities (Troyer and Ferrendelli, 1976). The soluble enzyme did not appear merely to represent readily solubilized membrane-bound enzyme. In cultures of neonatal or fetal brain cells from rat or chicken, guanylate cyclase was mainly associated with neurons rather than with glia (Goridis and Morgan, 1973; Goridis et al., 1974; Zwiller etal., 1976). Guanylate cyclase was present in most regions of rat brain with lowest levels in pons medulla and spinal cord (Nakazawa and Sano, 1974; Nakazawa et al., 1976). T h e high levels of cyclic GMP and cyclic AMP in cerebellum (Schmidt et al., 1972; Kuo et al., 1972; Steiner et al., 1972) probably primarily reflect not high levels of cyclases, but instead the low levels of cerebellar phosphodiesterases. Cyclic GMP levels in mouse cerebellum were 2-fold higher in the molecular layer containing the dendrites of Purkinje cells than in the granular layer (Rubin and Ferrendelli, 1976). Cyclic GMP levels were markedly reduced in cerebellum from “nervous” mutant mice in which Purkinje cells are nearly completely absent (Ma0 et al., 1975). No interrelationships between levels of cyclic AMP and cyclic GMP were apparent in studies with cerebellar slices from rat, mouse, guinea pig, and rabbit, suggesting that cyclic AMP and cyclic AMPgenerating systems are present in different morphological loci (Ohga and Daly, 1977a). Other data have been indicative of an association of both cyclic AMP and cyclic GMP systems with Purkinje cells. The effect of various putative neurotransmitters and depolarizing conditions on

CYCLIC NUCLEOTIDES I N T H E NERVOUS SYSTEM

139

accumulations of cyclic GMP has been studied extensively in incubated slices of brain tissue, particularly in slices from cerebellum where responses of cyclic GMP-generating systems are quite large (cf. Ferrendelli, 1975). The data have not provided definitive insights into the site of generation of cyclic GMP in brain tissue. In bovine superior cervical ganglion, an immunofluorescent assay indicated that acetylcholine-elicited accumulations of cyclic GMP occurred primarily in postganglionic neurons (Kebabian et al., 1975a). In rat pineal glands, however, denervation greatly decreased the norepinephrine-elicited accumulation of cyclic GMP, a result indicative of a presynaptic site for cyclic GMP formation in this sympathetically innervated tissue (O’Dea et al., 1976). Denervation, however, was reported by another group to result in an increase of guanylate cyclase in pineal gland (Strada el al., 1976). In summary, definitive statements as to the localization of cyclic GMP-generating systems in nervous tissue are not, as yet, warranted. Clearly, a significant portion of cyclic GMP systems would appear to be associated with neurons at presynaptic loci. However, the presence of cyclic GMP-generating systems at post-synaptic loci in neurons and in glia cannot be excluded. I t would appear that a significant portion of the enzyme is associated with cytosol, while the remainder is a membrane enzyme.

B. REGULATION OF GUANYLATE CYCLASES T h e activity of guanylate cyclases in the nervous system appears to be under a complex set of controls, including exogenous factors such as excitatory neurotransmitters and calcium ions, intracellular macromolecules, and levels of GTP and ATP. Guanylate cyclase from brain tissue required manganese for optimal activity, and has been reported to be inhibited by ATP (White and Aurbach, 1969; Boehme, 1970; Goridis and Morgan, 1973; Nakazawa and Sano, 1974; Olson et al., 1976: Troyer and Ferrendelli, 1976). Magnesium ions could partially activate the enzyme. Calcium ions usually inhibited membrane-bound guanylate cyclases and stimulated soluble guanylate cyclases. However, although guanylate cyclase activity from a neuroblastoma cell line was predominantly particulate, a stimulation of total enzyme activity by calcium pertained (Zwiller et al., 1976), The calcium-dependent activator protein had no effect on activity of guanylate cyclases (Olson et al., 1976). Other macromolecular factors may, however, be important to the activity of guanylate cyclases. Thus, the presence of a macromolecular factor was required in order for azide to activate soluble guanylate cyclases from

140

JOHN W. DALY

brain (Mittal et al., 1975). Azide alone was able to activate membranebound guanylate cyclases from brain and to activate cyclic GMPgenerating systems in brain slices (Kimura et al., 1975). Activation of soluble or membrane-bound guanylate cyclase by putative neurotransmitters or neuromodulators has proven difficult to demonstrate in brain and other tissue. Thus, activation of cyclic GMP-generating systems in brain slices or ganglia by cholinergic and noradrenergic agonists, by glutamate, by histamine, by adenosine, and by depolarizing agents or electrical stimulation may represent not a direct activation of the cyclase by a neurotransmitter, but instead may reflect only neurotransmitter or depolarization-elicited increases in influx of calcium and an intracellular activation of the guanylate cyclase by calcium. Indeed, virtually all instances of neurotransmitter or depolarization-elicited increases in cyclic GMP in brain slices or ganglia have been dependent on extracellular calcium (cf. Ferrendelli et al., 1976; Ohga and Daly, 1977b). Feedback control and adaptive changes in cyclic GMP-generating systems have not really been investigated. In vivo treatment with harmaline elevated cyclic GMP levels in brain and led to an increase in soluble but not particulate guanylate cyclase in rat cerebellum (Spano et al., 197513).A protein factor inhibitory to cyclic GMP-dependent protein kinases was decreased in cerebellum by harmaline treatment (Szmigielski and Guidotti, 1976). Denervation of the pineal gland and hence cessation of noradrenergic input led to an increase in guanylate cyclase activity (Strada et al., 1976). A decrease in responses of cyclic GMP-generating systems to norepinephrine occurred after denervation (ODea et al., 1976). It was proposed that the decrease was due to loss of presynaptic cyclic GMP systems. A brief survey of the effects of different neurotransmitters on cyclic GMP generation in brain and ganglia tissue is relevant to the role of cyclic GMP in the nervous system. 1. Norepinephrine

The effects of norepinephrine on levels of cyclic GMP in cerebral cortical slices from various species have been inconsistent. Either no effect or only a marginal stimulation by norepinephrine has been reported (Kinscherf et al., 1976; Ohga and Daly, 1977a; Schwabe et al., 1977). In cerebellar slices from various species norepinephrine did elicit a significant accumulation of cyclic GMP (Ferrendelli, 1975; Ferrendelli et al., 1975; Kinscherf et al., 1976; Ohga and Daly, 1977a,b; Schmidt et al., 1976). In mouse cerebellar slices, the response to norepinephrine was blocked by both a- and P-adrenergic antagonists, while in guinea pig cerebellar slices only the P-adrenergic antagonist was effective. In pineal

CYCLIC NUCLEOTIDES IN THE NERVOUS SYSTEM

141

gland, norepinephrine appeared to elicit accumulations of cyclic GMP at presynaptic sites via interaction with an a-adrenergic receptor (O’Dea et al., 1976). 2. Other Biogenic Amines Dopamine and serotonin have not been reported to elicit significant accumulations of cyclic GMP in brain tissue (Ferrendelli et al., 1975). Dopamine did appear to slightly antagonize acetylcholine-elicited accumulations of cyclic GMP in bovine superior cervical ganglion (Kebabian et al., 1975b). Histamine has been reported to elicit accumulations of cyclic GMP in cerebral cortical slices but not in cerebellar slices of rabbit, rat, and guinea pig (Kuo et al., 1972; Lee et al., 1972; Ohga and Daly, 1977a,b; Schwabe et al., 1977). Histamine increased cyclic GMP levels in bovine superior cervical ganglion (Kebabian et al., 1975b).

3. Adenosine Adenosine has been reported to have either no effect (Ferrendelli et al., 1973, 1975) or a small stimulatory effect (Ohga and Daly, 1977a,b; Schwabe et al., 1977a) on cyclic GMP levels in brain slices. Further studies are clearly required to establish adenosine as a valid modulator of cyclic GMP-generating systems in brain tissue. The response in guinea pig cerebellar slices was antagonized by theophylline. 4. Acetylcholine Acetylcholine and cholinergic agonists elicited accumulations of cyclic GMP in brain slices from cerebral cortex and cerebellum of rabbit and rat (Kuo et al., 1972; Lee et al., 1972; Palmer and Duszynski, 1975; Palmer et al., 1976) and in bovine superior cervical ganglion (Kebabian et al., 197513) via interaction with muscarinic receptors. In other studies with cortical and cerebellar slices from various species, acetylcholine and cholinergic agonists had no effect on cyclic GMP levels (Ferrendelli et al., 1973; Kinscherf et al., 1976; Ohga and Daly, 1977a). Carbamylcholine had no effect on cyclic GMP levels in rat superior cervical ganglion (Hanbauer et al., 1975a). Acetylcholine and cholinergic agonists have been reported to elicit accumulations of cyclic GMP in peripheral neurons (Kebabian et al., 1975b; Horn and McAfee, 1976). 5 . Prostaglandins Prostaglandins did not, in preliminary experiments, have effects on levels of cyclic GMP in brain slices (Ferrendelli, 1975; Ohga and Daly, 1977a).

142

JOHN W. DALY

6. Amino Acids Glutamate, at quite high concentrations, elicited accumulations of cyclic GMP in cerebral cortical slices from rabbit, guinea pig, cat, and mouse, but not in cerebral cortical slices from rat (Kinscherf et al., 1976). Glutamate elicited accumulations of cyclic GMP in cerebellar slices from guinea pig, rat, and mouse, but not in rabbit, Other groups have reported no effect of glutamate in guinea pig (Ohga and Daly, 1977a) or rat cerebellar slices (Schmidt et al., 1976). T h e glutamate response has been studied in some detail in mouse cerebellar slices (Ferrendelli et al., 1974, 19’75; Ferrendelli, 1975). The mechanism remains unclear, although it appears likely that glutamate-elicited activation of cyclic GMP-generating systems is related to its function as a n excitatory neurotransmitter in the nervous system. Kainic acid, an excitatory analog of glutamate, was a potent activator of cyclic GMP generation in rat cerebellar slices but not in slices from cerebral cortex, hippocampus, midbrain, hypothalamus, or brain stem (Schmidt et al., 1976). Glutamate had no effect on activity of cerebellar guanylate cyclase (Biggio and Guidotti, 1976b). The inhibitory neurotransmitters glycine and y-aminobutyrate had no effect on cyclic GMP levels in slices from cerebral cortex or cerebellum of various species (Kinscherfet al., 1976; Ohga and Daly, 1977a; Schmidt et al., 1976), except in cerebellar slices from mouse where both glycine and 7-aminobutyrate at quite high concentrations elicited modest accumulations of cyclic GMP (Ferrendelli et al., 1974, 1975; Ferrendelli, 1975). It is unlikely that the stirnulatory effects of y-aminobutyrate at such high concentrations on cyclic GMP systems bear any relationship to its physiological functions in brain, where a large body of evidence implicates y-aminobutyrate as an inhibitor of cyclic GMP generation and glutamate as an activator of cyclic GMP generation (cf. Costa et al., 1975a,b). 7. Peptides Enkephalins, in the present of a peptidase inhibitor, increased cyclic GMP levels in rat striatal slices (Minneman and Iversen, 1976b). The analgesic antagonist, naloxone, blocked the response. In bovine superior cervical ganglion, angiotensin, a ganglionic depolarizing agent, had no effect on cyclic GMP levels (Kebabian et al., 1975b). C. CYCLICGMP PHOSPHODIESTERASES It has proven difficult to establish which of the various phosphodiesterases in nervous tissue are concerned with hydrolysis of cyclic AMP and

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

143

which are concerned with hydrolysis of cyclic GMP. Certain phosphodiesterases are relatively selective while others appear to hydrolyze both cyclic AMP and cyclic GMP (cf. Thompson and Appleman, 1971a,b; Pledger et al., 1974, 1975). Regional distribution of cyclic GMP phosphodiesterases has been investigated in rat (Nakazawa and Sano, 1974) and cat (Dalton et al., 1974) brain. Cyclic GMP phosphodiesterases like cyclic AMP phosphodiesterases were found in both particulate and soluble fractions from rat brain (Nakazawa and Sano, 1974). It has been suggested that a high molecular weight phosphodiesterase isozyme might be concerned primarily with cyclic GMP hydrolysis (Thompson and Appleman, 1971b), and more recently that the calciiim-dependent phosphodiesterase of brain might be concerned primarily with cyclic GMP hydrolysis (Kakiuchi et al., 1973). No definitive evidence has, however, been obtained to support such hypotheses. In part this reflects the fact that selective activators or inhibitors of cyclic AMP or cyclic GMP phosphodiesterases for studies in intact cells have not as yet been defined. Dipyridamole, RO 20-1724, and ZK 62771 are, however, relatively ineffective inhibitors of cyclic GMP phosphodiesterase activity (see above). In brain slices RO 20-1724 and ZK 62771 were relatively ineffective with regard to elevating cyclic GMP levels (Schwabe et al., 1976). Isobutylmethylxanthine proved to be the most effective inhibitor with regard to elevation of cyclic GMP levels in guinea pig cerebellar slices (Ohga and Daly, 1977b). Theophylline and SQ 20009 were very ineffective, while RO 20-1724, diazepam, and papaverine had intermediate effects. Until really selective inhibitors of cyclic AMP and cyclic GMP phosphodiesterases are developed, the role of the different phosphodiesterases in brain cyclic nucleotide metabolism will probably remain rather poorly defined. Cyclic AMP has, with certain phosphodiesterases, been shown to competitively inhibit hydrolysis of cyclic GMP (Goldberg et al., 1970; Williams et al., 1971; Weiss et al., 1974: Uzunov et al., 1976). D. CYCLICGMP-DEPENDENT PROTEINKINASES The properties and function of the relatively unstable cyclic GMPdependent protein kinases in nervous tissue are as yet poorly understood. Levels of cyclic GMPdependent protein kinase were higher in cerebellum than in cerebral cortex (Sold and Hofmann, 1974). The enzyme occurred in both soluble and membrane fractions in brain homogenates, required magnesium ions for activity, and was inhibited by calcium ions (Greengard and Kuo, 1970; Hofmann and Sold, 1972; Kuo, 1974). At least two isozymes were present in brain (Takai et nl., 1975). Binding sites for cyclic GMP in brain homognates, presumably at

144

JOHN W . DALY

least in part, reflect sites of a regulatory subunit of cyclic GMPdependent protein kinases (Sold and Hofmann, 1974; Gill and Kanstein, 1975; Takai et al., 1975). Indeed, in cerebellar homogenates the major portion of binding sites have been recently shown to be associated with cyclic GMP-dependent protein kinase (Lincoln et al., 1976). Cyclic GMP-elicited dissociation of regulatory and catalytic subunits of the enzyme was not readily demonstrable, and it may therefore be premature to conclude that the activation of cyclic GMP-dependent kinase is similar in this respect to the activation of cyclic AMP-dependent kinases. Histones are excellent substrates for both cyclic AMP and cyclic GMPdependent protein kinases, but different serine residues in histone were phosphorylated by the two classes of protein kinases (Takai et al., 1975). Cyclic GMP-dependent kinases did not activate phosphorylase 6 kinase. A “stimulatory” modulator protein present in mammalian tissues alters the substrate specificity of cyclic GMP-dependent protein kinases (Donnelly et al., 1973; Kuo et al., 1976a,b; Kuo and Kuo, 1976). With arginine-rich histones or protamine as substrate, kinase activity was stimulated by the modulator protein, while with a histone mixture the kinase activity was somewhat inhibited. An inhibitory modulator protein for cyclic GMP-dependent protein kinase has recently been reported in rat cerebellum (Szmigielski and Guidotti, 1976). T h e physiological substrates for cyclic GMP-dependent protein kinases in brain tissue are unknown. In smooth muscle, but not apparently in other tissues, cyclic GMP stimulated phosphorylation of two membrane proteins (Casnellie and Greengard, 1974; Greengard, 1976).

IV. Cyclic Nucleotides and the Function of the Central and Peripheral Nervous System

Neurochemists have now provided a wealth of data on the various parameters related to function of cyclic AMP and cyclic GMP systems in the nervous system. It has, however, in brain, as in other even less complex tissues, proven immensely difficult to correlate changes in cyclic nucleotide levels with changes in physiological function. In nervous tissue, such correlations have been attempted at the level of biochemical changes in homogenates, intact cells, or even brain itself, at the level of neurophysiological changes in neurons or glia, and at the levels of brain pharmacology and behavior (cf. Daly, 1977). In view of the complexity and difficulties involved in establishing such correlations with cyclic nucleotide systems, only a brief overview of current areas of research will be attempted in the present review.

CYCLIC NUCLEOTIDES I N THE NERVOUS SYSTEM

145

A. BIOCHEMISTRY Cyclic AMP in isolated systems clearly has effects on at least three biochemical parameters: ( 1) intermediary metabolism through activation of phosphorylase b kinase and inactivation of glycogen synthetase I, thereby increasing glycolysis and glycogenolysis; (2) neurotransmitter pathways through activation of biosynthetic enzymes such as tyrosine hydroxylase; (3) cyclic AMP systems through feedback control of adenylate cyclase, phosphodiesterase, and protein kinases. Certain of these cyclic AMP-dependent effects appear to involve alterations in RNA and protein synthesis, but such alterations have not been defined in brain. Ribosomes did serve as substrates of cyclic AMP-dependent protein kinase (Schmidt and Sokoloff, 1973). Dibutyryl cyclic AMP, administered intraventricularly, did alter labeling patterns of brain RNA in goldfish (Shashoua, 1971). Stabilization of microtubules by cyclic AMPdependent mechanisms has been proposed (see Section I1,H). Activation of Na+-K+-ATPaseby cyclic AMP-dependent mechanisms has been proposed (cf. Phillis, 1976) and would be relevant to inhibitory effects of amines and cyclic AMP on neurons and perhaps to proposed roles for cyclic AMP in the control of transmitter release (see below). 1. Intermediary Metabolism One of the problems inherent in attempts to correlate changes in glycolysis or glycogenolysis with changes in cyclic AMP levels is that small increases in cyclic AMP often appear to evoke full activation of the phosphorylase kinase and inactivation of glycogen synthetase. This problem has rendered metabolic correlations in brain tissue particularly difficult, since sacrifice of animals results in large increases in brain cyclic AMP. In hypothermic mice, postdecapitation increases in cyclic AMP and phosphorylase a and decreases in glycogen synthetase have been shown to follow similar time courses (Lust and Passonneau, 1976). Simi-larly, the increases in brain cyclic AMP elicited by convulsions, stab wound trauma, and ischemia were accompanied by the expected changes in metabolic parameters (Watanabe and Passonneau, 1974; Folbergrova, 1975; Lust and Passonneau, 1976; Mrsulja et al., 1976; Watanabe and Ishii, 1976). Elevations of brain cyclic AMP under such conditions probably are due in large part of adenosine-dependent activation of cyclic AMP-generating systems. In slices of rat caudate nucleus both isoproterenol and 2-chloroadenosine increased glycogenolysis (Wilkening and Makman, 1976). In neonatal chicks elevations of cyclic AMP in brain elicited by intravenous P-adrenergic agonists were correlated with increased .glycogenolysis (Edwards et al., 1974; Nahorski et al.,

146

JOHN W. DALY

1975a). In contrast, elevations of cyclic AMP elicited in chick brain by intravenous histamine were blocked completely by an Hi?-antagonist,but the increases in glycogenolysis elicited by histamine were only partially blocked. In sum, the results indicate that amine and adenosine-sensitive cyclic AMP mechanisms have significant roles in the regulation of central glucose metabolism. The sites-neurone or glia-at which such roles pertain are as yet unknown. 2. Neurotransmitter Metabolism Activation of tyrosine hydroxylase by cyclic AMP-dependent mechanisms has been extensively investigated. The mechanism involved appears to involve a cyclic AMP-dependent phosphorylation of a protein, which then activates tyrosine hydroxylase by increasing the enzyme’s affinity for tetrahydropteridine cofactor and reducing its affinity for inhibitory catecholamines (cf. Lloyd and Kaufman, 1975; Lovenberg et al., 1975). Activation of tyrosine hydroxylase by this mechanism has been studied primarily with synaptosome preparations, but also appears to pertain in vivo in brain (cf. Lovenberg and Bruckwick, 1975; Roth et al., 1975; Zivkovic et al., 1975, 1976). Activation of phenylalanine hydroxylase by cyclic AMP was, in contrast, due to direct phosphorylation of the enzyme (Milstien et al., 1976). In superior cervical ganglia, cyclic AMP-dependent activation of tyrosine hydroxylase required protein synthesis (Mackay and Iversen, 1972). In vivo P-adrenergic agonists induced tyrosine hydroxylase activity in ganglia apparently through a low, sustained increase in cyclic AMP, followed by enhanced synthesis of the enzyme (Hanbauer et al., 1975a,b). Effects of cyclic AMP on metabolism of other neurotransmitters or neuromodulators are much more poorly defined (cf. Daly, 1977).

3 . Adaptation of Cyclic AMP Systems T h e adaptive changes in cyclic AMP systems which occur as a result of changes in synaptic input to receptors controlling cyclic AMP generation in brain have been recently reviewed (Dismukes and Daly, 1976b; Daly, 1977). It is clear that as a result of reductions in neurotransmitter input to a cyclic AMP systems, the system in many cases attempts an adaptation by becoming “supersensitive” to the particular neurotransmitter. Conversely, increases in synaptic input to cyclic AMP systems in brain have often been followed by a compensatory “subsensitivity” of the system. “Supersensitivity” to the neurotransmitter may be due to increases in neurotransmitter-sensitive adenylate cyclase activity, reductions in associated phosphodiesterase activity, or increases in cyclic AMP-dependent protein kinase activity. Reductions in neurotransmitter input to cyclic

CYCLIC NUCLEOTIDES I N T H E NERVOUS SYSTEM

147

AMP systems in brain can be evoked by various manipulations including (1) lesions, either physical or chemical, of neuronal pathways: (2) inhibition of neurotransmitter synthesis: (3) blockade of neurotransmitter receptors: and (4) treatment with drugs or manipulations which reduce levels or turnover of a neurotransmitter. A few recent examples are illustrative. Pretreatment of rats with 6-hydroxydopamine to destroy noradrenergic terminals resulted in enhanced responsiveness of norepinephrine-sensitive cyclic AMP-generating systems in limbic forebrain (Blumberg et al., 1976; Vetulani el nl., 1976). Similar adaptation of norepinephrine-sensitive cyclic AMP-generating systems after 6-hydroxydopamine treatment has been found in cerebral cortex of Sprague-Dawley rats but not in cerebral cortex of Fisher 344 rats (Skolnick and Daly, 1977; see also Section 11, A) or in cerebral cortex of guinea pig (Dismukes et al., 1976b). Furthermore, although central dopaminergic mechanisms clearly become supersensitive after lesions of the nigrostriatal dopaminergic pathways, after chronic administration of dopamine antagonists, after inhibition of dopamine synthesis with a-methyltyrosine, and after reduction of dopamine levels with reserpine, levels or responses of dopamine-sensitive cyclases of‘ striatum did not appear to be increased after such treatments (Von Voightlander et al., 1973, 1975; Rotrosen et al., 1975; Biggio et al., 1976; Krueger et nl., 1976). Two other groups, however, reported the expected increases in dopamine-sensitive cyclases after lesions (Mishra et al., 1974: Gardner el al., 1976) or treatment with a dopamine antagonist (Iwatsubo and Clouet, 1975). The reason for lack of apparent adaptation of cyclic AMP-generating systems in certain species, strains, or brain regions requires further investigation. It is possible that adaptation may occur at the level of protein kinase or its physiological substrates (cf. Romero and Axelrod, 1975; Routtenberg et al., 19’75). Subsensitivity of cyclic AMP systems to a neurotransmitter may be due to decreases in neurotransmitter-sensitive adenylate cyclase activity, increases in associated phosphodiesterase activity, or decreases in cyclic AMP-dependent protein kinase activity. Increases in neurotransmitter input to cyclic AMP systems can be evoked by various means including: (1) electrical stimulation of nerwonal pathways; (2) augmentation of neurotransmitter synthesis; (3) stimulation of receptors by “false transmitters”; (4) treatment with drugs or environmental manipulations which cause augmented release of the neurotransmitter; and ( 5 ) treatment with drugs which prevent the inactivation of the transmitter by either reuptake or metabolism. For example, recent studies have demonstrated that chronic treatment of rats with compounds such as imipramine, desipramine, or chlorpromazine, which prevent inactivation of’ norepinephrine by blocking reuptake into presynaptic terminals, re-

148

JOHN W. DALY

sulted in a significant reduction in the responses of cyclic AMPgenerating systems to norepinephrine in slices of cerebral cortex (Frazer et al., 1974; Schultz, 1976) or limbic forebrain (Vetulani and Sulser, 1975; Vetulani et al., 1976). Chronic treatment of mice with amphetamine, a norepinephrine-releasing agent, led to a similar reduction in responsiveness of cerebral cortical norepinephrine-sensitive cyclic AMP-generating systems (Martres et al., 1975). Such adaptive changes in cyclic AMP systems after drug or environmental manipulations can be expected to provide valuable insights into the mechanism of action of the drug or to the effect of environmental manipulations on turnover of central neurotransmitters. For example, chronic electroconvulsive shock treatments, known to increase norepinephrine turnover, resulted in rat limbic forebrain in a reduction in responsiveness of norepinephrine-sensitive cyclic AMP-generating systems (Vetulani and Sulser, 1975; Vetulani et al., 1976). Surprisingly, electroshock treatment reduced the responsiveness to norepinephrine even in 6-hydroxydopamine-treated rats. In such animals presynaptic noradrenergic terminals have been destroyed so that shock treatment could not be causing enhanced release of nonexistent norepinephrine. The effects of drug treatment of animals on the responsiveness of brain cyclic AMP-generating systems have been studied in detail for ethanol and morphine. For ethanol, the adaptive changes in responsiveness of cyclic AMP-generating systems were consonant with enhanced turnover of norepinephrine during chronic treatment with ethanol and with reduced turnover of norepinephrine during withdrawal (French and Palmer, 1973; French et al., 1974, 1975). For morphine the results have been rather inconsistent (cf. Daly, 1977).Recent studies have not clarified the situation with one report of an enhanced level of striatal adenylate cyclase during chronic morphine treatment (Merali et al., 1976), and another report that chronic morphine treatment had no effect on basal or dopamine-sensitive adenylate cyclase in striatal homogenates (Bosse and Kuschinsky, 1976). Supersensitivity of prostaglandin-sensitive cyclic AMP-generating systems has been demonstrated in cultured neuroma cells after morphine treatment (Sharma et al., 1975; Traber et al., 1975). The responsiveness of prostaglandin-sensitive cyclic AMP-generating systems in cerebral slices was increased after rearing rats under conditions of environmental impoverishment (Dismukes and Daly, 1976a).

B. NEUROPHYSIOLOGY The effects of neurotransmitters known to increase cyclic AMP levels and of cyclic AMP analogs on ( 1 ) spontaneous and evoked firing of

CYCLIC NUCLEOTIDES IN THE NERVOUS SYSTEM

149

central neurons; (2) transmission or spontaneous firing in ganglion preparations; and (3) evoked or spontaneous transmitter release in peripheral neurons have been studied extensively. Effects of neurotransmitters known to increase cyclic GMP levels and of cyclic GMP analogs have been studied to a lesser extent. 1 . Cyclic A M P and Central Neurons I n cerebellar Purkinje cells and in a variety of other central neurons, norepinephrine-elicited accumulations of cyclic AMP have been linked to inhibitory effects on spontaneous firing of the neurons. These extensive studies have been recently reviewed (Bloom, 1975; Daly, 1977). Norepinephrine and cyclic AMP analogs caused an increase in membrane resistance and a hyperpolarization of central neurons, probably via a phosphorylation of a membrane protein associated with ionic channels. Such a phosphorylation might have caused hyperpolarization and an increase in membrane resistance by either activating Na’-K+-ATPase or by reducing passive membrane conductances for sodium ions or calcium ions. Catecholamines have been reported to activate Na+-K+-ATPase (Yoshimura, 1973; Gilbert et al., 1975; Logan and O’Donovan, 1976). Ouabain and other inhibitors of the ATPase have been reported to antagonize catecholamine-elicited inhibitions of central neurons (Phillis, 1976; Yarbrough, 1976). Although a variety of evidence implicates cyclic AMP in the inhibition of central neurons by biogenic amines, it should be noted that calcium ions are inhibitory to central neurons and that the norepinephrine-elicited inhibition of central neurons can be antagonized by “calcium antagonists,” leading to the postulate that the inhibitory effects of norepinephrine are dependent on extracellular calcium ions (cf. Phillis, 1976). Studies by another group (Freedman et al., 1975; Nathanson et al., 1976) have, however, provided evidence indicating that “calcium antagonists” either d o not block the inhibitory effects of norepinephrine on Purkinje cells, or in the case of lanthanum do so by direct inhibition of norepinephrine-sensitive adenylate cyclases. T h e inhibition of firing of caudate neurons by dopamine would appear, like effects of norepinephrine on central neurons, to involve a cyclic AMPdependent mechanism (Siggins et al., 1974; Bloom, 1975).Other amines, such as histamine and serotonin, inhibit firing of certain central neurons, but the relationship of these inhibitions to cyclic AMP mechanisms has not been established. Adenosine, 5’-AMP, and certain other adenine nucleotides have recently been found to be potent inhibitors of central neurons (Phillis et al., 1974, 1975; Phillis and Kostopoulos, 1975). Adenosine and 5’-AMP are, of course, effective stimulants of cyclic AMP generation in brain slices. In a slice of guinea pig olfactory cortex,

150

JOHN W . DALY

adenosine and adenine nucleotide stimulated cyclic AMP generation and inhibited postsynaptic potentials evoked by stimulation of the olfactory tract (Kuroda and Kobayashi, 1975; Okada and Kuroda, 1975; Kuroda et al., 1976a,b). Thus, the adenosine-elicited accumulations of cyclic AMP, presumably at postsynaptic sites, appeared to inhibit synaptic transmission in this brain region. Adenosine had no effect on synaptic transmission in a slice of superior colliculus. Neurotransmitters, in particular norepinephrine, have been proposed to have effects in brain on both neurons and glia. It should, therefore, be noted that norepinephrine and other putative neurotransmitters have no effect on membrane potentials of glial cells (Wardell, 1966; Krnjevic and Schwartz, 1967; Hosli et al., 1976). Thus, if norepinephrine does elicit accumulations of cyclic AMP in glial cells, cyclic AMP does not appear to affect glial membrane potentials as it does neuronal potentials.

2 . Cyclic GMP in Central Neurons Recent data suggested that the activation of firing of rat cerebral cortical neurons by acetylcholine occurred via interaction with a muscarinic receptor and was mediated by cyclic GMP (Stone et al., 1975). Cyclic GMP excited many of the cortical neurons. Interestingly the same neuron was often excited by cyclic GMP or acetylcholine and inhibited by dibutyryl cyclic AMP or norepinephrine. In studies by another group, both cyclic AMP and cyclic GMP inhibited firing of cortical neurons (Phillis et al., 1974). A role for cyclic GMP in the effects of acetylcholine on spinal motor neurons has been questioned (Krnjevic et al., 1976).

3. Cyclic A M P in Ganglia In rabbit and bovine superior cervical ganglia release of dopamine from interneurons and dopamine-elicited accumulations of cyclic AMP on postganglionic neurons appeared to be responsible for the slow inhibitory postsynaptic potentials (cf. Greengard and Kebabian, 1974). It appears likely that norepinephrine is the inhibitory neurotransmitter in rat superior cervical ganglion (see Section 11, B). In cat, dopamine appeared likely to be the inhibitory neurotransmitter (Machova and Kristofova, 1973). In the ganglia of molluscs, peptides which elicit accumulations of cyclic AMP and cyclic AMP analogs caused marked hyperpolarizations in ganglion neurons during silent periods between bursts of electrical activity (Treistman and Levitan, 1976). Intracellular injection of cyclic AMP into neurons of the mollusc, Helix pomatia, has been reported to cause depolarization and enhanced spontaneous firing (Liberman et al., 1975).

CYCLIC NUCLEOTIDES IN T H E NERVOCJS SYSTEM

151

4. Cyclic GMP in Ganglia In the superior cervical ganglion, acetylcholine via interaction with muscarinic receptors on postganglionic neurons appears to elicit an accumulation of cyclic GMP which is responsible for the slow excitatory postsynaptic potentials (cf. Greengard and Kebabian, 1974). A prior treatment of rabbit superior cervical ganglion with dopamine or dibutyryl cyclic AMP had prolonged facilitative effects on the slow excitatory postsynaptic potentials (Libet and Tosaka, 1970; Libet et al., 1976). Thus, activation of synaptic cyclic AMP-dependent mechanisms appeared to have long-term facilitative effects on cyclic GMP-dependent neurotransmission. In bullfrog sympathetic ganglia, cyclic GMP has been proposed to have a role in the inhibition of post-tetanic potentiation of postsynaptic potentials by the phosphodiesterase inhibitors isobutylmethylxanthine and diazepam (Suria, 1976).

5 . Cyclic Nucleotides in Peripheral N m e s As yet, conclusive effects of cyclic nucleotides on axonal conduction or on membrane properties of axons have not been demonstrated. In one report, cyclic AMP was found to be inhibitory to generation and transmission of action potentials in frog sciatic nerve (Van d e Berg, 1974), while in another report neither cyclic AMP, cyclic GMP, nor stimulation of cyclic nucleotide-generating systems had any effect on membrane properties or action potentials in desheated frog sciatic nerve (Horn and McAfee, 1976). 6. Cyclic A M P and Transmitter Release

A possible facilitative role for cyclic AMP in the release of acetylcholine or norepinephrine from peripheral neurons has been extensively investigated during the ten years since its initial proposal (Breckenridge et al., 1967; cf. review by Daly, 1977). Neither norepinephrine nor cyclic AMP analogs had any effect on spontaneous release of acetylcholine from the rat phrenic nerve diaphragm preparation (Miyamoto and Breckenridge, 1974). However, acetylcholine release in depolarized preparations was enhanced by a cyclic AMP analog, and it was proposed that cyclic AMP-dependent mechanisms might increase availability of calcium in stimulated preparations, thereby enhancing transmitter release. Recent data on facilitative effects of dibutyryl cyclic AMP on neuromuscular transmission in vizjo were consonant with this interpretation (Standaert et al., 1976). Another phosphodiesterase inhibitor, SQ 20009, has recently been reported to facilitate neuromuscular transmission (McNiece and Jacobs, 1976). The inhibition of neuromuscular

152

JOHN

W. DALY

transmission by adenosine (Ginsborg and Hirst, 1972; Miyamoto and Breckenridge, 1974) has been interpreted as due to inhibition of cyclic AMP-dependent protein kinase (Dretchen et al., 1976). Adenosine has recently been reported to inhibit release of norepinephrine from sympathetic neurons (Hedqvist and Fredholm, 1976). Stimulus-evoked release of norepinephrine in spleen and vas deferens was enhanced by cyclic AMP analogs and by phosphodiesterase inhibitors (Cubeddu et al., 1974, 1975; Wooten et al., 1973). However, in another report neither dibutyryl cyclic AMP nor theophylline appeared to have any effect on stimulus-evoked release of norepinephrine in vas deferens (Stjarne, 1976). Activation of Na+-K+-ATPase has been proposed to play an important role in control of transmitter release via interaction of norepinephrine with a presynaptic a-adrenergic receptor (Gilbert et al., 1975). The activation of ATPase by norepinephrine has been considered as involving cyclic AMP mechanisms (see above), but as yet no evidence linking presynaptic a-adrenergic receptors with cyclic AMP-generating systems has been provided (cf. Skolnick and Daly, 1975c, 1976a). C. CENTRAL PHARMACOLOGY, BEHAVIOR, AND CYCLICNUCLEOTIDES The effects of various centrally active drugs on levels of cyclic nucleotides in brain have been studied extensively with the hope that correlations of central activity and alterations in levels of cyclic AMP and cyclic GMP would become apparent. Levels of cyclic nucleotides have been measured in specific brain regions, in cerebrospinal fluid, and in urine after drug treatments, after environmental manipulations, in different strains of rats or mice, and in patients with mental disorders. I n addition, the effects of intraventricular or intracerebral administration of cyclic nucleotides on behavior, on pharmacological responses to drugs, and on central vegetative functions have been investigated. These attempts to relate cyclic nucleotide mechanisms to the pharmacology of centrally active drugs and to behavior have been comprehensively reviewed (Daly, 1977), and only a few salient results will be mentioned in the present review. 1. Neurotransmitters The various neurotransmitters known to stimulate cyclic AMP or cyclic GMP formation in brain have quite different gross effects on behavior. Catecholamines are central stimulants (Laverty, 1975). Intraventricular norepinephrine was more effective in eliciting spontaneous motor activity in a rat strain with a highly responsive norepinephrinesensitive subocortical cyclic AMP-generating system than in a strain with a less responsive cyclic AMP-generating system (Segal at al., 1975).

CYCLIC NUCLEOTIDES I N THE NERVOUS SYSTEM

153

Dopamine administered to the nucleus accumbens elicited increases in locomotor activity as did activation of adenylate cyclases in this brain region by injection of cholera toxin (Miller and Kelly, 1975; Iversen et al., 1975). The behavioral effects of histamine (Monnier and Hall, 1969; Asakawa and Yoshida, 1971; Schwartz et al., 1974),serotonin (Chase and Murphy, 1973), and cholinergic agonists (Pradhan and Dutta, 1971) are complex and difficult to categorize in specific terms. Histamine and serotonin have inhibitory effects on many central neurons, while acetylcholine usually has excitatory effects. Prostaglandin is a central depressant (Horton, 1964). Transient increases in brain levels of cyclic AMP elicited by prostaglandin El were not, however, correlated with the long-lasting sedation (Wellmann and Schwabe, 1973). Adenosine, administered parenterally or centrally, has profound sedative effects (Marley and Nistico, 1972; Haulica et al., 1973; Maitre et al., 1974). Glutamate, an excitatory neurotransmitter, elevated cerebellar levels of cyclic GMP and is a potent convulsant (Guidotti et al., 1975; Biggio and Guidotti, 1976b). The inhibitory neurotransmitter, glycine, elevated cerebellar levels of cyclic GMP, while another inhibitory neurotransmitter, y-aminobutyrate, reduced cerebellar levels of cyclic GMP and increased levels of cyclic AMP (Ma0 et al., 1974; Guidotti et al., 1975). Antagonists of glycine and y-aminobutyrate and drugs which deplete y-aminobu ty rate have convu lsant activity . 2. Phosphodiesterase Inhibitors Except for the central stimulants, theophylline and caffeine, which are, of course, very weak inhibitors of brain phosphodiesterases, the various phosphodiesterase inhibitors such as papaverine, diazepam, chlorpromazine, RO 20-1724, ZK 6271 1, and SQ 20009 are central depressants (cf. Beer et al., 1972; Schwabe et al., 1976). The effect of a variety of phosphodiesterase inhibitors on turning behavior in rats with unilateral lesions of the substantia nigra has been investigated (Fredholm et al., 1976). This turning behavior is thought to be dependent on activation of striatal dopamine-sensitive adenylate cyclase. Isobutymethylxanthine and a phenylazapurinone greatly potentiated dopa-elicited turning, while dipyridamole and theophylline were less effective and papaverine had no effect. In an earlier study, another group reported that diazepam, chlordiazepoxide, and a triazolopyrimidine did not potentiate apomorphine-induced turning (Arbuthnott et al., 1974). 3. Cyclic Nucleotide Analogs I n most studies intraventricular or intracerebral administration of cyclic AMP or of cyclic AMP analogs resulted in increases in locomotor activity, gross excitation, and often convulsions (cf. Daly, 1977). Di-

154

JOHN W. DALY

bu tyryl cyclic AMP shortened amobarbital-induced narcosis (Kraynack et al., 1976 and references therein) and increased the rate of development of tolerance to ethanol (Wahlstrom, 1975) and to morphine (Ho et al., 1973, 1975). Dibutyryl cyclic AMP antagonized the analgesia elicited by morphine. It is noteworthy that peripheral administration of a cyclic GMP analog, 8-bromo cyclic GMP, caused a 2-fold increase in levels of cyclic AMP in brain (Fernandez-Pol and Hays, 1976). Such effects on cyclic AMP levels could confound any interpretation of the pharmacology of cyclic GMP analogs. 4. Central Stimulants Various central stimulants such as apomorphine, dopa, amphetamine, and tricyclic antidepressants, whose activity is linked to enhanced stimulation of catecholamine receptors, have been reported to elevate cyclic AMP and often cyclic GMP in brain or in specific brain regions (cf. Daly, 197’7). The apomorphine-elicited increases in cyclic GMP levels in rat cerebellum were proposed to be due to enhanced excitatory input to the cerebellum due to activation of striatal dopaminergic receptors (Burkard et al., 1976). The increase in levels of cyclic AMP in rat cerebrospinal fluid elicited by dopa appeared primarily mediated by activation of central /3-adrenergic rather than dopaminergic receptors (Cramer and Kiessling, 1976). Convulsants such as isoniazid, picrotoxin, pentylenetetrazole, glutamate, and tremorigenic agents such as oxotremorine and harmaline elevated cyclic GMP levels, particularly in cerebellum. Oxotremorine and harmaline apparently increase excitatory input into the cerebellum, thereby elevating cyclic GMP levels (cf. Ferrendelli et al., 1970, 19’72; Guidotti et al., 1975; Biggio and Guidotti, 1976b; Opmeer et al., 1976). I t would appear the y-aminobutyrate, an inhibitory neurotransmitter, is inhibitory to cyclic GMPgenerating systems, perhaps via an inhibition of release or action of excitatory neurotransmitters (cf. Suria, 1976). 5. Central Depressants Ethanol had minimal effects on central levels of cyclic AMP, but caused marked reductions in levels of cyclic GMP (Redos et al., 1976a,b). Other central depressants such as barbiturates, papaverine, reserpine, and chlorpromazine reduced central levels, particularly cerebellar levels, of cyclic GMP (Ferrendelli et al., 1972: Kimura et al., 1974; Lust et al., 1976; Opmeer et al., 1976). Morphine reduced levels of cyclic GMP in cerebellum (Lust et al., 1976), apparently due to decreased mossy fiber input to the cerebellum (Biggio and Guidotti, 1976a). Morphine had, in various studies, somewhat inconsistent effects on levels of cyclic AMP in brain. However, morphine has been reported by a number of groups to

CYCLIC NUCLEOTIDES IN T H E NERVOLJS SYSTEM

155

elicit increases of cyclic AMP in striatum (cf. Bonnet, 1975: Clouet et al., 1975; Merali et al., 1975). Such morphine-elicited increases in cyclic AMP levels contrast with inhibitory effects of morphine on cyclic AMP generation in homogenates or cultured cells (see Section 11, A). 6. Behavioral Correlates It should be obvious that many drugs affect such a variety of cyclic AMP-dependent inhibitory pathways andlor cyclic GMP-dependent pathways in the central nervous system so as to make impossible the interpretation of the behavioral results. Certainly, this will be is the case for cyclic AMP analogs and for phosphodiesterase inhibitors. Activation of adenylate cyclase by injection of cholera toxin into specific brain regions would appear to be a somewhat more selective approach. Another type of approach has been to attempt to correlate behavioral parameters with (1) responses of cyclic AMP-generating systems in specific brain regions of different rat or mouse strains (Skolnick and Daly, 1974a, 1975b; Williams and Pirch, 1974; Sattin, 1975; Stalvey et al., 1976); ( 2 ) levels of cyclic AMP in brains of mouse strains (Barchas et al., 1974; Orenberg et al., 1975); and (3) cyclic AMP-dependent phosphorylation of synaptosome proteins from rat strains (Ehrlich and Brunngraber, 1976). In the last study, autophosphorylation of a synaptosome protein with a molecular weight of 49,000, presumably the regulatory unit cor cyclic AMP-dependent protein kinase, was found to be high in striatal synaptosomes from a strain of rat with high spontaneous behavioral activity and low in a less active strain of rat. Autophosphorylation of this protein in cortical synaptosomes was low in the behaviorally active strain of rat and high in the less active strain of rat. These results in conjunction with data on responsiveness of norepinephrine-sensitive cyclic AMP-generating systems from the two rat strains (Skolnick and Daly, 1974a, 1975b) suggest that at least in these two rat strains, brain regions with highly responsive norepinephrine-sensitive cyclic AMP-generating systems have associated a high capacity for autophosphorylation of a synaptosome protein. Although the various studies cited above have provided a number of interesting examples of apparent correlations between particular behaviors and functions of the cyclic AMP systems, much more detailed studies of this type will be required if significant insights into the specific roles of cyclic nucleotide systems in complex behaviors are to be established. 7. Clinical Correlates During the past six years many groups have reported correlations or lack of correlations between clinical state of patients with mental disorders, in particular manic-depressive syndrome, and cyclic AMP levels in

156

JOHN W. DALY

urine or cerebrospinal fluid (cf. Daly, 1977). Recent studies provide further data which show no correlation between mental dysfunction and cyclic AMP levels in cerebrospinal fluid (Smith et al., 1976) or urine (Moyes and Moyes, 1976). V. Summary

Biochemical studies have provided evidence that cyclic AMP systems are associated in brain and ganglia with postsynaptic sites on neurons. Neurophysiological studies indicate that such cyclic AMP systems are involved in inhibitory neurotransmission for noradrenergic, dopaminergic, serotoninergic, and histaminergic pathways. Adenosine appears to serve as an inhibitory neuromodulator, probably via interaction with the postsynaptic cyclic AMP-generating systems. Behavioral studies suggest that cyclic AMP mechanisms may be involved in regulation of certain complex behaviors. The results do not preclude significant roles for cyclic AMP at presynaptic sites regulating neurotransmitter formation and/or release or at glial sites regulating perhaps metabolic events. Biochemical studies have provided evidence that cyclic GMP systems are associated in brain and ganglia with both presynaptic and postsynaptic sites. Neurophysiological studies indicate that the postsynaptic cyclic GMP systems are involved in responses to excitatory neurotransmitters such as acetycholine and glutamate. Roles for cyclic GMP systems at presynaptic sites are less well defined but might involve effects on release of transmitters. Cyclic GMP-generating systems appear to be associated with neurones, but their presence in glia cannot be excluded. Behavioral studies on the role of cyclic GMP have been hindered by lack of specific agents for activation or inhibition of the cyclic GMP-generating systems. The results obtained during the past decade and a half thus provide a solid basis for further studies on the precise nature and role of cyclic AMP and cyclic GMP as regulatory messengers in the central and peripheral nervous systems. REFERENCES Adinolfi, A. M., and Schmidt, S. Y . (1974). Brain Res. 76, 21-31. Arbuthnott, G . S., Attree, T.J., Eccleston, D., Loose, R. W., and Martin, M. J. (1974).Med. Biol. 52, 350-353. Asakawa, T., and Yoshida, H. (1971).Jap.J. Pharmatol. 21,569-583. Audiger, Y . , Virion, A., and Schwartz, J.-C. (1976).Nature (London) 262,307-308. Barchas,J. D., Ciaranello, R. D., Dominic, J. A., Deguchi, T.,Orenberg, E., Renson,J., and Kessler, S. (1974).Adu. Biochem. Psychophannacol. 12, 195-215. Baudry, M., Martres, M. P., and Schwartz,J.-C. (1975).Nature (London) 253,362-363. Beavo, J. A., Hardrnan,J. G., and Sutherland, E. W. ( 1 970).J . Baol. Chem. 245,5649-5655.

CYCLIC NUCLEOTIDES IN THE NERVOUS SYSTEM

157

Beavo,J. A., Hardman, J. G., and Sutherland, E. W. ( 1 97 l).J. Biol. C h . 246,384 1-3846. Beer, B., Chasin, M., Clody, D. E., Vogel, J. R., and Horovitz, 2. P. (1972). Science 176, 428-430. Bensinger, R. E., Fletcher, R. T., and Chader, G. J. (1974). Science 183, 86-87. Berti, F., Trahucchi, M., Bernareggi, V., and Fumagalli, R. (1972). Phannacol. Res. Commun. 4, 253-259. Biggio, G., and Guidotti, A. (1976a). Pharmncologut 18, 212 (Ahstr.) Biggio, G., and Guidotti, A. (1976b). Brain Res. 107, 365-374. Biggio, G., Porceddu, M. L., and Gessa, G. L. (1976).J. Neurochem. 26, 1253-1256. Bloom, F. (1975). Rev. Physiol. Biochem. Phannacol. 74, 1-104. Bloom, F. E., Hoffer, B.J., Battenberg, E. R., Siggins, G. R., Steiner, A. L., Parker, C. W., and Wedner, H. J. (1972). Science 177,436-438. Blumberg, J. B., Vetulani, J., Stawarz, R. M., and Sulser, F. (1976).Eur.J. Phannacol. 37, 357-366. Blume, A. J., and Foster, C. J. (1975).J. B i d . Chem. 250,5003-5008. Blume, A. J., and Foster, C. J. (1976a).J. Bwl. C h . 251, 3399-3404. Blume, A. J., and Foster, C. J. (1976b).J. Neurochem. 26, 305-312. Boehme, E. (1970). E u r . J . Biachem. 14,422-429. Bonnet, K . A. (1975). Lye Sci. 16, 1877-1882. Bosse, A., and Kuschinsky, K. (1976). Naunyn-Schmiedeberg’s Arch. Pharmacol. 294, 17-22. Boudreau, R.J., and Drummond, D. I. (1975).J. Cyclic Nucleotide Res. 1, 219-228. Bray, J. J., Kon, C. M., and Breckenridge, B. M. (1971). Brain Res. 26, 385-394. Breckenridge, B. M., and Johnston, R. E. (1969).J. H i r t o c h . Cytochem. 17,505-511. Breckenridge, B. M., Burn, J . H., and Matschinsky, F. M. (1967). Proc. Natl. Acad. Sci. U.S.A. 57, 1893-1897. Brostrom, C. O., and Wolff, D. J. (1976).Arch. Biochem.Biophys. 172, 301-311. Brostrom, C . O., Huang, Y.-C., Breckenridge, B. M., and Wolff, D. J. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 64-68. Browning, E. T., Brostrom, C. O., and Groppi, V. E. (1976). Mol. Phannacol. 12,32-40. Burkard, W. P., and Gey, K. F. (1968). Helv. Physiol. Acta 26, 197-198. Burkard, W. P., Pieri, L., and Haefely, W. (1976).J. Neurochem. 27, 297-298. Carenzi, A., Chency, D. L., Costa, E., Guidotti, A., and Racagni, G. (1975).NeuropharmncolOQ 14, 927-940. Casnellie, J. E., and Greengard, P. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 1891-1895. Cedar, H., and Schwartz, J. H. (1972).J . Gen. Physiol. 60, 570-587. Cedar, H., Kandel, E. R., and Schwartz, J. H. (1972).J. Gen. Physiol. 60, 558-569. Chase, T. N., and Murphy, D. L. (1973). Annu. Rev. Phannacol. 13, 181-197. Chasin, M., Rivkin, I., Mamrak, F., Samaniego, G., and Hess, S. M. (1971).J. Biol. Chem. 246,3037-3041. Chasin, M., Mamrak, F., Samaniego, S. G., and Hess, S. M. (1973). J . Neurochem. 21, 1415-1427. Chasin, M., Mamrak, F., and Samaniego, S. G. (1974).J. Neurochem. 22, 1031-1038. Cheung, W. Y. (1970). Adv. Biochem. Psychophannncol. 3, 5 1 4 5 . Cheung, W. Y. (197 1 ) . Biochim. Biophys. Acta. 242, 395-409. Cheung, W. Y., Bradham, L. S., Lynch, T.J., Lin, Y. M., and Tallant, E. A. (1975a).Biochem. Biophys. Res. C a m u n . 66, 1055-1062. Cheung, W. Y., Lin, Y. M., Liu, Y. P., and Smoake, J. A. (1975h).In “Cyclic Nucleotides in Disease” (B. Weiss, ed.), pp. 321-350. University Park Press, Baltimore, Maryland. Clark, R. B., and Seney, M. N. (1976).J. Biol. Chem. 251, 4239-4246. Clark, R. B., Su, Y.-F., Ortmann, R., Cubeddu, L., Johnson, G. L., and Perkins, J. P. (1975). Metub., Clin. Exp. 24, 343-358.

158

JOHN W. DALY

Clement-Cormier, Y. C., Kebabian, J . W., Petzold, G. L., and Greengard, P. (1974). Proc. Natl. Acad. Sci. U.S.A. 17, 1 1 13-1 1’17. Clouet, D. H., Gold, G . J., and lwatsubo, K. (1975). Br. J. Phanacol. 54, 541-548. Collier, H. 0.J., and Roy, A. C. (1974a). Nature (London) 225, 159-161. Collier, H. 0. J., and Roy, A. C. (197413).Prostaglandinc 7, 361-376. Corbin, J. D., Keely, S. L., and Park, C. R. (1975).J. Biol. Chem. 250,218-225. Costa, E., Guidotti, A., Mao, C. C., and Suria, A. (1975a).Life Sci. 17, 167-186. Costa, E., Guidotti, A., and Mao, C. C. (1975b).Adu. Biochem. Psychopharmacol. 14,113-130. Costa, E., Kurosawa, A.,and Guidotti, A. (1976).Proc. Natl. Acud. Sci. U.S.A. 73,1058-1062. Cramer, H., and Kiessling, M. (1976). Arzneimittefforschung 26, 1106-1107. Cramer, H., and Lindl, T. (1974).Nature (London) 249, 380-382. Cramer, H., Johnson, D. G., Hanbauer, I., Silberstein, S. D., and Kopin, I. J. (1973. Brain Res. 53,97-104. Cubeddu, L., Barnes, E., and Weiner, N. (1974).J . Phmnacol. Exp. Thm. 191, 444-457. Cubeddu, X., Barnes, E., and Weiner, N. (1975). J. Pharmacol. Exp. Ther. 193, 105-127. Dalton, C., Crowley, H. J., Sheppard, H., and Schallek, W. (1974).Proc. SOC.Exp. Biol. Med. 145, 407-4 10. Daly, J. W. (1977). “The Role of Cyclic Nucleotides in the Nervous System.” Plenum, New York. Davis, J. N., and Lefkowitz, R. J. (1976). Brain Res. 113,214-218. De Belleroche, J. S., Das, I . , and Bradford, H. F. (1974). B i o c h . Pharmacol. 23,835-843. DeLorenzo, R. J., (1976). Biochem. Biophys. Res. Commun. 71, 590-597. DeRobertis, E., Arnaiz, G. R. D.-L., Butcher, R. W., and Sutherland, E. W. (1967).J. Biol. Chem. 242,3487-3493. De Vellis.J., and Brooker, G. (1974). Science 186, 1221-1223. Dismukes, K.. and Daly, J. W. (1974). Mol. Pharmacol. 10, 933-940. Dismukes, K., and Daly, J. W. (1975a).Life Sci. 77, 199-210. Dismukes, R. K., and Daly, J. W. (l975b). Exp. Neurol. 49, 150-160. Dismukes, R. K., and Daly, J . W. (1976a). Experientin 32, 730-731. Dismukes, R. K., and Daly, J. W. (1976b).J. Cyclic Nucleotide Res. 2, 321-336. Dismukes, R. K., Ghosh, P., Creveling, C. R., and Daly, J. W. (1975). Exp. Neurol. 49, 725-735. Dismukes, K., Rogers, M., and Daly, J. W. (1976a).J. Neurochem. 26, 785-790. Dismukes, R. K., Ghosh, P., Creveling, C. R., and Daly, J. W. (1976b). Exp. Neurol. 52, 206-2 15. Donnelly, T. E., Jr., Kuo, J. F., Reyes, P. L., Liu, Y. P., and Greengard, P. (1973).J. Biol. Chem. 248, 190-199. Dretchen, K. L., Standaert, F. G.. Morgenroth, V. H., and Skirboll, L. R. (1976). Pharmacologist 18, 193. (Abstr.) Duffy, M. J., and Powell, D. (1975). Biochim. Bwphys. Acta. 385, 275-280. Dumler, I. L., and Etingof, R. H. (1976). Biochim. Biophp Actcr 429, 474-484. Dumoulin, A., and Schultz, J. (1975). Experientia 31, 883-884. Edwards, C., Nahorski, S. R., and Rogers, K. J. (1974).J. Neurochem. 22,565-572. Ehrlich, Y. H., and Brunngraber, E. G. (1976). Fed. Proc., Fed. Am. SOC.Exp. Biol. 35, 1645. (Abstr.) Ehrlich, Y. H., and Routtenberg, A. (1974). FEBS Lett. 45, 237-243. Erlichman, J. R.,Rosenfeld, R., and Rosen, 0. M. (l974).J. Biol. Chem. 249, 50005003. Fernandez-Pol, J. A., and Hays, M. T. (1976). Life Sci. 19, 35-40. Ferrendelli, J. A. (1975). In “Cyclic Nucleotides in Disease” (B. Weiss, ed.), pp. 377-390. University Park Press, Baltimore, Maryland.

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

159

Femendelli, J. A., Steiner, A. I,., McDougal, D. B., Jr., and Kipnk, D. M. (1970).Biorhem. Biophy~s.Reg. Commirn. 41, 1061-1067. Ferrendelli, J. A., Kinscherf, D. A., and Kipnis, D. M. (1972). Bioclwm. Biophy~.Re$. Commun. 46, 2 I 14-2 120. Ferrendelli, J. A., Kinscherf, D. A., and Chang, M. M. (1973).Mol. Phnrmacol. 9,445-454. Ferrendelli, J . A., Chang, M. M., and Kinscherf, D. A. (1974).J. Nevrorhtm. 22,535-540, Ferrendelli, J. A., Kinscherf', D. A., and Chang, M.-M. (1975). Brain Res. 84, 63-73. Ferrendelli, J. A., Rubin, E. H., and Kinscherf, D. A. (1976).J. Netrruchem. 26, 741 -748. Fertel, R., and Weiss, B. (1974).Anal. Biorhm. 59, 386-398. Florendo, N . T., Barrnett, R. J., and Greengard, P. (1971). S c i m e 173, 745-747. Folbergrova, J. (1975).Brain Re.$. 92, 165-169. Forn, J., Krueger, B. K.. a n d Greengard, P. (1974). Srience 186, 11 18-1 120. Frazer. A., Panday. G., Mendels, J.. Neeley, S., Kane. M., and Hess, M. E. (1974). Neurophnrmmology 13, 1131-1 140. Fredholm, 13. B., Fuxe, K., and Agnati, L. (1976). Eur. J. Phormacol. 38, 31-38. Freedman, R., Hoffer, B. J., and Woodward, D. J. { 1975). Br. J . Phamnrol. 54, 529-539. French, S. W., and Palmer, D. S. (1973).Res. Cummtrn. Chem. Pathol. Phonnarol. 6,661-662. French, S . W., Reid, P. E., Palmer. D. S., Marod, M. E., and Ramey, C. W. (1974). Re.). Comnam. C h m . Paihol. Phamacol. 9, 57.5-578. French, S. W., Palmer, D. S., a n d N a r d , M. E. (1975). Can. J. Pfiysiol. Pharmnrol. 53, 248-255. Furlanut, M., Carpenedo, F., and Ferrari, M. (1973). Biochem. Pharmarol. 22, 2642-2644. Gaballah, S., and Popoff, C. (1971a).Brain Re.<. 25, 220-222. Gaballah, S., and Popoff, C. (1971b).J. Networhem. 18, 1795-1797. Gaballah, S., Popoff', C., and Sooknandan, G. (1971). Brain Reg. 31, 229-232. Gahwiler, B. H. (1976). Brain Reg. 107, 176-179. Gardner, E. L., Mishra, P. K., Pantazis, I . , KatLman, R., a n d Makman, M. H. (1976). Fed. Proc., Fed. Am. Sac. Exp. Biol. 35, 456. (Abstr.) Gilbert, J. C., Wyllie, M. G., and Dayidson, D. V. (1975). Nature (London) 255, 237-238. Gill, G. N., and Kanstein, C. B. (1975). Biorhm. Biophys. Re.s. Cmmirn. 63, 1 1 13-1 122. Gillespie, E. (1971).j. Ctll B i d . 50, 544-549. Gilman, A. G., a n d Schrier, B. K. (1972). Mol. Phnrmtrrol. 8,410-416. Ginsborg, B. L., and Hirst, G. D. (1972).J. Pltysiol. (London) 224, 629-645. Gnegy, M. E., Costa, E., and Uzunov, P. (1976a).Proc. Nntl. A c d . Sci. C!.S.A. 73,332-355. Gnegy, M. E., Costa, E.. and Llziinov, P. (1976b).Fed. Proc., Fed. Am. Soc. Exp. B i d . 35,609. (Abstr.) Gnegy, M., Uzunov, P., and Costa, E. ( 1 9 7 6 ~ )Phnrmctcolopt . 18, 185. (Abstr.) Ckddberg, N. D., Lust, W. D., ODed, R. F., Wei, S., and O'Toole, A. G . (1970). Ad?). Biochtm. Psychophnnnarol. 3, 67-87. Goldberg, N . D., O'Dea, R. F., and Haddox, M. K. (1973). Adit. Cyrlir Nircleotidp Rc.~.3, 155-223. Goldberg, N. D., Haddox, M. K., Nicol, S. E., Glass, D. B., Sanford, C. H.. and Kuehl. F.A,, Jr.. and Estensen, R. (1975). Adif. Cyclic Nirrleotidt Res. 5, 307-330. Goldstein, A. (1976). Science 193, 1081-1086. Coldstein, G., Scheid, M. S., Hammerling, V., Boyse, E. A., Schlesinger, D. H., and Nidll, H. D. (1975).Proc. NaU. A c d . Sci. U.S.A. 72, 11-15. Goodman, D. B. P., Rasmussen, H., DiBella, F., and Guthrow, C. E., Jr. (1970). Pror. Nntl. A c d . Sci. U S A . 67,652-659. Goridis, C., and Morgan, I. G. (1973). FEBS Lett. 34.71-73. Goridis, C., Massarelli, R., Sensenbrenner, M., and Mandel, P. (1974).J. Netrrorhem. 23, 135-138.

160

JOHN W. DALY

Greengard, P. (1976). Nature (Lodon) 260, 101-108. Greengard, P., and Kebabian, J. W. (1974). Fed. Proc., Fed. Am. SOC.Exp. Biol. 33, 10591067. Greengard, P., and Kuo, J. F. (1970). Adz!. Biochem. Psychopharmacol. 3,287-306. Greengard, P., McAfee, D. A., and Kebabian, J. W. (1972). Adv. Cyclic Nucleotide Res. 1, 337-355. Guidotti, A., Biggio, G., and Costa, E. (1975). Brain Res. 96, 201-205. Hanbauer, I., Kopin, J.. Guidotti, A,, and Costa, E. (1975a). J . Pharmacol. Exp. Ther. 193, 95- 104. Hanbauer, I., Lovenberg, W., Guidotti, A., and Costa, E. (1975b).BrainRes. 96, 197-200. Harris, J. E. (1976). Mol. Phannacol. 12, 546-558. Haslam, R. J., and Lynham, J. A. (1972). Life Sci. 11, Part 11, 1143-1154. Haulica, I., Ababei, L., Brainisteanu, D., and Topolinceanu, F. (1973). J . Neurochem. 21, 1019- 1020. Hedqvist. P., and Fredholm, B. B. (1976). Naunyn-Schmkdeberg's Arch. Pharmacol. 293, 217-224. Hegstrand, L. R., Kanof', P. D., and Greengard, P. (1976).Nature (London) 260, 163-164. Hervonen, H., and Rechardt. L. (1976). Hktochemishy 48,43-50. H a s , S. M., Chasin, M., Free, C. A., and Harris, D. N. (1975). Adu. Biochem. Psychopharmacol. 14, 153-167. Hidaka, H., Shibuya, M., Asano, T., and Hara, F. (1975).J. Neurochem. 25, 49-54. Ho, I . K., Loh, H. H., and Way, E. L. (1973).J . Pharmacol. Exp. Ther. 185, 347-357. Ho, 1. K., Loh, H. H., Bhargava, H. N., and Way, E. L. (l975).LiJe &a'. 16, 1895-1900. Hoffer, B. J.. Freedman, R., Woodward, D. J., Daly, J. W., and Skolnick, P. (1976). Exp. Neurol. 51, 653-667. Hofmann, F., and Sold, G. (1972).Biochem. Biophys. Res. Commun. 49, 1100-1107. Horn, A. S., and Phillipson, 0. T. (1976). Eur. J. Pharmacol. 37, 1-11. Horn, A. S., Cuello, A. C., and Miller, R. J . (1974).J. Neurochem. 22,265-2'70. Horn, J. P., and McAfee, D. A. (1976). Fed. Proc., Fed. Am. SOC. Exp. Biol. 35, 424. (Ahstr.) Horton, E. W. (1964).B r . J . Phannacol. Chemother. 22, 189-192. Hosli, L., Andres, P. F., and Hosli, E. (1976). Neurosci. Lett. 2, 223-228. Huang, M., and Daly, J. W. (1972).J. Med. Chm. 15,458462. Huang, M., and Daly, J. W. (1974). Life Sci. 14, 489-503. Hiiang, M., Shimizu, H., and Daly, J. (1971). Mol. Phannacol. 7, 155-162. Huang, M., Gruenstein, E., and Ddy, J. W. (1973a).Biochim.Biophys. Acta. 329, 147-151. Huang, M., Ho, A. K. S., and Daly, J. W. (1973b). Mol. Pharmacol. 9, 711-717. Inoue, Y., Yamamura, H., and Nishizuka, Y. (1973). Eiochtm. Biophys. Res. Camun. 50, 228-236. Iversen, L. (1975). S c i w e 188, 1084-1085. Iversen, L. L., Horn, A. S., and Miller, R. J. (1975).1n "Pre- and Postsynaptic Receptors" (E. Usdin and W. E. Bunney, Jr.. eds.), pp. 207-245. Dekker, New York. Iwatsubo, K., and Clouet, D. H. (1975). Biochem. Phannacol. 24, 1499-1504. Izumi, H., Oyama, H., and Ozawa, H. (1975a).Jap.J. Phannacol. 25, 375-382. Izumi, H., Oyama, H., and Ozawa, H. (1975b). Chem. Phann. Bull. 23,2362-2368. Iziimi, H., Oyama, H., and Ozawa, H. (1976).Chem. Phann. Bull. 24, 1064-1067. Johnson, E. M., Maeno, H., and Greengard, P. (1971).J. Biol. Chem. 246, 7731-7739. Johnson, E. M., Ueda, T., Maeno, H., and Greengard, P. ( 1972).J. Biol. Chem. 247,56505652. Johnson, R. A., and Sutherland, E. W. (1973).J. Biol. Chem. 248, 5114-5121. Joo, F.. Toth, I., and Jansco, G. (1975). Naturwiscenschaften 62, 397-398.

CYCLIC NUCLEOTIDES IN THE NERVOUS SYSTEM

161

Kakiuchi, S., and Rall, T. W. ( 1 968a). Mol. Pharmacol. 4, 379-388. Kakiuchi, S., and Ra!!, T. W. (1968b). Mol. Phunnacol. 4, 367-378. Kakiuchi, S., and Yarnazaki, R. (1970).B i o c h a . Bl0phy.v. Res. C a m u n . 41, 1104-1 110. Kakiuchi, S., Yamazaki, R.,Teshirna, Y., and Uenishi, K. (1973).Proc. Natl. Acad. Sci. U.S.A. 70,3526-3535. Kakiuchi, S., Yamazaki, R., Teshima, Y., Uenishi, K., and Miyamoto, E. (1975a).Biochm.J. 146, 109-120. Kakiuchi, S., Yamazaki, R., Teshima, Y., Uenishi, K., and Miyamoto, E. (1975b).Adv. Cyclic Nucleotide Res. 5 , 163-1 78. Kalisker, A., Rutledge, C. O., and Perkins, J . P. (1973). Mol. Pharmacol. 9,619-629. Kalix, P., and Roch, P. (1975). Naunyn-Schmizdebm&s Arch. Pharmacol. 291, 131-137. Kalix, P., McAfee, D., Schorderet, M., and Greengard, P. (1974).J. Pharmacol. Exp. Ther. 188,676-687. Kauffrnan, F. C., Harkonen, M. H. A., and Johnson, E. C. (1972). L f e Sci. 11,613421. Kebabian, J . W., and Greengard, P. (1971). Science 174, 1346-1349. Kebabian, J. W., and Saavedra, J. M. (1976). Sczence 193, 683-685. Kebabian, J. W., Petzold, G. L., and Greengard, P. (1972).Proc. Natl. Acad. Sci. U.S.A. 69, 2145-2 149. Kebabian, J. W., Bloom, F. E., Steiner, A. L., and Greengard, P. (1975a). Science 190, 157-1 59. Kebabian, J. W., Steiner, A. L., and Greengard, P. (1975b).J. Pharmacol. Exp. Ther. 193, 474488. Proc. Natl. Acad. Sci. U.S.A. Kebabian, J. W., Zatz, M., Romer, J. A., and Axelrod, J. (1975~). 72,3735-3739. Kebabian, J. W., Clement-Cormier, Y. C., Petzold, G. L., and Greengard, P. (1975d).Adzr. Neurol. 9, 1- 1 1. Kimura, H., and Murad, F. (1974).J. Biol. Chem. 248, 691045916. Kimura, H., and Murad, F. (1975a). Metub., Clin. Exp. 24, 439-445. Kirnura, H.. and Murad, F. (1975b). Life Sci. 17,837-844. Kimura, H., Thomas, F., and Murad, F. (1974). Biochim. Biophys. Actu 343, 519-528. Kimura, H., Mitta!, C. K., and Murad, F. (1975). Nuhire (London) 257,700-702. Kinscherf, D. A.. Chang, M. M., Rubin, E. H., Schneider, D. R., and Ferrendelli, J. A. (1976).J. Neurochem. 26,527-530. Kkh, V. M.. and Kleinsmith, L. J. (1974).J. Biol. Chem. 249, 750-760. Rlainer, L. M., Chi, Y.-M., Friedberg, S . L., Rall, T. W., and Sutherland, E. W. (1962).5. Biol. Chem. 237, 1239-1243. Kodama, T., Matsukado, Y., and Shirnizu, H. (1973). Brain Res. 50, 135-146. Kraynack, B. J., Cohn, M. L., Cohn, M., and Taylor, F. H. (1976).Pharmacology14,39-46. Krishna, G., Forn, J., Voight, K., Paul, M., and Gessa, G. L. (1970). Adv. Biochem. Psyc1wpharmacol. 3, 155-172. Krnjevic, K., and Schwartz, S. (1967). Ex@. Brain Res. 3, 306319. Krnjevic, K., Pull, E., and Werman, R. (1976). Can. J . Physiol. Phannacol. 54, 172-176. Krueger, B. K., Forn, J., and Greengard, P. (1975). In “Pre- and Postsynaptic Receptors” (E. Usdin and W. E. Bunney, Jr., eds.), pp. 123-147. Dekker, New York. Krueger, B. K., Forn, J., Walters, J. R., Roth, R. H., and Greengard, P. (1976). Mol. Phannacol. 12,639-648. Kuehl, F. A., Jr., Humes, J. L., Cirillo, V. J., and Ham, E. A. (1972). Adu. Cyclic Nucleotide Res. 1, 493-502. Kuo, J. F. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 4037-4041. Kuo, J. F. (1975). Metab., Clin. Exp. 24, 321-329. Kuo, J. F., and Greengard, P. (1969). Proc. N a d . Acad. Sci. U.S.A. 64, 1349-1355.

162

JOHN W. DALY

Kuo, J. F., Krueger, B. K., Sanes, J. R., and Greengard, P. (1970).Bwchim. Biophp. Acta 212, 79-9 1, Kuo, J. F., Lee, T.-P. Reyes, P. L., Waltun, K. G., Donnelly, T. E., Jr., and Greengard, P. (1972).J . Biol. Chem. 247, 16-22. Kuo, W.-M., and Kuo, J. F. (1976).J , B i d . Chem. 251,42834286. Kuo, W.-M., Shoji, J., and Kuo, J. F. (1976a). Biochim. Biophys. Acta 437, 142-149. Kuo, W.-M., Shoji, M., and K u o , J. F. (1976b).Bwchem.Biophys. Rex Commun. 70,280-286. Kuroda, Y., and Kobayashi, K. (1975). Proc. Jpn. Acad. 51, 495-500. Kuroda, Y., and McIlwain, H. (1974).J. Neurochem. 22,691-700. Kuroda, Y., Saito, M., and Kobayashi, K. (1976a).Bruin Res. 109, 196-201. Kuroda, Y., Hashirnoto, E., Nishizuka, Y., Hamana, K., and Jwai, K. (1976b). Biochem. Biophys. Res. Commun. 71,629-635. Laduron, P. (1976).J . Phann. Phannacol. 28, 250-251. Laverty, R. (1975). Pfog. Neurobiol. (Oxford) 3, 33-70. Lee, T.-P., Kuo, J. F., and Greengard, P. (1972).Proc.Natl. Acad. Sci. U.S.A. 69,3287-3291. Lernay, A., and Jarett, L. (1975).J. Cell Biol. 65, 39-50. Levey, C . S., Lehotay, D. C., Canterbury, J. M. Bricker, L. A,, and Meltz, G. J. (1975).J. Biol. Chem. 250, 5730-5733. Levin, R. M., and Weiss, B. (1976). Mol. Pharmtacol. 12,581-589. Levitan, I. B., and Barondes, S. H. (1974). Proc. Natl. A c d . Sci. U.S.A. 71, 1145-1 148. Levitan, I. B., Madsen, C. J., and Barondes, S. H. (1974).J. Neurobiol. 5,511-525. Liberman, E. A., Minina, S. V., and Golubtsov, K. V. (1975). Biofizzka 20, 451-456. Libet, B., and Tosaka, T. (1970). Proc. Natl. Acad. Sci. U.S.A. 67, 667-673. Libet, B., Kobayashi, H., and Tanaka, T. ( 1976).Fed. Proc., Fed. Am. SOE.Exp. Biot. 35,326. (Abstr.) Lincoln, T. M., Hall, C. O., Park, C. R., and Corbin, J. D. (1976).Proc. Natl. Acad. Sci. U.S.A. 73, 2559-2563. Lindl, T., and Crarner, H. (1974).Biochim. Biophys. Acta 343, 182-191. Lindl, T., and Cramer, H. (1975). Bwchem. Biophys. Res. Commun. 65, 731-739. Lindl, T., Heinl-Sawaya, C. B., and Crarner, H. (1975). Biochem. Phannacol. 24, 947950. Lindl, T., Heinl-Sawaya, M. C. B., and Cramer, H. (1976). Res. Commun. C l m . Pathol. Phannucol. 13,65-74. Liu, Y . P., and Cheung, W. Y. (1976).J . Biol. Chem. 251, 4193-4198. Lloyd, T., and Kaufrnan, S. (1975). Biochem. Biophys. Res. Commun. 66, 907-913. Logan, J. G., and O’Donovan, D. J. (1976).J. Neurochem. 27, 185-189. Lovenberg, W., and Bruckwick, E. A. (1975). In “Pre- and Postsynaptic Receptors” (E. Usdin and W. E. Bunney, Jr., eds.), pp. 149-169. Dekker, New York. Lovenberg, W., Bruckwick, E. A., and Hanbauer, 1. (1975).Proc. Natl. Acad. Sci. U.S.A. 72, 2955-2958. Lust, W.D., and Passonneau, J. V. (1976).J. Neurocha. 26, 11-16. Lust, W. D., Goldberg, N. D., and Passonneau, J. V. (1976).J. Neuroclm. 26, 5-10. Lynch, T. J., Tallant, E. A., and Cheung, W. Y. (1976).Bwchem. Biophys. Res. Commun. 68, 616-625, McAfee, D. A., Schorderet, M., and Greengard, P. (1971). Science 171, 1156-1 158. MacDonald, I. A., (1975). Biochim. Biophys. Acta 397, 244-253. Machova, J., and Kristofova, A. (1973).L f e Sci. 13, 525-535. Mackay, A. V. P., and Iversen, L. L. (1972). Brain Res. 48, 424426. McKenzie, S. G., and Bar, H. P. (1973). Can. J. Physwl. Pharmacol. 51, 190-196. McNiece, D. M., and Jacobs, R . S. (1976). Fed. Proc., Fed. Am. SOC.Exp. Biol. 35,696. (Abstr.) Maeno, H., and Greengard, P. (1972).J. Biol. Chem. 247, 3269-3277.

CYCLJC NUCLEOTIDES IN THE NERVOUS SYSTEM

163

Maeno, H.,Johnson, E. M., and Greengard, P. (1971).J. Biol. Chem. 246, 134-142. Maeno, H., Reyes, P. L., Ueda, T., Rudolph, S. A,, and Greengard, P. (1974).Arch.Biochrrn. Biophys. 164, 55 1-559. Maeno, H., Ueda, T., and Greengard, P. ( 1 9 7 5 ) . ~Cyclic . Nurleotide Res. I, 37-48, Mah, H. D., and Daly, J. W. (1976). Pharmmol. Rex Cornmiin. 8, 65-79. Maitre, M., Ciesielski, L., Lehmann, P., Kempf, E., and Mandel, P. (1974).Biorhem. Phormnrol. 23, 2807-28 16. Malkinson, A. M. (1975). Riochem. Biophys. Res. Commun. 67, 752-759. Malkinson, A. M., Krueger, B. K., Rudolph, S. A., Casnellie, J. E., Haley, B. E., and Greengard, P. (1975). Metnb., Clin. Exp. 24, 331-341. Mao, C. C., Guidotti, A., and Costa, E. (1974). Brain Re.\. 79, 510-514. Mao, C. C., Guidotti, A.. and Landis, S. (1975). Brain Res. 90, 335-339. Marley, E., and Nistico, G. (1972). B r . J . Pharmacol. 46, 619-636. Martres, M. P., Baudry, M., and Schwartz, J. C. (1975). Nature (London) 255, 731-733. Merali, A., Singhal, R. L., Hrdina, P. D., and Ling, G. M. (1975).LifeSci. 16, 1889-1894. Merali, Z.,Tsang, B., Singhal, R. L., and Hrdina, P. D. (1976).RCF.Comrnitn. Chem. Potlid. Pharmacol. 14, 29-38. Miller, R. J., and Kelly, P. H. (1975). Nature (London) 255, 163-165. Miller, ‘r.J., Horn, A. S., and Iversen, L. I,. (1974). Mol. Phannacol. 10, 759-766. Milstien, S., Abita, J.-P., Chang, N., and Kaufman, S. (1976). Proc. Natl. Amd. Sci. U.S.A. 73, 1591-1593. Minneman, K. P., and Iversen, L. L. (1976a).Nature (London) 260, 59-61. Minneman, K. P., and Iversen, L. L. (1976b). Nuture (London) 262, 313-314. Mishra, R. K., Gardner, E. L., Katzman, R., and Makman, M. H. ( I 974). Pror. N d . . 4 r r l . Sci. U.S.A. 71, 3883-3887. Mishra, R. K., Demirjian, C., Katzman, R., and Makman, M. H. (1975). Broin Re.v. 96, 395-399. Mittal, C. K., Kimura, H., and Murad, F. (1975).J. Cyclic Nurleotide Res. 1, 261-269. Miyamoto, E., and Kakiuchi, S. (1975). Eiochim. Bio/)hys. Actn 384, 458-465. Miyamoto, E., Kuo, J. F., and Greengard, P. (1969).J. Biol. Chm. 244, 65954402. Miyamoto, E., Pet~old,G. L., Harris, J. S., and Greengard, P.(1971).Biochem.Bioplivs. Re.$. Commun. 44, 305-3 12. Miyamoto, M. D., and Breckenridge, B. M. (1974).J. Cen. Physiol. 63, 609-624. Monnier, M., and Hall, A. M. (1969). Experienfia 25, 1297-1298. Moyes, I . C. A,, and Moyes, R. B. (1976). Postgrad. Med. J.. Suppl. 3, 1 1 0 - 1 15. Mrsulja, B. B., Lust, W. D., Mrsulja, B. J., pass on near^, J. V., and Klatzo, 1. (1976).J . Neurochem. 26, 1099-1 104. Murad, F., Rall, T. W., and Vaughan, M. (1969).Blochim. Biophys. Act0 192, 430-45.5. Nahorski, S. R., Rogers, K. J., and Smith, B. M. (1974).Li/e Sci. 15, 1887-1894. Nahorski, S. R., Rogers, K. J., and Edwards, C. (1975a).Brain RPS.92, 529-533. Nahorski, S. R., Rogers, K. J., Smith, B. M., and Auson, J. (1975b).Nnunyn-Schmiedeb~g’.\ Arch. Phannncol. 291, 101-1 10. Nakazawa, K., and Sano, M. (1974).J. Biol. Chem. 249, 4207-421 1. Nakazawa, K., Sano, M., and Saito, T. (1976). Bwchim. Biophys. Actn 444, 563-570. Nathanson, J., and Bloom, F. (1975). Nature (London) 255, 419-420. Nathanson, J., and Bloom, F. (1976). Mol. Phannurol. 12, 390-398. Nathanson, J. A., and Greengard, P. (1973). Scienre 180,308-310. Nathanson, J. A., and Greengard, P. (1974). Proc. Nntl. Acad. Sci. U.S.A. 71, 797-801. Nathanson, J. A., Freedman, R., and Hoffer. B. J. (1976).Nutiire (London) 261, 330-331. Nikodijevic. B., Nikodijevic, B., Wongyu, M. Y., Pollard, H., and Curoff, G . (1975). Pror. Natl. Acad. Sci. U.S.A. 72, 4769-4771.

164

JOHN W. DALY

Nikodijevic, O., Nikodijevic, B., Zinder, O., Yi, M.-Y. W., Guroff, G., and Pollard, H. B. (1976). Proc. Natl. Acad. Sci. U.S.A. 73,771-774. O’Dea, R. V., Zatz, M., and Axelrod, J. (1976). Pharmacologzst 18, 222. (Abstr.) Ohga, Y., and Daly, J. W. (1977a). Biochim. Biophys. Acta (in press). Ohga, Y., and Daly, J. W. (1977b). Biochim. Biophys. Actu (in press). Okada, Y., and Kuroda, Y. (1975). Proc. Jpn. Acad. 51,491-494. Oleshansky, M. A,, and Ne5, N. H. (1975). Life Sci. 17, 1429-1432. Olson, D. R., Kon, C., and Breckenridge, G. M. (1976). Life Sci. 18,935-940. Opmeer, F. A., Gumulka, S. W., Dinnendahl, V., and Schonhofer, P. S. (1976).NuunynSchmiedeberg’s Arch. Phannacol. 292,259-266. Orenberg, E. K., Renson, J., Elliott, G. R., Barchas, J. D., and Kessler, S. (1975). Psychophannacol. Commun. 1, 99-1 07. Otten, Y., Mueller, R. A,, Oesch, F., and Thoenen, H. (1974). Proc. Natl. Acad. Sci. U.S.A. 71,2217-2221. Palmer, G. C. (1973). Res. Cmm.un. Chem. Pathol. Phannuol. 5,603-613. Palmer, G. C. (1976). Neurophannacology 15, 1-17. Palmer, G. C., and Duszynski, C. (1975). Eur. J . Pharmacol. 32, 375-379. Palmer, G. C., and Manian, A. A. (1974a). Neurophannacology 13, 651-664. Palmer, G. C., and Manian, A. A. (1974b).In “The Phenothiazines and Structurally Related Drugs” I. S. Forrest, C. J. Carr, and E. Usdin, eds.), pp. 749-767. Raven, New York. Palmer, G. C., and Manian, A. A. (1976). Biochem. Phannacol. 25, 63-72. Palmer, G. C., Sulser, F., and Robison, G. A. (1973). Neurophunnacology 12, 327-337. Palmer, G. C., Manian, A. A., and Sanborn, C. R. (1976).Eur.J. Pharmacol. 38,205-210. Penit, J., Huot, J., and Jard, S. (1976).J. Neurochem. 26, 256-273. Perkins, J. P.,and Moore, M. M. (1973).J . Phannacol. Exp. Ther. 185, 371-378. Perkins. J. P., Moore, M. M., Kalisker, A.. and Su, Y.-F. (1975).Adu. Cyclic Nucleotide Res. 5, 641-660. Phillipson, 0. T., and Horn, A. S. (1976). Nature (London) 261, 418-420. Phillis,J. W. (1976).In “Taurine” (R. Huxtable and A. Barbeau, eds.), pp. 209-223. Raven, New York. Phillis, J. W., and Kostopoulos, G. K. (1975). Lye Sci. 17, 1085-1094. Phillis, J. W., Kostopoulos, G . K., and Limacher, J. J. (1974). J . Physiol. Phannacol. 52, 1226-1229. Phillis,J. W., Kostopoulos, G. K., and Limacher,J. J. (1975).Eur.J. Pharmacol. 30, 125-129. Pledger, W. J.. Stangel, G. M., Thompson, W. J., and Strada, S. J. (1974).Biochim.Biophys. Acta 370, 242-248. Pledger, W. J., Thompson, W. J., and Strada, S. J. (1975). Bwchim. Biophys. Acta 391, 334-340. Pradfian, S. N., and Dutta, S. N. (1971). Int. Rev. Nturobiol. 14, 173-231. Pull, I., and McIIwain, H. (1975).J. Neurochon. 24,695-700. Racagni, G., and Carenzi, A. (1976).Phannacol. Ref. Commun. 8, 149-158. Rangel-Aldao, R., and Rosen, 0. M. (1976).J. Biol. Chem. 251,3375-3380. Rappaport, L., Leterrier, J. F., Virion, A., and Nunez, J. (1976). Eur. J . Biochem. 62, 539-549. Reddington, M., Rodnight, R., and Williams, M. (1973). Biochem. J. 132,475-482. Redos, J. D., Hunt, W. A., and Catravas, G. N . (1976a). Life Sci. 18,989-992. Redos, J. D., Catravas, G . N., and Hunt, W. A. (1976b). Science 193, 58-59. Roberts, E., and Simonsen, D. G. (1970).Brain Res. 24, 91-1 11. Robison, G. A., Schmidt, M. J., and Sutherland, E. W. (1970).Adv.Bwchem. Psychopharmacol. 3, 11-30. Roch, P., and Kalix, P. (1975a). Neurophannacology 14, 21-29.

u.

CYCLIC NUCLEOTIDES IN THE NERVOUS SYSTEM

165

Roch, P., and Kalix, P. (1975b). Biochem. Pharmacol. 24, 1293-1296. Rogers. M., Dismukes, K., and Daly, J. W. (1975).J. N e u r o r h . 25, 531-534. Romero, J. A., and Axelrod, J. (1975). Proc. Nail. Acad. Sci. U.S.A. 72, 1661-1665. Roskoski, R., Jr., Frederick, C. E., Greif, B. J., and Garvey, M. L. (1976). Fed. Proc., Fed. Am. Soc. Exp. Biol. 35, 1384. (Abstr.) Roth, R. H., Morgenroth, V. H . , and Salzman, P. M. (1975). Naunyn-SchmLdeberg’s Arch. Pharmacol. 289,327-343. Rotrosen, J., Friedman, E., and Gershon, S. (1975).L f e Sci. 17, 563-568. Routtenberg, A., and Ehrlich, Y. H. (1975).Brain Res. 92, 415-430. Routtenberg, A., Ehrlich, Y. H., and Rabjohns, R. R. (1975).Fed. Proc., Fed. Am. SOC.Exp. Biol. 34, 293 (Abstr.). Rubin, E. H., and Ferrendelli, J. A. (1976).J. HistochPm. Cytochem. 24, 964-966. Salem, R., and De Vellis, J. (1976).Fed. Proc., Fed. Am. Soc. Ex$. Biol. 35, 296. (Abstr.) Sandoval, I. V., and Cuatrecasas, P. (1976a). Biochemdry 15, 3424-3432. Sandoval, I. V., and Cuatrecasas, P. (1976b). Nature (London) 262,511-514. Sattin, A. (1975). Lqe Sci. 16, 903-914. Sattin, A., and Rall, T. W. (1970). Mol. Pharmncol. 6, 13-23. Sattin, A., Rall, T. W., and Zanella, J. (1975).J. Pharmaco/. Exp. Thm. 192, 22-32. Schmidt, M. J.. and Sokoloff, L. (1973).J. Neurochem. 21, 1193-1205. Schmidt, M. J., Palmer, E. C., Dettbarn, W. D., and Robison, G. A. (1970).Den. Psychobiol. 3, 53-67. Schmidt, M. J., Hopkins, J. T., Schmidt, D. E., and Robison, G. A. (1972).Brain Res. 42, 465-477. Schmidt, M. J., Ryan, J. J., and Molloy, B. B. (1976).Brain Res. 112, 113-126. Schultz, J. (1974a).J . Neurockm. 22,685-690. Schultz, J. (1974b). Arch. Bwchem. Biophys. 163, 15-20. Schultz, J. (1975a).J. Neurocha. 24, 1237-1242. Schultz, J. (1975b).J. Neurochem. 24, 494-501. Schultz, J. (1976). Nature (London) 261,417-418. Schultz, J., and Daly, J. W. (1973a).J. Biol. C h m . 248, 843-852. Schultz, J., and Daly, J. W. (1973b).J. B i d . C h m . 248, 860-866. Schultz, J., and Daly, J. W. (1973c).J. Neurocha. 21, 573-579. Schultz, J., and Daly, J. W. (1973d).J. Neurochem. 21, 1319-1326. Schwabe, U., and Daly, J. W. (1977).J . Pharmncol. Exp. Ther. (in press). Schwabe, U., Miyake, M., Ohga, Y., and Daly, J. W. (1976).Mol. Pharmucol. 12,900-910. Schwabe, U., Ohga, Y., and Daly, J. W. (1977). Biachim. Biophys. Acta (submitted). Schwartz,J.-C., Baudry, M., Chast, F.. Pollard, H., Bischoff, S., and Krishnamoorthy, M. S. ( 1 974). I n “International Symposium on Metabolic Regulation and Functional Activity of the Central Nervous System” (E. Genazzani and H. Herken, eds.), pp. 172-184. Springer-Verlag, Berlin and New York. Schwartz, J. P., and Passonneau, J. V. (1974).Proc. Natl. Acad. Sci. U.S.A. 71, 3844-3848. Seeds, N. W., and Gilman, A. G. (1971). Scitnce 174,292. Segal, D. A., Geyer, M. A., and Weiner, B. E. (1975). Science 189, 301-303. Shanta, T. R., Woods, W. D., Waitzman, M. B., and Bourne, G. H. (1966). Hi,staclrmie 7, 177-190. Sharma, S. K., Klee, W. A., and Nirenberg, M. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 3092-3096. Shashoua, V. E. (1971).Proc, Natl. Acad. Sci. U.S.A. 68, 2835-2838. Sheppard, H., and Burghardt, C. R. (1976). Fed. Proc., Fed. Am. SOC. Exp. Bwl. 35, 424. (Abstr.) Sheppard, H., Wiggan, G., and Tsien, W. H. (1972).Adv. Cyrlic Nucleotide Res. 1, 103-1 12.

166

JOHN W. DALY

Shimizu, H., Creveling, C. R., and Daly, J . (1970). Proc. Natl. Acad. Sci. U.S.A. 65, 10331040. Shimizu, H., Tanaka, S., Suzuki, T., and Matsukado, Y. (1971).J. Neurochem. 18, 11571161. Shimizu, H., lchishita, H., and Odagiri, H. (1974).j. Biol. Chem. 249, 5955-5962. Shimizu, H., lchishita, H., and Miaokami, Y. (1975a).J. Cyclic Nucleotide Res. 1, 61-67. Shimizu, H., Ichishita, H., Tateishi, M., and Umeda, I. (1975b). Mu/. Pharmacol. 11, 223231. Shimizu, H., lchishita, H., and Umeda, I. (1975~). Mol. Pharmacol. 11, 866-873. Sieghart, W., Leitich, H., and Karobath, M. (1976). Brain Res. 109, 418422. Siggins, G. R., Battenberg, E. F., Hoffer, B. J., and Bloom, F. E. (1973). Science 179, 585-588. Siggins, G. R., Hoffer, B. H., and Ungerstedt, Y. (1974). Life Sci. 15, 779-792. Skolnick, P., and Daly, J. W. (1974a). Science 184, 175-177. Skolnick, P., and Daly, J . W. (1974b). Brain ReJ. 73,513-525. Skolnick, P., and Daly, J. W. (1975a).J. Neurochem. 24,451-456. Skolnick, P., and Daly, J. W. (1975b).J. Pharmacol. Exp. Ther. 193, 549-558. Skolnick, P., and Dal,y,J. W. (1975~).Mol. Pharmucol. 11, 545-551. Skolnick, P., and Daly, J. W. (1976a). Eur. J . Pharmacol. 39, 11-21. Skolnick, P., and Daly, J. W. (1976b). Lzfe Sci. 19, 492-504. Skolnick, P., and Daly, J. W. (1977). Eur. J . Ph,anacol. 41, 145-152. Skolnick, P., Huang, M., Daly, J.. and Hoffer, B. (1973).J. Neurochem. 21, 237-240. Sholnick, P., Daly, J. W., Freedman, R., and Hoffer, B. J . (1976).j. Pharmucol. Exp. Ther. 197,280-292. Sloboda, R. D., Rudolph, S. A., Rosenbaum, J. L., and Greengard, P. (1975). Proc. Nutl. Acad. Sci. U.S.A. 72, 177-181. Smith, C. C., Tallman, J . F., Post, R. M., Van Kammen, D. P., Jimerson, D. C., Brown, G. L., Brooks, B. R., and Bunney, W. E., Jr. (1976).Life Sci. 19, 131-136. Sold, G., and Hofmann, F. (1974).Eur. J . Biochm. 44, 443449. Spano, P. F., Kumakura, K., Tonon, G. C., Govoni, S., and Trabucchi, M. (1975a). Brain Res. 93, 164-167. Spano, P. F., Kumakura, K., Govoni, S., and Trabucchi, M. (1975b). Pharmacol. Res. Commun. 7,223-237. Spiker, M. D., Palmer, G. C., and Manian, A. A. (1976). Brain Res. 104, 401---406. Stalvey, L., Daly, J. W., and Dismukes, R. K. (1976). Lzfe Sci. 19, 1845-1850. Standaert, F. G., Morgenroth, V. H., Skirboll, L. R., and Dretchen, K. L. (1976). Pharmacologist 18, 193. (Abstr.) Steiner, A. L., Ferrendelli, J. A., and Kipnis, D. M. (1972).J. Biol. C h m . 247, 1121-1124. Stellwagen, E., and Baker, B. (1976). Nature (London) 261,719-720. Stjarne, L. (1976). Neuroscience 1, 19-22. Stone, T. W., Taylor, D. A., and Bloom, F. E. (1975).Science 187,845-847. Strada, S. J., Kirkegaard, L., and Thompson, W. M. (1976).Neurophannncology 15, 261266. Stiltgill, T. W., Schrier, M. B. K., and Gilman, A. B. (1975).J.Cyclic Nucleotide Res. 1,21-30. Sun, M.-C., McIlwain, H., and Pull, 1. (1976).J. Neurohzol. 7, 109-122. Suria, A. (1976). Neuropharmacology 15, 11-16. Sutherland, E. W., Rall, T. W., and Menon, T. (1962).j. Biol. Chem. 237, 1220-1227. Szmigielski, A., and Guidotti, A. (1976). Pharmacologzst 18, 222. (Abstr.) Takahashi, T., Matsuzaki, S., and Nunez, J. (1975).j. Neurochem. 24, 303-3 10. Takai, Y., Nishiyama, K., Yamamura, H., and Nishizuka, Y. (1975).J. Bid. Chem. 250, 4690-4695.

CYCLIC NUCLEOTIDES IN T H E NERVOUS SYSTEM

167

Tell, G. P., Pasternak, C;. W., and Cuatrecasas, 1.’ (1975). FEBS Lett. 51, 242-245. Teshima, Y., and Kakiuchi, S. (1974). Blochem. Biophy.~.Res. Commttn. 56, 489-495. Thompson, W. J., and Appleinan, M. M. (197la).J. Biol. Ckm. 246, 3145-3150. Thompson, W. J., and Appleman, M. M. (1971b). Biochemistry 10, 311-316. Traber, J., Gullis, R., and Hamprecht, B. (1975). Life Sci. 16, 1863-1868. Treistman, S. N., and Levitan, I. B. (1976). Nature (London) 261, 62-63. Troyer, E. W., and Ferrendelli, J . A. (1976). Fed. Pror., F d . .4m. Sor. Ex/?. Bid. 35, 4.56. (Abstr.) Ueda, T., Maeno, H.. and Greengard, P. (1973).J. Biol. C l t m . 248, 829543305. Ueda, T., Rudolph, S. A,, and Greengard, P. (1975).Arch. B i o c h . Bi0pliy.s. 170,492-503. Uno, I., Lleda, T., and Greengard, P. (1976).J. Bid. C h a . 251,2192-2195. Uzunov, P., and Weiss, B. (1971). Nrurophamzncololy 10,697-708. L!zunov, P., and Weiss, B. (1972a). Biocllim. Biopbys. Actn 284, 220-226. Uzunov, P., and Weiss, B. (1972b).Adz). Cyclir Niicleotide Res. 1, 435-453. Uzunov, P., Shein, H. M., and Weiss, B. (1974). Neirrophnrrnacology 13, 377-392. Uznnov, P., Lehne, R., Revnelta, A. V., Gnegy, M. E., and Costa, E. (1976).Biochirn.Bio~ihy.~. Actcl 422,326-334. Van d e Berg, J. S. (1974). Experientin 30, 1025-1026. Van Inwegen, R. G., Strada, S. J.,.and Robison, G. A. (1975). Lzf2 Sri. 16, 1875-1876. Vetnlani, J., and Sulser, F. (1975). Nature (London) 257, 495-496. Vetiilani, J., Staware, R. J., Dimgell, J. V., and Sulser, F. (1976). Nristnyn-Sclimipdeh~~’.\ Arch. Pharmmol. 293, 109-1 14. Von Hungen, K., and Roberts, S. (1973a). Eirr.J. Biorhem. 36, 391-401. Von Hungen, K., and Roberts, S. (1973b). Nature (London), N P UBioi. I 242, 58-60, Von Hungen, K.. and Roberts, S. (1974). In “Reviews of Neuroscience” (S. Ehrenpreis and I . J. Kopin, eds.), Vol. 1, pp. 231-281. Raven, New York. Von Hungen, K., Roberts, S., and Hill, D. F. (1975).Brcrin Res. 94, 57-66. Von Voightlander, P. F., Boukma, S. J.. and Johnson, G. A. (1973). Nrttrofihormnrology 12, 1081-1086. Von Voightlander, P. F., Losey, E. G., and Triezenberg, H. S. (lE)75).J. P/tnrmacol. E x j . Ther. 193, 88-94. Wahlstrom, C . (1975). L$e Sci. 17, 1655-1662. Walker, J. B., and Walker, J. P. (1973). Brain Res. 54, 386-390. Wardell, W. M. (1966). Proc R. SOC.L a d o n , SIT. I? 165, 326-361. Wasner, H. K. (1975). FEES Lett. 57,60-63. Watanabe, H., and lshii, S. (1976). Brain Re.s. 102, 385-389. Watanabe, H., and Passonneau, J. V. (1974). Brain Re$. 66, 147-159. Watterson, D. M., Harrelson, W. G . , Keller, P. M., Sharief, F.. and Vanaman, T. C. ( I9761.J. B i d . C k m . 251, 45014513. Webb, J. C., Berv, K. R., and Kopin, I. J. (1975). NmrropJwzrmarology 14, 643-648. Commun. 2, 27-38. Weinryb, I., and Michel, 1. M. (1976). P.yc/w~ka~marol. Weinryb, I., Chasin, M., Free, C. A., Harris, D. N., Goldenberg, H., Michel, I . M., Paik, V. S., Phillips, M., Samaniego, S., and Hess, S. M. (1972). J. Plinrm. Sci. 61, 1556-1567. Weiss, B. (1975). A h . Cyclic Nltcleotide Res. 5 , 195-21 1. Weiss, B., and Costa,E. (1968). Biochem. Pharmncol. 17, 2107-21 16. Weiss, B., and Greenberg, 1.. H. (1975). I n “Cyclic Nucleotides in Disease” (B. Weiss, ed.), pp. 269-3 19. University Park Press, Baltimore, Maryland. Weiss, B., Fertel, R., Figlin, R., and Uzunov, P. (1974). Mol. Phannarol. 10, 615-625. Weller, M., and Morgan, I. G. (1976). Biochim. Biophy.r. Actn 433, 223-228. Weller, M., and Rodnight, R. (197 1). B i o c l ~ mJ. . 124, 393-406.

168

JOHN W . DALY

Weller, M., and Rodnight, R. (1973a). Biochem. J. 132,483-492. Weller, M., and Rodnight, R. (197313).Biochem.J. 133,387-389. Weller, M., and Rodnight, R. (1975). Biochim. Biophys. Acta 389, 573-577. Wellmann, W,, and Schwabe, U. (1973). Brain Res. 59,371-378. White, A. A,, and Aurbach, G. D. (1969). Eiochim. Biophys. Acta 191,686. Wickson, R. D., Boudreau, R. J., and Drummond, G. U. (1975).Biochemistry 14,669-675. Wilkening, D., and Makman, M. H. (1975). Brain Res. 92,522-528. Wilkening, D., and Makman, M. H. (1976).J. Neuroclm. 26, 923-928. Williams, B. J., and Pirch, J. H. (1974). Brain Res. 68, 227-234. Williams, M., and Rodnight, R. (1975).J. Neurochem. 24,601-603. Williams, M., and Rodnight, R. (1976). BWchem.J. 154, 163-170. Williams, M., Pavlik, A., and Rodnight, R. (1974a). Tram. Biochem. Soc. 2,259-261. Williams, M., Pavlik, A., and Rodnight, R. (1974b).J. Neurochem. 22, 373-376. Williams, R. E. (1976). Science 192, 473-474. Williams, R. H., Little, S. W., Beug, A. G., and Ensinck, J. W. (197l).Metab., Clin. Exp. 20, 743-748. Wolff, D. J., and Brostrom, C. 0. (1974). Arch. Bzochem. Bmphys. 163,349-358. WolR, D. J., and Brostrom, C. 0. (1976). Arch. Biochtm. Biophys. 173, 720-731. Wooten, G. F., Thoa, N. B., Kopin, I. J,, and Axelrod, J. (1973). Mol. Pharmacol. 9, 178183. Yarbrough, G. G. (1976). Neurophamacology 15,335-338. Yoshimura, K. (1973).J. B i o c h a . (Tokyo) 74, 389-391. Zanella, J., Jr., and Rall, T. W. (1973).J. Pharmacol. Exp. Ther. 186, 241-251. Zivkovic, G., Guidotti, A., and Costa, E. (1975). Brain Res. 92, 516-521. Zivkovic, G., Guidotti, A,. and Costa, E. (1976).]. Cyclic Nucleotide Res. 4, 1-10. Zwiller, J., Ciesielskitreska, J., and Mandel, P. (1976). C.R. Soc. B i d . 170, 183-186.

FLUCTUATION ANALYSIS IN NEUROBIOLOGY' By Louis J. DeFelice Department of Anatomy

Emory University Atlanta, Georgia

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Scope . ........................... B. Early Studies.. ............................................... 11. Methods ......... ....................... A. Conductance Fluctuations ....................... B. Voltage and Current Noise . . . . . . . . . . . . . . . . . . . C. General Relationships ................................................ 111. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Nerve A x o n . . . . . . . . . . ..................... B. Drug-Induced Noise ................................................. C. Sensory Systems ...... ........................ D. Other Preparations ... ..... IV. Summary ............................... ..... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

169 169 171 175 1176 76 I80 181

183 183 191 200 205 206 206

1. Introduction

Electrophysiology has progressed continuously from whole body surface recordings, through compound-action potentials, single-unit recordings, transmembrane-action potentials, and graded and postsynaptic transmembrane responses of ever decreasing size. The minute random fluctuations which occur spontaneously in the steady state of biological membranes reveal further information about the elementary events associated with electrical conduction. The measurement and interpretation of electrical noise from biological membranes may be thought of as the next level of awareness in electrophysiology. A. SCOPE

In this chapter I want to describe the basic concepts and tools required for understanding fluctuation analysis in membranes and to re-

' This work was supported in part by the National Library of Medicine (LM02505). 169

170

LOUIS J. DE FELICE

view recent progress in this field. Several reviews are already available (Stevens, 1972; Verveen and DeFelice, 1974; Conti and Wanke, 1975; Neher and Stevens, 1977). There are two important areas of fluctuation analysis which will not be covered in this chapter. The mere characterization of membrane noise may be important as a data base for the description of the macroscopic random behavior of cells. For example, in very small cells, spontaneous firing or spontaneous transmitter release may be related to membrane noise. Membrane noise may also be implicated in the probabilistic response to a constant or regular stimulus, including selfoscillatory systems such as heart cells or pacemaker neurons. None of these questions will be dealt with here. The second area of fluctuation analysis to be omitted is the use of noise as an external stimulus to characterize a system. In linear systems, the response to any input is known once the response to a known source of noise has been measured. This is the transfer function of the system. The input noise must contain all frequencies of interest. It is convenient if each frequency component in the noise has the same average amplitude (white noise). The system to be characterized may range from a single membrane to a large group of cells. Measurement of the transfer function of a linear system implies an equivalent electrical network made of passive linear elements (the conventional R, L, and C of circuit theory). Noise is a convenient way to measure the impedance of any electrical equivalent circuit of a linear system. In nonlinear systems an analogous situation exists. There are two general ways to approach nonlinear systems. One assumes that for a particular state of the system, small perturbations from that state may be treated as linear. Other states are treated the same way, but the equivalent linear networks used in each state are not the same. This “piecewise” linear approximation is used to describe the impedance of nerve membranes. A second approach is the Wiener theory of nonlinear systems (Wiener, 1958). This method is not restricted to small perturbations. T h e full dynamic range of the nonlinear system may be probed with a noise stimulus. The output from the system to be described is expanded as a series of terms, each of which is defined by an operation between the input noise and the Wiener kernel for that term. T h e operational procedures are explained very clearly in Lee and Schetzen ( 1965) using crosscorrelation analysis. Once measured, the set of Wiener kernels completely defines the nonlinear system for an arbitrary input and is analogous to the transfer function for a linear system. Wiener kernels may also be measured using a fast Fourier transform algorithm (French and Butz, 1973). White noise analysis of nonlinear systems has become a useful tool in neurophysiology. This subject is discussed in detail by Marmarelis and Naka (1974) and Naka et al. (1974) and was the

FLUCTUATION ANALYSIS I N NEUROBIOLOGY

171

topic of a recent conference (McCann and Marmarelis, 1975). The Wiener theory has not been applied to excitable membranes, which are usually treated as quasilinear for small perturbations or by nonlinear differential equations for large perturbations. This chapter will be restricted to the study of indigenous membrane noise. Since the inherent spontaneous electrical fluctuations from biological membranes are necessarily the smallest which may occur, it has been generally assumed that linear system analysis may be applied. To what extent we may continue to ignore nonlinearities in spontaneous membrane noise remains to be seen. B. EARLYSTUDIES Fluctuation analysis in neurobiology has been applied to three general areas: excitable nerve membranes with voltage-controlled conductances, postsynaptic or extrajunctional membranes whose conductance is altered by chemical transmitters, and sensory receptors whose conductance is altered by light or mechanical stimulation. In many preparations, noise phenomena have been known for years. Some early observations are discussed here. In 1952, Brock et nl., in their study of electrical potentials from motoneurons, reported: “It has regularly been observed that, as the microelectrode entered into close contact with a neuron, the noise level increased several times . . , the peak voltages were as much as 0.5 mV.” This was apparently a chance observation, since the paper deals primarily with intracellular neuronal potentials. They regarded the extracellular noise as being actually produced by the surface membrane, and not by increased microelectrode resistance as the electrode approached the membrane. No further analysis of the effect was made. Similar observations are well known to present-day neurophysiologists, although there has been no systematic study of the phenomenon (Krnjevic, personal communication). Extracellular noise analysis is now regarded as a usefiil tool in analyzing transmitter-induced noise at cell junctions. This will be reviewed below. Brock et nl. (1952) concluded that the noise they observed might possibly represent normally occuring voltage fluctuations from a small area of neuronal membrane. Approximately ten years later, Verveen and Derksen began their pioneering noise studies on the node of Ranvier (see Verveen and Derksen, 1968, for an early review). These investigators measured the transmembrane voltage fluctuations directly, introduced the use of the power spectral density to describe membrane noise quantitatively, and started a new level of research in electrophysiology .

FIG. IA. An intracellular recording from the frog end-plate (left panel) and from a distance 2 mm away (right panel) in the same muscle fiber. The lower part of each panel shows the response to a nerve stimulus (the scales are 50 mV and 2 msec). The upper part of each panel shows the spontaneous activity (the scales are 3.6 mV and 47 msec). The spontaneous discharge is seen where the end-plate potential is large and where the spike originates. (From Fatt and Katz, 1952.)B. lntracellular recording from the frog end-plate. In each photograph the upper trace was recorded from a low-gain dc channel (10 mV

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

173

In the same volume of the Journal of Physiology as the Brock et al. paper, another chance observation of a different type of noise was reported: Fatt and Katz (1952), in their microelectrode studies of the acetylcholine (ACh) receptors at the neuromuscular junction, state: “. . . that end-plates of resting muscle fibers are the seat of spontaneous electrical discharges which have the character of miniature end-plate potentials.” These observations were described initially as “biological noise” by Fatt and Katz (1950) in a brief note to Nature. It apparently took some time for these authors to become convinced that the now familiar miniature end-plate potentials were characteristic of normal nerve-muscle preparations. The miniature end-plate potentials are not caused by the insertion of the microelectrode since in some cases the spontaneous discharge could be detected at a distance of 1-1.5 mm from the end-plate. Fatt and Katz compared their internally recorded spontaneous discharges with those recorded externally. Figure 1A is reproduced from their paper. The potential drop recorded externally arises from a flow of current across the postsynaptic membrane and through the external fluid around the electrode tip. These local currents are more rapid than the miniature transmembrane potentials since the latter are recorded across the membrane capacitance. Local extracellular currents more nearly reflect membrane conductance changes than d o transmembrane voltages. This same distinction is made below when comparing voltage noise and current noise spectral densities. Fatt and Katz speculated that the miniature discharges might be attributed to molecular leakage of ACh from nerve endings. This was rejected, largely from negative results obtained by external application of ACh. Moderate concentrations of ACh depolarized the end-plates by a few millivolts with no appreciable effect on the miniature discharges. They conclude: “. . . if individual molecular collisions between ACh and the end-plate builds u p a steady depolarization then the molecular units of this depolarization must be much smaller than the recorded miniature end-plate potential.” scale). The lower trace was recorded simultaneously from a high-gain ac channel. The top row shows control experiments (no ACh). The bottom row shows the depolariTation and the increased membrane noise which are induced during ACh application. Two spontaneous miniatures are also seen. (From Katz and Miledi, 1972.) C. Current recorded through a patch of membrance approximately 1 +m’ from the extrasynaptic region of denervated frog muscle fibers. A downward deflection represents inward current. The membrane was exposed to SubCh and voltage clamped at -120 mV at 8°C. Histograms of the single events give a value of 3.4 pA for the elementary current pulses. Channel conductance was estimated to have a mean value of 22.4 pS. The mean open time of a single-channel current was 45 msec. (From Neher and Sakmann, 1976b.)

174

LOUIS J. DE FELICE

Approximately 20 years after this discovery, Katz and Miledi (1970, 1972) reported the measurement of membrane noise which results from the molecular bombardment described above. The miniature end-plate potential is approximately 1/100 the size of the normal end-plate potential. T h e molecular shot effects are approximately 1/1000 the size of the miniatures. In fact, the shot effects are normally too small and too numerous to be recorded individually, and require statistical analysis of the spontaneous fluctuations resulting from a steady exposure of ACh molecules to the postsynaptic membrane. In a lecture to the Fifth International Congress of Biophysics (Copenhagen, August, 1975) B. Katz said that we have yet to see the elementary conductance change associated with a single agonistreceptor interaction. At that time, only the statistical analysis of a random sum of many events was available (Fig. 1B). Recently, Neher and Sakmann (1976b) have reported seeing individual elementary events which show many of the features derived from noise measurements (Fig. 1C). Observing both the noise and the elementary event from the same preparation enables one to compare the two methods and to decide between ambiguous models in the statistical method. These results will be reviewed in detail. The ACh-induced channel is probably the best characterized channel in biological membranes at the present time. AChinduced noise and the observation of individual elementary events have given indirect support to those cases where only noise data are available. In 1909, N. Campbell discussed a similar problem in relation to the measurement of the electrical charge carried by alpha-particles from spontaneously radioactive material. Initially, Campbell was concerned with a more accurate disclosure of von Schweidler’s theory of the discontinuous emission of rays by radioactive substances. The theory, Campbell (1909) said, lost much of its importance when Rutherford and Geiger measured the charge on an alpha-particle “by the more direct and probably essentially more accurate method of counting the number of particles one at a time.” Campbell felt, however, that a more thorough exposition of the theory of discontinuous processes could enable fluctuation experiments to “rival in accuracy” those obtained by individual measurement. He pointed out that radioactivity is not the only discontinuous process amenable to study by such a theory. His study of discontinuous phenomena resulted in Campbell’s theorem which relates the average effect of a random process to the average fluctuation from the mean effect. This phenomenon is often called shot noise. It is Campbell’s theorem which Katz and Miledi used so effectively to study the molecular bombardment of transmitter molecules on receptors at the neuromuscular junction.

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

175

It is difficult to isolate the initial investigation of membrane noise in receptor physiology. An early application of fluctuation analysis was made by Dodge et al. (1968). In 1959, Hagins reported to the Biophysical Society (Hagins, 1959; Wagner and Hagins, 1959) that the predicted minimum primary receptor current should be the order of several thousand electronic charges per absorbed quantum. This estimate was based on the idea that the thermal fluctuations in the receptor membrane must set a lower limit to the size of the signal which results in response to light stimulation. In 1965, Hagins summarized his work in this field. The measurements were done on squid photoreceptors using extracellular field potential records. He reasoned that the photovoltage of the retina milst be made u p of the superposition of many absorbed photons and should show a shot noise effect. The light-induced fluctuations were found to be proportional to light intensity. By measuring the current flow in response to incident light, he was able to calculate the frequency content of the extracellularly recorded noise records. The spectrum of the noise was consistent with randomly spaced pulses whose mean duration was about 100 msec. He calculated that approximately 60,000 electronic charges flowed per incident photon, although values four times higher were also obtained. Hagins (1965) concluded that the ions carrying the current were Na and that the magnitude of the Na conductance produced by 1 quantum response (QR) to light is about 20 pmho. A previous result (Hagins et al., 1962) showed that the average depolarization produced by a single event is about 20 pV. Hagins (1965) concluded that “isolated molecular events in a cell membrane can, in principle, produce electrophysiologically significant currents” and that, as regards the quanta1 effects known to exist at that time at the frog neuromuscular junction, the “. . . miniature end-plate potential . . . yielded flows of 10:’ charges. Since the miniature responses are known to be caused by quanta of ACh containing thousands of molecules, however, it is possible that the single transmitter molecules acting at the postsynaptic end-plate membrane produce conductances of the order of that estimated here for the QRs. Thus, it may be that the QRs exemplify the unit of change in the membrane conductance on the molecular scale.” This hypothesis has been well documented in recent years, and is the main subject of this chapter. II. Methods

The basic goal of noise measurements is to be able to deduce something about the elementary events which give rise to the observable fluc-

176

LOUIS J. DE FELICE

tuations. The three general areas to which fluctuation analysis has been applied are all concerned with conductance changes in a membrane, either in response to a voltage gradient, a transmitter molecule, or a sensory stimulus. Therefore, the elementary event will initially be considered as a change in membrane conductance. A. CONDUCTANCE FLUCTUATIONS

"++

Introduce the symbol " for the phrase "has the units of." As a first example, consider the conductance event g(t) corresponding to a rapidly opening membrane channel which then closes exponentially, thus: g(t) = ye-"e

where y -8- Sz-' (or mho) and 0 -0- sec. I will adopt the unit Siemans (S) for Sz-'. Consider identical changes g(t) occurring in parallel at uncorrelated times with an average frequency of occurrence of v -€+ sec-I. Since these events individually occur in one direction, they will sum to an average effect in the same direction. Since the events are occurring randomly, there will be instantaneous deviations from the average effect. However, these deviations also tend toward an average value. Let G(t) represent the instantaneous value at time t of the total conductance due to the random addition of all the elementary conductances g(t). Let ( ) represent a time average of the quantity in the brackets. Then Campbell's theorem states that

((G - ( G ) ) ' ) = v loffig2 (t) dt

Notice that (G) and ( ( G - (G))') are no longer functions of time. In words, these equations state that by simply knowing the size and shape of the elementary event and the mean frequency of their occurrence, their average effect, ( G ) , and the mean square deviation from the average, ( ( G - ( G ) ) ' ) may be calculated. The mean square deviation is called the variance. The ratio of the two quantities is independent of the mean frequency of the elementary events. The average effect and the mean square deviation from the average effect are measurable quantities from which properties of the elementary event g(t) may be deduced. There are two important considerations yet to be made. If the interval between elementary events is small (i.e., v large) compared to the shape of g(t), then the resulting amplitude fluctuations about the mean will be symmetric even though g(t) itself always occurs in one direction. Also, measuring a quantity proportional to the time integral over g(t)

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

17'7

cannot, in general, reveal information about the temporal form of g(t), but only its size. Our present example illustrates this point. Evaluate the then two integrals above for g(t) = v

,

/om

ye-""dt =

v8y

y2e-2t/Odt = - v8y2

The ratio of the second to the first expression is simply b.Other shapes for g(t) would change the factor of $, but there is not a unique relationship between shape and this factor. The shape of the elementary event has been integrated out and replaced by a number. Conductance fluctuations may be analyzed in a way which does give temporal information. One method is called autocorrelation. The autocorrelation of the conductance fluctuations is a way of comparing the random noise signal in time with itself in an average way. The random signal is initially compared with itself by multiplication of every point of G(t) with itself and averaging the result. The autocorrelation operation shifts one copy of the signal with respect to the original by a fixed time and makes the same comparison by again multiplicating and averaging the two signals. The result is plotted as a function of the shift. This operation results in the autocorrelation function C ( T ) .Symbolically,

c(7)= (G(t)G(t + 7 ) ) Notice that C -€+ F.It is standard to use the symbol T for thecorrelation function; this prevents the use of T as the time constant of the elementary event, for which 8 is used above. When G(t) is the result of the random linear sum of the exponential g(t) described above, the predicted autocorrelation function is

This is the correlation function of the noise due to elementary events, each with the shape ye-"e. The correlation function decays with the same time course as the elementary event and is therefore a way to determine g(t). At T = 0: v8y~ C(0) = 2

+ (v8y)'

which is the sum of the mean square deviation of the conductance (the variance) and the square of the mean conductance caused by the random summation of the elementary events. Since the frequency of the events,

178

LOUIS J. DE FELICE

v, is often a very large number, the first term, which varies directly as v, may be much smaller than the second term, which carries as v’. This fact has important practical significance. In order to amplify the measured noise signal to sufficient levels for analysis, the mean (dc) component usually has to be removed. If ac coupling is used, the shape of the entire correlation function will be distorted. The distortion must either be negligible or be taken into account. Examples may be found in DeFelice and Sokol (1976a,b). In the statistical literature (e.g., Bendat and Piersol, 19’71) it is common to use the symbol R ( T )to represent the total correlation function, including the dc term, and to reserve C ( T )for the total correlation function minus the dc term. This convenient notation is usually not followed by people working in membrane noise and has not been adopted here. I t is obvious that certain practical considerations have to be dealt with carefully when interpreting the shape of the measured correlation function. Assuming that all these considerations have been met, does the shape of the correlation function uniquely represent the shape of the elementary event? Unfortunately, it does not. As an example, consider the elementary event as a channel which fully opens or closes rapidly, but stays open or closed for random lengths of time. This will be called the random-switch model. In our previous example, each event g(t) was mathematically determined, but their time of occurrence was random, In the random switch, a n event is not defined by a deterministic formula. Nevertheless, each channel has average properties, such as the average duration of the open state, and the sum of N randomly opening and closing channels in parallel does define a noise signal with average properties. Let the open conductance of a single channel be y and the closed conductance be zero. Let the average open-chamel lifetime be 8 seconds. The mean frequency of closings is therefore v = 1/8. The autocorrelation function for noise due to the parallel sum of N random switches is given by

The meaning of is the same in both cases; y is the maximum openchannel conductance. However, 8 (and v) have a completely different interpretation. From the autocorrelation function alone’it is not possible to distinguish between the randomly occurring exponential and the random switch model of the elementary event. In Section 111 below, it will be seen that both models of the ACh-receptor interaction have been used to describe electrical noise at the neuromuscular junction. The measurement of isolated elementary events should help resolve conflicting views.

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

179

The above examples illustrate some of the more important points about the use of noise analysis to study underlying events. In both of the examples, “random” was used to imply Poisson-distributed events and the results obtained depend on this assumption. Other models and assumptions are reviewed in Verveen and DeFelice (1974) and Holden and Rubio (1976). Membrane noise has been studied more often by using the spectral density analysis than by using autocorrelation analysis. Temporal information about the elementary events is given indirectly in the frequency domain. From an analysis point of view, there is little difference between methods. However, there may be some practical difference which favors one or the other in particular circumstances. To obtain a spectral density, the noise signal may be passed through a series of narrow band filters. The mean square output of each filter, divided by the filter bandwidth, is plotted against the center frequency of each filter. The result is an estimate of the spectral density. Theoretically, the power spectral density contains the same information as the autocorrelation function, and one may be transformed into the other. As an example, consider the two autocorrelation functions described above for the random sum of exponential events or the random-switch model. The general relationship between the autocorrelation function C ( r ) and the power spectral density S(f) is: S(j) = 4

or d r C(T)COS

For convenience, we have written w = 2 ~ f Notice . that) -@- H z (= sec-’) and S(,f) S2 . sec (= R-’/Hz). Substitution o f t h e expression for C ( r ) for the first model described above results in

+

The second term is an impulse at the origin ofS(f), corresponding to dc information in the noise signal. The usual way of measuring power spectra excludes dc components since each sharp filter has zero transmission at zero frequency. Assuming perfect removal of the dc component:

180

LOUIS J. DE FELICE

This function is called a Lorentzian. The expression for random switch model is

S(f)

for the

The two spectral densities have the same shape but totally different interpretations. Let us restrict ourselves to the second (random-switch) model. The results interchange for W - + ud. The zero frequency limit of the power spectrum is

S(0) = Ny'd It is common to define d = 1/(2n-fc), wheref, is called the half-power or cutoff frequency. Therefore do =fFc. When w = l/d, i.e.,f =f r , then

S(fr) = Ny28 = BS(0) A convenient way to measure 8 is to find the frequency at which the power spectral density is down by half the zero frequency limit. B. VOLTAGE A ND CURRENT NOISE We have considered the elementary event to be a conductance change and the noise to be a fluctuation from the total mean membrane conductance: Conductance is a quantity which must be measured indirectly. It may actually represent a molecular conformational change in channels in the membrane. Such changes are expressed as an electrical event either as a voltage or a current. The two extremes of the measurement may be illustrated by a simple example. Consider a parallel RC circuit. The resistor is a source of noise due to the thermal agitation of the charge carriers in the resistor. The noise has a flat power spectrum. In this case, the voltage spectral density is given by

S ( f ) = 4kTR where K is Boltzmann's constant and T is the absolute temperature. VVHz. The capacitor is considered noiseless. T h e reNotice that S ( f ) sistor noise is measured with a perfect voltage meter (noiseless, with infinite input resistance) across the capacitor of the RC circuit. Since high frequencies of the noise will be shorted to ground more easily than lower frequencies, the output spectrum will not be white, even though the only elementary noise source which is present, thermal noise in the resistor, is white. In fact, the output spectrum from the RC circuit is

+

4kTR

s(f)

= 1

+

(@o)Z

FLUCTUATION ANALYSIS I N NEUROBIOLOGY

181

where 8 = RC. Notice that white noise from an RC circuit has a voltage spectral density whose shape is equivalent to the conductance spectral density from the models considered above. In order to measure the current spectral density, the RC circuit is shorted with a perfect current meter (noiseless, with zero input resistance). T h e meter is always a better pathway for the noise current than the capacitor. T h e time constant 8 is effectively made zero. In this case, the spectral density of the current noise is: S(f) = 4 T / R

Notice that Scf, -0- A'/Hz and that the current spectral density has the same shape of the actual noise source (white) unspoiled by the capacitor. Current noise gives more direct information about the actual noise source than does voltage noise. If the impedance of the system is known, the two contain equivalent information. In excitable membranes, the equivalent impedance may be a rather complex LRC circuit which itself depends on the physiological stimulus. For this reason, it is usually desirable to measure current noise under voltage clamp. This was first done by Poussart (1969, 1971) in his studies of noise from nerve membranes. C. GENERAL RELATIONSHIPS I t is convenient to use the subscripts V, I , and G to denote voltage, current, and conductance. Capital letters will be reserved for a property of the entire membrane, i.e., the combined property of many ionic channels. Lower case letters will be used to represent individual channels. Voltage spectral densities and current spectral densities are related by the general expression:

whereZ is the membrane's impedance. In the example of the parallel RC circuit described above:

z=-1 + i w 0 and

IZI2 =

R'

1

+ (80)'

The relations given for the expected voltage and current noise spectral densities from this circuit satisfy the general expression. Current spectral densities are related to conductance spectral densities by the general expression:

s, = (V -E)'s(;

182

LOUIS J. DE FELICE

where V is the potential across the membrane and E is the equilibrium potential associated with the particular current whose fluctuations are being measured. In our example of the random-switch model, the current spectral density due to the conductance fluctuations is given by:

s/ =

Ny'O(V - E )' 1 (60)'

+

The current fluctuations are measured about some mean current I . The mean current goes to zero when V = E . By fitting the above expression to measured current spectral densities under voltage clamp, the value of N y 2 may be estimated. If we assign the membrane property associated with an equilibrium potential to every channel in the membrane, then for each ionic pathway:

i = (V

-E ) y

where i is the current through the channel and y is the open-channel conductance. T h e average current through N channels in the membrane will be given by: I = iNl2

Therefore:

- E)I s/ = 2y6(V 1 + (we)? By fitting this expression to measured current spectral densities under voltage clamp, when the current I through the N channels is directly measured, the value of y may be estimated. Therefore, N and y are known independently. There are other equivalent expressions for current spectral densities which are based on these same ideas. For example, the expression for current spectral density for the random-switch model may also be written I'

s1

4e

= N 1 + (we)'

which leads directly to an estimate of N . If the area of the membrane is also well known, the channel density may be estimated. These examples cover the basic concepts of the noise measurement. Certain simplifying assumptions have been made which are not discussed in this chapter. Details may be found in the articles reviewed below. Some of the applications made to biological membranes have been quite direct. For example, at the neuromuscular junction the

FLUCTUATION ANALYSIS IN NEURORIOLOCY

183

ACh-receptor interaction is well modeled by a random-switch conductance change. The induced channel is assumed to open and close rapidly, but to remain open for random times. There is an average open time 0; for the ACh-receptor model, one of the most important features of tlie average open time is that it is voltage dependent: 8 = e(V). This model and tlie results obtained from it will be discussed in detail below. In the squid giant axon, there are two populations of channels due to K and Na ionic pathways. The kinetics of-each pathway may be described by a model derived from the Hodgkin and Huxley (1952) picture of axon excitability (Hill and Chen, 1972; Stevens, 1972). For example, the K system is described by a sum of four Lorentians which correspond to then' kinetic scheme of the Hodgkin-Huxley model. However, the principles involved are similar to those outlined above, namely, obtaining information about the elementary conductance change associated with a particular process or ionic pathway from an analysis of membrane noise.

111. Results

In this section I will summarize experimental results and conclusions that have been derived from an analysis of membrane noise. I will concentrate on biological membranes, with occasional reference to model systems where it seems appropriate. A recent review of ionic channels in lipid bilayers is available (Ehrenstein and Lecar, 1977). The results will be organized around the type of preparation studied, rather than the class of observed noise or the nature of the conclusion. No attempt will be made at an historical presentation. Each section will build from a selected key article which has emphasized interpretation of the data in terms of a specific model. A. NERVEAXON 1. The Node of Ranvier

The most thorough study at the time of writing has been done on the node of the myelinated nerve fiber of the sciatic nerve of Rana esculenta (Conti et al., 1976a,b).T h e data represent current fluctuations measured under conditions of voltage clamp. An important feature of this study is the separation of the components of current fluctuations due to Na and K by using the poisons tetrodotoxin (TTX: to selectively inhibit the Na membrane conductance) and tetraethylammonium (TEA: selective for K). I t was shown, for example, that the normal current spectral density

184

LOUIS J. DE FELICE

(no toxin added) minus the spectrum with TEA was equivalent to the spectrum with TTX minus the spectrum with both TEA and TTX. In principle, both of these conditions yield the current spectral density of the fluctuations in the K current alone. Na current fluctuations were measured using similar techniques. Conti et al. fitted their results from the Na system only to a theoretical model, although some K data are presented. The model they used is based on an interpretation of the quantitative description of membrane currents, similar to the Hodgkin-Huxley equations for the squid giant axon. It assumes independent Na channels with only two conducting states, open or closed. Each Na channel is controlled by three m gates and one h gate. Each gate may be likened to the random-switch model discussed in Section 11. The channel is multistate in that various conditions of the four gates may exist for the closed state. However, each channel has only one open state. The kinetics of the channels conform to the model of the macroscopic Na conductance. The macroscopic parameters necessary to fit the current spectral densities predicted by the model were measured from the same nodes used for the noise experiments. The steady-state Na current was measured simultaneously with the Na current fluctuations. These assumptions allow the calculation of the conductance of a single open Na channel. Referred to the resting potential, the value is 3/Nn

= 7.9

* 0.9 p s

The measurement of yNadid show slight voltage dependence, being higher at greater depolarizations. In the presence of 0.1 mM external nickel, the value of yNais about half that in normal nickel-free Ringer’s solution. The power spectra were fit by a theoretical sum of m-gate and h-gate fluctuations. The h-gate fluctuations contribute more to the low frequencies than the m-gate fluctuations. To some extent, these contributions to the total spectrum may be varied independently. Because of the greater confidence in the data at higher frequencies (due in part to the presence of a background noise, which is presumably not related to the conductance fluctuations being discussed here, and which varies inversely with frequency), the values of yNamay be regarded as having been calculated from the m-gate fluctuations. The value of rm obtained from macroscopic voltage clamp records was compared with the microscopic values of r, derived from current noise spectra. In both normal and in Ni-treated nodes, qualitative agreement was obtained. In principle, this comparison allows one to determine the exponent of the kinetic variable rn, i.e., the number of m gates which control each channel. NO

FLUCTUATION ANALYSIS I N NEUROBIOLOGY

185

firm conclusion was reached because of insufficient data, and m“ was used in the theoretical calculations. By selective interference with the inactivation process (h gates) using scorpion venom, anemonia toxin, or iodate, it was shown that the current noise spectra due to the m kinetics were little affected. The intensity of the spectra is enhanced because blocking the Na inactivation increases steady-state Na current. The suggestion is that h and m kinetics are separate, and that inactivation interacts with activation of Na channels in a simple on-off fashion. We note that the value of yNahas also been obtained from TTX binding experiments. From the number of TTX binding sites and the measurement of the gating currents, which yield the maximum charge movements involved in Na activation, one gets an upper bound on the possible number of channels. If the maximum Na conductance, GNa,is also known, an estimate of the conductance of a single channel is obtained. Almers and Levinson (1975) estimate yNaas 1 pS using what is probably a low value of GNa.This point is discussed by Conti et al. who derive from Almers and Levinson’s binding data a value of yNaof 8.6 pS by using more recent and higher values of The value of 8.6 pS agrees well with the value obtained from their current fluctuations experiments described above. Similar values have also been obtained for the squid giant axon Na channels (see below). The current values of yNa from nerve axon are much smaller than earlier estimates of 100 pS or greater (Hille, 1970). The new values of yNalead to the interesting conclusion that the higher value of GNain the node than in the squid giant axon is due to a greater density of Na channels in the node rather than a difference in the Na channel itself. There are about 10”Na channels per node, or about 2000 per square micrometer. I n the squid giant axon there are less than 500 Na channelslpm’. Nonner et al. (1975) have calculated that the maximum sodium conductance in the node of Ranvier is about 10 times that of the squid giant axon; thus, 15 nS/pm2 for the node and 1.2 nSlpm’ for the giant axon. They also report an insensitivity to external Na concentration of ions of the Na slope conductance for large depolarizations in giant axons. From their measurements of the displacement current in nodes they estimate about 5000 Na channels/ pm2. Using this value, the ratio of Na channel density in the node to the density in the giant axon agrees roughly with the ratio of 10 for the maximum conductance. Lastly, since the external solution used in squid axon experiments contains more external Na than the node, it is noteworthy that the two values of yNain the node and in the squid agree as well as they do. Conti et al. imply that the channels in the two cases are similar in nature. The channel is normally “filled” and may not be

cNa.

186

LOUIS J. DE FELICE

further influenced by more external Na ions. Similar situations exist in simple model membrane systems. DeFelice and Michalides (1972) studied noise from collodian membranes. It was shown that the resistance of the membrane was well described by a model in which the internal ionic concentration of the membrane remained independent of the external bathing solution. On the other hand, fairly large channels induced in bimolecular lipid membranes by proteins added to the surrounding medium act to a certain extent like watery channels. That is, their conductance is influenced by external ionic concentrations (Lattore et al., 1972). A direct test for the nerve axon would be to measure current noise under conditions of varying external ionic concentrations. Electrical noise from the node of Ranvier of Rana temporaria was originally studied by Verveen and Derksen (see Verveen and Derksen, 1968, for an early review). One of the more recent studies is by van den Berg et al. (1975). By internal application of both Cs and TEA, these authors report the existence of a noise component due to the Na conduction system. Their data represent measured voltage fluctuations under current clamp. In order to block the K system, cut ends outside the central measurement node are exposed to Cs and TEA ions. The central node is bathed in normal or TTX-Ringer. In normal internal conditions with the K system intact, the variance of noise voltage increases monotonically with depolarizations from rest. In the Cs-TEA-treated nodes, the variance (between 20 and 1000 Hz) has a maximum at -38 mV. This bell-shaped voltage dependence of the variance disappears with TTX or in Na ion replacement experiments. The maximum noise in Cs-TEA is only about 1/10 the value of the normal voltage noise from the node at that same voltage. Voltage noise spectral densities were measured in the range 10-1000 Hz for membrane potentials between -70 and -20 mV. These spectra are fitted to a model which includes a simple on-off conductance component (random switch) due to the open-close kinetics of Na channels. The magnitude of this component also has a maximum value near -40 mV. When TTX is added, this spectral component disappears. Steady-state Na current is known from other studies to peak at about -40 mV. Van den Berg et al. interpret their results as representing the h process of the Na conduction system. They reasoned that the m process should have a much lower magnitude if any of the models they used were interpreted with the help of other voltage clamp data on a similar preparation (Hille, 1970). They also report that qt is within 30% of the measured time constant from the power spectra (at 20°C). Conti et al. ( 1976a,b) have interpreted their current noise data (at higher frequencies) primarily in terms of the m process. The major assumption in inter-

FLUCTUATION ANALYSIS IN NEUKOBIOLOGY

187

preting voltage noise spectral densities in terms of conductance fluctuations is that the nodal impedance is not frequency dependent. The voltage clamp equations of Hille (1970) used by van den Berg et al. do predict a rather complicated frequency- and voltage-dependent smallsignal impedance (Clapham and DeFelice, 1976). At the present time, it is unclear whether or not the nodes used by van den Berg et al. (1975) displayed the resonance behavior expected from excitable axon membranes. Capacitance neutralization amplifiers were used to help eliminate the rather large capacitative component of the input impedance. Until recently, no voltage clamp data existed from the preparations used by van den Berg et al. Two reports are now available in which current noise has been measured (van den Berg, 1976: van den Berget al., 1977). Large differences between long-term and short-term (less than 600 msec) voltage clamp records are discussed which are pertinent to their earlier studies. Van den Berg et al. (1975) conclude from their experiments that the density of Na channels is about 20,00O/pm'. This is roughly 10 times the figure arrived at by Conti et al. (1976a). Each Na channel has a yNa'v 2 pS, or about f the value of Conti et nl. In an earlier paper, Siebenga et al. (1974) reported K noise from the node of Ranvier of Rana tempmaria from essentially the same preparation as that used by van den Berg et al. (1975). Voltage noise was measured between -70 and +30 mV under current clamp at room temperature. Na-replacement blocking agents and metabolic inhibitors had no measurable effect on the spectra. The observed noise was most affected by TEA, although in a rather complex manner. The spectra were interpreted in terms of a leakage conductance plus battery, which are in parallel with a similar branch for the K system. The magnitude of the observed noise was therefore dependent on the relative magnitude of the K and the leakage conductances. The random-switch channel model was used for the K system to interpret part of the observed spectra. N o voltage dependence of the time constant obtained from voltage spectral densities could be measured. The data were not fitted to the model quantitatively, so that no elementary properties of the K conductance system were deduced. However, in an earlier study by Siebenga et (11. (1973), values of yKbetween 10 and 40 pS are reported. Siebenga et al. investigated the impedance of their node preparation under current clamp up to 150 Hz and report that in this range the impedance is frequency independent. In any case, their voltage noise spectral data extend to beyond 1000 Hz. These impedance considerations are implicit in the interpretation made by van den Berg et al. (1975) in their interpretation of voltage noise from the Na system discussed above. Van den

188

LOUIS J. DE FELICE

Berg et al. (1977) report a value of yK = 2.9 & 0.4 pS from their recent current noise experiments. N o voltage dependence of 3/k on voltage was found for V between -30 and +5 mV. Begenisich and Stevens (1975) have reported results of current noise measurements from the K system of the node of Rana. Their preliminary results use the variance of the noise to estimate the single channel conductance to be If the K channels have only two states, open and closed, then yk represents the conductance of an open state. In this case, the voltagedependent conductance of the membrane is not due to a dependence of yk on voltage. If the K channels have many conductance states, then the value of y k above represents some sort of average over these states; in this case, yK could be expected to be voltage dependent. In the Begenisich and Stevens study, the Na and K systems were selectively blocked by the use of TTX and TEA. Current noise was measured under voltage clamp between 4 and 2000 Hz. The steady-state K conductance was measured for each noise experiment. Fluctuations around the mean membrane current varied directly as GK and were effectively removed by TEA. Between V = -48 and +16 mV, at 15"C, the value of yKdid not significantly depend on voltage. One interpretation of their results is that K channels have only two conductance states which differ by about 4 pS. Begenisich and Stevens compare their results with an estimate of yKfrom the squid giant axon. yK in the squid is more than three times larger than in the node (see below). They make the interesting suggestion that this discrepancy might be expected since seawater has a higher conductance than frog Ringer's solution. This would be in contrast to the Na-channel values of yNa,which are very similar in the node and the giant axon in spite of large differences in external solutions. Yi-der Chen has emphasized the theoretical use of membrane noise to distinguish between various kinetic models of the elementary ionic conductance changes. Chen (1976, 1977) and Chen and Hill ( 1973) have written extensively on the use of the voltage dependence of current spectral densities to decide between alternative kinetic schemes for the elementary conductance change. Chen (1976) has stressed the use of the variance of the noise to test for the two conduction-state random-switch model as opposed to the multiconduction-state model. Begenisich and Stevens (1975) tentatively concluded that the relative insensitivity of yK to membrane voltage argues favorably for the random-switch model. Chen concludes that this is unlikely and that a careful examination of the voltage dependence of yK may indicate multiconducting states.

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

189

It is worth stressing at this point that an ensemble of independently acting channels may well give rise to nonlinear electrical current-voltage relationships even though each open channel is a linear ohmic device. Thus, the macroscopic behavior of a membrane's conductance need not be reflected in every channel. An ensemble of two-state random switches, where the duration of one of the states is voltage dependent, can account for current-voltage relations similar to those seen in biological membranes. This has been shown rather directly in an electronic model by Mauro and Rossetto (1976). A similar phenomenon exists in lipid bilayers modified by a proteinaceous substance which forms channels in the membrane (Ehrenstein et al., 1970). 2. Squzd Giant Axon Membrane noise from squid p a n t axon has been studied by Conti, DeFelice, and Wanke (Wanke et al., 1974; DeFelice et al., 1975; Conti et al., 1975). Voltage noise, current noise, and impedance were measured from the same preparations. Large areas of the axon membrane were isolated by means of air gaps. Initially, internal air gaps were used to isolate areas of about 0.03 cm'. For technical reasons related to amplifier noise, small areas may have advantages. This is one advantage the node preparations have. Subsequent experiments in the giant axon showed that membrane areas up to 0.3 cm' could be used. These larger areas were isolated with external air gaps. This provided a fairly standard axial wire space-clamped preparation, not unlike those used for macroscopic kinetic measurements. Noise and impedance were studied in the frequency range 1-1000 Hz between temperatures of 2°C and 26°C and over a membrane voltage range of - 100 to -40 mV. Normal rest potential at 6°C is about -60 mV. Both TTX and TEA were used to separate K and Na components of conductance fluctuations. The initial observations emphasized the experimental relationship between S v, S,, and IZ '. Z is the small-signal impedance of the excitable nerve membrane. The passive RC component of the membrane is only part of the total impedance. The rest is due to the voltage-dependent, time-varying conductances in the membrane. I t was established that in well-isolated and well-controlled preparations, voltage noise spectra are dominated by the shape of the total membrane impedance. Similar behavior is expected for any excitable membrane, since the conductance kinetics responsible for excitation are reflected in the membrane smallsignal impedance as voltage-dependent components of the equivalent circuit. By comparing current noise spectra in normal seawater and in the presence of TEA and TTX, both K and Na components of conductance fluctuations were observed. The two ionic components were compared

I

190

LOUIS J. DE FELICE

with a theoretical model which assumes that the channels conform to a Hodgkin-Huxley-like description for macroscopic behavior. The parameters for this description were obtained from unpublished data on the same preparations. This model was used to fit the shape of the current noise spectra. From the dependence of noise amplitude on voltage, both NK and NNa could be deduced. The elementary conductances of K and Na channels were obtained from the experimental values of the K and Na conductances for these axons = 70 mS/cm’, maximum GNa= 130 mS/cm2).The results are:

(c,

NK= 60Ipm‘ NNa = 330Ipm‘

YK

= 12 p s

yNa = 4 pS

There are fewer K channels than Na channels per unit area, but every K channel has a larger conductance. These data from the squid were used in the comparisons made above with the K and the Na system for the node. Electrical noise has also been studied from giant axons by Fishman, Poussart, and Moore (Fishman, 1973, 1975; Fishman et al., 1975a,b). In their experiments, small areas of membrane ( 10-4-10-5 cm’) were isolated on the external surface of squid giant axons using sucrose gap techniques. The “isolated” patch (10-100 m a ) is shunted by a relatively low resistance pathway (1 MR) through the sucrose solution in the Schwann cell interstitial space. Voltage and current noise spectral densities are similar because of the sucrose shunt pathways. The authors claim that the patch method improves channel noise resolution in comparison to large-area axial-wire measurements. Fishman et al. have not made a quantitative comparison between their data and a model. They report that the shapes of their K-noise spectra (Fishman et al., 1975a) are not compatible with either first-order reaction kinetics (single Lorentzian) or the probabilistic versions of conduction fluctuations based on the Hodgkin-Huxley model (Hill and Chen, 1972; Stevens, 1972). They conclude that the use of two-state conductance models in axons is questionable. New impedance data from this group (Fishman et al., 1977) purport to verify their interpretation of the observed spectral densities from the patch-clamped axon membrane. The noise analysis of kinetic systems and its application to channel kinetics has been thoroughly discussed by Chen (1977). The entire question rests on whether the measured spectra truly reflect conductance fluctuations. Since conductance fluctuations are only observed indirectly through current or voltage fluctuations, interpretation is often difficult. Fishman et al. (1976) have recently verified a noise component due to the Na system which had been suspected from some of their earlier data.

FLUCTUATION ANALYSIS IN NEURORIOLOGY

191

B. DRUGINDUCED NOISE 1. The Neuromirscular Junction The most complete noise data, in terms of a correspondence between the macroscopic and the microscopic, exist for the neuromuscular junction. The interaction between the ACh molecule and the ACh receptor in the postsynaptic membrane has been studied from the natural release of' packaged ACh, through the effect of individual ACh molecules iontophoretically applied to the ultimate observation of single-receptor interactions. This study is summarized in Fig. 1 (see Section I , p. 172). The most quantitative study of ACh-induced noise has been done by Anderson and Stevens (1973). Their voltage clamp analysis of the conductance fluctuations caused by the interaction of the ACh molecule and its receptor has led to a fairly complete biophysical description of the elementary conductance change associated with that event. This analysis and its relationship to other work on the postjunctional conductance increase induced by ACh is nicely summarized in Stevens (1975). Anderson and Stevens used the sartorius nerve-muscle preparation of Rana pipiens. The excitation-contraction coupling was disrupted by either ethylene glycol or glycerol treatment. High Ca and Mg Ringer's solution was used to improve input resistance and resting potential. Two intracellular glass microelectrodes, about 50 P apart, were necessary for their voltage clamp circuit; one electrode provided the holding current needed to clamp the voltage at a desired level, the other electrode monitored the voltage. A third external electrode, positioned near the intracellular electrodes, was used for releasing ACh to the end-plate. Under voltage clamp, the spontaneous miniature end plate currents were used as criteria for recording. End-plates with low spontaneous rates of firing of the miniature end-plate currents (less than l/sec at 18°C) were used. The current noise which was induced between the spontaneous miniatures was studied. The relationship between current noise, voltage noise, and impedance o f the postsynaptic membrane was discussed. Voltage noise and impedance were also measured, but only current noise data under voltage clamp is shown and analyzed. The analysis made use of difference spectra derived by subtracting control data without applied ACh from spectra in the presence of a steady-state application of ACh. This was done at each level of membrane potential, even though in most cases extraneous noise was only a few percent of ACh-induced noise. The primary data are an increase in the mean membrane current, and the fluctuations about the mean (the variance) upon application of ACh to the end-plate (cf. Fig. 1B). Relative to the equilibrium potential

192

LOUIS J. DE FELICE

for evoked end-plate currents, the variance and the mean are linearly related. That is, if one calculated conductance fluctuations from the observed current fluctuations, the variance and the mean are related by a constant. The current spectral density of the induced ACh noise were measured in the frequency range 1-300 Hz. The spectra were obtained over a voltage range of - 150 to 50mV. T h e shape of the spectra were well described by a single Lorentzian. Although the increase in membrane current and the fluctuations about the mean gradually diminish during constant ACh iontophoresis, the spectral shape (notably the half-power frequency) does not change. Spectra were taken from noise samples before appreciable desensitization during steady application of ACh had occurred. The ACh molecules are viewed as binding rapidly to the receptors in the postsynaptic membranes. This transmitter-receptor combination is fluctuating randomly between open and closed states. The elementary conductance change is considered to be a rectangular pulse with constant amplitude and random duration, i.e., equivalent to the random-switch model discussed earlier in Section 11. The rate-limiting step in the ACh-receptor interaction was considered to be the relaxation of an open channel. For low ACh concentrations, the model predicts the observed spectral density. The only arbitrary constant in the measured spectrum is 7, the single-channel conductance of the open ACh-receptor complex. The single-channel open conductance was found to be y = 32 pS

This value is sensibly independent of voltage. However, the rate-limiting relaxation is voltage dependent. On the average, a channel stays open longer at hyperpolarizations (the mean open time at - 150 mV is about 15 msec; at 50 mV it is about 4 msec). The mean open time (0) depends exponentially on membrane potential. The constants which fit this dependence are interpreted as the change in a dipole moment associated with the transition from an open to a closed configuration. Anderson and Stevens com pared their microscopic noise data with macroscopic voltage clamp data from the same preparation. The kinetics of ACh-receptor interaction may be determined from the spontaneous miniature end-plate currents. Miniature end-plate currents decay exponentially. The time constants depend exponentially on membrane potential in precisely the same way that the noise spectra time constants depend on membrane potential. This statement is also true of end-plate currents. It is important to realize that, while the macroscopic relaxation is truly exponential, the microscopic relaxation is viewed as a random switch with exponentially distributed durations of the open state.

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

193

The estimates of y are also independent of temperature. However, increasing the temperature does decrease the mean open time of a channel. The rate constants depend exponentially on temperature. A Q l o of 2.77 near rest is reported, but this value may vary between 2.4 and 3.6 depending on membrane potential. The spectral density of the spontaneous conductance fluctuations may be derived from the decay of end-plate currents. Unfortunately, the formal relationship between the macroscopic and the microscopic is not unique. The two-state conductance model used by Anderson and Stevens could be replaced by a channel which opened instantaneously and then decreased its conductance exponentially. All of their data would be equally well described under this assumption, although the values of parameters used to fit the data would take on different meanings in the two models. Evidence for the random switch model was quite indirect when Anderson and Stevens first proposed it for the end-plate. The assumption of the two-state model was partly made on the basis of analogies with work on artificial membranes (e.g., Ehrenstein, et al., 1970), where individual conductance changes are seen as rectangular pulses. As we now know, individual conductance changes have been observed for the receptor interaction which are compatible with this model (Neher and Sakmann, 1976b; Fig. 1C). Assuming that there are l o xACh channels at the end-plate, the single channel conductance of 30 pS would imply a total possible conductance of 3 mS. The observed peak conductance of 4 pS implies that only 0.13% of the total number of channels are open at the peak conductance. This result is important for certain simplifying assumptions made in the statistical analysis. The biophysical model of the postjunctional conductance increase was discussed by Stevens (1975) in a Cold Spring Harbor Symposium. The main features of this model are given here. Either one or two ACh molecules bind the ACh-receptor. This induces a conformational change which allows Na, K, and C1 ions to flow through the postsynaptic membrane. There are two principal conformational states; the difference between the two corresponds to an elementary conductance change of about 25 pS (independent of whether one or two ACh molecules open the channel). The two conductance states are separated by an energy barrier and do not easily flip from one to the other. The binding of ACh reduces the energy barrier and allows the channel to open and close randomly. These two states correspond to a change in a dipole moment perpendicular to the membrane of about 50 D. The electric field-dipole interaction is the mechanism for altering the energy barrier between the two conducting states. Depending on the field strength, the channels

194

LOUIS J. DE FELICE

normally stay open between 1 and 10 msec. The current which flows through an open channel corresponds to about 2 x lo7 iondsec, or i=3pA ACh binding is rapid compared to the conformation change and the channels behave independently. This fairly detailed picture of the ACh-receptor interaction does not rely entirely on noise analysis. Stevens first describes the postjunctional conductance increase (similar to Hagins’ “quanta1 response”) in response to brief ACh transients. However, the ACh transients are neither under experimental control nor are they directly measurable. Noise analysis, in the presence of a constant concentration of ACh, was considered advantageous although the two kinds of information converge. Also, information about the size of the elementary open-channel conductance y is only obtained from the noise analysis. The discovery of ACh noise and the first use of spectral analysis to help describe the underlying elementary events was made by Katz and Miledi (1972).The work is reminiscent of the suggestion made by Hagins et al. (1962) referred to in Section I, namely, that single transmitter molecules might act perceptibly at the postsynaptic junction. Katz and Miledi’s original photograph of the effect still remains the most illustrative, and is reproduced in Fig. 1B. Their analysis of ACh voltage noise is related to underlying conductance fluctuations, although to a lesser degree than transmembrane current noise. They were able to estimate the elementary conductance change to be the order of 100 pS. The current pulse associated with this change was approximately lo-’’ A lasting for about 1 msec. This produces minute voltage depolarizations the order of 0.3 pV which sum in random fashion to the effect shown in the baseline of Fig. 1B (lower panels). It was evident to Katz and Miledi that the time course of the voltage fluctuations were being filtered by the impedance of the postsynaptic membrane. By using focal extracellular recordings they, in effect, were able to measure current noise albeit indirectly. This allowed the more direct observations of conductance fluctuation kinetics. From these records they were able to show that carbachol produces a more brief current pulse than ACh and is therefore less effective. The comparative kinetics of drug-receptor interactions has been one of the major contributions of the technique of noise analysis. Katz and Miledi were led to a rather specific picture of the interaction. The macroscopic effects of prolonged and more stable depolarizations due to carbachol as compared with ACh are because carbachol is not hydrolysed by AChesterase. Noise analysis studies the effectiveness of the drug-receptor

FLUCTUATION ANALYSIS IN NEURORIOLOGY

195

interaction and not local drug concentrations. Katz and Miledi also showed that at low temperature the duration of the elementary current pulse increases. This was also found to be true for chronically denervated preparations. Katz and Miledi (1974) have also studied the effect of procaine on the action of ACh at the neuromuscular junction. I t was suggested that the elementary potential change, calculated from the ratio of the variance of the noise to the mean depolarization, is substantially reduced by procaine. The spectral analysis of the ACh-induced voltage noise in the presence of procaine is more complex than a single relaxation spectrum. This is related to a similar complex shape observed in the end-plate current of procaine-treated muscle. Ruff (1976) concludes that anesthetic molecules reversibly block ion conduction while preventing gates from closing. One view of ionic channels which are induced in the end-plate region by drugs is that the channel, once open, has a conductance which is independent of the drug used to open it. For example, Neher and Sakmann (1976a) have studied the effect of carbachol and suberyldicholine (SubACh) at the end-plate (and in regions outside the end-plate in denervated preparations, see below) and conclude that the single-channel conductances for these two are not substantially different from the value of y obtained for ACh-induced channels. In contrast, Colquhoun et nl. (1975) report that four cholinomimetic agonists each have different values for y. One of the drugs studied by Colquhoun et nl. in the endplates of summer Rnna Pipiens cutaneous pectoris muscle was also SubACh. In Colquhoun et al., the mean single-channel conductances were calculated from the variance of current noise and the average membrane current under voltage clamp. N o spectral densities of the noise are shown. I t is stated, however, that ACh and SubACh have simple Lorentzian spectra while the other two agonists have more complicated spectra well fitted by the sum of two Lorentzians. The method used assumes that the current fluctuations represent random opening and closing of channels similar to the model of Anderson and Stevens (1973). ACh and SubACh were found to have single channel conductances which differ by about 20%. The values are 25.0 k 0.9 pS and 28.6 2 1 .O pS, respectively. SubACh has a longer mean open lifetime than ACh, a result also found by Neher and Sakman (1975) and Katz and Miledi (1973). However, two agonists known to cause contraction in frog muscle have substantially lower values of y than either ACh o r SubACh. These are 3-(m-hydroxphenyl) propyltrimethylammonium or HPTMA and 3-phenylpropyltrimethylammonium or PPTMA (see Colquhoun et nl., 1975, for details). The values of the single-channel conductances are 18.8 k 0.8 pS for HPTMA and 12.8 k 1.1 pS for PPTMA. The analysis assumes low-

196

LOUIS J. DE FELICE

receptor occupancy. If this is not the case, the values given for y are lower limits for the single-channel conductances. The receptor of the normal cholinergic synapse has also been altered and changes in noise spectra observed. Landau and Ben-Haim (1974) studied voltage noise from frog sartorius muscle end-plates. T h e preparation was treated with dithiothreitol (DTT) which is known to decrease the postsynaptic sensitivity to ACh by altering the receptor molecule. Elementary voltage events were analyzed from the voltage noise data assuming a shot model similar to that used by Katz and Miledi (1972). The amplitude of the elementary pulses derived from the data were decreased by about half in DTT from normal values of 0.18 p V . They conclude that the elementary conductance event is probably reduced both in amplitude and duration. As in Katz and Miledi, extracellular noise was measured to help separate conductance fluctuations from membrane impedance. Their results show that changes in observed noise occur not only by modifying the agonist, but also by changing the structure of the receptor molecule. This problem has been studied in greater detail by Ben-Haim et al. (1975). Noise has also been studied during the decay phase of miniature end-plate currents prolonged by ethanol (Quastel and Linder, 1976). The amplitude of the noise corresponds to unit events similar to those produced by applying ACh directly to the end-plate.

2. Extrajunctional Receptors Chronically denervated frog muscle fibers have been studied recently under voltage clamp by Dreyer el al. (1976) and Neher and Sakmann (1976a). I t is well known that the entire extrasynaptic muscle membrane becomes sensitive to ACh following denervation. These extrajunctional receptors may have different macroscopic pharmacological properties than the normal junctional receptors. Neher and Sakman studied the effects of ACh, carbachol, and suberyldicholine (SubCh) on the denervated and normal cutaneous pectoris muscles from Rana esculenta and tempmaria. Experiments were done 40-70 days after denervation. T T X was used in the extracellular bath. The two-microelectrode clamping circuit was similar to that used by Anderson and Stevens. Autocorrelation functions were calculated from the current noise data and these were compared with simple exponential functions to extract one or more time constants from the process. The first point (7 = 0) of the autocorrelation function is the variance (see Section 11). By knowing the mean current (Z) associated with the current fluctuations, the elementary current L = C(O)/Z and the single-channel

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

197

conductance y = d(V - E ) were calculated. T h e equilibrium potential ( E ) for the channels was determined separately to be near 0 mV. The conductance of a single extrajunctional ACh-induced channel is: y = 1 5 2 1.8 pS

This is somewhat lower than the junctional open-channel conductance measured both for these preparations (23 & 2 pS) and the preparation used by Anderson and Stevens (1973). The average open time ( 8 ) of an extrajunctional ACh channel is about 1 1 msec at -80 mV and 8°C. This is reported as being three to five times the average open time of a normal junctional channel. Under the same conditions, carhachol extrajunctional channels are open less on the average (4 msec) and SubACh are open longer (19 msec). Both carbachol- and SubACh-induced extrajunctional channels are open three to five times longer than junctional channels. The extrajunctional singlechannel conductances for carbachol and SubACh are similar to the ACh value given above. An interesting result obtained by Neher and Sakmann is that former end-plate regions of chronically denervated fibers often (1 1 out of 15 experiments) show two time constants in the autocorrelation functions of the drug-induced current noise. The faster time constant corresponds roughly to the normal fiber. The slow component was similar to that derived from the extrajunctional denervated experiments. They conclude that this could indicate two populations of channels at former end-plate regions of denervated fibers. Neher and Sakmann make the assumption in their analysis that only a small percentage of the channels are in the open state during their current noise measurements. This is fairly well documented in the druginduced noise in normal fibers but has not been unequivocally established for extrajunctional denervated preparations. If the assumption were not valid, and high-receptor occupancy occurs in the extrajunctional experiments, the lower values of y given for this region may have to be raised. It is more difficult to see how the larger values of 8 which were obtained for extrajunctional regions could be explained by highreceptor occupancy. A genuine difference in the kinetics of extrajunctional versus normal ionic channels appears well established. A new study of junctional and extrajunctional ACh receptors that compliments the noise analysis and deals with the question of cooperativity, has been made by Dreyer et al. (1977). ACh-induced current fluctuations from muscle cells in tissue culture has been studied by Sachs and Lecar (1973, 1977). Chick skeletal muscle

198

LOUIS J. DE FELICE

cells were used. These cells possess ACh receptors spread over their entire surface. Current spectral densities were also obtained using carbachol as the agonist. A standard two-microelectrode voltage clamp was used on cells transformed to an approximately spherical shape by treatment with vinblastine. A useful analysis of noise sources from microelectrode voltage clamped cells is also given. Sachs and Lecar conclude that, in this preparation, y = 39 pS at 25°C

but increases with temperature with a Ql0of 1.7. The current spectral densities are well described by single Lorentzians. The relaxation time of the channel conductance is more markedly temperature dependent. The mean open duration is about 3 msec at 30°C. The Ql0is about 5 in the direction of increased 8 with decreased temperature. The relaxation time for carbachol is shorter than for ACh, in agreement with earlier work on the end-plate. 3. Glutamate Receptors Crawford and McBurney (1976) have studied voltage noise in the giant muscle fibers of the walking legs of the spider crab Maia squinado. These fibers may be cannulated and a large Ag-AgC1 wire inserted into the fiber. The fibers have a low membrane resistance (about 360 R * cm2). Relatively large areas (0.6 cm') gave preparations with an input resistance of about 600R. This is considerably lower than other preparations discussed so far. Extracellular noise was also recorded from these preparations using micropipettes. These experiments are similar to those of Katz and Miledi (1972) described above. The spontaneously occurring miniature excitatory junctional potentials (ejp's) have a mean amplitude of about 5 pV, which is about 1/100 the value of the frog end-plate spontaneous miniatures. This is due mainly to the difference in input resistance in the two fibers. O n applying L-glutamate to the external bath, transmembrane voltage recordings show a steady depolarization accompanied by an increase in voltage noise. The mean and the variance of this glutamate-induced noise were linearly related. The ratio of the noise variance to the mean depolarization was 2.2 X IO-'"V. A shot model was used which assumes numerous and random exponentially decaying elementary voltage events. The mean size of an event was therefore calculated to be about 5 x lo-'" V. Comparison of this value with the size of the ejp's led to the conclusion that each ejp contains 5000-10,000 elementary events. This is considerably larger than the number of events probably occurring during quanta1 release of ACh at the frog end-plate.

FL UC T UAT I ON ANALYSIS IN NE U R O B I O L O C Y

199

The power spectrum of the intracellularly recorded voltage noise was fitted by a single Lorentzian with a time constant 8 o f 8 msec. This cutoff frequency (20 Hz) was thought to be mainly due to the membrane’s impedance. Extracellularly recorded noise, which more nearly reflects current fluctuations through the membrane, was fitted to a Lorentzian with a time constant of 1.4 msec. Similar voltage noise does not occur in the presence of y-aminobutyric acid (GABA) in place of the t-glutamate. The effect of GABA on crayfish muscle is known to be the induction of changes in chloride conductance. The driving force for chloride is small near the resting potential (about -50 mV) of the fibers. Crustacean muscle fibers are known to have excitatory junctions distributed over their entire surface. Focal electrode recording showed that only where depolarizing extracellular miniatures were recorded did L-glutamate produce extracellularly recorded noise. Thus “silent” regions of the muscle surface could be found. An estimate was made of the current associated with the glutamateinduced elementary conductance event. A lower limit of 0.5 pA was established. I t was concluded that the size and duration of the glutamate elementary event is very nearly equal to values obtained for ACh elementary events at the frog end-plate. Crawford and McBurney believe that the quanta1 current delays are probably limited by the closure of conductance channels and not by the relaxation of the concentration of transmitter molecules in the synaptic cleft, which they consider to be very fast. There is no known enzyme system at crustacean neuromuscular junctions for the rapid denaturing of the transmitter. Crawford and McBurney point out that the observation of transmitter noise is a useful criterion for the identification of a putative molecule as the transmitter substance. Dionne and R ~ i f f(1976) has made a similar point by using drug-induced noise to measure the equilibrium potential of a postsynaptic membrane current. A current noise-voltage clamp analysis of glutamate noise has been done by Anderson et ul. (1976). The metathoracic extensor tibiae of the adult locust Scistocercn greguria was used. This preparation is composed of fibers about 1500 pm long by 100 pm in diameter. In Cl-free medium, the input resistance is about 5-8 MQ. Chlorine-free bathing solutions were used to restrict the transmitter responses to the depolarizing type. Hyperpolarizing effects are due to the opening of CI channels and are avoided in the absence of the C1 ion. Spectral analysis of the current noise fluctuations plus a knowledge of the equilibrium potential for the glutamate-induced channels (0 mV) and the mean current through these channels were used to calculate the elementary open-channel conductance assuming a model similar to An-

200

LOUIS J. DE FELICE

derson and Stevens (1973). y was found to be independent of membrane voltage and current, but dependent on temperature. The mean value of the open-state conductance of the glutamate induced channel was y = 231 pS

at 23°C

Recent measurements give a final mean value of 122 pS (S. G. CullCandy, personal communication). Measured values varied between 100 and 260 pS. y decreases as the temperature is lowered. For example: y = 39 pS

at 8.5"C

This corresponds to a Qloof 2.12. Recall that ACh-induced channels in the frog end-plate were found to have much smaller conductances and are insensitive to temperature. Temperature also prolongs the open state of the glutamate-induced channel from 1.59 msec at 23°C to 3.23 msec at 8.5"C. This corresponds to a Qlo of 1.45 which is lower than that observed for ACh channels. 8 also depends on membrane voltage. T h e duration of an open-single-channel decreases with hyperpolarization, from 2.12 msec at -50 mV to 1.56 msec at - 110 mV at 23°C. ACh channels have increased 8 with hyperpolarization. The glutamate analog, quisqualic acid, also produced a mean depolarization and increased current noise. In general, the time constants 8 were nearly twice as long in quisqualic noise compared with glutamate noise, but the single-channel conductances were almost identical. Macroscopically, quisqualic acid is known to be several times more effective on glutamate receptors than the natural agonist. Anderson et al. close with the interesting remark that, relative to the dipole model discussed above for the ACh-receptor complex in the frog end-plate (Stevens, 1975), the glutamate-receptor complex in invertebrate muscle has opposite polarity. C. SENSORY SYSTEMS The unifying concept in sensory mechanisms is that external forms of energy ultimately result in the conductance change in either the receptor cell membrane or in a related cell in the sensory pathway. One is usually concerned with average changes, or large transient changes such as the production of generator potentials, that result from the interaction of a physiological stimulus and the sensory system. Individual QRs have been reviewed by Fuortes and O'Bryan (1972). The steady-state fluctuations from some average change reflect the underlying elementary events in a way that is analogous to the internal forms of communication between cells already discussed.

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

20 1

1 . Photoreceptors Simon et al. (1975) and Lamb and Simon (1976b) have investigated voltage noise from bipolar cells and cones and rods in the retina of the turtle Pseudemys scriptu eleguns. Cells in the isolated eyecup of the turtle were stimulated optically using either spots, annuli, or narrow slits. Transmembrane voltage was recorded with intracellular glass microelectrodes. The original observations showed that when the eyecup is exposed to a spot of light, the bipolar cells hyperpolarize and show a decrease in voltage noise. One value showed a change in noise variance from 1.53 to 0.045 X V 2 for a corresponding hyperpolarization of 6.5 mV. The ratio of the change in the variance between light and dark to the mean change in potential was about 0.24 mV. This was interpreted as the average size of the elementary voltage event in the bipolar cell membrane. Bipolar cells receive information from the cones in the form of a transmitter which is continuously released in the dark but is suppressed by light. Therefore, the hyperpolarization and decrease in noise in bipolar cells might be expected. The value 0.24 mV for the elementary event is approximately the size of a miniature end-plate potential at the frog neuromuscular junction. However, the current associated with this event in the bipolar cell is probably much smaller than at the neuromuscular junction since bipolar cells have much larger input resistances. Exposure of the turtle retina to an annulus of light showed either a slight depolarization or no effect in bipolar cells. The slight depolarization was also accompanied by a decrease in noise. Since horizontal cells are implicated in annulus stimulation, these data suggested to the authors that horizontal cells release a hyperpolarizing transmitter in the dark and that this release is reduced in the light. Cones and rods also show a similar decrease in noise when illuminated by bright steady light. However, the magnitude of the observed intrinsic dark noise varied widely between cells. The noisiest cones had a variance of 0.4 x V 2while quiet cones could be as low as 0.0 1 X lo-’’ V 2 .This difference is explained by Lamb and Simon as being due to the relative amounts of intercellular coupling between cones. The noisiest cones are those with narrow receptive fields. Quiet cones have relatively wide fields, indicating that they are well coupled to other cells. Their model assumes that similar random processes occur in all cones in the retina and that only the degree of intercellular coupling varies. Three different passive electrical models of coupling were developed. The essential problem to be solved is to calculate the sum of all the elementary noise sources in a coupled network of cells at some point in the network. It was required to measure intercellular coupling indepen-

202

LOUIS J. DE FELICE

dently. This was done by measuring the spatial profile of the voltage response to light. For a slit stimulus, the peak flash response showed an exponential decay with distance. The length constant (A) of the exponential decay was used as a measure of intercellular coupling. The models investigated predicted a strong dependence of input resistance and voltage noise variance on A. For well-coupled cells, the mean square voltage noise varies inversely as A?, Through the model one is able to determine the magnitude of the voltage noise from a cell as though no coupling existed. These calculated values agreed with occasional noise records from truly isolated cells (small A) which are much noisier than well-coupled cells (large A). It was noted by the authors that noise from isolated cells might mistakenly be taken as an increase in microelectrode noise. Unlike bipolar cell noise, intrinsic cone and rod noise was an unexpected observation. Small spots of light 6 pm in diameter could cause a 10-fold decrease in voltage noise from an isolated cone. Lamb and Simon propose that photoreceptor noise results from the opening and closing of light-sensitive ionic channels. In darkness, isolated cones have a change in noise variance of 0.4 x V' for an estimated 4 mV hyperpolarization from the level with all channels open. This implies an elementary event of about 100 pV, or 40 simultaneous events to produce a 4 mV change. Taking the input resistance of the isolated cone to be 200 M a (from other work) they estimated the current associated with an elementary event to be:

i = 0.5 pA The driving force for this current was taken to be 40 mV, the same as the voltage required to reverse the macroscopic light response. Thus an estimate of the elementary conductance change would be about y = 10 ps

This is only approximate. Since the value is nearly the same as the values found for other single-channel conductances, the authors propose that photoreceptor noise results from the opening and closing of individual membrane ionic channels. The voltage spectral density of the noise from turtle cones was also investigated (Lamb and Simon, 1976a; Simon and Lamb, 1977). The spectra were fitted to the product of two Lorentzians with time constants O1 and 02. The shorter (0,) is associated with the receptor cell's capacitive time constant. The longer (0,) is associated with the temporal property of the elementary conductance events. Values were scattered. A typical

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

203

example used 12 and 25 msec for 0, and 0 2 . 0, averaged 8 msec. 0, ranged from 17 to 60 msec with a mean value of 40 msec. Two interpretations were given for the temporal property of the elementary conductance events. 0, could correspond to the average time that molecules blocking the conductance channels are activated, or the average time a channel exists in the open and closed states. Simon and Lamb compare their noise data with specific kinetic models and photoisomerization data. They consider it likely that the measured elementary event (100 p V , 40 msec) is the same as that evoked by one photon. An event could either be the production of one blocking molecule or the closure of one ionic channel. Lastly they propose that the role of intercellular coupling in photoreceptors may be to improve the signal-to-noise ratio for diffuse stimuli. The induction of photoreceptor voltage noise in the dark in Drosophila has been studied by Minke et nl. (1975). A mutant was used which has a large initial photoresponse with increased stimulus but decays during stimulus to the level observed with dim stimuli. The decay of the receptor potential was investigated. By applying noise analysis, the authors show that at the saturated steady-state level the response to strong and weak stimuli is made up of elementary events of approximately the same size. Thus, the hypothesis (Dodge et nl., 1968) that light adaptation is due to reduced elementary events was considered to be improbable. 2. Mechanoreceptms Voltage noise has been studied by DeFelice and Alkon (1977) from the hair cells of the nudi-branch mollusc Hermimendn drassicornk in response to rotation of the statocyst. The statocyst is usually 70-100 p in diameter and is composed of 12-13 cells 40-50 pm in diameter and 5-10 pm thick. The cells form the periphery of the statocyst. A ciliated surface projects inward toward a cluster of 3-10 pm particles (statoconia) suspended in the cyst fluid. The receptor cells, unlike vertebrate hair cells, have axons. The axons leave directly from the cell bodies and join to form the static nerve (Alkon and Bak, 1973; Alkon, 1975). The statoconia are in constant motion. Rotation of tile statocyst may be used to promote an average movement of the statoconia toward some cells and away from others. Intercellular recordings may be made during rotation of the statocysts from cells either in front of or behind the vector of rotation. In the voltage noise studies (DeFelice and Alkon, 1977), spontaneous firing was suppressed by hyperpolarizing the soma under single-electrode current clamp. Cells with cut and uncut axons were

204

LOUIS J. DE FELICE

used. Noise was measured under constant rotational acceleration u p to 1 g. Cells in front of the force vector depolarize and the variance of the voltage noise increases. Cells behind the force vector hyperpolarize and the noise decreases. A 1 mV change in mean membrane potential produces about a 100 p V change in the rms of the noise. The data are consistent with a shot model of the statoconia-hair cell interaction. At -80 mV, the elementary voltage event which results from this interaction is about 20 p V in a cell with an input resistance of 80 M a . This corresponds to a current of 2.5 x A. Although the exact driving force for the current is unknown, an estimate of the elementary conductance change associated with event was found: y=5ps The voltage spectral density shows that most of the power is located at low frequencies (below 50 Hz). This probably reflects the impedance of the membrane more than the temporal events associated with the conductance change. Correlation functions of the noise were fitted to a single exponential at low frequencies. A hyperpolarizing response of 6 mV increases 0 from 65 to 166 msec. This nearly 3-fold increase may be interpreted in part as an increase in average membrane resistance as the statoconia are drawn away from the ciliated surface. In 1974, Wiederhold (1974) suggested that observed fluctuations in mechanoreceptor membrane voltage result from the statoconia striking the cilia of the receptor cell, and directly measured the input resistance in response to physiological stimulation. Recently, a thorough study has been made of the response ofAplysia calafornica statocyst receptor cells by Gallin and Wiederhold (1977), who studied the intracellular electrical responses of the ciliated mechanoreceptor cells as a function of tilting. Tilting in the excitatory direction (statoconia move toward the cell) caused a depolarizing potential and large membrane potential fluctuations. The voltage noise was reduced or absent during opposite tilting. The voltage spectral density of the noise is reported as being primarily below 3 Hz. Removing the synaptic input of the receptor cells did not abolish the increase in noise due to tilting, and a decrease in input resistance is reported during the depolarizing response (see also Wiederhold, 1974). By Na replacement experiments it was argued that the basic mechanism underlying the receptor potential in this mechanoreceptor is an increase in Na permeability. Wiederhold (1977) has shown that the effect of tilting on membrane resistance is more complicated than previously thought. Slope resistance depends on the membrane potential and includes anomalous and delayed rectification. The results of his studies are critical to a complete interpretation of voltage noise from hair cells.

FLUCTUATION ANALYSIS IN NEUROBIOLOGY

205

3. Stretch Receptors Membrane noise from the slowly adapting stretch receptor neuron of the lobster has been studied by Sjolin and Grampp (1975). A singlemicroelectrode voltage clamp technique was used. Current noise was measured in the 2-400 Hz range. Depolarization from the resting potential (-60 mV) caused the current noise to increase. This increase was abolished with TTX or by replacement of the extracellular Na but was unaffected by TEA. For moderate depolarizations, the computed correlation functions of the current noise were approximated by single exponentials. Near the firing threshold, a sinusoidal component was observed indicating a subthreshold periodicity, perhaps from peripheral regions not under good voltage control. Time constants between 1.6 msec (near rest) and 7.9 msec (near threshold) were obtained from the exponential component of the noise. These time constants were considered to reflect Na inactivation and were compared to the nonadapting firing mode in the preparations. Sjolin and Grampp report that a distinct Na noise was not found in the rapidly adapting stretch receptor neuron of the lobster. In 1966, Firth (1966) investigated the role of membrane thermal and shot noise on the firing pattern of the crayfish stretch receptor. These noise sources are not large enough to explain the observed interval fluctuations. This simple system has an unusually small variation in the impulse train, much less than motoneurons where the fluctuation in the interval is due in large part to random synaptic input. The paper is of interest because of the analysis of the relation of membrane noise and its relation to firing patterns of the cell. A useful expression is given for the variance of the voltage noise expected from a long discrete cable composed of axoplasmic resistors and an RC equivalent for the membrane. D. OTHERPREPARATIONS Fluctuation analysis has been applied to many biological membranes besides those described above. ,For example, noise has been measured during ionic transport in the frog skin (Lindemann and Van Driessche, 1977; Van Driessche and Lindemann, 1976) and possibly in active transport systems (Segal, 1972, 1974). DeFelice and DeHaan (1975, 1977) have measured voltage noise from heart cells in tissue culture. Fluctuation analysis of membrane noise may be used to measure electrical coupling between cells without the injection of extraneous current. Periodicities were observed in the correlation functions derived from voltage noise from heart cell membranes whose spontaneous beating has been suppressed by TTX. This periodic component is related to the normal oscillatory properties of the heart cell membrane.

206

LOUIS J. DE FELlCE

IV. Summary

This chapter has been developed around the use of a relatively new technique in electrophysiology to study basic membrane phenomena in excitable nerve axons, muscle cells, and sensory receptors. The survey of literature has emphasized those studies which compare membrane noise with specific models of conductance change. Thus llf noise, which appeared frequently (and in some cases exclusively) in earlier studies, has not been discussed. The explanation of l/f noise remains an important but as yet unsolved problem of nerve physiology. The general impression one obtains at this stage of the literature is that certain parameters, such as single-channel conductance or the density of channels, may be reliably obtained from noise data. This can apparently be done even when the correct kinetic model of the elementary conductance change is not known. Where individual channel conductances have been observed directly, good agreement exists between these events and those derived from noise measurements. The single event remains high on the list of experimental objectives and may one day be observed in the excitable nerve membrane. The use of noise measurements to distinguish between kinetic models of conductance change appears less well determined at the present time, although this is potentially the most important application of fluctuation analysis. The information we have obtained from fluctuation analysis has been useful in discussing possible physical mechanisms of conductance change in membranes. This information must ultimately conform to the molecular models of ionic conduction in biological membranes that are now emerging. REFERENCES Alkon, D. L. (1975).J . Gen. Physiol. 66, 507-530. Alkon, D. L., and Bak, A. (1973).J. Gen. Physiol. 61, 619-637. Almers, W., and Levinson, S. (1975).J. Physiol. (London) 247,483-509. Anderson, C. R., and Stevens, C. F. (1973).J. Physiol. (London) 235, 655-692. Anderson, C. R., Cull-Candy, S. G., and Miledi, R. (1976).Nnture (London) 261, 151-153. Begenisich, T., and Stevens, C. F. (1975). Biophys. J. 15, 843-846. Ben-Haim, D., Dreyer, F., and Peper, K. (1975). Pjiregws Arch. 355, 19-26. Bendat, J. S., and Piersol, A. G. (1971). “Random Data: Analysis and Measurement Procedures.’’ Wiley (Interscience), New York. Brock, L. G., Coombs, J. S.,and Eccles, J. C. (1952).J. Physiol. (London) 117, 431-460. Campbell, N. (1909). Proc. Cambridge Phil. SOC. 15, 117-136. Chen, Y.-D. (1976). Bio,bhy~.J. 16, 965-971, Chen, Y.-D. (1977).J. Chtm. Phys. (in press). Chen, Y.-D., and Hill, T. L. (1973). Biophys. J . 13, 1276. Clapham, D. E., and DeFelice, L. J. (1976). Pjhregers Arch. 366, 273-276.

FLUCTUATION ANALYSIS IN NEUROBlOLOGY

207

Colquhouri, D., Dionne, V., Steinbach,J. H., and Stevens, C . F. (1975).Nnt~rre(Lo7ldon) 253, 204-206. Conti, F., and Wanke, E. (1975).Qiccirl. R w . BifJ/JhyS. 8, 451-506. Conti, F., DeFelice, I. J.. and Wanke, E. (1975).J . Pliysiol. (Lwufon) 248, 45-X2. Conti, F.,Hille, B., Neitnicke, B., Nonner, W..and SCinipfli, K. (1976a).J. Phvtiol. (Londml) 262,699-727.

Conti. F., Hille, B., Neumcke, B., Notinel-, W., a n d Stampfli, R. (1976b).J. Pliysi~l.(Lmmdm) 262, 729-742. Crawford, A . C., and McBurney, R. N. (1976).J. Phyiol. ( L o d m , ) 258, 205-225. DeFelice, I.. J.. and Alkon. D. L. ( 1 9 7 7 ) . Bio/diy.\. Sor. A M r . 17, 1%. DeFelice. L. J., and DeHaan, R. L. (1975).Biofihys. Sor. Ahrtr. 15, 13Oa. DeFelice, L. J.. and DeHaan. R. L. (1977). Proc. IEEE Sper. I.tsite H i d . Sig7in/.\ 65, 796-799. DeFelice, L. J., and Michalides, J. P. L. M. (1972).J. Membr. Eiol. 9, 261-290. DeFelice, L. J.. and Sokol, B. A . ( I 9%). HiqViy.\, J . 16, X27-83X. DeFelice, L. J., and Sokol, B. A . (1976h).J. Mernbr. Biol. 26, 405406. DeFelice, L. J., Wanke, E., and Conti, F. (1975).Fed. Pror., Fed. Am. Soc. E ~ JBid. . 34, 1338-1342. Dionne, V., and Ruff‘, R. L. (1976).H i o p h y . Soc. A h / r . 16, 212a. Dodge, F. A., Knight, B. W., and Toyoda, J. (1968).Science 160, XX-90. Dreyer, F., Walther, C., and I’eper, K. (1976).Pflitegor.\ Arch. 366, 19-26. Dreyer, F., Miiller, K.-D., t’eper, K., and Sterz, R. (1977).P f l i i ~ g mA r d i . (in press). Ehrenstein, G., and Lecar, H. (1977). Qitor/. R e r ~Bio/hy.\. . 10 (in press). Ehrenstein, G . , Lecar, H., and Nossal, R. (1970).J. G m . P / i y . ~ i ~55, / . 119-133. Fatt, P., and Katr, B. (1950).Nalrire (London] 166, 597-598. Fatt, P., and Kat7, B. (l952).J. Phy.sio/. (Lmdmz) 117, 109-128. Firth, D. R. (1966). BiophyY.J. 6, 201-215. Fishman, H. (1973). Pror. Nut/. Acad. Sri. L’.S.A. 70, 876-879. Fishman, H. (1975).J. M e m h . Biol. 24, 265-277. Fishman, H . M., Moore, L. E., and Poussart, D. J. M. (1975a).J. Membr. B i d . 24,305-328. Fishman. H. M., Poussart, D. J. M., and Moore, L. E. (1975b).J. Mpmbr. B i d . 24,2Xl-304. Fishman, H. M., Moore, I. E., I’oiissart. D. J . M., and Siebenpi, E. ( l 9 7 7 ) . j . M e m h . B d . 32, 255-290. Fishman, H., Moore, L. E., and Poussan, D. J . M . (1976). H i o l . H i r I I . 151, 408409, French, A . S., and But7, E. C;. ( 1 973). I n t . J . Con/ro/ 17, 529-539. Fuortes, M. G. F., and OBI-yan. P. M. (1972). Hnndb. Sensmy Physiol. 8, 279-320. Gallin, E. K . , and Wiederhold, M. L. (1977).j. Physiol. (Londmi) (in press). Hagins, W. A . ( I 959). Bio/dy.r. SOC.A h t r . Pap. 0 1 . Hagins, W. A . (1965). Cold Spring Harbor Sym/~.@mi/. Bid. 30, 4 0 3 4 I X. Hagins, W. A., Zonana, H. V.. and Adams, R. G . (1962). Nnlrtre (London) 194, 844-847. Hill, T. L., and Chen Y.-D. ( 1972).Biofdiys. J . 12, 948-969. Hille, B. (1970). Prog. Siophys. Mol. Biol. 21, 1-32, Hodgkin, A . L., and Huxley, A . F. (1932).J. Physiol. (Londmi) 117, 500-544. Holden, A . V., and Rubio, J . E. (1976).H i d . Cybe17iet. 24, 227-236. Katz, B., and Miledi, R. (1970). Nntiire (London) 226, 962-963. Katz, B., and Miledi, R. (l972).J. Pliysiol. (London) 224, 665-699. Katz, B., and Miledi, R. (1973).J. Pliysiol. (London) 230, 707-717. Katz, B., and Miledi, R. ( 1 9 7 4 ) . j . Pliysiol. (Lmidon) 249, 269-284. Lamb, T. D., and Simon, E. J. (1976a). Proc. I’hysiol. Sor. rip. 5P-7P. Lamb, T. D., and Simon, E. J. (1976h).J. Pliysiol. (Londmz) 263, 257-2x6. Landau, E. M., a n d Ben-Haim. D. (1974). Srienrt 185, 944-946. Latorre, R., Ehrenstein, G., and Lecar, H. ( 1 972).J . Gen. Phpiol. 60, 7 2 - 8 5 .

208

LOUIS J. DE FELICE

Lee, Y. W., and Schetzen, M. (1965).Int. J . Control 2, 237-254. Lindemann, B., and Van Driessche, W. (1977). Science 195, 292-294. McCann, G. D., and Marmarelis, P. Z., ed. (1975). “Proceedings of‘the First Symposium on Testing and Identification of Nonlinear Systems.” Cal. lnst. Technol. Press, Pasadena, California. Marmarelis, P. Z.,and Naka, K. 1. (1974). f E E E Trans. Biomed. Eng. 21, 88-101. Mauro, A., and Rossetto, M. (1976). I n “Electrobiology of Nerve, Synapse and Muscle’‘ (J. P. Reuben, D. P. Purpura, M. V. L. Bennett, and E. R. Kandel, eds.), pp. 37-44. Raven, New York. Minke, B., Wu, C.-F., and Pak, W. L. (1975).Nature (London) 258, 84-87. Naka, K. I . , Marmarelis, P. Z., and Chan, R. Y. (1974).J. Neurophysiol. 38,92-131. Neher, E., and Sakmann, B. (1975).Proc. Natl. Acad. Sci. U.S.A. 72,2140-2144. Neher, E., and Sakmann, B. (1976a).J. Physiol. (Lundon) 258, 705-729. Neher, E., and Sakmann, B. (l976b).Nature (London) 260, 799-802. Neher, E., and Stevens, C. F. (1977).Annu. Rev. Biophys., Bioeng. 6 , 345-381. Nonner, W., Rojas, E., and Stampfli, R. (1975). P’uegers Arch. 354, 1-18. Poussart, D. J. M. (1969). Proc. Natl. Acad. Sci. U.S.A. 64, 95-99. Poussart, D. J. M. (1971).Biophys.J. 11, 211-234. Quastel, D. M., and Linder, T. M. (1976). SOC.Neurosci. A h t r . If Pap. 1031. Ruff, R. (1976).Biophys. J. 16, 433-439. Sachs, F., and Lecar, H. (1973).Nature (London), New Biol 246, 214-216. Sachs, F., and Lecar, H. (1977). Biophys. J . (in press). Segal, J. R. (1972).Biophys.J. 12, 1371-1390. Segal, J. R. (1974).Biophys.J. 14, 513. Siebenga, E., Meyer, A., and Verveen, A. A. (1973).Pfluegers Arch. 341,87-96. Siebenga, E., d e Goede, J., and Verveen, A. A. (1974).Pfluegers Arch. 351,25-34. Simon, E. J., and Lamb, T. D. (1977).In “Photoreception” (H. B. Barlow and P. Fatt, eds.). Academic Press, London. In press. Simon, E. J., Lamb, T. D., and Hodgkin, A. L. (1975).Nature (London) 256,661-662. Sjolin, L., and Grampp, W. (1975).Nature (London) 957, 6 9 6 4 9 7 . Stevens, C . F. (1972). Bi0phys.J. 12, 1028-1047. Stevens, C. F. (1975). Cold Spring Harbor Symp. Quant. Biol. 40, 169-173. van den Berg, R. J. (1976). Abstr. Dutch Federatirw Meet., 17th, Amsterdam. van den Berg, R. J., de Goede, J., and Verveen, A. A. (1975).PJuegers Arch. 360, 17-23. van den Berg, R. J., Siebenga, E., and d e Bruin, G. (1977).Nature (London) 265,177-179. Van Driessche, W.. and Lindemann, B. (1976).Pfluegers Arch. 362, R28. Verveen, A. A., and Derksen, H. E. (1968). Proc. fEEE 56,906-916. Verveen, A . A., and DeFelice, L. J. (1974). frog. Biofihys. Mol. Biol. 28, 189-265. Wagner, H. G., and Hagins, W. A. (1959). Biophys. Soc. Abstr. Pap. 02. Wanke, E., DeFelice, J. J., and Conti, F. (1974). Pflitegers Arch. 347, 63-74. Wiederhold, M. L. (1974). Brain Res. 78,490-494. Wiederhold, M. L. (1977).J. Physiol. (London) (in press). Wiener, N. (1958). “Nonlinear Problems in Random Theory.” M l T Press, Cambridge, Massachusetts.

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM By W. H. Gispen, J. M. van Ree, and 0 . d e W i e d Rudolf Mangur Institute for Pharmacology Medical Faculty, University of Utrecht Utrecht, The Netherlands

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . .

........................................................... A. ACTH 4-10 and Learned Behavior . . . .

11. ACTH 4-10

B. ACTH 4-10 and the Induction of Excess C. Electrophysiological Correlates of ACTH 4-1 0 . . . . . . . . . . . . . . . . . . . . . . . . . D. ACTH 4-10 and Blood Flow E. Brain Uptake of ACTH 4-10 4-10,. . . ... . .. . .............................. F. Neurochemical Response to A G. ACTH 4-10 and Morph ............... 111. P-MSH . ... ... . . . . ... . ... IV. P-Lipotroprin 61-91 . . . . . . A. Morphinelike Activity . . B. Grooming Activity . . . . . C. Avoidance Behavior . . . . . . . . D. Electrophysiology . . . . . . E. Cyclic Nucleotides . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . F. Biological Significance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References .............................................................

209 21 1 21 1 217 218 220 22 1 222 229 230 232 233 234 235 236 237 238 239 242

1. Introduction

The lipotropic hormone (6-lipotropin, P-LPH) was first isolated in 1964 from sheep pituitary glands (Li, 1964) and subsequently from the pituitary of a variety of species (see Li and Chung, 1976a). It has been detected in the circulating blood of sheep (Lohmar and Li, 1968) and located in discrete cells of the anterior and intermediate lobes of the pituitary (Moon et al., 1973). Analysis of the amino acid sequence (Fig. 1) of the hormone revealed that this 9 1-residue polypeptide displays only minor species variances (Li and Chung, 1976a). Because the amino acid sequence of a part of the @-LPH molecule (residues 4 1-58) resembles that of the hormone P-melanocytestimulating hormone (P-MSH), Li and his colleagues (1965) sug209

210

W. H. GISPEN, J. M . VAN REE, AND D. DE WIED

P-LI POTROP I N (PORCINE) H-Glu Leu T h r Gly Glu Arg Leu Glu Gln Ala Arg Gly Pro Glu Ala Gln Ala Glu Ser A la Ala Ala Glu Ala Glu Ala Val Leu Gly Tyr Glu Leu Glu Ala Arg Ala Ala Glu LYS LYS P-MSH

(z) (a)

7-LPH (1-58)

Asp Ser Gly Pro Tyr Lys Met Glu His Phe Arg T r p Gly Ser Pro Pro Lys Asp ACTH

Thr (z)

4-10

LYS Met-enkephalin Arg

Val Leu Pro T h r Gln Ser Lys Glu Ser T h r Met Phe Gly Gly Tyr ~~

~

Leu a-endorphin Phe Lys Asn Ala Ile Val Lys Asn Ala His Lys Lys Gly Glu-OH (gj

P-endorphin C-fragmen t

FIG. 1. Amino acid sequence of the lipotropic hormone.

gested that P-LPH was a possible prohormone of P-MSH. Other peptides could also be formed during the activation of P-LPH and stored in the pituitary. Indeed, it was recently shown that in addition to P-LPH, the peptides P-LPH 1-58 (y-LPH), P-LPH 1-38 (N-fragment), P-LPH 41-58 (P-MSH, P-LPH 61-87 (C'-fragment), and P-LPH 61-91 (C-fragment) occur in considerable amounts in bovine pituitary glands as intact polypeptides (Bradbury et al., 1976a, 1976~).Each of these peptides seems to be formed by cleavage of the lipotropin chain at the carboxyl side of paired basic residues (Lys-Lys and Lys-Arg) and subsequent removal of the basic amino acids. Thus, the specificity of the pituitary enzymes involved resembles that of other enzymes which generate hormones from their prohormones (e.g., insulin from proinsulin, Kemmleret d.,1972). Studies dealing with the specificity of the pituitary enzymes which activate lipotropin reveal that the first cleavage of P-LPH is at the arginyl-tyrosine bond at positions 60-61, and is apparently caused by a trypsinlike enzyme (Bradbury et al., 1976a). Another pituitary enzyme with a different specificity from trypsin seems to be involved in the cleavage of the bond at positions 39-40 (Lys-Lys) and may therefore be implicated in the in viuo release of P-MSH from lipotropin. At the moment, it is not clear whether specific activation of P-LPH leading to the release of LPH 61-91 and P-MSH takes place in one or more separate pituitary compartments. T h e data nevertheless indicate that P-LPH may serve as a prohormone for other pituitary peptides. Surprisingly, the Dhysiological significance of B-LPH has remained rather " " "

I

L

I

LIPOTROPIN A N D T H E CENTRAL NERVOUS SYSTEM

21 1

obscure. I t possesses lipolytic activity (Yamashiro and Li, 1974) and very recently, morphinelike activity was observed after intracerebroventriciilar administrations (Ronai et al., 1976). In contrast, no activity cortld be found when the hormone was tested in in 1 h - o preparations reliably affected by morphine (Bradbury et al., 1976b; Cox et al., 1976a; Grafet al., 1976b; Lazarus et al., 1976). This is in accordance with the hypothetical role of P-LPH as a prohormone for other biologically active peptides. The present chapter deals with the interaction of the lipotropin fragments and related peptides with the central nervous system. For historical reasons P-LPH 41-58 is designated as P-MSH and P-LPH 47-53 as ACTH 4-10. It. ACTH 4-10 A. ACTH 4-10

AND

LEARNED BEHAVIOR

Hormones of the pituitary-adrenal system play an essential role in homeostatic functions (see de Wied and Weijnen, 1970; de Wied et al., 1972: Gispen et al., 1975~).Their regulatory role in central nervoits functioning and behavior was discovered in rats whose pituitary-adrenal axis was disrupted or functionally suppressed. Over the years, studies have shown that adrenocortotropic hormone (ACTH) and steroids can influence behavior independently. In this review we deal exclusively with the influence of ACTH on the behavior which is brought about by direct peptide-brain interaction and in the majority of instances can be elicited by ACTH 4-10 (de Wied, 1974). Since a-MSH is identical to [Ac-Ser'] ACTH 1-13-NH2, it is not surprising that a-MSH has behavioral effects similar to those of ACTH. For an extensive review of the behavioral effects of a-MSH in mammals, the reader is referred to the paper by Kastin, Sandman, Miller and their co-workers (Kastin et al., 1975). Since a number of studies revealed that either peptides derived from the pituitary (ACTH, MSH, vasopressin, oxytocin, prolactin) or fragments of these peptides may influence behavior extraendocrinally and probably have a direct effect on the brain, such peptides were designated as neuropeptides and the significance of the pituitary for production of these neuropeptides has been discussed previously (de Wied, 1969, 1974; de Wied et al., 1974a; de Wied and Gispen, 1977). 1. Hypophysectomized Rats Hypophysectomized (hypox) rats were used in a number of the studies; such rats are depleted of neuropeptides of pituitary origin and are severely deficient in acquiring a conditioned avoidance response (Applezweig and Baudry, 1955; Applezweig and Moeller, 1959; d e

212

W. H .

GISPEN, J.

M. VAN

REE,

AND D. DE WIED

Wied, 1969). Hypophysectom y interferes with passive (Anderson et al., 1968; Weiss et al., 1970; Lisshk and Bohus, 1972), one-way (Gispen, 1970; de Wied and Gispen, 1977), and two-way active avoidance behavior (see de Wied et ad., 1972). Although general hormone replacement therapy without ACTH (de Wied, 1964, 1971) or an improved food intake (Harris, 1973) can restore the deficient performance to some extent, treatment with ACTH 4-10 alone is sufficient to normalize the avoidance behavior of hypox rats (de Wied, 1974; Bohus and de Wied, 1977). The expression of the behavioral activity of these peptides seems to depend on their presence during the behavioral task to be performed. Termination of treatment with ACTH 4-10 leads to a rapid deterioration of the performance of hypox rats (Bohus et al., 1973). 2. Intact Rats The effect of ACTH on acquisition of aversively motivated behavior is apparently difficult to demonstrate in intact rats presumably due to its dependence on the strength of the aversive stimulus (Murphy and Miller, 1955; Levine and Jones, 1965; Beatty et al., 1970; Ley and Corson, 1971; Kelsey, 1975). Most of our knowledge of the regulatory role of ACTH in aversively motivated behavior was therefore gained in studies concerning the extinction or retention of such behavior. ACTH and ACTH 4-10 delay extinction of avoidance behavior in a two-way shuttle box (de Wied, 1966), in a one-way platformjumping test (Bohusetal., 1968) and in a one-way pole-jumping apparatus (van Wimersma Greidanus, 1970). The effects can be elicited in a dose-dependent manner after both systemic or intracerebral (de Wied, personal communication) administration. I n studies which deal with ACTH fragments and passive-avoidance behavior, a more complicated picture emerges. ACTH 4-1 0 facilitates retention of passive-avoidance behavior in rats when administered 1 hour prior to a 24-hour retention trial in a step-through, one-trial, passive-avoidance procedure (Ader et al., 1972). However, the improvement of passive-avoidance behavior by ACTH depends on the dose and shock intensity used (Lissak and Bohus, 1972; McGaugh et al., 1975; Gold and van Buskirk, 1976). In a study on the effect of ACTH treatment during both training and testing in a one-trial “step-out” passiveavoidance situation, it was argued that the peptide would in fact enhance state-dependent learning (Gray, 1975). In another aversively motivated behavior, it was shown that ACTH 4-10 delayed extinction of the conditioned taste aversion response induced by pairing sugar water and the unpleasant experience caused by LiCl injection (Rigter and Popping, 1976).

LIPOTROPIN A N D T H E CENTRAL NERVOUS SYSTEM

213

The observations on aversively motivated behavior of rats indicate that ACTH 4-10 can affect both acquisition and extinction processes. The latter seem to be more sensitive to peptide treatment since according to Bohus and de Wied (1977) extinction behavior, thus in the absence of punishment, is more labile than acquisition behavior and thus more sensitive to modulatory influences. Furthermore, it was argued that ACTH 4-10 presumably maintains fear-motivated responses by preserving the motivating value of environmental stimuli. In line with this hypothesis, it was shown that ACTH 4-10 attenuates carbon dioxide-induced amnesia for a passive-avoidance response when administered prior to the retention test (Rigteret al., 1974). As the antiamnesic effect of ACTH 4-10 is independent of the nature of the amnesia-inducing procedure used (Rigter et al., 1975a), it was concluded that ACTH 4-10 facilitates retrieval of stored information. In addition to the known effect of ACTH on aversively motivated behavior, effects on positively reinforced behavior have been seen in many studies. Cuth et al. ( 1 97 1) reported that ACTH increases the rate of bar pressing for water. The eEectiveness of the treatment depended on motivation factors. Leonard (1969) found that ACTH partially antagonizes the deleterious effect of sodium barbitone on running time for a food reward in a multiple T-maze. Gray studied the influence of ACTH on the acquisition and extinction of a partially reinforced runway response with food as reward. Partial reinforcement is postulated to induce a behavioral state akin to frustration nonreward (Gray, 1967). Interestingly, partially reinforced rats receiving ACTH (Gray et nl., 1971) or ACTH 4-10 (Garrud, 1975) behaved like continuously reinforced rats both during acquisition and extinction. In addition, in an experiment on bar-pressing for food reward it was found that the differential effects of high and low reward on bar press rates were both attenuated by ACTH 4-10 (Garrud, 1975). Recently, Isaacsonet al. (1976) suggested that in a response rewarded with water, ACTH 4- 10 improves correct performance through better use of environmental cues by peptide-treated rats without a general effect on learning itself. With respect to the extinction of rewarded behavior, it was clearly demonstrated that, in hungry rats, ACTH (Gray, 1971) and ACTH 4-10 (Garrud et al,, 1974) delayed extinction of a straight runway response for food. ACTH 4-10 delayed extinction of a sexually motivated approach response of male rats in a straight runway (Bohus et al., 1975). Copulation reward during extinction appeared to be essential for the expression of the behavioral activity of ACTH 4-10. In addition, it was reported that ACTH 4-10 increased the motivation of female rats to seek contact with a sexually active male (Meyerson and Bohus, 1976). In an attempt to

2 14

W. H . GISPEN, J. M . VAN REE, AND D. DE WIED

pool the known effects of ACTH 4-10 on animal behavior within one physiological mechanism, Bohus and de Wied (1977) favor a mechanism involving motivation but indicate that more data are needed to assess the validity of such a hypothesis. 3. ACTH 4 -1 0 in Humans a . Volunteers. Repeated auditory stimulation with clicks produced arousal followed by habituation, as was shown by a shift from electroencephalogram (EEG) desynchronization to hypersynchronization (Endrijczi et al., 1970). Treatment with ACTH 1-10 (1-2 mg i.v.) restored the initial arousal-induced EEG desynchronization. These data suggest a “disinhibitory” action of ACTH on stimulus-induced EEG synchrony (Endroczi et al., 1970). Both EEG and reaction time were measured in subjects participating in a disjunctive reaction-time task (Miller el al., 1974). Treatment with ACTH 4-10 shifted EEG activity which is assumed to reflect a higher vigilance level. Furthermore, rating scales revealed that ACTH 4-10 treated subjects felt less tense and performed better in the Benton Visual Retention Test, suggestive of increased visual short-term memory, Together, the data imply that ACTH 4-10 raised the level of attention (Miller et al., 1974). Further studies by the same authors support this notion. In a study on emotionality, immediate memory, concept learning, and field dependence, the authors concluded that ACTH 4-10 treatment resulted in a faster intradimensional shift and a slower extradimensional shift in a concept learning task as a consequence of increased dimensional attention (Sandman et al., 1975). Also, in a continuous performance task, in which subjects had to detect the letter X in a series of letters appearing on an oscilloscope screen, ACTH 4-10 improved the performance significantly. Subjects so treated made fewer errors of omission and commission and had improved attention (Miller et al., 1976). There are similar observations by Gaillard who reported that ACTH 4-10 reduced the number of lapses of attention in a self-paced serial reaction-time task. In addition, peptide-treated subjects made significantly less errors than did placebo-treated volunteers (Gaillard and Sanders, 1975a,b). b. Patients. Retarded adult males were recruited from a workshop for the trainable retarded (Sandman et al., 1976).They were subjected to the same concept learning task as used with healthy volunteers (Sandman et al., 1975). In contrast to the study in volunteers, retarded patients treated with ACTH 4-10 displayed improved performance on both intra- and extradimensional shift. Furthermore, their performance in the Benton Visual Retention Test was also improved. Sandman et al. interpreted these results also as an indication of improved attention. A

LIPOTROPIN A N D T H E CENTRAL NERVOUS SYSTEM

2 15

recent open-safety study in elderly patients failed to show drug-related effects on electrocardiogram (EKG), blood pressure, or EEG parameters after ACTH 4-10 (Ferris et al., 1976). However, the authors reported a significant change in mood, considered to be indicative of a mild antidepressant activity of ACTH 4-10. Thus, ACTH 4-10 is the first neuropeptide to have been studied in human behavior using double-blind experimental designs. Although far from complete, the available data show increased visual attention and/or motivation in healthy volunteers, a finding which may be of significance to elderly or retarded people. 4. Strzr cture -A ctivity The behavioral effects of peptides structurally related to ACTH and MSH were first ascribed to the presence of the sequence ACTH 4-10 (Ferrari et nl., 1963: de Wied, 1966). There is an extensive series of‘ studies by Greven and de Wied of the structural requirements of ACTH 4-10 necessary to delay extinction of avoidance behavior (Greven and de Wied, 1973, 1977: de Wied et al., 1975). I t was found that ACTH 4-7 was the smallest sequence to have essentially the same potency as ACTH 4-10 (Greven and de Wied, 1973). If, however, other fragments, e.g. ACTH 7-10 or ACTH 11-24 or the derivative [AC”] ACTH 11-13NH2, were given at 10 times higher dose levels, it appeared that these fragments also induced the behavioral response (Greven and de Wied, 1977). Elongation of ACTH 7-10 to ACTH 7-16 resulted in a peptide as active as ACTH 4-10. It was concluded that there is a redundancy of information within the ACTH molecule with respect to its behavioral activity, supporting the report of two active sites for MSH activity by Eberle and Schwyzer ( 1975). Information located distal from the C-terminal of ACTH 4-7 may be present in a dormant form and may need to be potentiated by chain elongation in order to be expressed (Greven and de Wied, 1977). Although other ACTH-central nervous system (CNS) structure-activity relationships are not always identical to the one found for ACTH avoidance behavior (excessive grooming, Cispen et al., 1975a, 1976a; Wiegant and Gispen, 1977; opiate receptor binding, Tereniiis et nl., 1975), the general principles of dormant activity and induction of such activity by chain elongation still seem to apply. Thus, if’ information is indeed encoded in a multiple form, comparison between peptides on the basis of primary structures alone is hazardous. An important breakthrough in our experimental approach toward the nature of the brain cell-peptide interaction may be the suggestion that at the brain receptor site, ACTH 4-10 assumes an a-helix conformation with the Met‘ and the Arg” in close proximity (Greven and de Wied, 1977). Seemingly contradictory structure-activity results based on pri-

216

W. H . GISPEN, J. M. VAN

REE,

AND D. DE WIED

mary structures could then be explained by more definite knowledge of the stereoconformation at the receptor site. When the Phe‘ residue in ACTH 4-10 was replaced by its D-enantiomer, the peptide facilitated rather than delayed extinction of an active avoidance response (Bohus and de Wied, 1966; d e Wied et al., 1975). The reversal of the behavioral effect was only found for analogs with the Phe’ residue in the D-configuration (de Wied et al., 1975). Since Phe’ seemed to be crucial for the behavioral effect, replacement of this residue with other amino acids was undertaken in subsequent experiments. It was concluded that the electron donor properties of the amino acid residue in position 7 correlate to some extent with the behavioral potency of the peptide (de Wied et al., 1975). It has been suggested that [~-Phe’]ACTH 4-10 contains a new intrinsic activity with regard to extinction behavior and may thus act at a level different from that of ACTH 4-10 (see Bohus and de Wied, 1977; de Wied and Gispen, 1977). Using knowledge gained from these detailed studies on structure and behavioral activity. Greven and de Wied (1977) were able to synthesize analogs of ACTH 4-10 which are 10:’ and even lo6 times more potent D-I,YS~, Phe“] ACTH than the parent molecule. T h e peptide [Met4 (02), 4-9 has a 1000-fold potentiated behavioral activity, whereas the MSH activity of the same molecule was reduced by about the same factor, while the steroidogenic activity was significantly reduced. As the in vitro half-life of the various substituted ACTH 4-9 analogs in plasma or brain extracts correlated well with their behavioral potency (Witter et al., 1975), a partial explanation for the effectiveness of the various substitutions may be that they provide protection against enzymatic breakdown.

5 . Site of Action; Role of the CSF The search for brain structures sensitive to ACTH fragments has been guided by the notion that the limbic system is important for acquisition and extinction of avoidance behavior. Bilateral destruction of the nucleus parafascicularis facilitated extinction of shuttle box avoidance behavior. Rats with lesions in this area did not respond to a-MSH with a delay of extinction of a pole-jump avoidance response (van Wimersma Greidanus et al., 1974; van Wimersma Greidanus and de Wied, 1976). Further lesion studies also implicated the anterodorsal hippocampus as a site of action of ACTH 4-10 for the delay of extinction of avoidance behavior (van Wimersma Greidanus and de W e d , 1976). The results thus obtained were expanded by experiments using microinjection of ACTH 4-10 or ACTH 1-10 in various brain areas and subsequent monitoring of their behavioral effectiveness. These peptides caused inhibition of extinction of the pole-jumping avoidance response when they

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM

2 17

were applied locally to the mesencephalic-diencephalic region at the level of the posterior thalamus and in the ventricles (van Wimersma Greidanus and de Wied, 1971). In view of the many negative results obtained in other brain regions, the data were taken as an indication of the importance of the parafascicular area. I t is quite possible that, in brain, the neural substrate of these neuropeptides is restricted to a functional rather than to an anatomical unit. It may thus be concluded that the limbic system needs to be intact to permit neuropeptides related to ACTH to exert their behavioral effects (van Wimersma Greidanus and de Wied, 1976). An interesting finding in the study by van Wimersma Greidanus and de Wied (197 1) was that whenever the ACTH 1-10 was applied into the ventricular system, there was a delay of extinction. Recently, it has become more and more evident that the brain ventricular system is an avenue for neuropeptides to reach their site of action (Johnson and Epstein, 1975; de Wied and Gispen, 1977). For instance, intraventricular administration of antibodies against the neurohypophyseal peptide hormone, vasopressin, resulted in a memory deficit most likely due to elimination of the physiologically circulating vasopressin which facilitates memory formation (van Wimersma Greidanus et al., 1975a,b). A variety of behavioral and neurochemical effects have been reported after intraventricular administration of ACTH-like peptides, thus indirectly supporting the notion that the cerebrospinal fluid (CSF) may also serve as a transport system for ACTH-like neuropeptides. There is some indication that ACTH-like immunoassayable material circulates in the CSF (Allen et al., 1974). The actual mode of transport of pituitary hormones to the brain is not known. Various possibilities have been suggested, such as retrograde transport along the pituitary stalk or the existence of basilar cisterns which may be connected with the CSF in the adenohypophysis close to the hormone-producing cells (Allen et al., 1974) or transport via the bloodstream. The latter hypothesis is supported by the recent finding of Ambach and Palkovits (1975) that the nucleus periventricularis receives blood from the anterior hypophyseal artery. There is an urgent need for studies aimed at the elucidation of a pituitary-CSF connection, since such a direct link could explain the role of the pituitary as an important source of behaviorally active neuropeptides. B. ACTH 4-10

A N D THE

INDUCTIONOF EXCESSIVE GROOMING

When peptides derived from ACTH, MSH, or LPH are administered intraventricularly, a peculiar phenomenon is observed: a stretching and yawning syndrome (SYS) is induced in a variety of animals (Ferrari et al.,

218

W. H. GISPEN, J. M. VAN REE, AND D. DE WIED

1963; Gessa et al., 1967). Given intrathecally in man, ACTH produces the same type of stretching and vomiting. The latter symptom is thought to be due to impurities in the preparation (Floris, 1963). In rodents, the onset of the syndrome is preceded by a display of excessive grooming (Ferrari et al., 1963; lzurni et al., 1973; Gispen et al., 1975a; Rees et al., 1976). The induction of grooming is independent of the presence of the adrenals, pituitary, or gonads (Gispen et al., 1975a). Evidence is accumulating, however, that the induction of excessive grooming and SYS are mediated by two different CNS mechanisms (Gispen et al., 1975a, 1976~). The biological significance of the behavior elicited after intraventricular administration of ACTH is unclear. Admittedly, the response is not seen exclusively after intraventricular administration of LPH-like peptides (MacLean, 1957; Beagly, 1976; Izumi et al., 1973). Some authors suggest that in addition to SYS, ACTH also induces sexual excitement in rodents (Bertolini et al., 1968, 1969; Baldwin et al., 1974). In pigeons, the behavioral response was compared with displacement behavior as it occurs naturally in these birds (Delius et al., 1976). Grooming is one of the behaviors in rodents often interpreted as representing displacement activities (Fentress, 1968; Hinde, 1970). It is not clear, however, whether the grooming response as discussed here and which lasts for at least 1 hour after the injection of ACTH and is in most instances interrupted only by stretching and yawning is in fact related to the grooming seen in a transitional behavioral state (Fentress, 1968). ACTH 1-24, a-MSH, and P-MSH are equipotent in inducing excessive grooming (Gispenet al., 1975a). In rabbits, ACTH 4-10 was not nearly as active as ACTH 1-24 (Baldwin et al., 1974), whereas in rats (Gispen et al., 1975a) and mice (Rees et al., 1976), ACTH 4-10 was totally inactive even in high doses. If, however, [~-Phe']ACTH 4-10 was used, excessive grooming was elicited even with low doses (Gispen et al., 1975a; Rees et al., 1976). This dormant activity of ACTH 4-10 could be potentiated by C-terminal elongation of the fragment (Gispen et al., 1975a) or by shortening of the fragment to ACTH 4-7 which is as active as [~-Phe'] ACTH 4-10 (Wiegant and Gispen, 1977). Thus, in essence the shortest peptide fragment with full activity on extinction of avoidance behavior (de Wied et al., 1975) is also the shortest which induces excessive grooming (Wiegant and Gispen, 1977). C. ELECTROPHYSIOLOGICAL CORRELATES OF ACTH 4-10 1. Central Nervous System Injections of ACTH increased the electrical activity of the rat brain as was concluded from an increase of voltage, occasional spiking, and

LIPOTROPIN A N D T H E CENTRAL NERVOUS SYSTEM

219

paroxysmal runs of' low frequency, high voltage waves. The effect could be demonstrated in intact, hypox, and adrenalectomized rats: it was therefore assumed to reflect a direct effect on the brain (Torda and Wolf'. 1952). Kawakami et al. (1966) suggested a correlation between circadian peaks in corticosterone and ACTH secretion and changes in multiple unit activity (MUA) in the hypothalamus. Most of the effects of ACTH on brain electrical activity are opposite to those evoked by the administration of adrenal steroids (Feldman et al., 196 1: Pfaff et al., 197 1). The mechanism by which ACTH induces these neuroph ysiological effects is independent of the presence of the adrenal cortex. Koranyi et al. (1 97 1 ) showed that systemic injection of ACTH resulted in a decrease of spontaneous electrical activity and responsiveness especially of the medial preoptic area of the free-moving cat. Steiner reported that ACTH directly activated neurons in hypothalamic and mesencephalic areas of rat brain (Steiner et al., 1969: Steiner, 1970). In addition, ACTH increased the spike frequency of diencephalic cells (van Delft and Kitay, 1972) and the excitability of cells in the spinal cord (Nicolov, 1967). I n addition to adrenal-mediated effects of ACTH in rat diencephalic MUA, some short latency excitatory influences were seen which could also be elicited in adrenalectomized rats (Sawyer et al., 1968). Intraventricular infusion of ACTH 1-24 or ACTH 4-1 0 in rabbits led to increased electrical activity only in the lateral preoptic diagonal band of Broca and the periventricular preoptic area. Other areas showed little or no change in MUA. The effect was related to the behavioral and neuroendocrine activities seen after intraventricularly administered ACTH (Baldwin et nl., 1974). Further support for a direct effect of ACTH-like peptides on brain electrical activity comes from a series of experiments on the effect of n-MSH (= Ac-Ser' ACTH 1-13-NHr) on EEG parameters in rabbits (DysterAas and Krakau, 1965), rats (Sandman et al., 197 I), frogs (Denman el al., 1972), and man (Kastin et al., 1975). In view of the limitations of the present chapter the reader is referred to Kastin et al. (1975) and Miller et nl. ( 1 977) for further information. In a free-moving dog, hippocampal electrical activity was sensitive to ACTH 4-10 and [~-Phe']ACTH 4-10. Although both peptides shifted the activity in the theta range to lower frequencies in an operant conditioning situation, the D-analog was less effective (Urban et al., 1974). In rats, when hippocampal theta activity was induced by electrical stimulation of the reticular formation, the administration of ACTH 4-10 produced a shift in peak frequency from 7.0 to 7.5 HL (Urban and de Wied, 1976). In another study, effects of ACTH 4-10 and [~-Phe']ACTH 4-10 on averaged visually evoked responses in cortical area 17 were recorded (Wolthuis and de Wied, 1976). When measured at a wide vari-

220

W. H.

GISPEN, J.

M . VAN REE, AND D. DE WIED

ety of light intensities, the amplitudes of the late components of the visually evoked potentials were significantly diminished after ACTH 4-10. In this situation the effect of [~-Phe']was in the same direction but again weaker (Wolthuis and de Wied, 1976). In humans, ACTH 4-10 treatment resulted in a statistically significant increase in the power output of the 12-k Hz and the 7-12 Hz frequency bands of occipital EEG, and such subjects did not show habituation of the EEG response to pattern arousal (Miller et al., 1974). This observation is consistent with the findings of Endroczi et al. (1970) on the effects of ACTH 1-10 and ACTH 1-24 on human EEG. In summary, there is ample evidence that ACTH 4-10 directly affects brain electrical activity. Since its effects can in some instances be mimicked by increasing the stimulus strength (Urban and de Wied, 1976), it may be that, in general, ACTH 4-10 facilitates transmission in its neural substrate, i.e., limbic structures (Urban and de Wied, 1975) or visual cortex (Miller et al., 1974; Wolthuis and de Wied, 1976). ACTH 4-10 might increase the state of arousal in such structures, which may determine the motivational influence of environmental stimulus and thereby the probability of the stimulus-specific behavioral response being generated. 2. Peripheral Nenious System Strand et al. studied the sciatic nerve-gastrocnemicus muscle preparation and observed that ACTH increased muscle action potential amplitude and contraction height and delays fatigue in normal, adrenalectomized, and hypox rats (Strand et d., 1973-1974). The extraadrenal nature of the peptide-nerve-muscle interaction was corroborated by the finding that a-MSH and ACTH 4-10 also augmented the action potential amplitude. [~-Phe'] ACTH 4-10 had no effect (Strand and Cayer, 1975). The effect of the peptides is best seen in fatigued and particularly in the hypox, fatigued rat. These observations led Strand et al. to perform clinical trials on a patient with myasthenia gravis; a short beneficial effect of ACTH 4-10 treatment was indeed observed (Strand ct al., 1975). The data support the hypothesis that such peptides may act as modulators of nerve function (Krivoy, 1970).

D. ACTH 4-10

AND

BLOODFLOW

I t has been demonstrated in rabbits and cats that ACTH 1-39 and [Gly] ACTH 1-18-NH2 depressed blood pressure (KorLnyi and Endroczi, 1967; Nakamura et d., 1976). The same response could be shown in both intact and adrenalectomized rats, suggesting an extraadrenal action of ACTH on blood pressure. Since both a-MSH and human

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM

22 1

P-MSH were inactive, it was further concluded that the depressor effect may also be independent of the melanocyte-stimulating properties of the molecule (Ueda et al., 1970). ACTH 4-10 affects the heart rate changes which accompany emotional behavior (Bohus, 1975). Its influence was seen only in situations were certain learning paradigms are involved as in classical conditioning or in passive avoidance behavior. According to Bohus (1975), in this case the primary effect of ACTH 4-10 was on the CNS mechanisms controlling cardiovascular functions. Goldman et al. (1975, 1977) have reported that a-MSH affected regional blood flow of the rat brain. They point out that a variety of drugs which have specific effects on animal behavior act only to redistribute the flow of blood to regions of the brain. There is reason to believe that these alterations in flow reflect the involvement of specific regions of the brain in the effect of these drugs (Goldmanet al., 1975). The flow in most areas was reduced within 10 minutes after intravenous administration of a-MSH and only the occipital cortex was spared. The effect was transient for most areas, but perfusion of pons and medulla, cerebellum, hippocampus, and parietal cortex was still low after 20 minutes. According to these authors, this rapid response of brain blood flow may be related to the effects of ACTH/MSH-like peptides on mental performance (Goldman et al., 1977).

E. BRAINUPTAKEOF ACTH 4-10 A considerable effort has been made to demonstrate that exogenously administered ACTH-like peptides indeed reach brain structures so as to enable them to exert their regulatory role on behavior. Injection of ['231] a-MSH resulted in a rapid accumulation of the radioactivity in structures such as the pineal gland which are exempted from the blood-brain barrier (Dupont et al., 1975). Much lower concentrations were found within the brain. Only after intraventricular injection did some localization appear in a thalamic nucleus (Pelletier et al., 1975). In view of the loss of biological activity after iodination of a-MSH, these studies could be only of a preliminary nature and were therefore followed u p by a study of the distribution of [jH] a-MSH in rat brain (Kastin et al., 1976). In the latter study some accumulation of radioactivity was found in the occipital cortex, cerebellum, and the pons medulla area as compared with other brain parts. Although identification of the labeled material was only tentative, the data support the notion that systemically administered peptides may pass the blood-brain barrier. Another group of investigators used the analog 'H[Met-02-D-LysXPhe"] ACTH 4-9 whose 1000-fold higher behavioral potency is most likely related in part to its increased metabolic stability (Witter et al.,

222

W. H . GISPEN, J. M. VAN REE, AND D. DE W E D

1975). After intravenous administration, only about 0.5 X of the dose was recovered in the brain as the intact peptide. The recovery was even lower after subcutaneous and oral administration (Verhoef and Witter, 1976). A detailed study of the distribution of this behaviorally potent ACTH 4-10 analog after intraventricular administration revealed that hippocampal and thalamic nuclei had a low or medium uptake of radioactivity (Verhoef et al., 1977a). These areas had previously been found essential for the expression of the behavioral activity of ACTH-like peptides (van Wimersma Greidanus and de Wied, 1971; van Wimersma Greidanusetal., 1975~). However, very high accumulation of radioactivity occurred only in septal nuclei (Verhoef et al., 1977a). It‘ was shown in further experiments that hypophysectomy specifically increased this uptake in the septum, emphasizing the physiological significance of this uptake (Verhoef et al., 1977b). The specificity of the septal uptake system was clearly demonstrated using hypox rats treated with ACTH 1-24, ACTH 4-10, or ACTH 11-24. It was found that only the peptides sharing the 4-9 sequence with the radioactive 4-9 analog (ACTH 1-24 and ACTH 4-10) competed with the septal uptake, whereas the structurally unrelated sequence, 11-24, was ineffective. The specificity of the uptake was further evidenced by the negative data obtained after pretreatment with other neuropeptides including P-LPH 61-76. Interestingly, the septal uptake of the 4-9 analog was not affected by [~-Phe’]ACTH 4-10 despite its structural resemblance (Verhoef et al., 1977b). It is conceivable, as was pointed out earlier, that ACTH 4-10 and [ ~ - P h e ACTH ~] 4-10 affect behavior at different brain levels. This may account for the ineffectiveness of the [~-Phe’]ACTH 4-10 analog but also emphasizes the functional significance of the septal uptake of ACTH 4-9. Whether or not this uptake in fact resembles peptide septum-receptor interaction is the object of further research (Verhoef et al., 1977b). F. NEUROCHEMICAL RESPONSE TO ACTH 4-1 0

1. Binding to Cell Membranes: Peptide-Cell Communication If the interaction of ACTH 4-10 with brain cells resembles that of peptides with their peripheral target cells, the generally accepted hypothesis is that ACTH 4-10 exerts its regulatory influence as first messenger in the manner described for the classical “second messenger” of Sutherland (1972). Binding to a specific receptor at the outside of the plasma membrane triggers adenyl cyclase (and/or guanyl cyclase) resulting in an alteration of the intracellular level of free cyclic nucleotides. The activity of specific cyclic nucleotide-dependent protein kinases is

LIPOTROPIN A N D THE CENTRAL NERVOUS SYSTEM

223

then altered and a regulatory influence can be measured at varioils levels in the cell. No specific binding sites for ACTH 4-10 have been demonstrated, but there is abundant indirect evidence to support the existence of such sites (see below). The paucity of receptors or the limited regions of ACTH-sensitive cells (Motta et a[., 1965; Steiner, 1970; van Delft and Kitay, 1972) may be major obstacles to such studies. Alternatively, it is possible that the conventional receptor methodology as applied in the study of specific morphine- (Terenius, 1973) or specific thyrotropinreleasing hormone (TRH)- (Burt and Snyder, 1975) binding to brain membranes is hampered by low affinity of‘ the peptide for its brain receptor . 2. Cyclic Nucleotides and Protein Ptwsfihmylntion I t has been reported by several investigators that addition of ACTH to rat brain broken cell preparations in uitro does not affect adenyl cyclase activity (Burkhard and Gey, 1968; von Hungen and Roberts, 1973). Furthermore, it was found that ACTH, in a concentration of 10-‘M did not alter the levels of cyclic adenosine 3’-5’-monophosphate (CAMP)in slices of cerebral cortex, cerebellum, and hypothalamus of’ rabbit, rat, cat, and monkey (Forn and Krishna, 1971). In contrast to these findings. ACTH 1-10 in a dose of lo-” M increased the levels of’ cAMP in slices from rat posterior thalamus by about 40% (Wiegant and Gispen, 1975). Recent data suggest that, in rat striatal slices, ACTH 1-24 increases cAMP and decreases cyclic guanosine 3’-5’-monophosphate (cGMP) levels concomitantly (Wiegant, unpublished observations). Intrathecal administration of high doses of ACTH to rabbits elevated the levels of cAMP in their CSF (Rudman and Isaacs, 1975). A preliminary report on the effect of a-MSH on rat brain cAMP levels in 7 1 i w states that a-MSH increased the content in occipital cortex of both intact and hypophysectomized rats. Changes in other brain areas were also noted but these did not occur in a consistent manner (Christensen et nl., 1976). Another study implicates a modulatory role ofthe pituitary-adrenal axis in rat cerebral metabolism of CAMP. Although in z i 7 m administration of ACTH to hypophysectomized rats restored the decreased activity of adenyl cyclase and subsequent cAMP formation, the authors suggested a steroid-CNS rather than a peptide-CNS interaction (Nakagawa and Kuriyama, 1976). Thus at present no data are available which unequivocally establish an ACTH-CAMP mediated effect in brain. It should be kept in mind, however, that not all peptide-brain membrane interactions necessarily involve cyclic nucleotides. The effect of cyclic nucleotides and that of ACTH fragments on

224

W. H. GISPEN, J. M. VAN REE, AND D. DE WIED

endogenous phosphorylation of rat brain synaptosomal plasma membranes in vitro has been compared by Zwiers et al. (1976). I t was shown that even low doses of CAMPand ACTH 1-24 altered phosphorylation, whereas cGMP, ACTH 1-10, and ACTH 11-24 were without effect. However, CAMP-dependent phosphorylation was more or less restricted to an increase in three protein bands (MW 78,000-53,000), whereas ACTH 1-24 diminished the incorporation of radioactive phosphate into five bands of lower molecular weight (Zwiers et al., 1976). Although certainly not conclusive, the data seem to indicate that there may be CAMP-independent effects of ACTH on brain phosphoproteins themselves thought to be important to brain cell physiology (Greengard, 1976) and behavior (Routtenberg et al., 1975; Perumal et al., 1975).

3 . RNA Metabolism Pituitary-adrenal hormones may have a regulatory effect on brain macromolecule metabolism as can be concluded from the effectiveness of various stressors to alter the incorporation of labeled precursors in vivo or their content in macromolecules (Jakoubek et al., 1970; Rees et al., 1974; Dunn et al., 1976; Dunn and Rees, 1976; Schotmanet al., 1977a). The absence of pituitary also affects the metabolism of these macromolecules as can be concluded from the severe deficits in brain RNA and protein metabolism in hypox rats (see Gispen and Schotman, 1973). Yet little is known about the effect of exogenously administered ACTHlike peptides on brain RNA. Giving a single high dose of a purified ACTH preparation (5 I.U./lOO gm s.c.) resulted in a transient inhibition of ['HI uridine into mouse (Jakoubek et al., 1972) and rat (Gispen and Schotman, 1976) brain RNA, followed by a small decrease in total RNA. In adrenalectomized rats, there was a marked increase in uridine incorporation (+40%) after similar treatment with ACTH, suggesting a counteraction between ACTH and corticosteroid at the genome level (Gispen and Schotman, 1976). Such counteraction may explain the small effects reported for intact rats or may even totally obscure such effects (Dunn, 1976). The data are sufficient to allow the conclusion that ACTH 4-10 or ACTH 1-10 fails to influence either the labeling of messengerlike and ribosomal brain RNA (Gispen et al., 1970a; Schotman et al., 1972) or the aggregation of brainstem polysomes (Gispen et al., 1971). Furthermore, treatment with ACTH 1-10 did not affect brainstem RNA labeling or content of intact or of adrenalectomized rats (Gispen and Schotman, 1976). No effect of ACTH 4-10 on labeling and content of mouse cerebral RNA was found by Reading and Dewar (1971) or by Dunn (1976). The activity of brain RNAse was not affected by ACTH 4-10 either

LIPOTROPIN A N D T H E CENTRAL NERVOUS SYSTEM

225

(Reading and Dewar, 197 1). While the larger sequence ACTH 1-24 may influence brain protein synthesis at the transcriptional level, the short sequence ACTH 4-10, if it were to have such an effect, should exert its action at the translational level (see below). However, in view of the complexity of the methodology involved and of the regulation of mammalian protein synthesis, the significance of this differential effect of ACTH 1-24 and ACTH 4-10 is uncertain (see also Dunn and Gispen, 1977). 4. Protein Synthesis The effects so far reported for neuropeptides on brain protein metabolism are small and the methodology used is open to question. In most studies, the relevant data on precursor pool size, specific activity of the precursor, and of the labeled protein are not presented. The data are, however, treated here as though the incorporation studies indeed dealt with protein synthesis rate as was implied by most authors. a. Hypox Rats. Chronic treatment of hypox rats with ACTH 1-10 (Schotman et al., 1972) or ACTH 4-10 (Versteeg et al., 1972; Reith, 1975) enhanced the incorporation of ["HI leucine into brainstem proteins. At the end of the short incorporation period used, leucine still retained virtually all the acid-soluble radioactivity (Schotman et al., 1974), whereas the acid-insoluble radioactivity was confined to cytoplasmic proteins (Schotman et al., 1972). The same peptide treatment did not increase aggregation of polyribosomes in brainstem tissue (Gispen et al., 1971). If, however, such hypox rats were subjected to additional daily training in a two-way shuttle box avoidance apparatus, the amount of polyribosomes in their brainstem was increased (Gispen and Schotman, 1970; Gispen et al., 1971). With a 5-minute incorporation pulse and in view of recent calculations by Lajtha et al. on the turnover of rat brain proteins (Lajtha et al., 1976), one should expect to have labeling not only of the rapidly turning over proteins but of the slowly turning over proteins as well (Reith et al., 1977). Such a conclusion is consistent with the observation that a whole spectrum of soluble and membrane-bound proteins was labeled (Reith et d., 1974a, 1975b). ACTH 1-10 treatment enhanced the incorporation of radioactive leucine into all proteins. Minor increases in labeling of water-soluble, high-molecular-weight proteins were superimposed on this rather general effect (Reith et al., 1975b). It has been suggested that the enhancement of brain protein synthesis by ACTH-like peptides is specifically related to the behavioral activity of these peptides since, under similar conditions, [~-Phe']ACTH

226

W. H . GISPEN, J. M. VAN

REE,

AND D. DE WIED

1-10 decreased the labeling o f brainstem proteins. In addition, the sequence ACTH 11-24 had no effect on brainstem protein synthesis (Schotman et al., 1972; Reith, 1975). b. Intact Rodents. In intact rats, ACTH increased the incorporation of amino acids into protein in brain and spinal cord (Semiginovsky and Jakoubek, 1971). Chronic treatment of young rats with high doses of ACTH resulted in a complex pattern of neurochemical changes including a biphasic effect on brain protein content (Palo and Savolainen, 1974). A single high dose of ACTH stimulated the incorporation of various 14C-labeled amino acids into mouse brain by 20-100% 6-24 hours after injection of the precursor. The effect seems to be specific for brain since no change was seen in liver or kidney (Rudman et al., 1974). Using another strain of mice and a shorter incorporation period, others found that ACTH 1-24 increased the incorporation of [:3H]lysine into brain and liver proteins, whereas ACTH 4-10 was effective only on brain proteins (Dunn et al., 1976). There is further evidence for this differential effect of ACTH 1-24 and ACTH 4-10 from similar data on chronic treatment of rats with ACTH 4-10 (Reading and Dewar, 1971; Reading, 1972; Lloyd, 1974). Interestingly, such treatment with [~-Phe']ACTH 4-10 did not influence the incorporation of labeled precursor into brain proteins. Intraventricular administration of ACTH 1-24 or [o-Phe'] ACTH 4-10 to mice enhanced ["HI lysine incorporation into brain proteins (Reeset al., 1976). ACTH 4-10 was ineffective (Reeset nl., 1976). In view of the fact that after intraventricular administration of ACTH-like peptides only the grooming-inducing peptides affected brain proteins, it can be argued that the change in cerebral protein synthesis reported by Rees et al. (1976) is related to grooming and to SYS rather than to a peptidebrain protein interaction per se. There is thus evidence to indicate that in both hypox and intact rodents, ACTH 4-10 can enhance brain protein synthesis in viuo but these effects of the neuropeptides are rather slight. It appears, however, that in general, even under severe circumstances such as hypophysectomy (Schotman et al., 1972), undernourishment (Stern et al., 1976), or neurotoxic treatment (Schotman et al., 1977b), brain protein metabolism responds only with changes of the order of 20-3096. The changes of 10-25% in precursor incorporation found after treatment with ACTH 4-10 and related peptides are then as important as could have been expected. To study the possibility that an altered brain uptake of amino acids after peptide treatment (see effect on blood flow) might have contributed to the increased protein labeling, a-amino isobutyric acid was

LIPOTROPIN AND T H E CENTRAL NERVOUS SYSTEM

227

used as a metabolically inert amino acid analog (Rudman et nl., 1974; Schotmanet al., 1976). ACTH or ACTH 1-10 did not alter the penetration of this amino acid into brain tissue, suggesting that the observed increase indeed related to some brain cell mechanism. There is further evidence for such a direct effect on brain protein metabolism from studies on the effect of peptide addition on the in vitro incorporation of amino acid into brain slice protein (Reith et al., 1974b: Lloyd, 1974). When such slices were taken from hypox rats, ACTH 1-10 and ACTH 1-24 enhanced the incorporation of [“C] leiicine (Reith et ul., 1974b). Control experiments showed that incubation with the peptides did not alter amino acid uptake (Schotman et al., 1976) or the extracellular space of slices (Reith, 1975). I t is therefore unlikely that effects on these parameters underlie the increased protein labeling in zlitro. In contrast with the in uizio observations, it was found that [~-Phe’]ACTH 1-10 did not affect slice protein synthesis in zdro (Reith et al., 1975b). I t was found that, in slices from intact rat brain, ACTH 4-10 stimulated the incorporation of [ ‘ T I leucine into protein (Lloyd, 1974). Surprisingly, in 7 1 i . i ~ treatment with ACTH was reported to decrease amino acid incorporation into rat brain slices in vitro (Jakoubek et al., 1970). This discrepancy has not been explained, but has been confirmed by the recent unpublished observations of Schotman who showed that intraventricular administration of ACTH 1-24 resulted in a decreased incorporation of [“C] leucine into proteins in a cell-free brain system. Apparently, a procedure involving in ziivo injection followed by an in zdro test must be studied more extensively before meaningful conclusions can be drawn. In conclusion, there is ample evidence that ACTH 4-10 enhances cerebral protein synthesis and that the mechanism underlying this effect involves an extraadrenal, direct CNS effect of the peptide. 5. Neurotransmitters I t was reported more than 20 years ago that ACTH increased acetylcholine synthesis in the brains of hypox and intact rats (Torda and Wolf, 1952). In subsequent years, neurophysiological data were collected which suggested that ACTH and melanotropic peptides could modulate neurotransmission (Krivoy, 1970), but surprisingly little is yet known of the transmitters affected. Stressful stimuli which stimulate that release of ACTH and corticosteroids induced a general increase in brain noradrenaline (NA) turnover (see Versteeg and Wurtman, 1976). Treatment of intact rats with ACTH increased NA turnover in various brain regions (Hokfelt and Fiixe, 1972). Subsequent studies suggested that this effect results from an extraadrenal action of ACTH, since adrenalectomy (high-endogenous

228

W. H.

GISPEN, J.

M . VAN

REE,

AND D. DE WIED

ACTH) increases and hypophysectomy (absence of ACTH) decreases brain NA turnover (Versteeget al., 1972; Versteeg and Wurtman, 1976). I t was recently reported that administration of ACTH to rats 2 days after adrenalectomy resulted in a significant decrease of NA uptake in several brain regions. This suppressive action of ACTH on hippocampal and neocortical NA uptake lasted for at least 12 hours (Endroczi, 1975). In intact rats pretreated with a-MPT to block NA synthesis, ACTH 4-10 accelerated the decline of NA, indicating an increased turnover (Versteeg, 1973). It was also seen in this study that, under similar conditions, treatment with [ ~ - P h eACTH ~] 4-10 was ineffective. A similar approach but using ACTH 1-10 led to negative results in hypox rats (Versteeget al., 1972), and it was suggested that steroid hormones might play a permissive role. It was subsequently found that ACTH 4-10 increased the incorporation of [“HI tyrosine into total cerebral catecholamines, and that adrenalectomy blocked this peptide-induced response (Versteeg and Wurtman, 1975). ACTH 4-10 and [ ~ - P h e ACTH ~] 4-10 increased the turnover of NA in rat forebrain and hindbrain, while ACTH 4-10 increased dopamine (DA) turnover in midbrain (Leonard, 1974).Others showed that a-MSH (Ac-Ser’ ACTH 1-13-NHz) increased the turnover of NA, but not of DA in intact rats (Kostrzewa et al., 1975). Recent experiments in mice showed that ACTH 1-24, ACTH 4-10, or [ ~ - P h e ~ ] ACTH 4-10 all increased the conversion of [”H] tyrosine to DA, but not to NA (Dunn et al., 1976) and again, adrenalectomy blocked the effect of the ACTH-like peptides. The fluorescence intensity of DA neurons in the substantia nigra was increased after treatment of rats with a-MSH and ACTH 1-24 (Lichtensteiger and Lienhart, 1975). Interestingly, peripheral administration of DA antagonists (haloperidol, fluphenazine) or DA-receptor blockade in the neostriatum suppressed ACTH-induced excessive grooming (Wiegant et al., 1977a). Since ACTH 1-24 was ineffective when injected into the neostriatum but effective when applied in the substantia nigra, it was speculated that a nigrostriatal DA pathway was involved (Wiegant et al., 1977a). Few investigations have measured the effect of ACTH on the enzymes involved in catecholamine matabolism. In hypox rats, the activity of dopamine-P-hydroxylase (DBH) was decreased throughout the brain and the effect was only reversed by treatment with ACTH or ACTH 1-10 in hypothalamus and brainstem (van Loon and Mascardo, 1975). In intact rats, ACTH increased hypothalamic and decreased cortical DBH (van Loon and Mascardo, 1975). Other preliminary data suggest that ACTH and ACTH 1-10 administered in vivo can increase the in vitro activity of striatal tyrosine hydroxylase (Dunn et al., 1976). With respect to brain 5-hydroxytryptamine (5-HT) there is one report of an acceleration of 5-HT turnover in hypox rats after a-MSH (Spirtes et al.,

LIPOTROPIN A N D THE CENTRAL NERVOUS SYSTEM

229

1975). In intact rats, ACTH 4-10 and [~-Phe’]ACTH decreased 5-HT content and turnover (Leonard, 1974), whereas a-MSH had no effect (Spirtes et al., 1975). Rigter et al. (1977) studied the relation between ACTH 4-1 0, retrieval of a passive-avoidance response, and hippocampal 5-HT content. They had shown previously that application of footshock during the acquisition of a one trial passive-avoidance test was associated with a rise in hippocam pal 5-HT concentration during the retention test 24 hours later (Leonard and Rigter, 1975; Rigter et al., 1975b). Retrograde amnesia can be produced by treating the rats with COB. Treatment with ACTH 4-10 1 hour prior to the retrieval test alleviated the C02-induced amnesia, T h e antiamnesic effect of ACTH 4-10 was paralleled by a rise in the hippocampal 5-HT concentration while preacquisition treatment with ACTH 4-10 did not affect hippocampal5-HT (Rigter et al., 1977). T h e biochemical response therefore seems more likely to be the result of a behavioral change than of a peptide action. In summary, the data on brain neurotransmitters are scarce and often confusing. No clear-cut evidence is available regarding the mechanism of action (modulator putative neurotransmitter, synthesis, release, etc.), the nature of the neurotransmitter system involved, and the specificity of the observed effects. G. ACTH 4-10

AND

MORPHINE

1. Brain Opiate Receptors In the course of search for the identity of the endogenous ligand for morphine receptors, Terenius tested a variety of peptides for their capacity to inhibit binding of [“HI dihydromorphine to rat brain opiate receptors. Substance P, lysine vasopressin (LVP), desglycinamide lysine vasopressin (DG-LVP), bradykinin, TRH, and prolyl-leycyl-glycine amide (PLG) were inactive (Terenius, 1975: Terenius et al., 1975), while ACTH 1-28 and ACTH 4-10 had an affinity for these receptors (Terenius, 1975). Structure-activity studies pointed to an active site within ACTH 4-10 with some indication that a second affinity sight might be present in a sequence which is distal to the C-terminal (Terenius et al., 1975; Gispenet al., 1976a). [~-Phe’]ACTH 4-10 and ACTH 4-10 were equally active. ACTH 4-7, the shortest sequence able to delay avoidance extinction (de Wied et al., 1975) or to induce excessive grooming (Wiegant and Gispen, 1977), was inactive in the dose used. Analysis of the binding characteristics of ACTH-like peptides revealed a relatively low selectivity of these peptides for agonist or antagonist binding

230

W. H .

GISPEN, J.

M . VAN

REE,

A N D D. DE WIED

sites, comparable to nalorphine, a partial agonist-antagonist (Terenius et al., 1975; Terenius, 1976). In view of the relatively low affinity of ACTH 4-10 for the opiate receptor (ICsoin the order of 10-'-10-"M), it seems unlikely that it is an important endogenous opiate receptor ligand under physiological conditions. Since fragments from LPH 6 1-9 1 have a high affinity for such receptors, the low-affinity site of LPH 47-53 reflects the redundant manner in which information is encoded in LPH. The fact that ACTH 4-10 and ACTH 1-24 have affinity for CNS opiate receptors may, as discussed below, explain the observed interaction of these peptides with morphine at various levels of nervous system function (Wiegant et al., 1977b). 2. Morphine-Induced Analgesia Morphine has profound effects on pituitary-adrenal function (Selye, 1936; Briggs and Munson, 1955; de Wied et al., 1974b: van Ree et al., 1977a,b), and evidence is accumulating to suggest an interaction of ACTH with the CNS effects of morphine, e.g., counteraction of morphine-induced spinal reflex activity in vivo and in vitro (Zimmermann and Krivoy, 1973). In studies on the analgesic action of morphine in rodents, it was observed that ACTH could counteract this response (Winter and Flataker, 1951), although evidence was obtained that such an antagonism occurred only in the presence of the adrenal gland (Paroli, 1967; Gispen et al., 197513). Furthermore, glucocorticoids have been shown to alter certain actions of morphine on the CNS (Gispen et al., 1975b; Brown and Garret, 1972; Zimmermann et al., 1974). In a recent study, ACTH-like peptides, devoid of corticotrophic activity, were used to investigate the counteraction by peptides of morphine-induced analgesia as measured by a hot-plate technique (Gispen et al., 1976a). The peptides inhibited the analgesic response by 5 0 4 0 % with [~-Phe'] ACTH 4-10 more potent than ACTH 4-10 (Gispen et al., 1976a). The peptides were without a detectable action on the response behavior of saline-treated control rats, (Gispen et al., 1970b, 1973, 1976a). Thus, the structure-activity relationship observed for ACTH-morphine interaction in vivo resembles that for ACTH affinity to rat brain opiate binding sites in vitro. 111. PMSH

It is most likely because of the sequence 47-53 (ACTH 4-10) that P-MSH has effects on the CNS which in many cases are identical to those observed with ACTH or ACTH 4-10. We will thus briefly review relevant data on CNS effects of P-MSH.

LIPOTROPIN A N D T H E CENTRAL NERVOUS SYSTEM

23 1

In mice, chronic subcutaneous treatment with P-MSH results in hyperexcitability and hypersensitivity which could not be mimicked by a-MSH (Sakamoto, 1966; Segawa et al., 1973). In contrast, repeated injections of P-MSH in rats induce excessive drowsiness (Sakamoto, 1966). For a variety of species, intracranial application of P-MSH elicits the SYS. (Ferrari et al., 1963). In rats, this syndrome is preceded by display of excessive grooming (lzumi et al., 1973; Gispen et al., 1975a). In the majority of instances, the effects of P-MSH on learned behavior resemble those brought about by a-MSH. Although most studies deal with a-MSH, many state that similar results have been obtained with P-MSH in the same paradigm (Kastin et al., 1975; van Wimersma Greidanus, 1977). In short, a- and P-MSH increase the rate of avoidance acquisition in hypophysectomized (de Wied, 1969; Gispen and Schotman, 1970) and intact rats (Stratton and Kastin, 1974). Furthermore, these melanotropic peptides inhibit the extinction of active and passive avoidance behavior (de Wied, 1966, 1969; Sandman et al., 1971; Dempsey et a/., 1972; Greven and de Wied, 1973; van Wimersma Greidanus, et al., 1975d; van Wimersma Greidanus, 1977) and appetitive behavior (Sandman et d.,1969; Kastin et a/., 1974). On the basis of the few studies on neurochemical correlates of P-MSH activity, it seems conceivable that P-MSH influences brain protein metabolism in a manner similar to ACTH 4-10. P-MSH 6-24 hours after being injected increases the rate of incorporation of [“C] valine into mouse brain proteins without affecting the acid-soluble radioactivity (amino acid precursor pool) (Rudman et al., 1974). Furthermore, after treatment with P-MSH and conditioning in the shuttle box, hypox rats who had acquired the response as a result of peptide treatment had more of the large polysomal aggregates in their brainstem than did the controls (Gispen, 1970). In a study of the ventral root potential evoked by stimulation of the dorsal root of cat spinal cord it was found that intravenously administered P-MSH facilitated the ventral root response (Guillemin and Krivoy, 1960; Krivoy and Guillemin, 1961), and this action could not be mimicked by ACTH or a-MSH. I t has been demonstrated that P-MSH altered the recovery period of nerve cells, causing these cells to remain longer in a hyperexcitable state (Krivoy et nl., 1963). Further study of the facilitating effect of P-MSH on postfiring recovery of synaptic transmission in cat spinal cord revealed that a-motor neurons are specially sensitive to P-MSH and the the action of the peptide is probably post-synaptic (Krivoy and Zimmermann, 1977). The existence of P-MSH-degrading enzymatic activity in the brain may further support the hypothesis of a physiological role for this peptide in CNS function (Long et al., 1961).

232

W. H . GISPEN, J. M. VAN REE, AND D. DE WIED

Krivoy proposed that P-MSH should be regarded as a modulator of synaptic transmission (Krivoy, 1970; Zimmermann and Krivoy, 1973); the report by Strand and Cayer (1975) of an increased amplitude of sciatic nerve potentials (hypox rats) and gastrocnemicus muscle (intact and hypox rats) after treatment with P-MSH fits the concept of a modulator. Krivoy and co-workers have shown that morphine reduced the amplitude of evoked mono- and polysynaptic reflex activity in the decerebrated cat (Krivoy et al., 1973). When cats were injected with P-MSH before morphine, the depressant action of morphine was not seen (Zimmermann and Krivoy, 1973), probably as a result of the stimulatory effect of P-MSH itself (Krivoyet al., 1974). If this counteraction was not caused by competition of P-MSH and morphine for the same receptor site, it is not surprising that P-MSH did not alter morphine-induced analgesia as manifested in reduced response of rats to inescapable footshock (Gispen et al., 1975b).

IV. P-Lipotropin 61 -91

Mild digestion of P-LPH by trypsin led to specific cleavage at the peptide bond 60-61 and as a consequence, to release of the C-terminal part of P-LPH representing residues 6 1-9 1 (C-fragment, @endorphin) (Bradbury et al., 1976d). P-LPH 61-91 was isolated in large quantities as an intact polypeptide from pig and camel pituitary glands (Bradbury et al., 1976; Li and Chung, 1976b) and has since been shown to be endogenous to the brain (Bradbury et al., 1976~).However, the calculated amount of P-LPH 61-91 in a single brain is approximately one-eighth the amount present in a single pituitary (Bradbury et al., 1 9 7 6 ~ ) . In addit.ion, smaller peptides which share amino acid sequences with P-LPH 61-91 have been isolated from whole brain (Hugheset al., 1975) or from hypothalamus-neurohypophysis extracts (Guillemin et al., 1976). These peptides have in common at least the sequence Tyr(61)Phe(64) of the NH2-terminus of P-LPH 61-9 1. Tyr-Gly-Gly-Phe-Met (sequence 61-65, Met"-enkephalin) and Tyr-Gly-Gly-Phe-Leu (Leu"enkephalin) were isolated from pig and bovine brain tissue (Hughes et al., 1975; Simantov and Snyder, 1976). Longer peptides were obtained from a crude extract of porcine hypothalamus-neurohypophysis, e.g. P-LPH 61-76 (a-endorphin) and P-LPH 61-77 (y-endorphin) (Guillemin Pt al., 1976; Ling and Guillemin, 1976) and P-LPH 61-87 (C'fragment) occurred as an intact polypeptide in hypophyseal extracts (Bradbury et al., 1 9 7 6 ~ ) .

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM

233

A. MORPHINELIKE ACTIVITY The striking feature of these peptides, and one which in part led to the discovery, is that they mimic certain actions of morphine. The affinity to stereospecific receptors of narcotic analgesics in brain and peripheral tissue preparations has been used as a guide to isolate most of these peptides. Their morphinelike action was assessed in vitro with preparations of the guinea pig ileum longitudinal muscle and of mouse vas deferens (Coxet al., 1975; Hughes, 1975; Hughes et al., 1975; Guillemin et al., 1976) and in displacement studies using specific binding to highaffinity opiate binding sites in brain subcellular fractions (Terenius and Wahlstrom, 1975; Pasternak et al., 1975). Such endogenous substances with opioid activity have been designated generally as endorphins. To date, P-LPH 61-91 appears to be the most active of the endorphins, especially in the binding assay (Ling and Guillemin, 1976; Bradbury et al., 1976d). All the endorphins presently available act as full morphine agonists on the guinea pig ileum and mouse vas deferens; the same holds for the binding assay with respect to the enkephalins. P-LPH, however, has the binding properties expected for a morphine antagonist or mixed agonist-antagonist (Bradbury et al., 1976d; Birdsall et al., 1976). In view of the in vivo activity of this peptide (see below), generalizations derived from studies with morphinelike drugs are of little value in predicting the characteristics of a peptide in this respect. When endorphins were administered in vivo, their morphinomimetic action could be determined. Antinociceptive effects of intraventricularly injected Met"-enkephalin have been reported for rats and mice, but these effects were rather weak and transient (Belluzzi et al., 1976; Buscher et al., 1976), probably due to rapid enzymatic degradation of Met"-enkephalin in vivo (Hambrook et al., 1976). Analogs of enkephalin resistant to degradation by brain enzymatic activity can indeed produce long-lasting analgesia (Hambrook et al., 1976; Pert et al., 1976, Pert, 1976). Profound analgesia was found after intraventricular injection of P-LPH 61-91 in cats, rats, and mice (Feldberg and Smyth, 1976; Graf et al., 1976a; Loh et al., 1976; van Ree et al., 1976), suggesting that this longer peptide is more resistant to the metabolizing enzymes. O n a molar basis, P-LPH 61-91 appeared to be about 100 times more potent than morphine in cats (Feldberg and Smyth, 1976). The antinociceptive activity of this peptide was slightly lower (30-40) and 15-35 times more potent than morphine in rats and mice (Graf et al., 1976a; Loh et al., 1976; van Ree et al., 1976). In addition, it was found that this peptide produced analgesic effects when it was injected intravenously in mice (Tseng et al., 1976). After five intraventricular injections of P-LPH 61-91 in a 3-day treatment schedule, hardly any analgesia could be detected using a hot-

234

W. H . GISPEN, J. M. VAN REE, AND D. DE WIED

plate test procedure. In these “tolerant” animals, the analgesic response to intraventricularly administered morphine was also diminished, indicating that cross-tolerance between P-LPH 61-91 and morphine had occurred (van Ree et al., 1976). Cross-tolerance between morphine and enkephalin has been established in the same manner. Rats made tolerant by implantation of morphine pellets showed a diminished analgesic response to both intraventricularly administered morphine and Met’enkephalin (Blasig and Herz, 1976). The enkephalin-induced inhibition of cortical single neurons was abolished in morphine-tolerant animals (Zieglgansberger et al., 1976). Furthermore, cross-tolerance between morphine and Met”-enkephalin appeared to occur in peripheral preparations obtained from morphine-tolerant animals (Waterfield et al., 1976). P-LPH 6 1-91 and morphine share similar dependence properties as assessed by naloxone-induced withdrawal signs (Wei and Loh, 1976a,b; Loh et al., 1976). Continuous infusion of P-LPH 61-91 directly into periaqueductal gray fourth ventricular spaces or repeated intraventricular injections of purified pituitary material induced physical dependence as evidenced by the occurrence of a typical morphinelike withdrawal syndrome after treatment with naloxone (Wei and Loh, 1976a,b; Loh et al., 1976; Blasig and Herz, 1976). Physical dependence was also present after infusion of considerable amounts of Met’-enkephalin into brain tissue. In contrast, Leu”-enkephalin appeared to be devoid of dependence-producing properties (Wei and Loh, 1976a,b). Assuming that tolerance and physical dependence may be due to a biochemical response subsequent to the activation of the recognition of the receptor complex, these data provide further evidence that morphine and endorphins interact with similar receptor sites as had been proposed on the basis of affinity studies. Structure comparisons between morphine and enkephalin reveal that the primary attachment of enkephalin to the receptor may involve the aromatic hydroxyl moiety of the tyrosine residue (Roques et al., 1976; Jones et al., 1976). Indeed, substitution of the tyrosine residue in endorphins leads to substantial loss of affinity for the receptor (Ling and Guillemin, 1976; Chang et al., 1976).

B. GROOMING ACTIVITY As already mentioned, rats displayed excessive grooming after intraventricular administration of ACTH-like peptides (Ferrari et al., 1963; Izumi et al., 1973; Gispen et al., 1975a). This behavioral response was completely suppressed by specific opiate antagonists (Gispen and

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM

235

Wiegant, 1976), suggesting that these peptides interact with brain opiate receptors. There is more evidence for a specific interaction from the findings that: ACTH-like peptides have an appreciable affinity for brain opiate binding sites in d r o (Terenius, 1976; Terenius et al., 1975); they counteract the analgesic effect of morphine (Gispen et nl., 1976a); and they exhibit a morphinelike action in preparations of mouse vas deferens (van Ree, unpublished observations), The data obtained with the binding assay, the mouse vas deferens, and the morphineinhibiting action in 7 i i z w show that the active core of ACTH with respect to affinity and intrinsic activity to opiate receptors is located in the sequence 4-10. However, relatively large quantities of ACTH-like peptides are needed to demonstrate these activities. I f the induction of excessive grooming by ACTH-like peptides is mediated by opiate receptors, endorphins should be more potent in this respect. Indeed, P-LPH 61-91 elicited grooming activity and was somewhat more potent than ACTH-like peptides. An intraventricular injection of as little as 10 ng of P-LPH 61-91 significantly induced excessive grooming (Gispen et al., 1 9 7 6 ~ )P-LPH . 61-76 was much less active than P-LPH 61-91. The sequence P-LPH 61-91 was slightly less potent than P-LPH 61-76, while Met"- and Leu"-enkephalin could hardly elicit grooming activity over the wide dose range tested. These results agree with the data given above concerning the analgesic action of these peptides in z)ivo. The rapid enzymatic deactivation of enkephalin might also be responsible for the low activity of these shorter peptides in inducing excessive grooming. C. AVOIDANCE BEHAVIOR Interestingly, peptides derived from P-LPH 61-91 and ACTH share another specific interaction with CNS processes. Extinction of- active avoidance behavior (pole-jumping test) can be delayed by a single systemic injection of ACTH 4-10 (see above). Using a similar test procedure, it was found that a subcutaneous injection of 3 pg Met"enkephalin, 0.3 pg LPH 61-76, or 0.3 pg P-LPH 6 1-91 had behavioral effects identical to those of 3 pg ACTH 4-10 (de Wied, 1977). Peptides derived from P-LPH 61-91 are thus approximately 10 times more potent in inhibiting the extinction of active avoidance behavior. I t is unlikely that this particular action of the peptides is due to their morphinelike activity, since morphine is unable to induce similar effects (de Wied, unpublished observations). Several years ago, peptides with high behavioral potency as assessed by their effect on the extinction rate of active avoidance behavior were isolated from pituitary glands (de

236

W. H . GISPEN, J. M. VAN REE, A N D D. DE W E D

Wied et al., 1970; Lande et al., 1973). Tryptic digestion of one of these peptides yielded three major components after subsequent electrophoresis. The behavioral activity appeared to be restricted to one component. Although the amount of material was not sufficient to permit complete chemical identification or structure elucidation, the amino acid composition of the behaviorally active component was similar to that of P-LPH 61-69. One of the inactive components resembled the residues found in P-LPH 70-79. Therefore, it might be assumed that the behaviorally active peptide isolated from the pituitary gland shares at least the sequence 61-79 with P-LPH. Evidence has been presented that ACTH 4-10 and related neuropeptides are involved in motivational processes (de Wied, 1974). Since neuropeptides related to P-LPH 6 1-91 have behavioral effects identical to those of ACTH 4-10, and since P-LPH 61-91 and ACTHlike neuropeptides have an affinity for brain opiate binding sites, it is tempting to speculate that peptides of pituitary origin which mimic or antagonize the acute action of morphine may modulate pain motivation (van Ree and de Wied, 1976).

D. ELECTROPHYSIOLOGY Iontophoretic application of Met"- Leu"-enkephalin to single neurons in the medial brainstem (i.e., reticular neurons) of decerebrated cats revealed that these peptides depressed the firing rate of most neurons (Gent and Wolstencroft, 1976a,b). P-LPH 61-76 exerted the same overall inhibitory effect as enkephalin. Similar results were obtained when MetJ-enkephalin was applied to single neurons of rat brainstem (Bradley et al., 1976; Gayton and Bradley, 1976). Morphine mimicked the response of enkephalin in both species. The specific opiate antagonist naloxone inhibited the action of both morphine and enkephalin in the rat (Bradley et al., 1976). Because reticular neurons are implicated in the neuronal transfer of noxious stimulation (Casey, 197 1; Wolstencroft, 1964; Haigler, 1976), it may be that enkephalin is physiologically involved in the pathway mediating pain perception. This hypothesis is supported by the finding that the excitatory response of thalamic neurons to peripheral noxious stimuli was completely blocked by iontophoretic application of Met'-enkephalin (Hill et al., 1976a,b). Although iontophoresis of Met"-enkephalin depressed the firing rate of single neurons regardless of which brain area was studied (cerebral cortex, thalamus, medulla), the short latency, synaptically evoked firing of cuneate nucleus and ventrobasal thalamus sensory neurons was not blocked by Met"-enkephalin (Hill et al., 1976b).

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM

237

Evidence has been presented that enkephalin can affect the postsynaptic cell membrane (Zieglganberger and Fry, 1976; Satoh et al., 1976). This peptide delayed and finally blocked the late response of cortical neurons to transcallosal stimulation. In tracellular recordings from cat spinal neurons revealed that iontophoresis of enkephalin antagonized l-glutamate-induced depolarisation without causing detectable changes in membrane potential or membrane resistance. The data from microiontophoretic application of enkephalin to single neurons, together with the rapid deactivation of these peptides when they are administered in vivo, point to the possibility that these substances are inhibitory neurotransmitters (Gent and Wolstencroft, 1976a; Bradley et al., 1976; Hill et al., 1976a; Frederickson et al., 1976; Kosterlitz and Hughes, 1975; Zieglgansberger and Fry, 1976). The pre- or postsynaptic receptors of these putative transmitters could be sites where the opiate alkaloids exert their pharmacologically morphinomimetic action.

E. CYCLICNUCLEOTIDES There is increasing evidence LO suggest that cyclic nucleotides are involved as mediators of the postsynaptic action of some neurotransmitters (Greengard, 1976). The hypothesis that enkephalins may act as brain neurotransmitters and the observation that these peptides may be endogenous ligands for “opiate receptors” have raised the question whether enkephalins influence the metabolism of cyclic nucleotides in preparations in which morphine-induced changes have been seen. In cultured neuroblastoma X glioma hybrid cells which display many properties characteristic of neurons and which have a high density of opiate receptor binding sites, morphine elicits an increase in cGMP content, a decrease in cAMP content, an inhibition of basal adenylcyclase activity, and a blockade of prostaglandin El-induced rise in cAMP (Traber et al., 1975a; Sharma et al., 1975a,b; Gullis et al., 1975). Similar effects were observed when enkephalins were added to the incubation medium (Gullis et al., 1976; Brandt et al., 1976; Klee et al., 1976; Klee and Nirenberg, 1976). Both these effects of morphine and of enkephalins can be completely antagonized by naloxone. In this particular assay, as in the mouse vas deferens preparation (Hugheset al., 1975), the enkephalins showed a very high potency in comparison with morphine. Prolonged exposure of the cells to morphine caused an increase in adenyl cyclase activity, which was interpreted to reflect a biochemical correlate of tolerance and dependence (Traber et al., 1975b; Klee et al., 1975). The same effects were observed when the cells were incubated with Met“-enkephalin for 12 or more hours (Klee et al., 1976; Lampert et al., 1976).

238

W. H ,

GISPEN,

J. M. V A N

REE,

AND D. DE WIED

Morphine increased the in vitro accumulation of cGMP and decreased that of CAMP in slices of rat neostriatum, an area containing a very high density of opiate binding sites (Minneman and Iversen, 1976a,b; Kuhar et al., 1973; Snyder, 1975). The addition of enkephalins to these slices had similar effects. High concentrations of the enkephalins were needed, however, even after the addition of a peptidase inhibitor, in contrast to the neuroblastoma X glioma hybrid cells preparations. Again, these data are suggestive of the instability of the enkephalins in mammalian tissue. It is nevertheless concluded that the enkephalins affect the brain cyclic nucleotides in a manner similar to morphine. Whether such changes can account for the postulated action of enkephalins on postsynaptic cell membranes is still open to question. F. BIOLOGICAL SIGNIFICANCE

The origin of endorphin is not yet known with certainty. The identity of Met"-enkephalin with P-LPH 6 1-65 indicates a possible pituitary origin for this peptide. However, no evidence was found for the existence of Met"-enkephalin in pituitary material. I t is possible that larger endorphins, specially P-LPH 6 1-91, are physiologically active and that Met"enkephalin is merely a degradation product with high receptor affinity. Indeed, when P-LPH was incubated with extracts of rat brain, peptide fragments with opiatelike activity were generated (Lazarus et al., 1976). This finding suggests that the enkephalin found in brain extracts arose from breakdown of larger sequences during isolation procedures. The first cleavage of P-LPH to liberate P-LPH 61-91 can occur in the pituitary. This fragment might then reach the brain via the CSF or other routes and interact with specific receptors in the brain. At the moment, one can only speculate about the physiological significance of P-LPH 61-91 and related peptides. Most of the present research seems to be concentrated on the enkephalins. Whether the morphinelike action of these peptides really is their most important biological function is questionable as, even in very low doses, the endorphins affect brain processes involved in conditioned avoidance behavior. In contrast, relatively high amounts of the peptides were needed for morphinelike activity in vivo, especially when the enkephalins were tested. The lower potency of enkephalins cannot solely be explained by their assumed higher sensitivity to enzymatic degradation, since such a mechanism is also present in the avoidance behavior test. Other peptides with morphinelike activity and with characteristics which differentiate them from peptides related to P-LPH 61-91 have been isolated from the pituitary (Cox et al., 1976b; Gentleman et al.,

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM

239

1976: Goldstein, 1976) suggesting that enkephalin need not be the only endogenous ligand for opiate receptors. The present confusion extends to the physiological role of the endorphins as well. N o conclusions could be drawn from experiments involving naloxone treatment to block the receptors for these ligands in test situations reliably affected by morphine. Naloxone did not change the sensitivity threshold for footshocks (Goldstein et al., 1976) or affect temperature control under cold stress (Goldstein and Lowery, 1975). Yet, using inescapable electric footshock (EFS) to measure the animal’s response to pain, it was found that hypox rats-thus depleted from pituitary endorphinsaisplayed an increased responsiveness to the aversive stimulus (Gispen et nl., 1970b). Although developed for the purpose (Evans, 1961), it is unlikely that EFS exclusively measures the animal’s response to pain: Other factors such as motivation, locomotor activity, and integrity of reflex pathways also influence the response to the EFS (Gispen et al., 1973). It is interesting that nalaxone blocked the hyperthermia produced by conditional stimuli (La1et al., 1976) and food- and water-seeking behavior in hungry or thirsty rats (Holtzman, 1974, 1975). A study of the regional distribution of the opiate binding sites showed many such sites in structures of the limbic system (Kuhar et al., 1973; Snyder, 1975). Therefore, it can be inferred that peptides which have specific affinity for opiate binding sites play an essential role in limbic system functions. This assumption is supported by the effects of naloxone on food- and water-seeking behavior and by the marked effect of endorphins and ACTH 4-10 on extinction of active avoidance behavior. The activity of limbic structures might be involved in all these test situations.

V. Concluding Remarks

Since the discovery of P-LPH in 1964 by C. H. Li, this polypeptide has been a hormone in search of a function. At the moment, the picture emerges that peptides derived from P-LPH strongly influence nerve cell functioning. ACTH 4-10 (=P-LPH 47-53) and related peptides (ACTH, P-MSH, a-MSH) are thought to facilitate acquisition and retention of learned behavior by increasing the motivation value of environmental stimuli. In man, the effects are defined in terms of increased attention, in both volunteers and retarded patients. At the cellular level, these peptides exert a trophic response which had previously been interpreted as reflecting the activity necessary to alter neuronal connectivity. This latter process then may underlie the storage of newly acquired information. Thus, a behavioral experience may activate the pituitary to

240

W. H.

GISPEN, J.

M. VAN REE, AND D. DE WIED

release ACTH 4- 10-containing neuropeptides which reach the brain by a still unknown mechanism and then exert their action. Peptides derived from the C-terminus of P-LPH have a similar or higher potency than ACTH 4-10 in behavioral tests. Therefore, the physiological role of these peptides is more likely to be the regulation of behavioral mechanisms than the surprising morphinelike activity. If the morphinelike activity is meaningful in this respect it may be that these peptides affect pain motivation rather than pain perception. I t still remains to be proven that the effects of LPH fragments in vivo are due to peptide action alone. It may be that part of the biological activity results from an interference with normal pituitary function due to the effects of the fragments on neuroendocrine mechanisms. Thus, some of the effects of P-LPH fragments may serve only as trigger signals, whereas a mediating principle (hormone) in fact underlies the functional response. The recognition that both ACTH 4-10, P-MSH, and several endorphins are identical to certain sequences in the peptide P-LPH may be crucial for understanding their mechanism of action. Increasingly, P-LPH is viewed as a prohormone for P-MSH and endorphins. Prohormones are not uncommon in mammalian physiology: A peptide-prohormone usually generates an active peptide by enzymatic cleavage of the inactive parental molecule. P-LPH seems to fit such a precursor role since, in many tests, the molecule does not mimic the effect of its active fragments. LPH has no affinity for the opiate receptor, has no morphinelike activity, and is inactive in the behavioral tests studied so far. In the case of P-LPH, it is postulated that a behavioral experience triggers the production of LPH fragments. Knowledge of the enzymatic activity involved (enzymes, cellular/regional localization, regulation, etc.) is a prerequisite for the understanding of the biological significance of P-LPH. It seems not unlikely that 0-LPH generates at least two classes of neuropeptides (ACTH 4-1OIP-MSH and endorphins). In view of the similarities between the activities at least as they are known at present of these two classes (affinity for opiate receptors, effects on avoidance and grooming behavior) it can be argued that their existence merely reflects the redundant way in which information is encoded in biologically active peptides. According to this view the C-terminal peptides would contain the true active site, whereas the N-terminal peptides would reflect the dormant activity sites. However, we would like to propose that the difference between the two classes resides not in their potency but rather in the direction in which they influence brain mechanisms. There is a variety of evidence to support the notion that two opposite messages are encoded in LPH, one, N-terminal from T y P (ACTH 4-lO/P-MSH), and the other, C-terminal, with the T y P as N-terminal (endorphins). A known

LIPOTKOPIN AND THE CENTRAL NERVOUS SYSTEM

24 1

but still debated parallel is the case of oxytocin which does not itself affect MSH-release but appears to influence this release in opposite ways by means of its two fragments, ring (tocinamide) and tail (PLG). In summary (see Table I), peptides containing ACTH 4-10, in general, stimulate firing rate, excitability of neurons and neurotransmission, increase cAMP levels, and counteract some morphine-induced CNS effects (inhibition of spinal reflex activity, analgesia). Effects opposite to these have been reported for LPH-endorphins, i.e., the inhibitory influence on neurophysiological processes and cAMP levels and the morphinomimetic action (in vitro, analgesia, tolerance, etc.). Thus, proper enzymatic cleavage could release both an excitatory and an inhibitory influence on nerve tissue. The actual processing of such signals should not necessarily be as “black and white” as the release of opposing information suggests. The differences known to exist between the two classes of LPH fragments in, e.g., metabolic stability, intrinsic activity, and receptor affinity can be regarded as tools which allow, upon cleavage of LPH, a release of information to neurons in a subtle manner. I f indeed, as already suggested, there is more than one type of opiate or peptide receptor in the central nervous system, the two classes of LPH fragments need not affect behavioral processes through identical receptors but may influence identical brain structures (limbic system?) via binding to different receptor sites. Indeed after intraventricular administration of low doses of P-MSH or P-LPH 6 1-9 1 to rats, a complex behavTABLE I SUMMARY OF CNS ACTIVITIES OF LPH FRAGMENTS‘‘ P-LPH CNS activity

P-MSHIACTH 4-10

Electrical activity cAMP levels Opiate receptor affinity in ztitro Opiatelike activity in 71Ii~o Avoidance extinction Excessive grooming

t

Endorphins

t

+ 01-

+++ ++

+++ + +++

+++

The evidence summarized in this table is discussed in the various sections of-this chapter. The arrows indicate increase or decrease of activity. The + rating system compares the potency of the two classes of Fragments for a given parameter. 0 stands for no effect and - for counteraction of effect. ‘I

242

W. H . GISPEN, J. M. VAN REE, AND D . DE WIED

ioral response is observed, which is qualitatively and quantitatively different for the two classes of LPH peptides despite an overlap with respect to excessive grooming behavior. It is clear that, in 1975/1976, neurobiology has arrived at the brink of a new neuropeptide era. From the abundance of preliminary communications in recent months it is not difficult to predict that, in the near future, much research will be undertaken to elucidate the role of endorphins in the CNS, behavior, and even abnormal behavior. Previously, the delay of extinction of a learned response, thus the perseverating performance in the absence of punishment or reward, has already been considered to reflect a maladapted behavioral response to an altered environment. Exogenously administered ACTH 4-1 0 enhances this process and it was therefore proposed that one of the endogenous factors involved in the onset or maintenance of neuroses related to a dysfunction in the production of pituitary ACTH 4-10-containing neuropeptides (Gispen et al., 1976b). Attention has recently been called to a possible role of LPH-endorphins in schizophrenia and psychoses (Terenius et al., 1976; Bloom et al., 1976, Jacquet and Marks, 1976). Amidst the turbulent development of current LPH-neuropeptide research into putative neurotransmitters, modulators, peptidergic neurons, pain perception/ motivation, etc., the major effort should not be a single-minded approach but a broad and critical one based on the enormous potential of such peptides in brain and behavior processes. REFERENCES Ader, R., Weijnen, J. A. W. M., and Moleman, P. (1972).Psychon. Sci. 26, 125-128. Allen, J. P., Kendall, J. W., McGilvra, R., and Vancura, C. (1974).J . Clin. Endocrinol. Metab. 38,586-593. Ambach, G., and Palkovits, M . (1975).ActaMmphol. Acud. Sci. Hung. 23, 21-49. Anderson, D. C., Winn, W., and Tam, T. (1968).J . Comp. Physiol. Psychol. 66, 497-499. Applezweig, M. H., and Baudry, F. D. (1955).Psychol. Rep. 1, 417-420. Applezweig, M. H., and Moeller, G. (1959).Acta Psychol. 15,602-603. Baldwin, D. M., H a m , C. K., and Sawyer, C. H . (1974).Brain Rex. 80, 291-301. Beagly, W. K. ( 1 976).J. Comp. Phys. Psychol. 90, 790-798. Beatty, D. A., Beatty, W. A., Bowman, R. E., and Gilchrist, J. C. (1970).Physiol. Behaw. 5, 939-944. Belluzzi, J. D., Grant, N., Garsky, V., Sarantakis, D., Wise, C. D., and Stein, L. (1976). Nature (London) 260, 625-626. Bertolini, A., Vergoni, W., Gessa, G. L., and Ferrari, W. (1968).Life Sci. 7, Part 2, 12031206. Bertolini, A., Vergoni, W., Gessa, G. L., and Ferrari, W. (1969). Nature (London) 221, 667-669. Birdsall, N. J. M., Hulme, E. C., Bradbury, A. F., Smyth, D. G., and Snell, C. R. (1976).111 "Opiates and Endogenous Opioid Peptides" (H. W. Kosterlitz, ed.), pp. 19-26. North-Holland Publ., Amsterdam. Blasig, J., and Herz, A. ( 1976). Naunyn-Srhmiedeberg's Arch. Pharmacol. 294, 297-300.

LIPOTROPIN A N D T H E CENTRAL NERVOUS SYSTEM

243

Bloom, F., Segal, D., Ling, N., a n d Guillemin, R. (1976). S c i m e 194, 630-632. Bohus, B. (1975). Prog. Brain Res. 42, 275-283. Bohus, B., a n d de Wied, D. (1966). Scieilrr 153, 318-320. Bohus, B., and de Wied, D. (1977).I n “General, Comparative a n d Clinical Endocrinology of the Adrenal Cortex” 0. Chester Jones and I. W. Henserson, eds.), Vol. 3, Academic Press, New York. In press. Bohus. B.. Nyakas, C., a n d Endroczi, E. (1968). Int. J. Neitrv,fJhannnco/.7 , 307-314. Bohus, B., Gispen, W. H., a n d d e Wied, D. (1973).Neirromdocriwlogy 11, 137-143. Bohtis, B., Hendrickx, H. H. L., van Kolfschoten, A. A,. and Krediet, T. G. (1975). I n “Sexual Behavior: Pharmacology and Biochemhtry” (M. Sandler and G. 1.. Gessa, eds.), pp. 269-275. Raven, New York. Bradbury, A. F.. Smyth, D. G., and Snell, C. R. (1976a).Biorhem.Biophys. Res. Commnn. 69, 950-956. Bradbury, A. F., Smyth, D. G., Snell, C. R., Birdsall, N. J. M., and Hulme, E. C. (1976h). Nntnre (London) 260, 793-795. Bradbury, A. F., Feldherg, W. F., Smyth, D. G., a n d Snell, C. R. ( 1 9 7 6 ~ )In. “Opiates and Endogenous Opioid Peptides” ( H . W.Kosterlitz, ed.), pp. 9-17. North-Holland l’uhl., Amsterdam. Bradbury, A. F., Smyth, D. G., a n d Snell, C. R. (1976d). Cihn Foirnd. Symp. 41, 61-75. Bradley, P. B., Briggs, I., Gayton, R. J.. and Lamhert, I,. A. (1976). Nntiire (Lmidun) 261, 425-426. Brandt. M., Gullis, R. J., Fischer, K., Buchen, C., Hampi-echt, B., Moroder, L., and Wiinsch, E. (1976).Nntiire (London) 262, 91 1-312. Briggs, F. N., and Miinson, P. L. (1955).Endocrinnlog~57, 205-219. Brown, J. H., a n d Garrett, R. L. (1972).Arch. Int. Plrormocodyn. T l w . 196, 176-183. Buscher, H . H . , Hill, R. C., Romer, D., Cardinaux, F., Closse, A,, Hauser, D., and Pless, J. (1976).Notiire (Lvndmi) 261, 423-425. Burkhard, W. P., a n d Gey, K. F. (1968).Helil. Pliysiol. Plrnrmcrcol. Acfn 26, 197-198. Burt, D. R., a n d Snyder, S. H. (1975). Bririn RPL 93, 309-528. Casey, K. L. (1971). S c i a r e 173, 77-80. Chang, J . K., Bosco, T. W., Pert, F. A., a n d Pert, C. B. (1976).Lilt Sci. 18, 1473-1482. Christensen, C. W., Harston, C. T., Kastin, A. J., Kostriewa. R. M., and Spirtes, M. A. (1976). Phrnr~rcvl.,Biorhem. B P / I ~5,~suppl. . I , 1 li-120. Cox, B. M., Opheim, K. E., Teschemacher, H., and Goldstein, A. (1975). Life Sri. 16, 1777-1 782. Cox. B. M., Goldstein, A,, a n d Li, C . H. (1976a).Proc. Nntl. Acnd. Sci. 1T.S.A. 73, 61-91. Cox, B. M., Gentleman, S.,Tsiing-Ping Su, and Goldstein, .4. (1976b). Rrnin Res. 115, 285-296. Delius, J. D., Craig, B., a n d Chaudoir, C. (1976).Z. TiPr/l.\yclrol. 40, 183-193. Denipsey, G. L., Kastin, A. J., a n d Schally, A. V. (1972). Horn. B~hnir.3, 333-337. Denman, P. M., Miller, L. H., Sandman, C. A., Schally, A. V..and Kastin, A. J . (1972).J. Camp. Physiol. Psycliol. 80, 59-65. d e Wied, D. (1964). A m . J. Physiol. 207, 255-259. d e Wied, D. (1966). Proc. Sor. Exf,. Rid. Mrd. 122, 28-32, d e Wied, D. (1969).I n “Frontiers in Neuroendocrinology” (W. F. Ganong and L. Martini, eds.), pp. 97-140. Oxford Univ. Press, London a n d New York. de Wied, D. (1971). I n “Normal and Abnormal Development o f Brain a n d Behavior” (G. B. A. Stoelinga and J . J. van der Werff ten Bosch, eds.), pp. 315-322. Leiden Univ. Press, Leiden. d e Wied, D. (1974).In “ T h e Neurosciences, Third Study Program” (F. 0.Schmitt a n d F. G. Worden, eds.), pp. 653-666. Rockefeller Llniv. Press, New York.

244

W. H . GISPEN, J. M. VAN REE, AND D. DE WIED

de Wied, D. (1977).Life Sci. 20, 195-204. d e Wied, D., and Gispen, W. H . (1977).In “Peptides in Neurobiology” (H. Gainer ed.), pp. 391442. Plenum, New York. d e Wied, D., a n d Weijnen, J. A. W. M., eds. (1970).Prog. Brain Res. 32. de Wied, D., Witter, A., a n d Lande, S. (1970).Prog. Brain Res. 32, 213-218. d e Wied, D., van Delft, A. M. L., Gispen, W. H., Weijnen, J. A. W. M., a n d van Wimersma Greidanus, T. B. (1972).In “Hormones and Behavior” (S. Lexine, ed.), pp. 135-171. Academic Press, New York. de Wied, D., Bohus, B., and van Wimersma Greidaniis, T. B. (1974a).Prog. Bruin Res. 41, 417-428. de Wied, D., van Ree, J. M., and de Jong, W. (I974b).fn “Narcotics and the Hypothalamus” (E. Zimmermann a n d R. George, eds.), pp. 251-266. Raven, New York. d e Wied, D., Witter, A., and Greven, H. M. (1975).Biochm. P h m c o l . 44, 1463-1468. Dunn, A. J. (1976).In “Molecular and Functional Neurobiology” (W. H. Gispen, ed.), pp. 347-387. Elsevier, Amsterdam. Dunn, A. J., and Gispen, W. H. (1977).Biobehau. Rn!.1 (in press). Dunn, A. J., and Rees, H. D. (1976).In “ T h e Neurobiology of Sleep a n d Memory” L. McGaugh and R. R. Dmcker-Colin, eds.), pp. 33-55, Academic Press, New York. Dunn, A. J., luvone, P. M., and Rees. H. D. (1976).P h a m c o l . , Biochem. Behan. 5 , suppl. 1 , 139-145. Dnpont, A., Kastin. A. J., Labrie, F., Pelletier, G., Purani, R., and Schally, A. V. (I975).J. Endocriml. 64, 237-24 1. Dyster-Aas, H. K., and Krakau, C. E. T. (1965).ActaEndocrinal. (Copenhagen) 46,409419. Eberle, A., and Schwyzer, R. (1975).H e h . Chzm.Acta 58, 1528-1535. Endroczi, E. (1975). Prog. Brain Res. 42, 125-134. Endroczi, E., Lissak, K., Fekete, T., and de Wied, D. (1970). Prog. Brain Res. 32, 254-262. Evans, W. 0. (1961). Psychophamcologia 2, 318-325. Feldberg, W. S., and Smyth, D. G. (1976).]. Phydol. (London) 260, 30-31 P. Feldman, S., Todt, J. C., a n d Porter, R. W. (1961).Neurology 11, 109-1 15. Fentress, J. C. (1968).Anim. Behail. 16, 135-153. Ferrari, W., Gessa, G. L., a n d Vargiu, L. (1963).Ann. N.Y. Acad. Sci. 104,330-345. Ferris, S. H., Sathananthan, G., Clarke, C., Moshinsky, J., Irwin, P., and Gurshon, S. (1976).Phunnarol. Biochm. Behair. 5 , suppl. 1 , 73-78. Floris, E. (1963).Boll. SOC. Ital. Biol. Sper. 39, 558-560. Forn, J., a n d Krishna, G. (1971). Pharmacology 5, 193-204. Frederickson, R. C. A., Nickander, R., Smithwick, E. L., Shuman, R., and Norris, F. H. (1976). In “Opiates a n d Endogenous Opioid Peptides” ( H . W. Kosterlitz, ed.), pp. 239-246. North-Holland Publ., Amsterdam. Gaillard. A. W. K., and Sanders, A. F. (1975a).Prog. Brain Res. 42, 209. Gaillard, A. W. K., and Sanders, A. F. (1975b).Psychapharmacologia 42, 201-208. Garrud, P. (1975).Prog. Brain Res. 42, 173-186. Garnid, P., Gray, J. A., a n d de Wied, D. (1974).Physzol. Behav. 12, 109-1 19. Gayton, R. J., and Bradley,-1’. B. (1’976).In “Opiates and Endogenous Opioid Peptides” ( H . W. Kosterlitz, ed.), pp. 213-216. North-Holland Publ., Amsterdam. Gent,J. P., and Wolstencroft, J. H. (1976a).Nature (London) 261, 426-427. Gent, J. P., and Wolstencroft, J. H. (1976b).In “Opiates and Endogenous Opioid Peptides” (H. W. Kosterlitz, ed.), pp. 217-224. North-Holland Publ., Amsterdam. Gentleman, S., Ross, M., Lowney, L. I., Cox, B. M., and Goldstein, A. (1976).In “Opiates and Endogenous Opioid Peptides” ( H . W. Kosterlitz, ed.), pp. 27-34. North-Holland Publ., Amsterdam. Gessa, G. L., Pisano, M., Vargiu, L., Crabai, F., and Ferrari, W. (1967).Rat. Can. Biol. 26, 229-236.

u.

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM

245

Gispen, W. H. (1970). Ph.D. Thesis, Univ. of Utrecht, Utrecht. Gispen, W. H., and Schotman, P. (1970). Prog. Brain Res. 32, 236-244. Gispen, W. H., and Schotman, P. (1973). Prog. Brain Res. 39,443459. Gispen, W. H., and Schotman, P. (1976). Neuroemfocrinology. 21,97-110. Gispen, W. H., and Wiegant, V. M. (1976). Neurosci. Lett. 2, 159-164. Gispen, W. H., de Wied, D., Schotrnan, P., and Jansz, H. S. (1970a).J. Neurochem. 17, 751-761. Gispen, W. H., van Wimersrna Greidanus, T. B., and de Wied, D. (1970b).Physiol.Beharp. 5, 143-146. Gispen, W. H., de Wied, D., Schotman, P., and Jansz, H. S. (197 I).Brain Res. 31,34 1-35], Gispen, W. H., van der Poel, A. M., and van Wimersnia Greidanus, T. B. (1973). Physiol. Behail. 10, 345-350. Gispen, W. H., Wiegant, V. M., Greven, H. M., and de Wied, D. (1975a). Life Sci. 17, 645-652. Gispen, W. H., van Wirnersma Greidanus, T. B., Waters-Ezrin, C., Zimmermann, E., Krivoy, W. A,, and de Wied, D. (1975b). Eur. J . Pharmacol. 33, 99-105. Gispen, W. H., van Wirnersma Greidanus, T. B., Bohus, B., and d e Wied, D., eds. (1975~). Prog. Brain Res. 42. Gispen, W. H., Buitelaar, J.. Wiegant, V. M., Terenius, L., and de Wied, D. (1976a). Eur. J. Pharmucol. 39, 393-397. Gkpen, W. H., Reith, M. E. A., Schotman, P., Wiegant, V. M., Zwiers, H., Bohus, B., and de Wied, D. (1976b). In “Research in Neurosis (H. M. van Praag, ed.), pp. 192-202. Bohn, Utrecht. Gkpen, W. H., Wiegant, V. M., Bradbury, A. F., Hulme, E. C., Srnyth, D. G., Snell, C. R., and de Wied, D. (1976~). Nature (London)264, 794-795. Cmld, P. E., and van Buskirk, R. (1976).Behau. Biol. 16, 387400. Goidman, H., Dagirmanjian, R., Drewe, W. G., and Murphy, S. (1975). Life Sci. 17,477482. Goldman, H., Skelley, E. N., Sandman, C. A,, Kastin, A. J., and Murphy, S. (1977). Prof. Bicentennial Neiiropeptide Meet., Philadelphia (in press). Goldstein, A. (1976). Science 193, 1081-1086. Goldstein, A., and Lowery, P. (1975). Life Sci. 17, 927-932. Goldstein, A,. Pryor, G. T., Otis, L. S., and Larsen, V. (1976). Life Sci. 18, 599-604. Graf, L., Szekely, J. I., Ronai, A. Z.,Dunai-Kovacs, Z.,and Bajusz, S. (1976a). Nature (London) 263,240-24 1. Graf, L., Ronai, A. Z., Bajusz, S., Cseh, G., and Szekely, J. I . (1976b). FEBS Lett. 64, 181-184. Gray, J. A. ( 1 967). Adz!. Sci. 23, 595-605. Gray, J. A. (1971). Nature (London)229, 52-54. Gray, J. A., Mayes, A. R., and Wilson, M. (1971). Neurophnrmacology 10, 223-230. Gray, P. (1975).J . Comp. Physiol. Psychol. 88, 281-284. Greengard, P. (1976). Nature (London) 260, 101-108. Greven, H. M., and de Wied, D. (1973). Prog. Brain Res. 39,429-442. Greven, H. M., and de Wied, D. (1977). In “Control, Chemistry and Effects of MSH” (F. Tilders, D. Swaab, and T. B. van Wimersrna Greidanus, eds.), pp. 140-152, Karger, Basel. Guillemin, R., and Krivoy, W. (1960). C.R. Acad. Sci., SPY.D. 250, 11 17-1 119. Guillemin, R., Ling, N., and Burgus, R. (1976). C.R. Acad. Sci. Sw.D. 282, 783-785. Gullis, R., Traber, J., and Hamprecht, B. (1975). Nature ( L d o n ) 256,57-59. Gullis, R. J., Buchen, C., Moroder, L., Wunsch, E., and Hamprecht, B. (1976). In “Opiates and Endogenous Opioid Peptides” (H. W. Kosterlitz, ed.), pp. 143-151. NorthHolland Publ.. Amsterdam.

246

W. H. GISPEN, J. M. VAN REE, AND D. DE WIED

Guth, S., Levine, S., a n d Seward, J. P. (1971).Physiol. Behau. 7, 195-200. Haigler, H. J. (1976).Life Sci. 19, 841-885. Hambrook, J . M., Morgan. B. A., Rance, M. J.. and Smith, C. F. C. (1976).Nature (London) 262,782-783. Harris, R. K. (1973).J . Comp. Physiol. Psyclwl. 82, 254-260. Hill, R. G., Pepper, C. M., and Mitchell, J. F. (1976a).Nature (London)262, 604-606. Hill, R. G., Pepper, C. M., and Mitchell, J . F. (1976b).In “Opiates and Endogenous Opioid Peptides” (H. W. Kosterlitz, ed.), pp. 225-230. North-Holland Publ., Amsterdam. Hinde, R. A. (1970). “Animal Behaviour. A Synthesis of Ethology and Comparative Psychology,” 2nd Ed., Ch. 15. McGraw-Hill, New York. Hokfelt, T., and Fuxe, K. (1972). In “Brain-Endocrine Interaction. Median Eminence: Structure and Function” (K. M. Knigge, D. E. Scott, a n d A. Weindl, eds.), pp. 181223. Karger, Basel. Holtzman, S . G. (1974).J. P h a m c o l . Ezp. Ther. 189, 51-60. Holtzman, S. G. (1975). Life Sci. 16, 1465-1470. Hughes, J. (1975).Brain Res. 88, 295-308. Hughes, J., Smith, T. W,, Kosterlitz, H . W., Fothergill, L. A., Morgan, B. A., and Morris, H. R. (1975).Nature (London) 258, 577-579. Isaacson, R. L., Dunn, A. j.,Rees, H . D., and Waldock, B. (1976). Physiol. Psychof. 4, 159-162. Izumi, K., Donaldson, J., and Barbeau, A. (1973).Life Sci. 12, 203-210. Jacquet, Y. F., and Marks, N. (1976). Science 194,632-634. Jakoubek, B., Semiginovsky, B., Kraus, M., and Erdossova, R. (1970).Life Sci. 91, 11691179. Jakoubek, B., Buresova, M., Hajek, I., Etrychova, J., Pavlik, A., and Dedicova, A. (1972). Brain Res. 4 3 , 4 1 7 4 2 8 . Johnson, A. K., and Epstein, A. N. (1975). Bruin Res. 86, 399-418. Jones, C. R., Gibbons, W. A., and Garsky, V. (1976).Nature (London)262, 779-782. Kastin, A. J., Dempsey, G. L., LeBlanc, B., Dyster-Aas, K., a n d Schally, A. V. (1974).Horm. Behav. 5, 135-139. Kastin, A. J., Sandman, C. A., Stratton, L. D., Schally, A. V., and Miller, I.. H . (1975).Prog. Brain Res. 42, 143-150. Kastin, A. J., Nissen, C., Nikolics, K., Medziradsky, K., Coy, D. H., Teplan, I . , and Schally, A. V. (1976).Brain Res. Bull. 1, 19-26. Kawakami, M., Koshino, T., and Hattori, Y. (1966).Jpn. J . Physiol. 10, 551-569. Kelsey, J. E. (1975).J . Comp. Physiol. Psychol. 88, 271-280. Kemmler, W., Peterson, J. D., Rubenstein, A. H., and Steiner, D. F. (1972). Dinbete.\ 21, 572-583. Klee, W. A., and Nirenberg, M. (1976).Nature (London)263,609-611. Klee, W. A., Sharma, S. K., and Nirenberg, M. (1975).L f e Sci. 16, 1869-1874. Klee, W. A., Lampert, A,, and Nirenberg, M. (1976).In “Opiates and Endogenous Opioid Peptides” (H. W. Kosterlitz, ed.), pp. 153-159. North-Holland Publ., Amsterdam. KorLnyi, L., and Endroczi, E. (1967).NeuroendocrinoloRy 2, 65-75. Koranyi, L., Beyer, C., and Guzman-Flores, C . (1971). Physiol. Behau. 7, 331-335. Kosterlitz, H. W., and Hughes, J. (1975).L f e Sci. 17, 91-96. Kostrzewa, R., Kastin, A., and Spirtes, M. (1975).P h a m c o l . Biochem. Behail. 3, 1017-1023. Krivoy, W. A. (1970).Prog. Brain Res. 32, 108-1 19. Krivoy, W. A., and Guillemin, R. (1961). Endocrinology 69, 170-175. Krivoy, W. A., and Zimmermann, E. (1977). Eur. J . Phnnnucol. (in press). Krivoy, W. A., Lane, M., and Kroeger, D. (1963).Ann. N.Y. Acad. Sci. 104, 312-329. Krivoy, W. A., Kroeger, D., and Zirnrnermann, E. (1973). Br. J , P h a m c o l . 47, 4 5 7 4 6 3 .

LIPOTROPIN A N D THE CENTRAL NERVOUS SYSTEM

247

Krivoy, W. A., Kroeger, D., Taylor, A., and Zimmermann, E. (1974).Eiir. J . P h u m r o l . 27, 339-345. Kuhar, M. J., Pert, C. B., and Snyder, S. H. (1973). Nuttrre (London) 245, 447-450. Lajtha, A . , Latzkovits, L., and Toth, J . (1976). Eiorliim. Aiopliys. Acki 425, 511-520. Lal, H., Miksic, S., and Smith, N. (1976).Life Sci. 18, 971-976. Lampert, A., Nirenberg, M., and Klee, W. A. (1976). Pror. Nut/. Arud. Sei. U.S.A. 73, 3 165-3 167. Lande, S . , d e Wied, D., and Witter, A . (1973), Prog. Brain Res. 39, 421-427. Lazarus, L. H., Ling, N., and Guillemin, R. (1976). Pruc. Notl. Arod. Sci. 1I.S.A. 73, 21562159. Leonard, B. E. (1969). Int. J . Neziri~phnmnrol.8, 427-435. Leonard, B. E. (1974). Arch. Int. Phurmarodyn. T h o 207, 242-253. Leonard, B. E., and Rigter, H. (1975). Phnrmrol. RiorhPm. Rehnrl. 3, 775-780. Levine, S., and Jones, L. E. (1965).J. Comp. Physiol. Psyrhol. 59, 357-360. Ley, K . F., and Corson, J. A. (1971). Exprrientirr 27, 958 Li, C. H. (1964). Nrrtnre (London) 201, 924. Li, C. H., and Chung, D. (1976a). Nntnre (London) 260, 622-624. Li. C. H., and Chung, D. (1976b).PI-uc.Nntl.,Acnd. Sri. U.S.A. 73, 1145-1 148. Li, C. H., Barnafi, L., Chrktien, M., and Cliung, D. (1965).Nii/ure(Lmtdon) 208,1093-1094. Lichtensteiger, W.. and Lienhart, R. (1975).Arnte E f f .Nnrrotic Anulgesir Sites M e r h . Actiun; Satellite Symp. Int. Congr. Phnnnrol, 6th, Nnkknln, Espuo, Finland pp. 26-27. (Abstr.) Ling, N., and Guillemin, R. (1976).Pror. Nntl. Arnd. Sci. 1r.S.A. 73, 3308-3310. Lissak, K., and Bohus, B. (1972).Int. J . Psychohiol. 2, 103-1 15. Lloyd, C. (1974). P1i.D. Thesis, Univ. of Edinburgh, Edinhurgh. Loh, H. H., Tseng, L. F., Wei, E., and Li, C. H. (1976).Proc. Null. Arid. Sri. 1I.S.A. 73, 2895-2898. Lohmar, P., and Li, C. H. (1968). Endorrinolug 82, 898-904. Long, J., Krivoy, W. A . , and Guillemin, R. (1961).Endocrinology 69, 176-181. McCaugh, J. L., Gold, P. E., van Buskirk, R., and Haycock, J. (1975). Prog. Brnin Re.y. 42, 151-162. MacLean, P. D. (1957). A M A Arch. Neiirol, P,\yrhin/. 78, 128-142. Meyerson, B. J., and Bohus, B . (1956).Phnmnrol., Riorhem. Behn71.5, 5351-545. Miller, L. H., Kastin, A . J., Sandman, C., Fink, M., and van Veen, W. J. (1974). Phnini~rcol., Bioclzm. Behart. 2, 663-668. Miller, L. H., Harris, L. C . , Kastin, A. J., and van Rieren, H. (1976). Pharmnrol., Biorlwm. Belinrl. 5, suppl. 1, 17-21, Miller, L. H., Sandman, C . , and Kastin, A. J. (1977).In "Frontiers of Hormone Research" (F. J. H. Tilders, D. F. Swaab, and T. B. van Wimersma Greidanus, eds.). Vol. 4, pp. 153-161, Karger, Basel. Minneman, K. P., and Iversen, L. L. (1976a).Nutiire (Lrmdon)262, 313-314. Minneman, K. P., and Iversen, L. L. (1976b).I n "Opiates and Endogenoils Opioid Peptides" (H. W. Kosterlitz, ed.), pp. 137-142. North-Holland Aibl., Amsterdam. Moon, H . D., Li, C . H., and Jenning, B. M. (1973).Anut. Rrc. 175, 529-538. Motta, M., Mangili, G., and Martini, L. (1965).Endocrinology 77, 392-395. Murphy, J. V., and Miller, R. E. (1955).J . Comp. Pl?ysiol. Psychol. 48, 47-49. Nakagawa, K., and Kuriyama, K . (1976).J . Nenrochem. 27, 609-612. Nakamura, M., Matsuda, S., Ueda, M.. and Tanaka, A. (1976). E.vpprimliu 32, 3683611. Nicolov, N. (1967).Folia Med. (Ploirdi.cf)9, 249-255. Palo, J., and Savolainen, H. (1974). Brain Re.\. 70, 313-320. Paroli. E. (1967).Arch. Itnl. Sci. F a n c ~ r o l .13, 234-237.

248

W. H. GISPEN, J. M . VAN REE, AND D. DE WIED

Pasternak, G . W., Goodman, R., and Snyder, S. H. (1975). Life Sci. 16, 1765-1769. Pelletier, G., Labrie, F., Kastin, A. J., and Schally, A. V. (1975)Phannacol., Biochem. Behait. 3, 67 1-674. Pert, A. (1976). In “Opiates and Endogenous Opioid Peptides” ( H . W. Kosterlitz, ed.), pp. 87-95. North-Holland Publ., Amsterdam. Pert, C. B., Bowie. D. L., Fong, B. T. W., and Chang, J. K. (1976). In “Opiates and Endogenous Opioid Peptides” (H. W. Kosterlitz, ed.), pp. 79-86. North-Holland Publ., Amsterdam. Perumal, R., Gispen, W. H., Wilson, J. E., a n d Glassman, E. (1975). Prog. Brain Res. 42, 20 1-207. Pfaff, D. W., Teresa, M., Silva, A., and Weiss, J. (1971), Science 172, 394-395. Reading, H . W. (1972). L3iochem.J. 127, 7 P. Reading, H . W., and Dewar, A. J. (1971). Meet. Int. Soc. Neurochem., 3rd, Budupest p. 199. Rees, H. D., Brogan, L. L., Entingh, D. J., Dunn, A. J., Shinkman, P. G., Damstra-Entingh, T., Wilson, J. E., and Glassman, E. (1974). Brain Res. 68, 143-156. Rees, H. D., Dunn, A. J., a n d luvone, P. M. (1976).L f e Sci. IS, 1333-1340. Reith, M . E. A. (1975). Ph.D. Thesis, Univ. of Utrecht, Utrecht. Reith, M. E. A., Gispen, W. H., and Schotman, P. (1974a).In “Central Nervous System” (E. Genazzani and H. Herken, eds.), pp. 236-240. Springer-Verlag. Berlin a n d New York. Reith, M. E. A., Schotman, P., and Gispen, W. H. (1974b). Brain Res. 81, 571-575. Reith, M . E. A., Schotman, P., a n d Gispen, W. H. (1975a). Nntrosci. Lett. 1, 55-59. Reith, M. E. A,, Schotman, P., and Gispen, W. H. (1975b). Neurobiology 5, 355-368. Reith. M . E. A., Schotman, P., and Gispen, W. H . (1977). In “Mechanisms, Regulation and Special Functions of Protein Synthesis in the Brain” (S. Roberts, A. Lajtha, and W. H. Gispen, eds.). Elsevier, Amsterdam. (In press). Rigter, H., and Popping, A. (1976).Psychophnrmacologia 46, 255-261. Rigter, H., van Riezen, H., and de Wied, D. (1974). Physiol. Behav. 13, 381-388. Rigter, H., Elbertse, R., and van Riezen, H. (1975a). Prog. Brain Res. 42, 163-171. Rigter, H., van Eys, G., and Leonard, B. E. (1975b).Pharmacol.,Biochem. Behav. 3,781-785. Rigter, H., Ramaekers, F., and Leonard, B. E. (1977).Brain Res. (in press). Ronai, A. Z., Szekely, J. I., Graf, L., Dunai-Kovacs, Z., a n d Bajusz, S. (1976). L f e Sci. 19, 733-738. Roques, B. P., Carbay-Jaureguiberry, C., Oberlin, R.. Anteunis, M., and Lala, A. K. (1976). Nature (London) 262, 778-779. Routtenberg, A., Ehrlich, Y. H., and Rabjohns, R. R. (1975). Fed. Proc., Fed. Am. SOC.Exp. B i d . 34, 17. Rudman, D., and Isaacs, J. W. (1975). Endocrinology 97, 1476-1480. Rudman, D., Scott, J. W., Del Rio, A. E., Houser, D. H., and Sheen, S. (1974).Am.J. Physiol. 226,687-692. Sakamoto, A. (I966).Nntiire (London) 211, 1370-1371. Sandman, C. A., Kastin, A. J., and Schally, A. V. (1969).Exfierientia25, 1001-1002. Sandman, C. A,, Kastin, A. J.. and Schally, A . V. (1971). Physiol. Behav. 6,45-48. Sandman, C. A., George, J. M., Nolan, J. D., van Riezen, H., and Kastin, A. J. (1975). Phy.siol. Behail. 15, 427-43 1.‘ Sandman, C. A., George, J., Walker, B., Nolan, J. D., and Kastin, A. J. (1976). Pharmacol., Biochem. Behav. 5, suppl. 1 , 23-28. Satoh, M . , Zieglgansberger, W., and Herr, A. (1976). Brain Res. 115, 99-1 10. Sawyer, C. H., Kawakami, M., Meyerson, B., Whitmoyer, D. I . , and Lilley, J. I. (1968). Brain Res. 10, 213-226. Schotman, P., Gispen, W. H., Jansz, H . S., and de Wied, D. (1972).BraznRes. 46,349-362. Schotman, P., Gipon, L., and Gispen, W. H. (1974).Brain Res. 70, 377-380. Schotman, P., Reith, M. E. A., van Wimersma Greidanus, T . B., Gispen, W. H., and de

LIPOTROPIN AND THE CENTRAL NERVOUS SYSTEM

249

Wied, D. (1976).1n “Molecular and Functional Neurobiology” (W. H . Gispen, ed.), pp. 309-344. Elsevier, Amsterdam. Schotman, P., Reith, M. E. A., and Gispen, W. H. (1977a).Brain Res. Bull. 2, 83-87. Schotman, P., Gipon, L., Jennekens, F. G. I . , and Gispen, W. H. (1977b).Neuropat/zo/. Appl. Neurobiol. 3, 125-136. Segawa, T., Kawakasi, K., a n d Yajima, H. (1973).Jpn.J. Pliannacol. 23, 121-123. Selye, H. (1936). Br. J . Exp. Pathol. 17, 234-248. Semiginovsky, B., a n d Jakoubek, B. (1971). Brain Rex 35, 319-323. Sharma, S. H., Klee, W. A., and Nirenberg, M. (1975a). Proc. Natl. Acad. Sci. U.S.A. 72, 3092-3096. Sharma, S. H., Nirenberg, M., and Klee, W. A. (1975b). Proc. Natl. Acad. Sci. U.S.A. 72, 590-594. Simantov, R., a n d Snyder, S. H. (1976). Life Sci. 18, 781-788. Snyder, S. (1975). Nature (London)257, 185-189. Spirtes, M . A., Kostrzewa, R. M., and Kastin, A. J. (1975). P h a m c o l . , Biochtm. Behav. 3, 1011-1015. Steiner, F. A. (1970). Prog. Bruin Res. 32, 102-107. Steiner, F. A,, Ruf, K., a n d Akert, K. (1969). Brain Res. 12, 74-85. Stern, W. C., Miller, M., Forbes, W. B., Leahy, J. P., Morgane, P.J., and Resnick, 0.( 1 976). Brain Res. Bull. I, 27-31. Strand, F. L., and Cayer, A. (1975).Prog. Brain Res. 42, 102-107. Strand, F. L., Stobay, H., a n d Cayer, A. (1973-1974). Neuroendocrinology 13, 1-20. Strand, F. L., Stoboy, H., Freidebold, W. A.. Krivov, W. A., Heyck, H., and van Riezen, H . (1975).Lamet 11, 919. Stratton, L. 0..a n d Kastin, A. J. (1974). H u m . Behail. 5, 149-155. Sutherland, E. W. (1972). Science 177,401-408. Terenius, L. (1973).Artu Phannarol. Toxicol. 32, 317-320. Terenius, L. (1975).J. P h a m . Pharmacol. 27, 450-452. Terenius, L. (1976). Eu7.J. Pliannacol. 38, 21 1-213. Terenius, L., and Wahlstrom, A. (1975). Acta Physiol. Scand. 94, 74. Terenius, L., Gispen, W. H., and d e Wied, D. (1975). Eur. J. Phannacol. 33, 395-399. Terenius, L., Wahlstrom, A., Lindstrom, L., and Widerlov, E. (1976). Nmrosci. Lett. 3, 157-162. Torda, C., a n d Wolf, H. G. (1952).Am. J . Physiol. 169, 140-149. Traber, J., Fischer, K., Latzin, S., and Hamprecht, B. (1975a). Nature (London) 253, 120122. Traber, J., Gullis, R., and Hamprecht, B. (1975b).Lye Sci. 16, 1863-1868. Tseng, L. F., Loh, H. H., and Li, C. H. (1976).Nature (London)263, 239-240. Ueda, M., Matsuda, S., Kawakami, M., and Takeda, H. (1970).Jpn.J. Pharmucol. 20,585. Urban, I . L., and d e Wied, D. (1975). Brain Rer. 97, 362-366. Urban, I . L., a n d d e Wied, D. (1976).Exp. Brain Res. 24, 325-334. Urban, 1. L., Lopes d a Silva, F. H., Storm van Leeuwen, W., and d e Wied, D. (1974).Brain Res. 69, 361-365. van Delft, A. M. L.. and Kitay, J. 1 . (1972).Neuromdocrinology 9, 188-196. van Loon, R., and Mascardo, R. N. (1975).Proc. Annzi. Meet. SOC.Neurosci., 5th p. 444. van Ree, J. M., a n d de Wied, D. (1976). In “Opiates and Endogenous Opioid Peptides” (H. W. Kosterlitz, ed.), pp. 443-445. North-Holland Publ., Amsterdam. van Ree, J. M., de Wied, D., Bradbury, A. F., Hulme, E. C., Smyth, D. G., and Snell, C. R. (1976). Nature (London)264, 792-794. van Ree, J. M., Versteeg, D. H . G., Spaapen-Kok, W. B., and d e Wied. D. (1977a).Neuroendocrinology (in press). van Ree, J. M., Spaapen-Kok, W. B., a n d d e Wied, D. (1977b).Nei/roendorrinology (in press).

250

W. H. GISPEN, J. M. VAN REE, AND D . DE WIED

van Wimersma Greidanus, T. B. (1970). Prog. Brain Res. 32, 185-191. van Wimersma Greidanus, T. B. (1977). I n “Frontiers in Hormone Research” (F. J. H . Tilders, D. F. Swaab, and T. B. van Wimersma Greidanus, eds.), Vol. 4, pp. 129-139, Karger, Basel. van Wimersma Greidanus, T. B., and d e Wied, D. (1971). Neuroendocrinology 7, 291-301. van Wimersma Greidanus, T. B., and d e Wied, D. (1976). Pharmacol., Bioclwm. Behav. 5 , suppl. 1, 29-33. van, Wimersma Greidanus, T. B., Bohus, B., and d e Wied, D. (1974).Nezrroendocrinology 14, 280-288. van Wimersma Greidanus, T. B., Dogterom, J., and d e Wied, D. (1975a). Life Sci. 16, 637444. van Wimersma Greidanus, T. B., Bohus, B., and de Wied, D. (1975b). Prog. Bruin Res. 42, 135- 141. van Wimersma Greidanus, T. B., Bohus, B., and d e Wied, D. (1975~).In “Anatomical Neuroendocrinology” (W. E. Stumpf and L. D. Grant, eds.), pp. 284-289. Karger, Basel. van Wimersma Greidanus, T. B., Lowry, P. J., Scott, A. P., Rees, L. H., and d e Wied, D. (1975d). Horm. Behav. 6, 319-327. Verhoef, J., and Witter, A. (1976). Pharmacol., Biochem. Behav. 4, 583-590. Verhoef, J., Palkovits, M., and Witter, A. (1977a). Brain Res. 126, 89-104. Verhoef, J., Witter, A., and de Wied, D. (1977b). Brain Res. (in press). Versteeg, D. H. G. (1973). Brain Res. 49,483485. Versteeg, D. H. G., and Wurtman, R. J. (1975). Brain Res. 93, 552-557. Versteeg, D. H. G., and Wurtman, R. J. (1976). In “Molecular and Functional Neurobiology” (W. H. Gispen, ed.), pp. 201-234. Elsevier, Amsterdam. Versteeg, D. H. G., Gispen, W. H., Schotman, P., Witter, A., and d e Wied, D. (1972). Adv. Biochem. Psychopharmacol. 6, 219-239. von Hungen, K., and Roberts, S. (1973). Eur.J. Biochem. 36, 391-401. Waterfield, A. A., Hughes, J., and Kosterlitz, H. W. (1976).Nature (London) 260,624-625. Wei, E., and Loh, H. (1976a). In “Opiates and Endogenous Opioid Peptides” (H. W. Kosterlitz, ed.), pp. 303-3 10. North-Holland Publ., Amsterdam. Wei, E., and Loh, H. (1976b). Science 193, 1262-1263. Weiss, J. M., McEwen, B. S., Silva, M. T., and Kalkut, M. (1970). Am. J . Physiol. 218, 864-868. Wiegant, V. M., and Gispen, W. H . (1975). E x f . Bruin Res. 23, Suppl., 219. Wiegant, V. M., and Gispen, W. H. (1977). Behav. Bwl. 19, 554-558. Wiegant, V. M., Cools, A. R., and Gispen, W. H. (1977a).Eur.J. Pharmucol. 41,343-345. Wiegant, V. M., Gispen, W. H., Terenius, L., and d e Wied, D. (1977b). Psychoneuroadocrinology 2, 63-69. Winter, C. A., and Flataker, L. (1951).J. I’hannacol. Exp. Ther. 101, 93-105. Witter, A., Greven, H. M., and d e Wied, D. (1975).J. Pharmacol. Exp. Ther. 193,853-860. Wolstencraft, J. H. (1964).J. Physiol. (Londm) 174, 91-108. Wolthuis, 0. L., and d e Wied, D. (1976). Pharmacol., Bioclwm. Behuii. 4, 273-278. Yamashiro, D., and Li, C. H. (1974). Proc. Natl. A c d . Sci. U.S.A. 71,4945-4949. Zieglgansberger. W., and Fry, J. P. (1976).In “Opiates and Endogenous Opioid Peptides” (H. W. Kosterlitz, ed.), pp. 23 1-246. North-Holland Publ., Amsterdam. Zieglgansberger, W., Fry, J. P., Herz, A,, Moroder, L., and Wunsch, E. (1976). Brain Res. 115, 160-164. Zimmermann, E., and Krivoy, W. (1973). Prog. Brain Res. 39, 383-394. Zimmermann, E., Branch, B., Taylor, A,, Young, J., and Pang, C. (1974). Fed. Proc., Fed. Am. Sac. Exp. Biol. 33, 293. Zwiers, H., Veldhuis, D., Schotman, P., and Gispen, W. H. (1976).Nmrochm. Res. 1,669-677.

TISSUE FRACTIONATION IN NEUROBIOCHEMISTRY: A N ANALYTICAL TOOL OR A SOURCE OF ARTIFACTS By Pierre Laduron Department of Biochemical Pharmacology Janrren Phormaceutica Beerre, Belgium

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 1 I I . Analytical Approach to Tissue Fractionation in the Brain . . . 253 A. Regional Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 254 B. Tissue Fractionation . 262262 C. Subfractionation by Centrifuging in Density Gradient ................... D. Application of Tissue Fractionation to the Subcell~ilarLocalization of 266 D-Amino Oxidase in the Rat Brain .................................... 269 I l l . Interpretation of Tissiie Fractionation Studies .................... 269 A. Biochemical Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 B. Morphological Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Synaptosomes and Synaptic Vesicles . . . . . . . . 276 280 1V. Conclrision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction

Most of the fundamental discoveries in biochemistry and, also, in molecular biology have been made possible through the introduction of cell fractionation techniques; indeed the study of a given enzyme or the study of nucleic acid metabolism cannot be carried out without fractionating cells. Therefore the biochemist begins his work, as often as not, by homogenizing something. However, from this starting point, two possible ways are open to him depending on whether the tissue fractionation is used as a preparative tool or an analytical one. As considerable confusion about this fundamental distinction still persists throughout the literature, I would like to clarify this point. In order to study the physicochemical or kinetic properties of a given enzyme, different preparations may be used: a total homogenate, a particulate fraction (nuclear, mitochondrial, or microsomal), a supernatant, or a purified enzyme. Here, fractionation is used as a preparative tool; 25 1

252

PIERRE LADURON

indeed it is only required in order to obtain an enzyme preparation (purified or not) which possesses sufficient activity to be tested under various experimental conditions. In this case, a fraction of the total homogenate (nuclear, mitochondrial, supernatant, etc.) may be discarded without impairing the interpretation of kinetic results. Therefore the loss of enzymatic activity is here of no consequence. In contrast to this, if one intends to determine the subcellular localization of this enzyme or of any other component, all the fractions must be analyzed in order to obtain a complete balance sheet. Throughout this analytical approach, nothing is discarded. Thus, a distribution pattern of a given component can be obtained from the analysis of all the subcellular fractions obtained. Moreover, numerous marker enzymes must also be determined in order to assess the subcellular composition of a given fraction. In many reports, the preparative approach has been followed but the conclusions were drawn as though the fractionation had been used as an analytical tool. A typical example is the study of brain synaptosomes. Synaptosomal fractions have been commonly used to study, in nerve terminals, processes like amine uptake, respiration, glycolysis, or protein synthesis. In fact, the analytical approach clearly shows that these preparations are always contaminated by other subcellular structures, which considerably restricts the interpretation of results. Nevertheless, the synaptosome alone was made responsible for one of these processes, as though a contaminant was automatically of no consequence. As discussed later, in some cases, the contaminant of a so-called purified fraction has been responsible for a phenomenon erroneously attributed to the main component of this fraction. For the analytical approach, numerous criteria, mostly biochemical ones, must be applied to assess the composition of a given fraction. Sometimes more refined techniques, such as subfractionation, are required for attaining this aim. From the emphasis placed on this essential distinction between the preparative and analytical approach in tissue fractionation, the reader will have understood that only the latter will be considered in this chapter. Brain fractionation has nearly always been performed according to the procedure originally developed by Whittaker et al. (1964) and De Robertis (1964). This method, essentially a preparative approach, has been extremely useful in developing the concept of synaptosome and synaptic vesicle which is the basis of neurotransmission in the central nervous system (CNS). However, it is now becoming more and more evident that such a methodology is not refined and analytical enough for the study of complex subcellular localizations like those of receptors, certain membrane- or particle-bound enzymes, and drugs. Therefore, I

TISSUE FRACTIONATION IN NEUROBIOCHEMISTRY

253

believe that the neurobiochemist should look back at what was achieved many years ago in connection with the liver. The awareness of the importance of this work immediately convinces one that neither form of tissue fractionation should exceed the limits of the work that pioneers previously performed in exploring liver cells. I t is the great merit of de Duve and his group to have developed and proposed a critical approach to tissue fractionation. My sole purpose in this chapter is to show that such a methodology can also be applied successfully to the study of brain tissue. Hence, as an introduction to this chapter, the reader is referred to several important reviews in which all the principles of tissue fractionation have been discussed in detail (de Duve, 1967, 1971, 1975; Beaufay, 1972). It is beyond doubt that one of the most remarkable contributions of this group of researchers is to have developed a fractionation procedure which is now classically known as the five-fraction procedure (de Duve et al., 1955). I t is very surprising that, whereas for 20 years this fractionation by differential centrifugation has been successfully applied to practically all the peripheral organs, it has very rarely been used for the brain. Even in the field of terminology, there is a marked difference between that of biochemists of the CNS and the others. At the present time, one uses the classical abbreviations to name the fractions, N, M, L, P, and S, while PI, Pt, and S or P I , Pz, PJ, and S are commonly used by neurobiochemists. In order to obtain more mutual understanding between different groups of researchers using more or less the same methodology it will be necessary to review not only the technical procedure but also the terminology. The comparison of brain fractionation presented according to the three different methods already mentioned will show the considerable advantage of using the five-fraction procedure. Brain fractionation needs to become more and more critical and analytical in order to solve the problems posed by the subcellular localization of receptors in the brain. Indeed after the characterization of a stereospecific binding which does not require the use of purified preparations, the second step is to determine their precise localization in the brain mainly at the subcellular level (Leysen and Laduron, 1977; Laduron and Leysen, 1977). Similarly, a better understanding of the mechanism of action of centrally acting drugs requires the same approach.

II. Analytical Approach to Tissue Fractionation in the Brain

Although peripheral tissues, like the liver or kidney, are extremely heterogeneous with respect to the different cell types, even within each organ, they do not possess, as the brain does, specific areas in which a

254

PIERRE LADURON

given neurotransmitter or a given enzyme appears to be much more concentrated than elsewhere, sometimes even to the exclusion of any other site. For instance, dopamine is more specifically recovered in the striatum and limbic system or d-amino acid oxidase in the medulla oblongata and cerebellum, while noradrenaline is the main catecholamine in the hypothalamus. Because of this extreme diversity in brain areas or nuclei, it should be theoretically possible to fractionate each of these structures separately. However, in numerous cases, the material is too small for subcellular studies to be performed. Nevertheless, it seems advisable, where possible, to select a given brain area before starting fractionation studies. Hence in most cases, the approach will be as follows: (1) regional distribution and selection of a given area: (2) fractionation by differential centrifugation; (3) subfractionation by equilibration in density gradients. A. REGIONAL DISTRIBUTION Tissue fractionation in the brain starts with this regional distribution whereby a given component, neurotransmitter, enzyme, labeled precursor, or drug will be estimated in different regions of the brain. For this purpose, dissection can be performed according to the classical procedure of Glowinski and Iversen (1966) or by a more refined or sophisticated method. In these experiments, the enzyme activity or any other activity must be estimated from a total homogenate. Nevertheless it is sometimes desirable to measure the enzyme activity after low-speed centrifugation, especially when the nuclear fraction contains endogenous inhibitors of the enzyme.

B. TISSUE FRACTIONATION 1. Homogenization It would appear perhaps somewhat superfluous to specify how a tissue must be homogenized, because this is a technique commonly used by all neurobiochemists. It is, however, useful to recall certain fundamental principles. Homogenizers with high shearing forces such as Ultraturrax or the Waring Blender, must be excluded definitively because they cause too much damage to the subcellular structures. In most cases, a homogenizer with a conical Teflon pestle, like the Duall, appears to be the most adequate. Clearance between pestle and tube of 0.25 mm is almost ideal. However, this clearance increases with the age of the homogenizer. We

TISSUE FRACTIONATION IN NEUROBIOCHEMISTRY

255

have often observed that the disruption of nerve terminals was much more pronounced when using a new homogenizer than a few months later, especially if the device was used to homogenize tissue with a hard framework like bovine adrenal medullae. Hence, we recommend reserving one homogenizer for brain fractionation only. Homogenization of brain tissue has to be performed as gently as possible. This point remains particularly vulnerable since “gentleness” is rather a subjective concept. As pointed out by de Duve (1967), “homogenization is still largely an art.” The main difficulty lies in the choice of a convenient rotating speed or shearing force able to break down the cells without causing too much damage to the subcellular particles. Another difficulty in the CNS concerns the presence of nerve endings or synaptosomes. This structure, which is not a subcellular particle but part of a neuron, seems to be more resistant to breakage than cell membranes. Owing to the presence of this entity in the homogenate one can say that homogenization of brain tissue, in a sense, always remains incomplete. In some cases, it may be useful to homogenize by means of a greater shearing force, thus more violently, in order to liberate the different components and subcellular structures trapped within nerve endings. As the success of fractionation is largely dependent on the homogenization process, which as a starting process already appears somewhat empirical, special care must be devoted to evaluating the quality of the homogenization. Morphological criteria may sometimes be useful although this approach is often very limited. In fact as stressed by de Duve “the most useful instrument for estimating the quality of a homogenate is the centrifuge.” Indeed, it is the distribution pattern of various marker enzymes that make it possible to measure to what extent the cells are disrupted or the subcellular organelles are damaged. For instance the fact that an enzyme which is normally particle-bound is recovered in excess in the supernatant will indicate excessive damage to the subcellular structures. Conversely if an enzyme normally located in the supernatant is present in large amounts in the nuclear fraction, an insufficient disruption of cells is indicated. I t is generally believed that excessive disruption is usually caused by allowing the pestle to rotate too fast. This is partially true. On the other hand, by using a low speed, the manner of handling the homogenizer may also cause a greater or lesser degree of damage to the particles. Indeed, trying to overcome the resistance of the tissue by pushing the pestle too violently to the bottom of the homogenizer or by milling the homogenate up and down too violently may lead to even more pronounced disruption. Certain tissues such as bovine adrenal medulla are particularly difficult to homogenize presum-

256

PIERRE LADURON

ably owing to their resistant framework of connective tissue. I n contrast to this, rat adrenals are much softer, to the extent that they may be considered the adrenal of choice for tissue fractionation studies (Laduron et al., 1976a). Indeed in contrast to what happens in bovine or pig adrenal medulla, the chromaffin granules in rat adrenal homogenates are practically undamaged by the homogenization process, as is shown by the very low catecholamine content in the supernatant fraction.

2 . Choice of Medium Sucrose solutions are mostly used in tissue fractionation experiments. Although in the peripheral tissues the concentration in sucrose is always 0.25 M, a slightly hypertonic solution (0.31 M) is generally used to homogenize brain tissue (Whittaker, 1969). As far as I know, no explanation for this difference has ever been provided. When using biochemical criteria to evaluate the quality of the homogenate, we were unable to find any reason for preferring sucrose solutions of 0.31 M (Laduron, unpublished results). Therefore for the purpose of standardizing somewhat the fractionation techniques, it is preferable to homogenize brain tissue in 0.25 M sucrose. Nevertheless, sucrose itself is not always appropriate, namely for determining the intracellular localization of “soluble” enzymes. A good example is tyrosine hydroxylase: A long-standing controversy about its subcellular localization has been going on since the moment of its discovery in 1964 until recently. Laduron and Belpaire (1968) have shown that the enzyme is exclusively recovered in the supernatant when isotonic KCl(0.16 M ) is used as medium for homogenizing bovine adrenal medullae. In contrast to this a more complex distribution was obtained using isotonic sucrose. This is illustrated in Fig. 1. It is noteworthy that isotonic KCl solution can only desorb tyrosine hydroxylase from various structures to which it was nonspecifically bound. When subjected to the same conditions, inosine diphosphatase, a membrane-bound enzyme or dopamine-/3-hydroxylase, an enzyme contained within the chromaffin granules, remained linked to their respective structures (Fig. 1). Unfortunately, in the brain, isotonic KCl solution causes a severe agglutination phenomenon whereby practically all the subcellular particles sediment together in the nuclear fraction even when brain tissue is homogenized in very large volumes. Recent experiments have shown that sucrose solutions might be buffered either with bicarbonate 0.05 M , Tris-HC10.01 M, or imidazol HClO.01 M, p H 7.4, without altering the sedimentation properties of the particles. In some cases, it may also useful to add ethylenediaminetetraacetic acid (EDTA) or (EGTA) in order to prevent agglutination of subcellular particles or eniyme inactivation (Tulkens et d.,1974).

TISSUE FRACTIONATION

S w

A 31

IN NEUROBIOCHEMISTRY

257

B TYROSINE HYDROXYLASE

z

u W

a0 ? z

C W

au 1

da

U W

‘LD_O

A possible but perhaps theoretical advantage of using buffered sucrose medium at a different pH is that such a procedure might prevent further dissociation of a drug from its receptor site when previously injected into the animals. Indeed any study of the subcellular localization of a given drug in the brain implies that its specific binding occurring in viuo may not be dissociated during the homogenization process. There is no doubt that a problem related to subcellular fractionation sometimes calls for a specific solution. In this respect the choice of the medium may often be critical. 3 . Fractionation by Dz$erential Centrifugation: The Five-Fraction Scheme When a tissue has been homogenized, the first step in studying the subcellular distribution of a given element consists of fractionating the total homogenate by differential centrifugation. This approach yields considerable information and even, in some cases, appears to be quite sufficient to determine the exact intracellular compartment to which a

258

PIERRE LADURON

given component belongs. It also represents a compulsory step before performing centrifugation in a density gradient. Although the five-fraction procedure was introduced 20 years ago by de Duve et al. (1955) as a model for the fractionation of rat liver, with the exception of a very few reports (Sellinger et al., 1964; Laduron et al., 1975; Leysen and Laduron, 1977; Laduron and Leysen, 1977) this procedure has virtually never been used for subcellular studies in the brain. It is often argued that such a procedure is convenient only in the liver, whereas it has been successfully adopted for many other tissues and cells like the kidney, intestine, thyroid, adrenal medulla, bone, lymphoid tissue, spleen, fibroblasts, HeLa cells, neutrophil polymorphonuclear leukocytes, placental tissue, aortic smooth muscle, enterocytes, and protozoa (cf. de Duve, 1975). I would like to show how much more profitable it would be for the neurobiochemist to follow this procedure rather than that developed by Whittaker (1969). In order to dispel any suspicion that I am affiliated to one school rather than to another merely for subjective reasons, I append Fig. 2 to illustrate the differences in the distribution patterns of some marker enzymes in the brain using three different methods. The first procedure was proposed by Gray and Whittaker (1962). It is the scheme with three fractions which, according to their terminology, are the P I fraction (nuclei, large myelin fragments, and tissue debris), the P p fraction (mitochrondria, synaptosomes, and some microsomes), and the S, (microsomes, some small mitochondria, and synaptosomes). The second method is an adaptation of the first one in the sense that the microsomes P3 have been separated from the cell sap S (Whittaker, 1969). Finally, the scheme with five fractions according to de Duve et al. ( 1 955) which will be described in detail in the next paragraph makes it possible to obtain two different mitochondria1 fractions. Figure 2 is constructed according to the manner proposed by de Duve et al. (1955);the relative specific activity (RSA) which is the ratio of the percentage of total activity to the percentage of total proteins is plotted against the percentage of total proteins. It should be noted that if the homogenates were not fractionated the percentages of enzyme activity and total proteins would both be loo%, so that the RSA would have the value of 1. Hence, an RSA higher than 1 means that a given component is enriched in a given fraction with a concomitant RSA lower than 1 in one or several other fractions. Awareness of this fact is of prime importance in the understanding of the distribution pattern of an element after differential centrifugation. As shown in Fig. 2, the procedure in which three fractions were obtained from rat brain homogenate does not enable one to differentiate the subcellular localization of cytochrome

T I S S U E FRACTIONATION

I N NEUROBIOCHEMISTRY

B

A

a

259

C

P-GALACTOSIDASE

,

0

N-ACETYL P-D-GLUCOSAMINIDASE

4

5

INOSINE DIPHOSPHATASE

- v

!

%$I

5'-NUCLEOTIDASE I

I

0

20

LO

60

80

100

0

20

LO

60

80 100

0

20

LO

60

80

100

PERCENTAGE OF TOTAL RECOVERED PROTEINS

FIG.2 . Distribution pattern of five marker enzymes in rat brain after lractionation by differential centrifugation according to: A , the three-traction scheme of Gray and Wliittaker (1962): B, the four-fraction scheme of Whittaker (1969): and C, the five-fraction scheme of d e Duve et al. (195.5). adapted for brain tissue by Laduron et al. 1975).

oxidase, a mitochondrial enzyme from the two lysosomal enzymes, /3-galactosidase and N-acetyl-P-D-glucosaminidase. Moreover, these results indicate that inosine diphosphatase should have a nuclear localization. More interesting is the distribution pattern of 5'-nucleotidase which does not appear to be enriched anywhere just as i f no fractionation had been carried out. Using the four-fraction procedure gives a greater enrichment of both microsomal enzymes without marked effect on the lysosomal ones. On the other hand, the considerable advantage of the five-fraction procedure is immediately apparent since not only the mitochondrial and lysosomal enzymes exhibit different profiles but, interestingly, the two microsomal enzymes differ as to their distribution patterns. This latter point suggests that as in the liver, they do not share

260

PIERRE LADURON

the same microsomal structure, 5'-nucleotidase being a typical plasma membrane enzyme and inosine diphosphatase being associated with the endoplasmic reticulum (Beau fay et al., 1974). Therefore only the more refined procedure makes it possible to obtain an enrichment of lysosomal enzymes in the L-fraction. The same results were reported using rat brain striatum (Laduron et al., 1975). However, as shown in Fig. 5, enrichment in the L-fraction is not the only attribute of lysosomal enzymes since D-amino acid oxidase, a peroxisomal enzyme, behaves in a similar manner (cf. Fig. 5). In conclusion, as the five-fraction procedure must be considered the fractionation method par excellence in practically all cases, there is no doubt that its use for the purpose of brain studies could yield a considerable amount of information about the intracellular distribution of elements as different as enzymes, neurotransmitters, drugs, and receptors. Moreover, it can easily be performed with whole brain or with various brain areas which are more or less specific as regards to one of these elements.

4. Technical Procedure for Obtaining Five Fractionsfrom Brain Tissue As already discussed, the homogenization of brain samples may be performed in 0.25 M sucrose as there are no reasons to justify the preferential use of a sucrose concentration of 0.31 M . T h e procedure is schematically presented in Fig. 3. I would like to underline certain points which require particular attention. Homogenization must be performed in a relatively large volume of medium in order to prevent, as far as possible, adsorption phenomenon or coprecipitation. Hence, I recommend using a dilution 1 : 10 ( 1 gm of tissue in 10 vol) for the first homogenization. After centrifuging at low speed, the pellet must be homogenized again but in 5 vol. Biochemical analysis of fractions obtained by differential centrifugation shows that this second homogenization is necessary not only to wash the nuclear fraction but also to further homogenize the unbroken cells. This may be easily estimated by measuring the activity of an enzyme normally present in the cell sap; if homogenization is not sufficient a large amount of the enzyme activity will appear in the nuclear fraction. Both supernatants obtained by centrifuging at low speed are pooled and represent the cytoplasmic extract or fraction E. A sample (generally 1/10) from this fraction is kept for the purpose of determining the recovery of the enzyme activities of proteins and of other components (this point will be discussed in detail in Section HI). Fraction E is further centrifuged to yield the fraction M and L. Both particulate fractions are washed once by suspending the pellet in 5 vol of

TISSUE FRACTlONATlON IN NEUROBIOCHEMISTRY

26 1

BRAIN TISSUE Homogenized in 0 ml 02SM sucrose

m 5 W I D 25M

Supwnalanl 310000g-min

Supernalant

FIG. 3. Scheme of fractionation in brain tissue.

0.25 M sucrose and then centrifuging again. Note that the washing solutions are never discarded but combined with the first supernatant, thus incorporated in the following fractions. ‘The most delicate point in this fractionation is the separation of both mitochondria1 fractions (M and L) (de Duve et al., 1955; Laduron et al., 1975). When the M fraction is obtained for the first time, the supernatant is easy to remove owing to the well-packed pellet (all the supernatants must be removed by suction with an adequate pipette but not by decantation). The second pellet thus obtained after washing is formed of two layers, one dark and relatively fixed on the bottom of the tube and another above it, white and less fixed, the latter generally being called the “fluffy layer.” In fact this fluffy layer must be partially removed without disturbing the well-packed sediment. A situation more or less identical also occurs with the L fraction. The P fraction, which appears as a com-

262

PIERRE LADURON

pact sediment, is easier, so that in the most cases it is not necessary to wash this fraction, especially as the sample is already relatively diluted (1:25). All the fractionation procedures must be carried out at 0°C and the different fractions are kept at - 10°C until enzyme assays are performed.

C . SUBFRACTIONATION BY CENTRIFUGING I N DENSITY GRADIENT Besides the five-fraction procedure which separates particles or groups of particles sedimenting at different rates, other procedures may be successfully used for exploring the nerve cell. One of these is the equilibration in density gradient or isopycnic centrifugation which separates particles equilibrating at different densities. This subfractionation procedure is often needed to determine the exact subcellular localization of a given component. Nevertheless, differential centrifugation must be considered as the first approach. Indeed, if the element to study is entirely recovered in the supernatant fraction, further subfractionation is not necessarily required. On the other hand, if the component is particle-bound but not obviously associated with an organelle, but not clearly characterized after differential centrifugation, equilibration in a density gradient may be useful to locate the subcellular particles to which it belongs. In certain cases only the use of both procedures may solve the problem. I would like to illustrate this consideration by examination of the subcellular distribution of dopamine-sensitive adenylate cyclase in the rat brain striatum. Indeed, this enzyme was found to have the same equilibrium density as Na+/K+Mg2+adenosine triphosphatase (ATPase) in gradients of sucrose suggesting that eventually both enzymes might be associated with the same subcellular structures (Laduron et al., 1976b). However, the results obtained by differential centrifugation made it quite unlikely, since the ATPase was especially enriched in the microsoma1 fraction and the dopamine-sensitive adenylate cyclase in the mitochondria1 fractions. Hence these two different techniques were essentially complementary, which is not surprising since one is based on the sedimentation coefficient and the other on the density equilibrium. The density-gradient centrifugation experiments, in which the material is subjected to incomplete sedimentation in a stabilizing density gradient (thus an analytical procedure also based on the sedimentation rate) (Beaufay, 1966), will not be considered here since they have not yet been applied in brain studies. Several types of density gradients can be used. They are either discontinuous or continuous. For analytical purposes, the first one must definitively be abandoned. As stressed by de Duve (1971) “the discon-

T I S S U E FRACTIONATION IN NEUROBIOCHEMISTRY

263

tinuous gradient is essentially a device for generating artificial bands. . . it is also a very dangerous procedure in that it creates the illusion of clear-cut separation.” Unfortunately, despite this, many authors persist in using this type of gradient in subcellular distribution studies. In order to prepare the gradient, several different solutes can be used. Sucrose gradient is beyond doubt the most common but Ficoll, colloidal silica, or diatrizoate have been also used (Abdel-Latif, 1966; Lagercrantz and Pertoft, 1972; Tamir et nl., 1974). Certain devices were constructed to produce a variety of S-shaped gradients (Beaufay, 1966). In our laboratory, linear gradients are currently prepared using a very simple device consisting of two syringes which are connected together and are filled with two different concentrations of sucrose (0.6 M and 1.6 M). Various subcellular fractions may be submitted to equilibration in density gradients. I t is preferable not to layer a total brain homogenate onto sucrose gradients owing to the presence of excessively large elements such as nuclei, unbroken cells, connective tissues, or small vessels which prevent one from obtaining a convenient resolution and which also make it difficult to interpret the results. Therefore a cytoplasmic extract is preferable (Bretz et nl., 1974). Nevertheless, in order to solve certain specific problems, a large mitochondria1 fraction or a microsomal fraction appears to be more suitable. A very important point is the amount of material which must be layered onto the gradient. Indeed, the distribution pattern of a given enzyme may be completely different even leading to an erroneous interpretation of its intracellular localization, when the gradient is loaded with a sample with an excessive concentration of proteins. An example of this process is illustrated in Fig. 4. Indeed, it was recently reported that when an M L fraction of bovine adrenal medulla was further diluted, the ATPase activity was found to shift from an area of the gradient consisting of chromaffin granules to a region containing microsomal fractions (Laduron et al., 1976a). In explanation of this phenomenon, we assumed that an adsorption phenomenon occurred, so that in the presence of a higher concentration of particles the ATPase presumably linked to plasma membranes might adhere to chromaffin granules, thus giving an artificial distribution pattern of this enzyme in the sucrosedensity gradient. Another phenomenon occurring under certain conditions, which has been discussed in detail by Beaufay (1966), is the drop sedimentation. Unless the sample has been sufficiently diluted, the presence of microsomes or of supernatant may sometimes represent a major drawback in interpreting the results obtained by density-gradient techniques. Recent

+

264

PIERRE LADURON

W

a

0 2 5 10

12

IL

16 18 2.0 SUCROSE MOLARITY

FIG. 4. Subfractionation of' an M + L fraction from bovine adrenal medullae after equilibration through a sucrose gradient: in A, 1 ml of dilution 1 : 2 was layered on the gradient; B, 1 mi of' dilution 1 : 4: and C, 1 ml of dilution 1 : 10, ATPase (0-O), cytochrome oxidase ( x - - - X ) , and catecholamine (A-A).

experiments in our laboratory seem to indicate that it might be useful to incorporate buffer solutions in the sucrose solutions when producing the gradient; for instance imidazole buffer 0.01 M, bicarbonate 0.01 M, or tris-HC1 0.01 M, pH 7.6. This did not markedly alter the equilibrium

TISSUE FRACTIONATION IN NEUROBIOCHEMISTRY

265

density of particles. In contrast to this, the addition of KCl through sucrose gradients considerably modified the behavior of particles so that lactate dehydrogenase (LDH), cytochrome oxidase, and Na+/K+Mg'+ ATPase were found to have the same equilibrium density, which is, however, not the case under normal conditions. It was previously believed that centrifuging through density gradients did not cause major alterations to the organelles. Since the introduction of the swinging bucket rotor, able to rotate at 65,000 rpm, Wattiaux et al. (1971) have shown that the mitochondria of rat liver suffered severe damage in sucrose gradients when submitted to higher hydrostatic pressure. Two main factors are responsible for this phenomenon: first, the gravitational force which, in turn, depends not only on the speed, but also on the radius of the axis of the rotor and on the height of the gradient column. For instance, as shown by Wattiaux et al. (197l ) , alterations to mitochondria were already detectable at a speed of 20,000 rpm when using a SW-40 rotor, while in a SW-65 rotor, they occurred only at about 45,000 rpm. The main damage is an increase in the equilibrium density for some mitochondria, while by further increasing the hydrostatic pressure, a large proportion of mitochondrial enzymes were recovered in a zone of lower density. Hence three peaks of mitochondria have been observed under these experimental conditions. This kind of mitochondrial deterioration can be prevented by centrifuging at a higher temperature (15°C) or by adding imipramine to the sucrose gradient (Wattiaux-De Coninck et al., 1973, 1974). Recently, we observed an identical phenomenon for mitochondria of rat adrenal glands, whereas the chromaffin granules appeared much more resistant to the hydrostatic pressure in the density gradient (cf. Fig. 4). It is possible that such a process might also occur when centrifuging mitochondria of the brain in density gradient. Therefore precautions have to be taken when centrifuging brain extracts. The centrifuge is not only an instrument for separating subcellular fractions, it is also a sensitive analytical tool which has been successfully used to characterize physical properties of organelles. Indeed, the sedimentation coefficient and the diameter of subcellular particles can be estimated from their sedimentation rates in 0.25 M-0.5 M sucrose gradients as reported for mitochondria, lysosomes, and peroxisomes in the rat liver (Beaufay, 1966), and for mitochondria and chromaffin granules in the bovine adrenal medulla (Laduron, 1969). Other physical parameters can be obtained from the analysis of equilibrium densities after equilibration in a gradient of sucrose in water or D20 and in a gradient of glycogen by varying the sucrose concentration. Indeed the physical characteristics of subcellular particles depend on the composi-

2 66

PIERRE LADURON

tion of the suspension medium (Beaufay and Berthet, 1963). In fact, the equilibrium density of an organelle in a sucrose gradient does not correspond truly to its own density since it also depends on the penetration of sucrose into the particle. The use of a gradient of glycogen (which does not penetrate into the subcellular particle) makes it possible to estimate the proper density of the organelle which does not always increase linearly in proportion to the sucrose concentration. For instance, certain physical parameters of chromaffin granules and mitochondria of bovine adrenal medullae have been calculated from results provided by gradient centrifugation, and the significanceof these data has been discussed in detail elsewhere (Laduron, 1969). It is to be hoped that such an approach will be followed in the future for the study of the subcellular fractions of brain tissue. Several attempts have been made to separate subcellular particles of the brain by centrifuging in a zonal rotor (Bretz et al., 1974). It is not clear whether this technique is superior to the conventional swinging bucket owing to the complete lack of comparative data. Nevertheless, in the rat liver it has been shown that the zonal rotor constructed by Beau fay provided a much better resolution of subcellular particles like mitochondria and peroxisomes (Beau fay, 1966).

D. APPLICATION OF TISSUE FRACTIONATION TO THE SUBCELLULAR LOCALIZATION OF D-AMINO OXIDASE IN THE RATBRAIN Recent observations on D-amino oxidase activity in the rat brain provide a good illustration of the three different steps which are required to determine the subcellular localization of an enzyme through an analytical approach. D-amino acid oxidase catalyzes the oxidative deamination of certain D-a-amino acids, but its physiological role remains enigmatic owing to the absence of D-form of amino acid in the brain. In the liver, its intracellular distribution is well known, since the enzyme is associated with the peroxisomes (Baudhuin, 1969). 1. Reg.lonal Distribution

For a long time, D-amino acid oxidase has been reported as being highly concentrated in the cerebellum and the medulla oblongata of the mammalian brain (Neims et al., 1966), although in other species like the dogfish and frog it is distributed fairly evenly throughout the entire brain (Goldstein, 1966). Table I gives the regional distribution of this enzyme in the rat brain. From these results, it is obvious that there are two areas of choice in which to conduct fractionation studies in the brain, the cerebellum and the medulla oblongata.

T I S S U E FRACTIONATION

IN NEUROBIOCHEMISTRY

267

TABLE I REGIONAL D I S T R I S ~ I OOF N D-AMINOACID OXIDASE I N RAT BRAIN Brain area Cerebellum Medulla oblongata Midbrain H ypothalamris Cortex Striatum Tu bercu lu m 01 factorit i m

Enzyme activity (pM/min/gm)

17.7 13.5 0.87 0.85 0

0 0

2. Differential Centrifugation By applying the five-fraction procedure for fractionating medulla oblongata, the D-amino acid oxidase activity was found to be more specifically enriched in the L fraction practically as a lysosomal enzyme (Fig. 5 ) . On the other hand, these experiments also ruled out its classification as mitochondrial or microsomal, since its distribution pattern differed markedly from that of the marker enzymes of these organelles. 3. Density-Gradient Cmtrzfirgation

In order to further determine the intracellular localization of D-amino acid oxidase in the rat brain medulla oblongata, densitygradient experiments were performed. Figure 6 shows the distribution pattern of the enzyme after equilibration in a sucrose gradient. N-acetyl-P-D-glucosaminidase and cytochrome oxidase were found to differ markedly from D-amino acid oxidase, indicating that the latter is neither lysosomal nor mitochondrial. Hence, since it did not appear to be associated with lysosomes one may conclude that there must be a special class of organelles containing D-amino acid oxidase and probably also catalase but in much lower amounts. These results are quite compatible with the presence of peroxisomes in the brain medulla oblongata, which confirms previous studies quoted by de Duve (1973). Using cytochemical methods, Citkowitz and Holtzman ( 1973) have demonstrated the presence of peroxisomelike structures in the adult rat dorsal root ganglion. Nevertheless, such structures seem to be more abundant in Schwann cells and satellite cells than in neurons. The foregoing results thus suggest that D-amino acid oxidase is contained in peroxisomes in the rat brain but do not bring definite information about the physiological role of this enzyme in the brain. However, if it is true that D-amino acid oxidase is capable of oxidizing glycine, one

1

CYTOCHROME OXIDASE

N-ACETYL- 8 - D GLUCOSAMINIDASE

3 > 2 E

> FI

,I

V

a

INOSINE DIPHOSPHATASE

2 0

LL V

W

D-AMINO ACID OXIDASE

a v)

w 3

LACTATE DE HY DROGENASE

2

I-

4 2

W

LT

1

0

M

N

0

20

L P LO

M

N

S

0

LO PERCENTAGE OF TOTAL PROTEINS

60

80

100

20

L P

S

60

80

100

FIG. 3 . Distribution pattern of o-amino acid oxidase and marker enzymes in medulla oblongata of rat brain after fractionation by ditferential centrifugation.

CY TOCHROME

3

I>

2

-

Y

2

W

2

I-

4W rx

1

0.25 0.6

08

1.0

12

1L

16

SUCROSE MOLARITY

F I G .6. Subfractionation of an M + L fraction ofmedrdla oblongata from rat brain after equilibration in a siicrose gradient. NPGase: N-acetyl-P-o-glucosaminidase.

TISSUE FRACTIONATION I N NEUROBIOCHEMISTRY

269

can suspect that it plays a role in the catabolism of this putative neurotransmitter. It is noteworthy that the distribution of endogenous glycine and glycine receptors closely parallel that of D-amino oxidase, except in the cerebellum (cf. Snyder, 1975). 111. Interpretation of Tissue Fractionation Studies

Up to now, electron microscopy examination has been the most widely used diagnostic tool for assessing the subcellular composition in brain fractionation studies. It is probably the reason for the relative paucity of progress made in this particular field. Before continuing it is necessary to make and defend the following criticism: Morphological criteria for evaluating the subcellular composition in brain fractionation are quite insufficient and when used alone, very often lead to erroneous interpretations. Why such an assertion? There are t w o main reasons for it: First, morphological examinations are not quantitative except when morphometric analysis (Weibel, 1963; Baudhuin, 1968) is performed, but it has never been done in brain studies. Second, numerous subcellular particles are not characteristic enough to be recognized with certainty. Indeed the lack of convenient markers to identify brain organelles represents the major drawback in this kind of study, especially when particles containing a neurotransmitter have to be detected. Although mitochondria and synaptosomes can easily be identified, there is still a very large amount of membranes, vesicles, or “pseudo-vesicles” for which there are no criteria of identification. Therefore the first and most important approach in assessing the composition of a given fraction is the biochemical one as it has been stressed by de Duve (1967) that “once the fractions are isolated, they fall almost entirely within the area of competence of the biochemist.” After achieving this biochemical analysis, it can be useful to visualize or confirm the results through a morphological approach. Therefore the electron microscopic examination must be considered as a complementary tool. Nevertheless, it is by combining both techniques, biochemical and morphological, that the most satisfactory results will be obtained. A. BIOCHEMICAL CRITERIA 1. Analytical Approach

The measurement of a given enzyme activity and the chemical analysis of a given component are always quantitative if the method used has been sufficiently and correctly tested. This point illustrates the con-

270

PIERRE LADURON

siderable advantage and the superiority of biochemical criteria with respect to morphological ones, thus making the former the analytical method of choice for assessing the subcellular composition of fractions obtained by differential or isopycnic centrifugation. In fact, as already pointed out, the first aim in tissue fractionation is not to isolate a given subcellular particle (which is often extremely difficult to do), but rather to determine the distribution pattern of various components, enzymes, transmitters, or chemical substances, which make it possible to recognize a variety of profiles corresponding to different subcellular entities. Hence, it very often happens that the biochemical analysis will show more especially that a given fraction is mainly contaminated by various subcellular structures rather than revealing that this fraction in enriched with a given element like mitochondrion, nerve ending, plasma membrane, etc. Therefore, knowing of the presence of contaminants is essential in the interpretation of tissue fractionation studies. There is another possible approach, which is certainly easier but much less profitable, and which consists of measuring only one or two marker enzymes and ignoring other constituents which may sometimes be more important. For instance, it is relatively often reported in the literature that only a mitochondria1 or/and a synaptosomal enzyme was determined in fractions obtained by centrifugation, as if other subcellular entities like lysosomes, peroxisomes, or microsomes including various different structures did not exist. In certain studies, no marker enzymes were determined and only the speed of the centrifuge was used as “marker” to assess whether a given fraction contains, for instance, many more synaptosomes than another one. In fact as many markers as possible have to be determined to ascertain the subcellular composition of various fractions. Even the presentation of results is of prime importance; the knowledge that a given fraction contains 10 or 50% of a given enzyme activity is not, in itself, particularly significant. This enzyme activity must be referred or compared to another element (generally the protein concentration) before one can evaluate to what extent its presence in greater or lesser amounts, with respect to the protein content, is really specific for a given fraction. For instance, if 40% of the enzyme is recovered in one fraction containing 60% of the total recovered proteins, the interpretation will be different from when this fraction contains only 20% of proteins. Indeed in the latter case there is a true enrichment, while in the former case a contamination cannot be excluded. Figure 1 illustrates the complexity of interpretating tissue fractionation studies. Tyrosine hydroxylase was first believed to be a particle-

TISSUE FRACTIONATION IN NEUROBIOCHEMISTRY

27 1

bound enzyme and even to be contained within chromaffin granules in bovine adrenal medullae (Nagatsu et nl., 1964). When bovine adrenal medullae were fractionated by differential centrifugation after homogenization in 0.25 M sucrose, the tyrosine hydroxylase activity was recovered in various fractions (cf. Fig. lB), suggesting a somewhat complex distribution pattern. Nevertheless, under these experimental conditions it differed from dopamine-P-hydroxylase taken here as a marker enzyme for chromaffin granules (Laduron and Belpaire, 1968). When isotonic KCI was used as the homogenization medium, more than 80% of the enzyme was found in the supernatant fraction, indicating that tyrosine hydroxylase is not associated with chromaffin granules but is localized in the cell sap. On the other hand, i f another fractionation procedure (Shiman et al., 197 1) was used, the tyrosine hydroxylase profile was found to parallel that of inosine diphosphatase considered as a marker of the endoplasmic reticulum (cf. Fig. 1C). Under these experimental conditions, the RSA in nearly all fractions except the fraction P was about equal to 1, indicating that the fractionation procedure was not optimal. This serves to emphasize the need for a very critical approach to the interpretation of tissue fractionation studies. The fact that two enzymes or a pharmacological drug and an enLyme have the same distribution pattern after equilibration through a density gradient does not imply that they are associated with the same subcellular entity. We recently found, for instance, that marker enzyme from four different organelles all equilibrated in exactly the same way in a sucrose gradient containing KCI. Biochemical criteria are based on two postulates which were formulated by de Duve (1967) as follows: The first one states that “all the members of a given subcellular population have the. same enzymatic composition,” while the second one, called “postulate of the single location,” assumes that “each enzyme is restricted to a single intracellular component in the living cell.” This kind of reasoning is applicable not only to enzymes but also to other elements like the neurotransmitter. There are some apparent exceptions to the second postulate which may be easily interpreted. In some cases, isoenzymes may be responsible for a bimodal distribution. Therefore “when we are faced with a complex distribution, the only one attitude is to assume that it is mostly due to an artefact” (de Duve, 1967). Let us take an example: Inosine diphosphatase is known to be an enzyme associated with the endoplasmic reticulum (Novikoff and Heus, 1963). Nevertheless in bovine adrenal medullae (Laduron, 1972) and in the rat brain (cf. Figs. 2 and 5 ) it presents a bimodal distribution in the nuclear and microsomal fraction. This does not mean that the enzyme is located in two different intracellu-

272

PIERRE LADURON

lar compartments. Two possible explanations for this can be put forward: First, the size of the membrane bearing the enzyme is different, so that the large ones sediment with the nuclear fraction and the small ones with microsomes. The second explanation provided by cytochemical data (Goldfischer et al., 1964) is that a part of endoplasmic reticulum is linked to the nuclei. In the adrenal medulla and the brain, this part of inosine diphosphatase sedimentating in the nuclear fraction may be relatively more important than in the rat liver where the endoplasmic reticulum is much more developed but is only located in the microsomal fraction (Beaufay et al., 1974). In the brain, there is another problem of distribution related to the presence of nerve endings. Indeed these structures contain various cytoplasmic constituents including “soluble” enzyme such as LDH. In fact as the homogenization of brain tissue is nearly always incomplete-this is indicated by the presence of nerve endings-many enzymes may partially sediment in particulate fractions although they are normally located in the cell sap. Again, the nerve endings further complicate the interpretation of brain fractionation, which makes it all the more necessary to use several marker enzymes.

2. Enzyme and Neurotransmitter as Marker A marker of any kind may indicate the site of the relative enrichment of a given subcellular structure in a fraction obtained by differential or isopycnic centrifugation. This analytical approach is certainly laborious and also difficult. Perhaps it is the fact that it is so exacting which makes many researchers hesitate to adopt an analytical approach in brain fractionation. What are the most important markers in brain studies? Table I1 gives some of the most useful marker enzymes for the brain fractionation. However, this list is evidently not exhaustive; it is even to be hoped that other and even more specific markers will be added to those proposed here. I would like to make some comments in this connection. First, in some particular cases, the use of one marker for a given organelle appears to be insufficient. For instance, we recently observed that two lysosomal enzymes from the rat brain striaturn, N-acetyl-Pglucosaminidase and arylsulfatase, have completely different distribution patterns in sucrose density gradients. This observation confirms the earlier results of Sellinger and Hiatt (1968) suggesting that both lysosomal enzymes belong to different types of cells, glial and neuronal. Another example is the distribution of monoamine oxidase (MAO) which does not exactly parallel that of cytochrome oxidase since the former is associated with the external membrane and the latter with the inner membrane of mitochondria (Schnaitman and Greenwalt,

TISSUE FRACTIONATION

IN NEUROBIOCHEMISTRY

273

TABLE I 1 USEFULMARKERENZYMES A N D NEUROTRANSMITTERS I N BRAINFRACTIONAT~ON Fraction Mitochondria Lysosome Peroxisome Endoplasmic reticulum Plasma membrane

Marker Cytochrome oxidase (Monoamine oxidase) N-acetyl p-Dglucosaminidase Aryl sulfatase D-amino acid oxidase Inosine diphosphatase 5’-nucleotidase Na +/K +Mg+ ATPase Dopamine-sensitive adenylate cyclase Receptor Noradrenaline Acetylcholine (Lactate deh ydrogenase) Histamine-N-meth yltransferase Catechol o-methyltransferase (Dopa decarboxylase) Sulfatide Lactate dehydrogenase Choline acetylase Glutamate decarboxylase Noradrenaline Dopamine +

Postsynaptic membrane Synaptic vesicle Supernatant fraction

Myelin Nerve endings

5-h ydroxytryptamine Dopa decarboxylase

1968). This explains why the M A 0 activity is always higher in fraction P than cytochrome oxidase since the external membrane of mitochondria may be partly disrupted in the course of the homogenization. Finally, it is often useful to distinguish whether an enzyme measured in a fraction after equilibration in a density gradient is really representative of its normal subcellular structure or whether this organelle is associated in nerve endings. Indeed an enzyme, i.e., microsomal or lysosomal, may be related to free subcellular structures or to structures themselves entrapped in nerve endings. Therefore, a bimodal distribution, although not always necessarily present, can be suspected for many constituents of the density gradient. Some recent results from our laboratory seem to indicate that dopamine-sensitive adenylate cyclase may be a valuable marker enzyme for the postsynaptic membrane in the rat brain striatum (Laduron et al., 1976b).

274

PIERRE LADURON

Although a very large number of marker enzymes seems to be necessary to characterize subcellular structures in density gradient experiments, five enzymes appear to be sufficient to identify the five fractions obtained by means of the differential centrifugation: cytochrome oxidase, N-acetyl-P-glucosaminidase, inosine diphosphatase, 5’-nucleotidase, and LDH. 3. Balance Sheet

Besides the use of marker enzymes to assess the composition of subcellular fractions, one of the most essential principles in tissue fractionation is the balance sheet (de Duve, 1967). It ensures that the fractionation procedure was conducted in a quantitative manner. In the procedure outlined in Fig. 3 ,nothing was discarded, not even the washing solutions, so that the sum of proteins or of enzyme activities must be equal to the amount detected in the unfractionated or starting homogenate. This is an essential control which may sometimes indicate that something must have gone wrong somewhere. For this purpose, an aliquot of the cytoplasmic extract (E) was kept, as indicated in Fig. 3. Generally, it is preferable to estimate the recovery of a given component from cytoplasmic extract rather than from total homogenate, the latter being too heterogeneous. The recovery ( R ) expressed in percent is equal to the sum of activity found in all the fractions divided by the activity in the starting material, nuclear fraction (N), plus cytoplasmic extract:

R =

N+M+L+P+Sx N+E

The same procedure can be used to calculate the recovery in gradient experiments. In this case, the layered material must be diluted in a volume corresponding to the total volume of gradient which also permits an easy determination of the relative activity in a given fraction. The recovery is equal to the sum of activity found in all the fractions expressed as a percentage of the total activity measured in the layered material. If the recovery is lower than loo%, that means either that the enzyme has been inactivated in the course of centrifugation or that an activator of the enzyme which is normally present in the starting material has been separated by centrifugation so that certain fractions exhibit lesser enzyme activity. Conversely, a recovery of more than 100% may mean that an inhibition of the enzyme is not located in the same subcellular fraction as the enzyme itself. An example of this kind has been reported for dopamine-/3-hydroxylase in adrenal medullae (Belpaire and Laduron, 1970). Indeed, in certain experimental conditions, a recovery of 300%

,

TISSUE FRACTIONATION I N NEUROBIOCHEMISTRY

275

for the dopamine-P-hydroxylase was found in fractions obtained by differential centrifugation. In fact it was due to a different subcellular localization of endogenous inhibitors of dopamine-P-hydroxylase and the enzyme, so that a very low enzyme activity was recovered in the cytoplasmic extract in which both were present simultaneously. This led to the determination of the intracellular localization of the endogenous inhibitors, which are mostly in the supernatant fraction. Another example was recently provided in our laboratory in fractionation experiments with a labeled compound. After equilibration in sucrose gradient, a very low recovery of the radioactivity (about 30%)was measured. By washing the gradient tube with an organic solvent it was found that a large amount of the labeled compound remained adsorbed to the tube wall. It should be remembered that by centrifuging through a density gradient, a radial dispersion of particles takes place so that the concentration in material decreases in the middle of the tube but concomitantly increases along the tube walls. B. MORPHOLOGICAL CRITERIA Apparently, morphological examination is a valid criterion for evaluating the composition of subcellular fractions. Unfortunately, this approach is not quantitative enough to provide an accurate analysis of material obtained by centrifugation techniques. Moreover, as previously pointed out, the lack of morphological marker for various structures is a very serious drawback. In addition to this, a technical difficulty mostly related to packing particles by centrifugation has been neglected in most studies and has often led to a false interpretation of the degree of purity. Indeed, morphological examination was generally performed on pellets obtained by centrifuging particles. Owing to their different sedimentation coefficients, subcellular particles do not sediment in a homogeneous manner. Therefore, stratification appears not only throughout the sediment but also along the surface of them, and particles of a given species occur with different frequencies near the center and near the periphery (Beaufay, 1972). When such pellets are examined (they always represent a very small portion of the total material), it is possible to obtain an electron micrograph, showing a very high degree of purity for a given subcellular structure but also one which creates the illusion of a clear-cut separation. Morphological examination cannot always distinguish what is the main structural component of a given fraction and what may be considered as contaminant. A good example of this lack of accuracy was provided by experiments in the adrenal medulla which contains approx-

276

PIERRE LADURON

imately 100 chromaffin granules for 10 mitochondria and one lysosome. By means of biochemical analysis, it was possible to identify a peak of lysosomes located in a sucrose gradient between the peak of mitochondria and chromaffin granules (Laduron, 1969). Owing to the very small amount of lysosomes in proportion to the other particles, electron microscopy was unable to detect this kind of particle which was masked by the large amount of contamination by mitochondria and chromaffin granules. In order to prevent the stratification of different kinds of particles by pelleting, a filtration procedure has been developed and the technique recommended by Baudhuin (1969) makes it possible to obtain a very thin pellet of material which is distributed in quite a homogeneous manner. Only under these conditions morphological examination of subcellular fractions can be devoid of major artifacts. C. SYNAPTOSOMES AND SYNAPTIC VESICLES

It is somewhat surprising to see that practically all the reviews dealing with brain fractionation are mostly focused on the isolation of synaptosomes as if the nerve endings were the only noteworthy structure in the brain (Whittaker, 1969; Marchbanks and Whittaker, 1969; Cotman, 1972; Marchbanks, 1974; Bradford, 1975). The synaptosome is certainly a special entity which plays a prominent role in neurotransmission processes. But this relatively narrow view has led to a dogmatic concept of brain fractionation which is not quite compatible with the progress required for the better understanding of the extraordinary complexity of brain function. Nevertheless, the synaptosomal concept has emphasized the presence of an unusual structure in brain homogenates. To claim that a synaptosome is not a subcellular particle may appear to be somewhat semantic but it has the great advantage of making a clear-cut distinction between the true subcellular structures and this “foreign” structure which consists of a part of the neuron. Synaptosomes can be biochemically recognized using enzymes or neurotransmitters as marker (cf. Whittaker, 1969; Marchbanks and Whittaker, 1969). Of the available enzymes, LDH is widely used. Figure 7 shows the distribution pattern of this enzyme in fractions of rat brain striatum obtained by differential centrifugation. This is normally a “soluble” enzyme which is also present in the cytoplasm occluded within the synaptosome, as illustrated in Fig. 7. After osmotic shock, it is almost entirely recovered in the supernatant (Marchbanks and Whittaker, 1969).

TISSUE FRACTIONATION IN NEUROBIOCHEMISTRY

277

FIG. 7 . Distribution pattern of dopamine and marker enzymes in rat brain striaturn.

In rat hypothalamus, recent experiments from our laboratory have shown a similar distribution for LDH and noradrenaline. Indeed, in this brain area, more than 50% of noradrenaline was found to be particulate-bound (Laduron, unpublished results). In contrast to this, Fig. 7 shows that dopamine was mostly recovered in the supernatant fraction, only 18% being associated with the M and L fraction as opposed to 55% of LDH in the same fractions. The present evidence strongly suggests that doparnine is probably not contained in synaptic vesicles because, if it were so, we would expect to find much more dopamine in synaptosomes than LDH, the latter being free in the nerve endings. Moreover, the foregoing results also indicate that a synaptosome is not as highly occluded as was generally believed (Whittaker, 1969), since

278

PIERRE LADURON

dopamine can be released from nerve endings throughout the fractionation process, which is normally thought to be easier for a small molecule, like dopamine, than for an enzyme like LDH. Finally, it must be noted that after equilibration in sucrose gradients, the distribution pattern of LDH and that of dopamine are quite similar in the rat brain striatum. Previous observations in the dog caudate nucleus have shown that most of the dopamine, in contrast to acetylcholine, occurred in the soluble supernatant fraction, thus suggesting that the dopamine exists in a free or easily released form throughout cell cytoplasm (Laverty et al., 1963). Another possible explanation is that the dopamine-containing synaptic vesicles, if they exist, might be extremely susceptible to breakdown by the homogenization process. It seems rather unlikely since they are protected within the synaptosomes. The large amount of tyrosine hydroxylase in the striatum and the rapid turnover of dopamine is nevertheless compatible with an extravesicular localization of this biogenic amine. If it is true, one might postulate a mechanism which could prevent dopamine from being degraded by O-methylation and deamination process, for instance, by means of a conjugating process with a protein or any chemical substance. From recent studies (Green et ad., 1969; Kuhar et al., 1971; Gfelleret al., 1971) it should be possible to separate different types of synaptosomes according to their content of neurotransmitter. Unfortunately, there are two reasons which make it inconclusive: First, in most cases, the synaptosomes were characterized by exogenous transmitter, assuming that the radioactivity was not localized elsewhere than in the synaptosomes. Second, the lack of essential marker enzymes did not enable us to exclude an uptake of labelled neurotransmitter in contaminants of nerve endings rather than in synaptosomes themselves. It is generally believed that exogenous neurotransmitter is able to mix with endogenous amine. However, this is not the case in all tissues (cf. Laduron, 1969). By using labeled noradrenaline as a marker in the salivary gland and heart of the rat, De Champlain et al. (1969) have shown that M A 0 is associated with vesicles containing noradrenaline. Here again, this was the typical example of erroneous interpretation owing to the use of an inadequate marker. Therefore, the exogenous material is not a good marker unless its subcellular localization has been accurately studied throughout an analytical procedure. Up to now, synaptic vesicles have generally been obtained by submitting synaptosomes to osmotic shock (Whittaker, 1969; Marchbanks and Whittaker, 1969). According to users of this procedure the suspension of synaptosomes in water is selective enough to disrupt only the nerve end-

TISSUE FRACTIONATION IN NEUROBIOCHEMISTRY

279

ings without affecting the integrity of synaptic vesicles. In my opinion, however, there is some doubt in regard to the latter point. Of all the subcellular particles the peroxisomes are well known as being particularly resistant in hypotonic solutions, a fact which is entirely compatible with the absence of water space in these organelles (Beaufay, 1966). Chromaffin granules and splenic nerve vesicles are very sensitive to osmotic shock (Laduron, 1969). In the brain, there is no clear-cut evidence that synaptic vesicles remain intact when synaptosomes are suspended in distilled water. Indeed, when a synaptosomal fraction is suspended in distilled water and then layered on sucrose gradients, a peak of‘ neurotransmitter can be recovered in a region of very low sucrose density (0.4 M ) . We recently performed the same experiment but using chromaffin granules from bovine adrenal medullae. Here again, after osmotic shock, a peak of catecholamine and dopamine-/3-hydroxylase was identified after equilibration in density gradients. Therefore we are forced to conclude that, as in the brain, some vesicles have been released by the breakdown of chromaffin granules in distilled water. The existence of such entities within the chromaffin granules is quite unlikely and, moreover, has never been observed by electron microscopy. Hence, the most probable explanation is that in the course of disruption by osmotic shock, neurotransmitter remains associated, in very small amounts, to the debris of chromaffin granules as presumably also to the debris of synaptosomes or of synaptic vesicles in the brain. ‘Therefore, it is sotnewhat doubtful whether intact synaptic vesicles can really be obtained by the osmotic shock technique, especially as morphological examination of such preparations will always be positive since the membranes, when in solution, form typical vesicles often identified as “synaptic vesicles.” When examining a microsomal preparation of rat liver (Beaufay ut al., 1974), a neuromorphologist would have to conclude that such a fraction contains a large amount of synaptic vesicles. Nevertheless, the procedure of subfractionation after submitting a large mitochondria1 fraction to osmotic shock may represent a useful tool for the determination of refined subcellular localization. Using this technique, we recently found that, in the rat striatum, cyclic nucleotide phosphodiesterase behaves exactly like LDH or dopamine, indicating that this enzyme belongs to cytoplasm of nerve endings (Laduron, unpublished results). In contrast to this, dopamine-sensitive adenylate cyclase equilibrated in a sucrose gradient practically at the same site with and without osmotic shock (Laduron et al., 1976b). This enzyme was found to be not associated with nerve endings containing dopamine.

280

PIERRE LADURON

IV. Conclusion

Tissue fractionation has, until now, largely contributed to our knowledge of cell physiology. In the CNS, despite considerable efforts, exploration of the neuronal cell is still in the early stages. If brain fractionation can contribute to this final goal, it will only be through an analytical approach. Basic research in CNS remains an attempt to understand the events caused by different signals. All the active molecules in the brain must act on receptors which are presumably incorporated in membranes. Therefore, in the future, the study of receptors and membranes will be the main concern of those engaged in brain fractionation. There are many receptors-those for neurotransmitters (dopamine, noradrenaline, 7-aminobutyric acid (GABA), histamine, serotonin, glycine, etc.)-but also receptors for pharmacological drugs (opiates, neuroleptics, antidepressants, tranquilizers, etc.). I t was one of the objectives of this chapter to try to convince neurobiochemists of the extreme difficulty in interpreting results provided by tissue fractionation techniques, Nevertheless, through its analytical form, this methodology may prove to be an invaluable contribution to the knowledge of brain physiology. REFERENCES Abdel-Latif, A. A. (1966). Biochim. Biophys. Actn 121, 403-406. Baudhuin, P. (1968). Thesis, Univ. of Louvain, Louvain. Baudhuin, P. (1969). In “Handbook of Molecular Cytology” (A. Limade Faria, ed.), p. 1179. North-Holland Publ., Amsterdam. Beaufay, H. (1966).Thesis, Univ. of Louvain, Louvain. Beaufay, H. (1972). In “Lysosomes” (J. T. Dingle, ed.), p. 1 , North-Holland Publ., A msterdam. Beaufay, H., and Berthet, J. (1963).Biochem. SOC.Symp. 23, 66-85, Beau fay, H., Amar-Costesec, A., Thines-Sempoux, D., Wibo, M., Robbi, M., and Berthet, J. (1974).]. CellBiol. 61, 213-231. Belpaire, F., and Laduron, P. (1970).Biocli~m.Pharmucol. 19, 1323-1331. Bradford, H. F. (1975). In “Handbook of Psychopharrnacology” (L. L. Iversen, S. D. Iversen, and S. H. Snyder, eds.), Vol. 1 , p. 191. Plenum, New York. Bretz, U.,Baggiolini, M., Hauser, R., and Hodel, C. (1974).J. CellBiol. 61, 466-480. Citkowitz, E., and Holtzman, E. (1973).J. Hislochem. Cytochm. 41, 34-41. Cotrnan, C. W. (1972).In “Research Methods in Neurochemistry” (N. Marks and R. Rodnight, eds.),Vol. 1. p. 45, Plenum, New York. De Champlain, J., Mueller, R. A., and Axelrod, A. (1969).]. Pharmucol. Exp. Ther. 166, 339-345. de Duve. C. (1967).In “Enzyme Cytology” (D.B. Roodyn, ed.), p. 1 . Academic Press, New York. de Duce, C. (1971).J. CellEiol. 50, 20-55. de Duve, C. (1973).J.Historlien. Cytoch~m.21, 941-948.

TISSUE FRACTIONATION

IN NEUROBIOCHEMISTRY

28 1

de Duve, C. (1975). Science 189, 186-194. de Duve, C., Pressman, B. C., Gianetto, R., Wattiaux, R., and Appelrnans, F. (1955). Biochem. J . 60, 604-61 7. De Robertis, E. (1964). “Histophysiology of Synapses and Neurosecretion.” Pergamon, Oxford. Gfeller,E., Kuhar, M. J., and Snyder,% H. (1971).Proc. Natl. Acnd. Sci. U.S.A. 68, 156-189. Glowinski, J., and Iversen, L. L. (1966). J. Neurochem. 13, 6 5 5 4 6 9 . Goldfischer, S., Essner, E., and Novikoff, A. B. (1964).J. Histochem. Cytochem. 12, 72-95. Goldstein, D. B. (1966). J. Neurochem. 13, 1011-1016. Gray, E. G., and Whittaker, V. P. (1962).J. Anat. 96, 79-88. Green, A. I., Snyder, S. H., and Iversen, L. L. (1969).J. Pharmacol. Exfi. T k . 168,264-27 1. Kuhar, M. J., Shaskan, E. G., and Snyder, S. H. (1971).J. Neurochem. 18, 333-343. Laduron, P. (1969). Thesis, Vander., Univ. of Louvain, Louvain. Laduron, P. (1972). Arch. Int. Pharmacodyn. Ther. 195, 197-208. Laduron, P., and Belpaire, F. (1968). Biochem. Pharmucol. 17, 1127-1 140. Laduron. P., and Leysen, J. (1977). Biochem. Pharmucol. 26, 1003-1007. Laduron, P., Verwimp, M. F., Janssen, P. F. M., and Gommeren, W. R. (1975).Biochimie 57, 2 5 3-260. Laduron, P., Aerts, G., De Bie, K., and Van Compel, P. (1976a).Neuroscieme 1, 219-226. Laduron, P., Verwimp, M., Janssen, P. F. M., and Leysen, J. (1976b).Life Sci. 18,433-440. Lagercrantz, H., and Pertoft, H. (1972).J. Neurochem. 19, 8 1 1-823. Laverty, R., Michaelson, 1. A., Sharman, D. F., and Whittaker, V. P. (1963).B r . J . Pharmacol. 21,482-490. Leysen, J., and Laduron, P. (1977). Life Sci. 20, 281-288. Marchbanks, R. M. (1974). I n “Research Methods in Neurochemistry” (N. Marks and R. Rodnight, eds.), Vol. 2, p. 79. Plenum, New York. Marchbanks, R. M., and Whittaker, V. P. (1969). In “The Biological Basis of Medicine” (E. E. Bittar and N. Bittar, eds.), Vo1.-5, p. 39. Academic Press, New York. Nagatsu, T., Levitt, M., and Udenfriend, S. (1964).J. Biol. Chem. 239, 2910-2917. Neims, A. H., Zieverink, W. D., and Smilack, J. D. (1966).J. Neurochem. 13, 163-168. Novikoff, A. B., and Heus, M. (1963).J. Biol. Chem. 233, 710-716. Schnaitman, C.. and Greenwalt, J. W. (1968). J . Cell B i d . 38, 158-175. Sellinger, 0. Z., and Hiatt, R. A. (1968). Brain Res. 7 , 191-200. Sellinger, 0. Z., Rucker, D. L., and d e Balbian, V. F. (1964).J. Neurochem. 11, 271-280. Shiman, R., Akino, M., and Kaufman, S. (1971).j. Biol. Chem. 246, 1330-1340. Snyder, S. H. (1975). Br. J. Pharmacol. 53, 473-484. Tamir, H., Rapport, M. M., and Roizin, L. (1974).J. Neurochern. 23, 943-949. Tulkens, P., Beaufay, H., and Trouet, A. (1974).J . Cell Biol. 63, 383-401. Wattiaux, R.,Wattiaux-De Coninck, S., and Ronveaux-Dupal, M. F. (197 1).Eur.J. Biochem. 22, 31-39. Wattiaux-De Coninck, S., Ronveaux-Dupal, M. F., Dubois, F., and Wattiaux, R. (197S).Eur. J. Biochem. 39, 93-99. Wattiaux-De Coninck, S., Dubois, F., and Wattiaux, R. (1974).Eur.J. Biochem. 48,407416. Weibel, E. R. (1963). Lab. Invest. 12, 131-139. Whittaker, V. P. (1969). In “Handbook of Neurochemistry” (A. Lajtha, ed.), p. 327. Plenum, New York. Whittaker, V. P., Michaelson, I. A., and Kirkland, R. J. A. (1964). Bi0chem.J. 90,293-305.

This Page Intentionally Left Blank

CHOLl ACETYLTRANSFERASE: REVIEW C HOLl NE E AC ETYLTRANSFERASE: A REV1 EW WITH SPECIAL REFERENCE TO ITS CELLULAR AND AND SU BC ELLULAR LOCALIZATION '

By Jean Rossier '

Institute for Biological Biological Studies The Salk Institute

I . Introduction I f . History 111. Assay A. Comparison between Fonnum's Assay and Schricr and Shuster's Assay B. Acetylcarnitine Contamination G. Pitfalls in the ChAc Assays 0. Assays of ChAc: General Conclusion IV. Distribution of ChAc in Nonneuronal T i i r e A. Human Placenta B. Corneal Epithelium C. Other N o n n c u r o ~ lT i i u e s V. Distribution of ChAc in Neumnal Tissue VI. Riritication of ChAc A. Purification of Vertebrate ChAc B. Stability C. Purification of Mammalian Brain ChAc D. lmmiinoahsorbenl E. AHinity Cliromatography F. Future Purification VII. Biophysical Studies V I I I . Mechanism of Action A. Kinetic Mechanism B. Active Site C . Substrate Specificity D. Inhibitors E. Dextran Blue: A Competitive Inhibitor with Respect to Acetyl-&A F. Regulation G. Effect 01' CI Ions on ChAc I X . Axonal Transport of ChAc X. lmmiinology A. Antigenicity of ChAc B. Cross-Reactivity C. Future of the lmrnunohbtochemical Locali7ation of ChAc

' Chargb de Recherche INSERM. 283

284 284 287 287 288 289 29 I 29 1 29 I 298 293 294 2% 297 297 297

298 300 305 503 504 304 305

306

son 309 310 310 312 314 315 315 318

284

JEAN ROSSIER

XI. Localization at the Cellular Level ....................................... A. Classical Histochemistry ............................................ B. ChAc and Cholinergic Pathways.. ................................... C. Ch Ac versus Acetylcholinesterase Localization ........................ D. ChAc Assays in Single Cells and in Discrete Areas .................... E. Histochemical Localization of ChAc: Future .......................... XII. Subcellular Localization of ChAc ....................................... XIII. Choline Transport and ChAc .......................................... A. Coupling of Choline Transport and ChAc ........................... B. Choline Transport and ChAc: A Possible Multienzymatic Complex ..... XIV. The Role of CI- in the Regulation of Ach Synthesis by ChAc . . . . . . . . . . . . . . XV. Pleiotropic Effect of Nerve Impulses on Ach Synthesis .................... References .........................................................

318 319 320 32 1 322 323 324 327 328 328 329 330 33 1

I. Introduction

Although discovered 33 years ago, choline acetyltransferase (ChAc) is still an unknown molecule. In contrast to other enzymes involved in neurotransmission, ChAc has never been purified to homogeneity. Also, a clear-cut method to localize this enzyme does not exist. The absence of pure enzyme and the lack of any reliable method to visualize cholinergic pathways still hamper all the studies involving the cholinergic system. The present review will mainly focus on the recent progress made in the purification and localization of ChAc.

II. History

In 1933 Kahlsohn and MacIntosh (1933), working with the cat, isolated perfused sympathetic cervical ganglion, demonstrated a need for glucose in the perfusion medium. In the absence of glucose, the ganglion quickly became unable to synthesize acetylcholine (Ach). This experiment first demonstrated the need for an energy source such as glucose for Ach synthesis. Nachmansohn worked in Meyerhoffs laboratory on the bioenergetics of muscular contraction. When he started his work on the electric organ of electric fish, he was very surprised to find phosphocreatine and adenosine triphosphate (ATP) present in an amount equivalent to that in striated muscles. He was, of course, very well informed on the central role played by those molecules in the bioenergetic phenomenon related to muscle contraction. After observing that during the electrical discharge the electric organ produced some heat, he proposed that these molecules may also have some importance in energy consumption dur-

CHOLINE ACETYLTRANSFERASE

285

ing nerve impulses. This heat was produced through the hydrolysis of phosphocreatine which actually was, as assumed, consumed in order to maintain a steady level of ATP. Therefore, the following scheme was proposed: ATP is used as an energy source during the electrical activities and the ATP molecules hydrolyzed are immediately resynthesized from the phosphocreatine pool. What was the real use of the ATP molecules? Here is shown the extraordinary intuition of Nachmansohn, who not only assumed that a large part of the ATP was consumed to restore the ionic gradient, but that another part was used to resynthesize the Ach destroyed during electrical stimulation. From a thermodynamic point of view this synthesis of Ach from acetate and choline requires energy. Bearing this in mind, Nachmansohn concluded that to acetylate choline, the acetate must first be energized by ATP and that this form of “active” acetate will then react with choline. This enzymatic reaction, involving ATP, acetate, and choline, was tentatively named cholineacetylase. At the same time, Lipmann had discovered the molecule acetylphosphate. He proposed that this molecule was the one used in all the enzymatic acetylation reactions. Therefore, for Lipmann, this “active” acetate was acetylphosphate. Since then it has been demonstrated that only in some bacteria does acetylphosphate have a central role in the acetylation reaction. This molecule is not used in eukaryotic cells. Back in 1943, when Nachmansohn tried to demonstrate the importance of ATP in the acetylation process of choline, his first results were negative. ATP added to rabbit brain extract was unable to promote any Ach synthesis. However, this first negative result was due to an artifact; ATP was hydrolyzed by the ATPases present in brain extract. After the ATPase inhibition with sodium fluoride, it was possible to measure a strong Ach synthesis promoted by the addition of ATP. In this system acetylphosphate was unable to promote any Ach synthesis. This experiment was of primary importance in the history of biochemistry. It marked the first time that, in solution, the energy required for an enzymatic acetylation reaction was contributed by the hydrolysis of ATP (Nachmansohn and Machado, 1943). At that time Nachmansohn’s results had appeared so strange that three famous journals first rejected the paper presenting them. Finally, a fourth journal-one devoted more to neurophysiology than to biochemistry-accepted his paper (Nachmansohn, 1972). Viewed from the present, it is easy to sympathize with the referees. In 1943 it was obvious that acetylphosphate was the “active” acetate used for enzymatic acetylation. The experiments of Nachmansohn rejected this concept and proposed a new one: Acetate was activated by ATP. The validity of this

286

JEAN R O S I E R

concept was finally recognized and accepted by all biochemists, even Lipmann himself. Nachmansohn has related recently in this autobiography how Lipmann accepted his idea: “Lipmann refused to accept my findings for two years. But at the end of 1945 he came to my laboratory with a box containing a number of phosphate derivatives to test their action in my system. In the beginning, he had hoped that acetylphosphate would work in my test system, but for some reason it did not work in his. After a number of experiments he realized that only ATP was effective of all the compounds tested, and he then accepted my view of its role as the energy sources for acetylation.” The general concept of activation of acetate by ATP was accepted when Lipmann (1945) observed that only in the presence of ATP did pigeon liver extracts acetylate sulfanilamide, However, the mechanism of this activation of acetate was still unresolved. Nachmansohn had observed that after dialysis the rabbit brain extract was no longer able to acetylate choline. This observation suggested the need for a coenzyme (Nachmansohn et al., 1943). Nachmansohn and Berman ( 1 946) tried to isolate this coenzyme from liver, muscle, and brain extracts. Their preparation, which they named Kochsaft, was then used to restore the choline acetylation activity in a brain extract which

FIG. I. Coenzyme A structure (CoA or CoASH).

CHOLINE ACETYLTRANSFERASE

287

had been previously dialyzed in order to eliminate the coenzyme. In 1945, Lipmann demonstrated that the pigeon liver sulfanilamide acetylation system also needed a cofactor. On the basis of comparisons between the two systems it was proposed that the cofactor may be the same. The name of coenzyme A (CoA) (A for acetylation) was proposed by Lipmann who, two years later, demonstrated that this molecule contained pantothenic acid bound to a nucleotide by a phosphate bridge. Figure 1 gives the general structure of this molecule (Lipmann et nl., 1947). The two enzymatic reactions which take place in brain extract are

+ acetate * acetyl-CoA + AMP + P P , Acetyl-CoA + choline C acetylcholine + CoA

A T P + CoA

(1)

(2)

The first reaction is catalyzed by the enzyme acetyl-CoA synthetase (EC 6.2.1.1.) (Berg, 1956), the second by ChAc (acetyl-CoA: cho1ine-Oacetyltransferase [EC 2.3.1.6]-[ChAc]). 111. Assay

The number of papers devoted to ChAc assay is extremely large. In a recent review Catherine Hebb (1972) detailed 13 of them. The great number of proposed methods may reflect the real difficulties encountered by workers in assaying ChAc. 1 will not describe the numerous assays proposed, but only discuss the two most often used, i.e., the Fonnum assay and the Schrier and Shuster assay. Other assays are often only variations of these two original techniques (Fonnum, 1969, 1975a; Schrier and Shuster, 196’7). A. COMPARISON BETWEEN FONNUM’S ASSAYAND SCHRIER A N D SHUSTER’S ASSAY Both use radioactive acetyl-CoA to label the acetyl moeity of Ach. The problem is to separate the product Ach from the substrate acetylCoA. This is done by an ion-exchange technique in both methods. Fonnum’s technique is based on liquid cation exchange using sodium tetraphenylboron dissolved in either ethylbutylketone or other adequate organic solvents. Sodium tetraphenylboron reacts with all the positively charged quaternary ammonium compounds. This reagent is soluble only in organic solvents and poorly soluble or insoluble in water. On the other hand, acetyl-CoA is poorly soluble in organic solvents. Therefore after incubation of enzyme solution with labeled acetyl-CoA and cold choline the reaction is stopped and the labeled product extracted by adding

288

JEAN

ROSSlER

tetraphenylboron dissolved in an organic solvent ethylbutylketone, acetonitrile, or butyronitrile. The labeled substrate, acetyl-CoA, is not extracted and remains in the aqueous phase. With the method of Schrier and Shuster, the product formed by incubation is separated from labeled acetyl-CoA by means of an anionexchange resin. Labeled Ach passes through the anion-exchange resin, whereas labeled acetyl-CoA is retained. Unfortunately, the presence of impurities or products of radiolysis in the commercial radioactive acetyl-CoA gives high blank values. These high blank values may be reduced by purifying acetyl-CoA before use by ion-exchange chromatography (Hamprecht and Amano, 1974). However, other compounds synthesized from acetyl-CoA during incubation still pass through the anion-exchange resin with Ach.

B. ACETYLCARNITINE CONTAMINATION The main contaminating compound is acetylcarnitine (White and Wu, 1973a; Hamprecht and Amano, 1974) which especially occurs when tissue containing a high amount of the enzyme carnitine acetyltransferase is incubated. This presence of acetylcarnitine in the effluents of the columns may give an artificially high ChAc level in tissue devoid of any ChAc activity. A good example of this is the claim of some workers who say that the liver contains the highest body level of ChAc, and that the brain has 10 times less ChAc activity than the liver (Mahoney et al., 1971). To get rid of acetylcarnitine contamination, the use of a rigorous control is necessary. Blanks done without choline, with a high amount of the hydrolytic enzyme AchE, with an inhibitor of ChAc, or with an antiserum against ChAc were recently proposed (Schrier and Klein, 1974; Peterson et al., 1973; Hamprecht and Amano, 1974; Roskoski el al., 1974; White and Wu, 1973a). The controls described here are also in use in Fonnum’s assay but are less important. The extraction technique of Ach proposed by Fonnum is much more specific for Ach than the column chromatographic procedure of Schrier and Shuster. In Fonnum’s procedure, Ach is isolated by cation exchange at neutral pH. Most acetylated products formed during incubation are not basic enough to react with tetraphenylboron. The presence of a carboxylic group on the acetylcarnitine molecule prevents the formation of a complex of acetylcarnitine with tetraphenylboron at neutral pH; this reaction will occur only at pH below 2. Therefore, the extraction procedure has to be performed in the presence of sufficient buffer at neutral pH. The procedure of Fonnum also has the advantage of decreasing the

CHOLINE ACETYLTRANSFERASE

289

blank due to the [‘%I acetyl-CoA and its radiolysis products which are not soluble in the organic solvents used in the extraction procedure. With the Fonnum technique, the blank value corresponds to 0.35% of the original amount of [“C] acetyl-CoA in the incubation medium. Finally, in the opinion of the author, who has through the years tested the two procedures, the Fonnum procedure is quicker and more reproducible. The rapidity of this assay was recently further increased (Fonnum, 1975a). Now the whole procedure may be done directly in the scintillation vial which right after the enzymatic incubation may be counted with the scintillation spectrometer.

c. PITFALLSI N 1 H E ChAc ASSAYS Fonnum’s assay for ChAc is simple, rapid, and reproducible, but some pitfalls exist in these assays and it may be worthwhile to inform future experimenters of possible sources of error. 1. Effect of Imiduzol The simple presence of imidazol in the incubation medium promotes the nonenzymatic synthesis of Ach when both substrates are present (Burt and Silver, 1973a; Hebb et al., 1975). The claim by Chao and Wolfgram (1974) that imidazol or imidazol-containing compounds like creatinine were possible physiological activators of ChAc seems therefore not very sound since without enzyme, imidazol itself induces, as a catalyst, the synthesis of Ach. 2. Interference of Potassium low The trade name of tetraphenylboron is Kalignost. This compound was primarily used to assay potassium (Kalium in German). As tetraphenylboron reacts with potassium, high concentration of potassium in the incubation mixture must be avoided when using Fonnum’s extraction technique. During the extraction, tetraphenylboron will form an insoluble salt with potassium and the subsequent precipitate will interfere with the recovery of Ach. 3 . Synthzsis of Ach by Carnitine Acetyltrunsferase

I t has already been seen that acetylcarnitine may be synthesized in the presence of carnitine acetyltransferase and that the radiolabeled acetylcarnitine may be coextracted with Ach. I t has also been found that in using appropriate procedures, it is possible to get rid of this contamination. Nevertheless, the presence of high amounts of the enzyme carnitine acetyltransferase in the homogenate promotes the synthesis of true Ach. Therefore part of the Ach synthesized during the incubation

290

JEAN ROSSIER

will not be formed through ChAc, but through another enzyme, carnitine acetyltransferase, which may be used as an alternative substrate choline instead of carnitine. Thus, tissues containing high amounts of carnitine acetyltransferase and no Ch Ac may enzymatically synthesize some Ach (White and Wu, 1973a). To reduce this synthesis, a low concentration of choline in the incubation medium may be used (Roskoski et al., 1974). These recent findings may explain some of the previous results obtained after muscle denervation. It is now agreed that all ChAc present at the motor end plate is of neuronal origin (Israel, 1970). Nevertheless, after total denervation residual Ach synthesis may be found in denervated muscle extract (Hebb et al., 1964; Israel, 1970; Tucek, 1973). It may be proposed that this residual activity is due to the high muscle content of carnitine acetyltransferase. 4. Stability

of ChAc

ChAc may become very unstable when diluted. To avoid denaturation it is recommended to dilute the enzyme in a medium containing a nonenzymatic protein like bovine serum albumin (Rossier et al., 1973a).

5. Effect of SH Reducing Agents SH reagents are potent inhibitors of ChAc (Mannervik and Sorbo, 1970). Therefore several investigators recommend the addition of reducing agent like cysteine or dithiothreitol in the assay medium. This addition must be avoided as reducing agents promote the thiolysis of acetylCoA. For example, at pH 7.5 and 39"C, acetyl-CoA (300 p M ) was completely hydrolyzed after a 20-minute incubation in the presence of dithiothreitol (20 m M ) (Morris, 1967). 6. Incubation Time Although the K , for acetyl-CoA is below 10 p M , Fonnum (1975a) recommends a final concentration of 400 p M in the assay medium. This high acetyl-CoA concentration is especially needed when long incubations are performed in the presence of Triton X-100. This detergent solubilizes not only the occluded ChAc activity, but also numerous enzymes consuming acetyl-CoA (Matsuda and Yoshida, 1976). These deacylase activities lead to a rapid decrease of the acetyl-CoA concentration with a parallel increase of CoA which is a potent ChAc inhibitor. Therefore short incubation periods ( 5 minutes) are recommended. 7. Effect

of

Su,crose

Recently Hebb et al. ( 1975) have observed that high concentrations of sucrose may give higher values in the ChAc assay. Sucrose or a con-

CHOLINE ACETYLTRANSFERASE

29 I

taminating compound present in the commercial sucrose available may be acetylated in the presence of acetyl-CoA. This compound may be extracted and will artifactually increase ChAc activity in samples containing a high concentration of sucrose. D. ASSAYS OF ChAc: GENERAL CONCLUSION

As described here, the ChAc assay is not a simple one. Presence of other enzymes like carnitine acetyltransferase in the homogenate, unexpected nonenzymatic synthesis of Ach and the synthesis of other acetylated compounds during the incubation are the prominent difficulties of this assay. Therefore, when investigators report low values of ChAc in extracts of tissue lacking innervation, these results must be treated with some caution. A general rule may be that ChAc-specific activities below 0.05 nmolelminute per mg protein are background values except when the two following controls have been carefully performed: ( 1) extended analysis of the final acetylated compound must show that this product is indeed Ach; and (2) specific ChAc inhibitors or specific antiserum against ChAc must block the Ach synthesis. Having in mind these criteria, the recent literature dealing with the distribution of ChAc will now be reviewed. IV. Distribution of ChAc in Nonneuronal Tissue

ChAc is well known to be a specific protein of the nervous system. This assumption is not far from reality but ChAc also exists in nonnervous tissue, and paradoxically, it appears that nonnervous tissue have been the best source of starting material for Ch Ac purification. ChAc with a specific activity around 1 pmole/minute per mg was obtained by Alpert et nl. (1966) as early as 1966 starting from the bacteria Lactobacillirs plantarum. The same year, Morris (1966), starting from immature human placenta had obtained an enzyme with a specific activity of 2.4 pmoleslminute per mg. I t has long been known that Ach is present in numerous plants (for review, see Hebb, 1963), yet the presence of ChAc in plants was only recently demonstrated by Barlow and Dixon (1973) who studied the kinetic characteristics of Ch Ac isolated from nettles. A. HUMAN PLACENTA I t is well known that the human placenta contains a high amount of ChAc (Bull et nl., 1961). The placenta of mammals other than primates does not contain ChAc (Hebb and Ratkovic, 1962). More recently, ex-

292

JEAN ROSSIER

tremely low concentration of ChAc have been detected in the placentas of all the mammals studied by Welsch (1974a). From my own experience with the assay of ChAc in placenta of rats and mice, I think that ChAc is not present in the placenta of mammals other than primates. The presence of ChAc with Ach and acetylcholinesterase in the primate placentas has led to some speculation on the role of this cholinergic mechanism in the placenta. Koelle (1967) first proposed that this cholinergic mechanism was involved in regulating the exchange of substances, perhaps amino acids, between mother and fetus. This hypothesis has also been presented by Welsch (1974a) and Harbison et al. (1975). The first protagonist of this theory, Koelle has recently carefully studied the distribution of acetylcholinesterase in the human placenta (Ruch et nl., 1976). The results indicated that most of the acetylcholinesterase content of the human placenta is due to its presence in blood. Histochemical examination of well-perfused placenta suggested that trapped intact erythrocytes and the stroma of hemolysed red cells are the probable source of placental acetylcholinesterase previously reported. The specific activity of ChAc in mature human placenta is around the same as that of the brain. Specific activity over 10 to 20 times higher is found in immature placenta between the 12th and 20th weeks in utero. Activity is present as soon as placental formation is accomplished and it peaks around the 20th week of pregnancy (Welsch, 1974b; Bull et al., 1961). I t appears difficult to correlate the peak of activity with one of the events of embryonic maturation. For the fetus the only important process taking place around the 20th week is the maturation of the immunological system (Papiernik, 1972). Therefore, it may be proposed that the secretion of Ach by immature placenta may be of importance in the maturation of the fetal immunological system. This may appear as a pure speculation but recently it has been demonstrated that lymphocytes possess an Ach receptor involved in the immunological response (Strom et al., 1974). B. CORNEAL EPITHELIUM The presence of ChAc in the corneal epithelium is well known since the work of Van Alphen (1957) and was confirmed by numerous studies (Williams and Cooper, 1965; Howardetal., 1973). It seems that a considerable variability of the corneal ChAc levels exists between species and even within a given species, e.g., the corneal ChAc activity of a particular rabbit could be 50 times that of another rabbit. This variability would be genetically determined, but only in part (Mindel and Mittag, 1977). It was proposed that the high amount of ChAc in cornea, around 6.5

CHOLINE ACETYLTRANSFERASE

293

nmoles/min per mg of protein, was related to the innervation of the cornea and some observations indicated that Ach in the cornea may play the role of a sensory mediator (Fitzgerald and Cooper, 1971). Nevertheless, more recent studies indicate that Ach may also play a still unidentified nonneuronal role in the cornea. The hypothesis that the ChAc, Ach and acetylcholinesterase content of corneal epithelium is not only due to the innervation of cornea by the opthalmic branch of nerve V, may be inferred by the following observations: (1) corneal epithelial cells maintained in culture and thus denervated contain significant amounts of ChAc, Ach, and acetylcholinesterase (Gnadinger et al., 1967, 1973); (2) in vivo, following denervation, corneal epithelium still contain residual Ch Ac activity (Van Alphen, 1957; Fitzgerald and Cooper, 1971); (3) by acetylcholinesterase histochemistry, it has been shown that corneal epithelium cells contain true acetylcholinesterase (Howard et al., 1975). The origin of ChAc in cornea was very recently reinvestigated by Mindel and Mittag (1977). The denervation was performed by retrobulbar ethanol injection. This treatment was followed by a marked decrease of the ChAc content of the iris and the ciliary body. Nevertheless, the levels of ChAc in cornea were not significantly affected by this method of chemical denervation. The. general conclusion of these studies is that two pools of Ach exist in the corneal epithelium, one related to corneal innervation but the function of the other is still hypothetical. Recently, Stevenson and Wilson (1975) have proposed that Ach in corneal epithelium cells may modify Na+ movements across this tissue and in turn play a part in controlling corneal thickness and transparency. C. OTHERNONNEURONAL TISSUES In vertebrates, the presence of ChAc in noninnervated tissues seems limited to the primate placenta and to the corneal epithelium. The presence of ChAc in red blood cells, which indeed contain acetylcholinesterase and an Ach receptor, had often been ruled out (Hebb, 1963; Potter, 1970). More recently Chuang et al. (1976) have found a very low ChAc activity in human and canine platelets. These platelets contain a measurable acetylcholinesterase activity and an Ach receptor: it was shown that platelets contain a choline permease system (Green et al., 1972). Nevertheless, the presence of elements of the cholinergic system is not proof of the presence of ChAc in platelets and the low ChAc activity found in platelets has yet to be confirmed. I t has been proposed that a ChAc-Ach-acetylcholinesterase system

294

JEAN ROSSIER

plays a significant role in insuring and controlling the motility of spermatozoa. Some of the data presented for this theory are convincing. Nevertheless, I think that the presence in sperm of true ChAc enzyme activity has not been completely established (Harbison et al., 1976, Bishop et al., 1976). The presence of ChAc in human cerebrospinal fluid was proposed before by Johnson and Domino (1971) and by Rimon el al. (1973). More recently an extended study was performed by Aquilonius and Eckernas (1976). Their conclusions are that the ChAc activity previously found in the cerebrospinal fluid is not promoted by ChAc but mainly by a small molecular weight compound. These still unidentified compounds are responsible for the nonenzymatically catalysis of Ach formation. Such compounds are presumably imidazol-containing molecules like creatinine or histamine. It has already been mentioned that imidazol catalyses the nonenzymatic synthesis of Ach (see Section 111, A). V. Distribution of ChAc in Neuronal Tissue

Table I gathers together ChAc specific activities in various parts of the nervous system of invertebrates and vertebrates. The highest concentrations of ChAc are seen in invertebrates. Optic lobes of squid and octopus contain the largest amount of ChAc (Berman-Reisberg, 1957; Florey and Winesdorfer, 1968). It must be noted that the specific activity of crude extract from squid head ganglia is 100-fold higher than crude extract of rat brain. This difference of two orders of magnitude was interpreted generally as a major difference in the turnover number of the enzymes. ChAc from invertebrates would have a greater turnover number than ChAc from vertebrates (Shuster and O’Toole, 1974). The recent progress in the purification of ChAc from vertebrates demonstrate that both enzymes would have a turnover number in the same range. Accordingly, to obtain a pure enzyme from mammalian brains will require a much greater purification. In vertebrates, the chick ciliary ganglion is the richest tissue in ChAc (Sorimachi and Kataoka, 1974b). A material which may provide a large amount of enzyme is the electric organ of the torpedo. Assays give values between 20 to 30 nmoles/minute per mg of protein (Bull et al., 1969). ChAc is, of course, present in brain tissue, but there its distribution is heterogeneous. The first study on the heterogeneous distribution of ChAc in the brain was made on dog brain by Hebb and Silver (1956). Since then, numerous reports have confirmed this heterogeneous distribution. Caudate nucleus contains large amounts of ChAc, whereas the cerebellum is the tissue poorest in ChAc.

295

CHOLINE ACETYLTRANSFERASE

TABLE I CHOLINE ACETYLTRANSIERASE I N VARIOL.S SPECIES

Species Bacteria Lactobarillus plantantm Plant Nettles 1n\ ertebrates Squid Fly Drosophiln Snail Vertebrate Fish Torpedo Cmldfish Bird Chicken Mammals Rat

Rabbit Beet'

Sheep Human

Organ

Actkity (pmoleiminimg of protein x 1 0 - 1 at 37°C)

10

1

Optic lobe Brain Brain Esophagal ganglia

100 10 I00

2.6

References

Alpert

ut

d.(1966)

Barlow and Dixon (397S)

Husain and blautner (1 973) Smallman (1956) Dewhurst rt 01. (1972) Ernson rt ( I / . (1971)

Electric organ Brain Retina

20-30 4 10

Bu 11 ~t ( I / . ( 1969) Hebb cf nl. (1972) Ross and bfacDougd1 (1976)

Ciliary ganglia

30

Sorimachi and Kataoka (1954h)

4

Sorimachi and Kataoka (197-h) Sorimachi and Kataoka (1974h)

Cervical ganglia I n terpedonai laris niicleus Caudate nucleus Ventral root spinal cord Cornea Retina Caudate nucleus Caudate nucleus Placenta: 20 weeks Placenta: term

12 3.8 2 6.5 3 3.2 3 10 1

Tucek (1967) Hebh (1963) Williams and Cooper ( I 96.5) Hebh ( 1963) Tircek (1967) BuIIrt nl. (1970) Morris (1966) M'elsch (1974a)

Recently it was demonstrated that the richest region in brain was the nucleus interpedoncularis (Kataoka et nl., 1973; Yalkovits et nl., 1974). Other small nuclei are also rich in ChAc, including all the motor nuclei of the cranial nerves. For example, the specific activity in homogenate of the motor nucleus of' the hypoglosse is around four times the value found in caudate nucleus homogenate (Kobayashi et nl., 1975).

296

JEAN

ROSSIER

VI. Purification of ChAc

The most critical problem in any enzyme purification is not the purification procedure, but the choice of an adequate starting material. The most adequate starting material would contain a high amount of enzyme and be available in large quantity. From a rapid analysis of Table I it appears that in invertebrates the most adequate material would be the squid optic lobes. Indeed, the first procedure describing partial purification of ChAc had used this starting material (Berman-Reisberg, 1957). Moveover, as shown in Table 11, purified ChAc with the highest specific activity was obtained using this invertebrate material (Husain and Mautner, 1973; Polsky and Shuster, 1976a). Nevertheless, neurochemists may have more interest in the vertebrate enzyme as the invertebrate enzyme may be quite different from the vertebrate. For example, it has been shown that the susceptibility of the invertebrate enzyme to inhibition by styrylpyridine is two orders of magnitude less than for the mammalian enzyme (Husain and Mautner, 1973). CoA is a potent inhibitor of the mammalian enzyme and a poor TABLE 11 PURIFICATION OF ChAc Specific activity (~molelminlmg of protein) Starting material

Starting

Final

References

Loctobacillzrs planturum Squid, optic lobes

0.0 10 0.1

Horseshoe crab brain Fly brain Guinea pig brain Beef, caudate nucleus

0.005 0.01 0.001 0.001

Human caudate nucleus

0.003

Human immature placenta Mouse brain Rat brain

0.010 0.001 0.001

0.6 1.3 67.0 58.0 1.0 0.07 0.02 0.059 0.3 16.0 0.012 0.072 2.4 0.07 0.7 1.2 2.3 20.0

Alpen el (11. (1966) Berman-Reisberg (1957) Husain and Mautner (1973) Polsky and Shuster (1976a) Emson et al. (1974) Mehrotra (1961) Kumagai and Ebashi (1954) Mannervik and Sorho (1970) Glover and Potter (1971) Ryan and McClure (1976) Singh and McCeer (1974a) White and Wu (197%) Morris (1966) Shuster and O’Toole (1974) Potter et nl. (1968) Wenrhold and Mahler (1975) Malthe-Sbrenssen et 01. (1973) Rossier (1976a)

CHOLINE ACETYLTRANSFERASE

297

one for the squid enzyme (Rossier, 1977). Finally, vertebrate and invertebrate ChAc are immunologically different (Rossier, 1977).

OF VERTEBRATE ChAc A. PURIFICATION

In vertebrate, the richest and the most convenient source would be the electric organ of the torpedo, but until now, no attempts have been done to purify ChAc from this material. The other rich source is immature placenta from 20 weeks of gestation. In using this material, Morris (1 966) had obtained an enzyme of specific activity around 2.4 pmoledminute per mg of protein. The purification steps included ammonium sulfate fractionation, gel filtration on various grades of Sephadex, and ion-exchange chromatography on DEAE-Sephadex. Purification to homogeneity was not obtained due in part to the instability of the purified enzyme. B. STABILITY The stability of the enzyme was partially increased by adding SH reducing agents. The instability of the enzyme seems not to be due to the action of proteolytic enzymes which are present in high amounts in the placenta as no protection was afforded by the addition of disopropylphosphofluoridate or DFP, a proteolytic enzyme inhibitor. It seems that ChAc like many other enzymes is unstable at low protein concentrations. Therefore, it is recommended that during purification procedure a protein concentration higher than 1 mglml should be maintained. This high concentration of protein is not easily maintained through all the purification procedures especially in the final steps and during gel filtration. Thus, several authors have successfully performed all the purification procedures in the presence of high concentrations of glycerol or sucrose (Husain and Mautner, 1973; Shuster and O'Toole, 1974; Chao and Wolfgram, 1973; Rossier, 1976a). Also, the presence of heavy metals chelating agents seems useful. ChAc is quickly inactivated in the presence of Cu'+ ions. Therefore, the use of ethylenediaminetetraacetic acid (EDTA) or another chelator more specific for Cu'+ ions such as 1-10 phenanthroline were successfully used to increase the enzyme stability (Rossier, 1976a). C. PURIFICATION OF MAMMALIAN BRAIN ChAc

The purification of the mammalian brain enzyme was started as far back as 1954 when two workers published a paper devoted to "Highly

298

JEAN ROSSIER

purified choline acetylase.” The purification factor was around 20 in starting from guinea pig brain (Kumagai and Ebashi, 1954). Since this first claim of the high purification of ChAc, many researchers have published papers where they claim to have obtained a pure enzyme. In 1968, starting from rat brain, Potter et al. (1968) described a method which gave an enzyme of specific activity around 0.7 pmole/ minute per mg of protein. The purification factor was around 1000. The first steps of their purification take advantage of the ability of the enzyme to absorb to membranes at low ionic strength and to be released at higher salt concentrations. This particular property of rat brain ChAc was first described by Fonnum (1968). The first step gave a purification factor of 10 times in using only two subsequent centrifugations. The other steps were ammonium sulfate fractionation, ion-exchange chromatography, and gel filtration. Later, the same group presented a method applicable to ox brain ChAc purification (Glover and Potter, 1971). More recently this method was scaled up to large quantities of starting material. A batch of 20 kg of ox brain was used (Street et al., 1973), but the specific activity of the final enzyme was one order of magnitude less than that obtained with rat brain.

D. IMMUNOABSORBENT In the last few years the use of immunological techniques to purify ChAc from rat brain was developed by the author and simultaneously by the Kjeller (Norway) group of Fonnum. Both groups had in mind to produce antiserum against ChAc from vertebrate. Therefore, they had to purify the antigen. For a long time they injected, without success, ChAc preparations with various specific activities. The antisera produced did not’precipitate or inhibit ChAc, but reacted strongly in immunodiffusion by giving precipitation lines with the purest ChAc preparation available. It became clear from these negative attempts that the antisera produced were directed against proteins other than ChAc. These proteins were still present in the purest ChAc preparation obtained by classical biochemical purification technique. Instead of throwing away these unsatisfactory antisera, both groups have used them to further purify ChAc. These antisera directed against contaminants of the Ch Ac preparation were insolubilized by various means including glutaraldehyde gel or BrCN-treated agarose gels (Rossier et al., 1973a; Rossier, 1976a; Malthe-Sbrenssen et al., 1973). The immobilized antisera, also named immunoabsorbent, were used in a chromatography column. Enzyme preparations were applied to these columns. Ch Ac activity did

299

CHOLINE ACETYLTRANSFERASE

not bind and was eluted quantitatively while proteins reacting with their respective antibodies were retained on the column. In using such a procedure as the final purification step, a specific activity of 20 pmoled minute per mg was recently achieved (Rossier, 1976a). A description of' the purification procedure is given in Table 111. The entire purification procedure required nine steps. The first is the homogenization of 2.7 kg of rat brain from a total of nearly 2000 rats. We will not detail the other six of the first seven steps which are classical biochemical techniques. The choice of appropriate technique led us to obtain a 3550-fold purification with a 14% recovery. The last two steps used immunoabsorbent. The first was obtained by using various sera from rabbits injected with aliquots from each of the previous purification steps. We prepared antisera against rat adrenal extract and rat gammaglobulins as it was detected by immunological technique that enzyme from step 7 contained serum gammaglobulins and a protein common to adrenals and brain. All the sera used in the immunoabsorhent columns were devoid of' any activity against Ch Ac itself, i.e., they did not inhibit nor precipitate ChAc activity. T h e garnmaglobulins of the sera were coupled to BrCN-activated Sepharose 4B. The enzyme passed through the column packed with this first immunoabsorbent and was purified 3.2 times. The passage on the immunoabsorbent column was completely efficient, since the en;lyme preparation now did not react anymore in immunodiffusion with aliquots of the sera used in the immunoabsorbent. TABLE I11 SCMMARY 01. T H E Pk'RII.ICATION

N

Step

1

Homogenate Supernatant fraction, 200 mn.l NaCl CM-50 Sephadex First ammonium sull'ite preci pitat ion Hydi-oxylapatite Second ammonium sulfate precipitation Sephadex G- 100 First i rn mu noabsorben t Second irnmu noa bsorbent

2 3 4

5

6 7 X

9

\'olnme (ml)

Total protein

(mg)

Total eivyme Re(@mole/ covery min) (57)

1x,900

314.000

15,200 190

30,400 20x0

30 26

690 I 3.5

2.6 16

114 12.8 1 .(i

IX 0.9

0.1

240

I00

Specific activity (@mole/ minimg) 0.00076

0.595 2.700 x.700 20.000

300

JEAN ROSSIER

A fraction of the previous preparation was injected into a goat which gave an active serum 3 months later. This serum reacted strongly in immunodiffusion by giving four precipitation bands with the antigen used for the immunization procedure. This serum did not inhibit or precipitate rat brain ChAc: therefore it was used to prepare a second immunoabsorbent column. ChAc activity passed through this column and other contaminants were retained. This final purification step gave 0.1 mg of protein and the overall procedure gave a purification of around 30,000-fold. SDS-gel electrophoresis revealed nine bands of variable intensity (Fig. 2). Four were major bands. We believe that ChAc may correspond to one of these major bands. This would indicate that ChAc represented only 20% of the protein from our purest preparation. Therefore the specific activity of pure ChAc from rat brain would be around 100 pmoledminute per mg of protein. If as assumed here, pure ChAc from vertebrate will have a specific activity of 100 pmoles/minute per mg of protein, it would be necessary to purify 150,000-fold the enzyme present in crude rat brain extract before reaching homogeneity. The degree of purification required for obtaining pure mammalian brain ChAc gives low pertinence to the recent claims of the ready production (purification factor being less than 1000) of pure ChAc from human and beef caudate nucleus (Singh and McGeer, 1974a; Chao and Wolfgram, 1973; Roskoski et al., 1975).

E. AFFINITY CHROMATOGRAPHY As stressed in the previous pages, the preparation of pure ChAc will require a huge purification factor. Pure enzyme will only be obtained by the use of affinity chromatography. This technique may give in one step a purification factor of more than 1000. Usually affinity chromatography columns are prepared by insolubilization of a specific reversible inhibitor of the enzyme. Unfortunately, in the case of ChAc, the number of reversible inhibitors are few. 1. Sulfhydlyl (SH) Reagent

Mannervik and Sorbo (1970) have shown that vertebrate ChAc contains a free SH group which may react with paramercuribenzoate (PMB). Therefore, the use of immobilized PMB on Sepharose 4B was tried as a purification step by Chao and Wolfgram (1973). The chromatography on this material was successful. ChAc is retained and may be eluted by disulfide reducing agents with a purification factor around three times. This low purification factor might be expected, as many

CHOLINE ACETYLTRANSF'ERASE

FIG. 2. SDS-gel electrophoresis. 20 pg of step 9 and 30 p g of steps 5 and 8 were submitted to polyacrylamide SDS-gel electrophoresis. In the same run, gels with markers, bovine serum albumin (BSA), ovalbumin (OVA), and cytochrome C (CYT.C) were performed. Their respective mobilities are indicated by arrows on the densitogram of ChAc, step 9. The mobility of the dimer of BSA and OVA is also indicated.

302

JEAN ROSSIER

proteins contain free SH groups which may react with PMB. The same column was also used by Husain and Mautner (1973). 2. Hydrophobic-Afinity Chromatography I have recently found that rat brain ChAc is a hydrophobic protein which may be bound to hexane linked to agarose columns (Rosier, 1977). Thus the use of this type of hydrophobic-affinity columns may be helpful in the future.

3. Styrylpyridine Inhibitors Styrylpyridine compounds inhibit ChAc but are not competitive. Nevertheless, Husain and Mautner ( 1973) have used these compounds in affinity chromatography in order to purify squid optic lobe ChAc. The purification factor was low (three times) and it may be assumed that this column was merely a hydrophobic-affinity column.

4. CoASH CoASH product of the ChAc reaction is also a very potent inhibitor. This compound competes reversibly with acetyl-CoA, the substrate of the reaction. This compound, bound to agarose, was used with great success to purify ChAc by Ryan and McClure (1976). In a preliminary communication, these authors reported a purification factor of 6500. From my own experience, it appears that such columns could be used in the first steps of the purification procedure. With aging and purification, the affinity of ChAc for its substrates decreases. For example, purified rat brain ChAc has a K , value for acetyl-CoA ranging between 20 and 50 pI4 (Spantidakis et al., 1976) but crude enzyme had a K , value one order of magnitude lower (Rosier, 1977). Therefore when affinity chromatography is used it would be worthwhile to use such techniques during the earlier stages of the purification procedure. 5. Dextran Blue The use of dextran blue columns to purify nucleotide metabolizing enzymes has recently been proposed by Thompson et al. (1975).Working in the same laboratory, Roskoski et al. (1975) have used such columns to purify human placental ChAc. Unfortunately their results were very poor. On the other hand, Hersh has obtained extremely good results starting from the same source of enzyme and using the same columns. ChAc bound to the column was eluted by a gradient of acetyl-CoA. The purification factor was more than I00 (Hersh, personal communication) .' ''

For an insight into the mechanism of dextrdn blue columns, see Section VI I I, E below.

CHOLINE ACETYLTRANSFERASE

303

F. FUTUREPURIFICATION For numerous applications, it is very important to obtain a pure enzyme from any source. Unfortunately, until now all the preparations presented as pure are still very crude. In the future, the extensive use of ChAc affinity chromatography will produce a large amount of pure enzyme. Nevertheless, the work will be difficult as a tremendous purification factor (more than 100,000) is involved. This work would not be impossible. Very recently Cozzari and Hartman (1977) have prepared from ox striatum an enzyme with a specific activity around 200 pmole/ min per mg. The preparation of' this pure enzyme required 40,000 ox brains. VII. Biophysical Studies

The study of ChAc in terms of physical properties is not extensive. The lack of biophysical insight is in part due to the absense of pure enzyme. Nevertheless the molecular weight of mammalian brain ChAc has often been measured by techniques using only minute quantitites of an impure preparation. The first reported molecular weight was 67,000, and this value was found by submitting rabbit brain ChAc to sedimentation analysis (Bull el al., 1964). By gel filtration, a molecular weight of 62,000 was found for human brain ChAc (White and Wu, 1973b); 50,000 for rat brain ChAc (Potter et nl., 1968), and 65,000 for bovine brain ChAc (Glover and Potter, 1971). An estimate of 100,000 has also been reported for bovine brain ChAc by White and Cavallito (1970a). This higher molecular form was further studied by Chao and Wolfgram (1973) who assumed that the molecular weight they found ( 120,000) was due to the existence of two nonidentical subunits of 5 1,000 and 69,000: later the same authors (Chao and Wolfgram, 1974) found that ammonium sulfate precipitation promoted the formation of aggregates of' approximately 1,600,000 daltons in size. More recently Chao ( 1975) claimed that the molecular weight of ChAc was 84,000 due to the presence of' six identical subunits of 14,000 daltons. This assumption was proposed after SDS-gel electrophoresis of a purified preparation. Confronted by this variety of results, the molecular weight of rat brain ChAc was reinvestigated recently by the author (Rosier, 1 9 7 6 ~ ) . After determination of various biophysical parameters, Stokes radius, diffusion coefficient, density, and sedimentation coefficient, the molecular weight 68,000 was found in using the method developed by Siege1 and Monty (1966). The data presented in Table I V allow the calculation of the frictional ration f/f '.

304

JEAN ROSSIER

TABLE IV BIOPHYSICAL CONSTANTS

Horse liver ADH ChAc Ovalbumin

5 cm"/gm

S x 10''i

Re (nm)

MW

.fY

0.789 0.784 0.774

5.08 4.6 3.55

3.51 3.39 2.75

78,000 68,000 45,000

I .20

~

1.22 1.15

~~

Partial specific volume (5).sedimentation coefficient (S),Stokes radius (Re), molecular weight (MW), and frictional ratio of horse liver A D H , ovalbumin, and ChAc. 3 are values found in this laboratory; other reference values are from Tanford (1961) for ovalbumin and from Green and McKay (1969) for horse liver ADH.

(fro)

flf'"

= R e / q 3 5 M l 4 ~ N(N = Avogadro's number)

This ratio expresses the comparison between the apparent radius of the molecule, the Stokes radius (Re) obtained by gel filtration experiments, and the radius of a hypothetical sphere constructed by taking into account the molecular weight and the density of the molecule. The value found (1.22) is in the range of values found for globular proteins like horse liver ADH and ovalbumin. Therefore, it may be assumed that rat brain ChAc is a globular protein. The existence in vivo of higher molecular forms of ChAc dimers or other aggregates which was proposed by Chao and Wolfgram (1973, 1974) is still in doubt. Only during purification procedures when inadequate treatments are used, such as insufficient resuspension of an ammonium sulfate precipitate or column elution of low protein concentration and in the absence of reducing agent, does the dimer or higher aggregates appear; these forms probably d o not exist in the native state (Rosier, 1976~).Similar conclusions were recently proposed by other investigators (Banns, 1976: Malthe-Sorenssen, 1977). VIII. Mechanism of Action

The mechanism of action of ChAc has been recently reviewed (Haubrich, 1976). Thus, I will abstract only the more recent studies and also present some findings related to the effect of C1- on the enzyme. A. KINETIC MECHANISM

All the recent studies agree that ChAc follows a Theorell-Chance mechanism (see Table V for references). This is a sequential mechanism

305

CHOLINE ACETYLTRANSFERASE

TABLE V KINETICPARAMETERSO F ChAc ~

~~

~

KO,' K,,, Acetyl CoA choline (@)

Rat

19 17.5 18

ox

16

Human placenta

10 20 72 110

Human brain "

11

(@)

39 770 410 800 750 1100 526 3120 510

Mechanism"

References

T.C. T.C. O.M. P. P. T.C. T.C. T.C.

Potter et al. (1968) Kaitd and Coldberg (1969) White and Wu ( 1 9 7 3 ~ ) White and Cavallito (1970b) Glover and Potter (197 1) Schubert (1966) Morris el al. ( 1 97 1) Sastry and Henderson (1972) White and Wu ( 1 9 7 3 ~ )

T.C., Theorell-Chance: P.P., Ping-pong: O.M., ordered mechanism.

where presumably acetyl-CoA first binds to the enzyme. It was proposed by White and Cavallito (1970b) and also more recently by Roskoski (1973) that a stable intermediate acetyl-enzyme may be isolated. On the other hand, Currier and Mautner (1974) were unable to repeat the experiments of Roskoski but instead of using ox brain like Roskoski, they repeated the experiments with invertebrate squid ChAc. Similarly, Malthe-Sbrenssen (1976) was unable to repeat the experiments of Roskoski using purified ox brain ChAc but the enzyme used by MaltheSbrenssen was two orders of magnitude more pure. I n his last report Roskoski (Roskoski et al., 1975) had much less confidence in the existence of a stable acetyl-enzyme intermediate. B. ACTIVE SITE

The existence of an imidazol group in the active site is accepted by all the investigators (White and Cavallito, 1970b; Currier and Mautner, 1974: Roskoski, 1974c; Malthe-Sdrenssen, 1976). The existence of a free SH group in the active site is still a controversial problem (Roskoski, 1974a,b; Currier and Mautner, 1976). It is well known that ChAc is very sensitive to SH reagents (Mannervik and Sorbo, 1970; Currier and Mautner, 1974). It was recently proposed by Currier and Mautner that the mode of action of the SH reagents was indirect. The SH reagents may first react with the product of the ChAc reaction, CoASH forming a mixed CoA disulfide which would be the active compound. For example, a recently introduced SH reagent, methylmethanethiolsulfonate (CH3-S-S03-CH3) will react with CoA and form the mixed disulfide

306

JEAN ROSSIER

CoA-S-S--CH3. This compound is the most potent ChAc inhibitor ever synthesized (see Tables VI and VII). Nevertheless, it seems from the data of Currier and Mautner ( 1976) that methylmethanethiolsulfonate has a direct effect on the enzyme. My view is that ChAc contains free SH groups. T h e purification of ChAc by a SH reagent bound to a column is one of the facts which support this hypothesis (Husain and Mautner, 1973; Chao and Wolfgram, 1973). The localization and the importance of these SH groups will be known only when pure enzyme is available. C. SUBSTRATE SPECIFICITY 1. Choline Choline may be only slightly modified. The two carbon chains cannot be elongated and any substitution on this chain will dramatically decrease the affinity (Hemsworth and Smith, 1970a,b; Currier and Mautner, 1974; Barker and Mittag, 1975). The ammonium group is a prerequisite, but the substitution of one methyl group by an ethyl or a propyl does not change the affinity. The substitution of two methyl groups by ethyl or by cyclization in a pyrrolidinium compound greatly affects the affinity (Barker and Mittag, 1975; Collier et al., 1976). It was reported that triethylcholine, substitution of the three methyl groups by TABLE V1 K I VALUESFOR CoA A N D ANALOG" Compound

K,(W)

CoASH DephosphoCoA 3 '-5'-ADP COA-S-S-CH:j (methyl CoA disulfide) COAS-S-CHZ-CH4 (ethyl CoA disulfide) CoA-SS-CHp-CH?-CH:j (propyl CoA disulfide) Dethio CoA CoA-S-CO-CH,Br (bromoacetyl CoA)

1.Sf> 500" 415"

0 . 1 I' 0.34"

0.3"

2" 0.lC

" All the values presented in this table were obtained with rat or ox brain enzymes. " Rossier el (11. (1977). ' Roskoski (1974).

2:

CHOLINE ACETYLTRANSFERASE

i

2:

7 -

307

308

JEAN

ROSSIER

ethyl, was not a substrate for ChAc (Hemsworth and Smith, 1970a; Prince, 1971); nevertheless, Barker and Mittag (1975) have shown that this compound is a weak ChAc substrate. The substitution of the hydroxyl group by an amino or by a thiol is possible but the acetylation of these compounds will already take place in absence of any enzyme (Currier and Mautner, 1974). The works abstracted here were done in test tubes with an enzyme obtained after tissue homogenization. Recently it was shown that in intact tissues, synaptosomes or sympathetic ganglia, choline analogs which are not ChAc substrate, homocholine, or a poor substrate, triethylcholine, were acetylated at the same rate as choline itself (Barker and Mittag 1976, Collier et al., 1977, Ilson and Collier 1975). These surprising and very important observations are discussed in Section XIII, B.

2. Acetyl-CoA In accord with previous observations on vertebrate and invertebrate ChAc (Berry and Whittaker, 1959; Berman-Reisberg, 1957), I have observed that rat brain ChAc may use propionyl-CoA and butyryl-CoA as alternative substrates. The affinity of ChAc for all the acyl-CoA tested (C:,, C4, Clu,and CI6)is undistinguishable from that of the natural substrate acetyl-CoA (Rossier, 1977). D. INHIBITORS An extensive and excellent review by Haubrich (1976) had been recently focused on the ChAc inhibitors. Here I will discuss only the recent developments of analogs of CoA as ChAc inhibitors. CoA as mentioned earlier is a product of the ChAc reaction, but is also a potent inhibitor. By a structure-activity analysis, the author has found that the 3’-phosphate on the ribose ring of the CoA is a prerequisite for any activity. The pantetheine part of the molecule seems less important; 4-phosphopanthetheine has no inhibitory properties (Rossier, 1977). The nucleotide part of the molecule is much more important; 3’5‘-ADP is a ChAc inhibitor. The most potent inhibitors are the newly developed mixed CoA disulfides (Currier and Mautner, 1976). These compounds have more affinity for ChAc than acetyl-CoA itself (Rossier, 1977, and Table VI). Of course the analogs of CoA described here are not specific for ChAc and may also inhibit other enzymes using CoA or acetyl-CoA. Nevertheless, these compounds are pure competitive inhibitors. This contrasts with the inhibitors described earlier, stryrylpyridines, halogenated acetylcholine, acryloylcholine, bromoacetonyl-

CHOLINE ACETYLTRANSFERASE

309

trimethylammonium, and juglone, which are noncompetitive or irreversible inhibitors. Also their specificity for ChAc is not established (Goldberg et al., 1971). Another deficiency of all the ChAc inhibitors is their lack of activity when used for in 7 k m studies (Hebb, 1972; Carson et al., 1976).

E. DEXTRAN BLUE:A COMPETITIVE INHIBITOR WITH RESPECT TO ACETYL-COA. It has already been mentioned in Section VI that dextran blue coupled to Sepharose was used with success to purify ChAc by affinity chromatography. These observations give new insight to the nature of the active site of ChAc. It was recently observed that most of the nucleotide metabolizing enzymes contained a common structural domain, “the nucleotide fold.” Amino acid sequencing has shown that the primary structure of this common structural domain was not changed by evolution and may have evolved from a common ancestor. One of the characteristics of this “nucleotide fold” is its high hydrophobicity (Rossman et al., 1974). ChAc may be considered as a nucleotide metabolizing enzyme; one of its substrates is acetyl-CoA, which contains a nucleotide moiety. Dextran blue is commonly used to evaluate the void volume of gel filtration columns. This molecule is composed of one chromophore, reactive blue, covalently bound to dextran, a polysaccharide. It was observed that all the enzymes containing the so-called “nucleotide fold” were inhibited by dextran blue (Thompson et al., 1975). Following the amazing work of Roskoski and co-workers (1975), I have observed that dextran blue is a potent ChAc competitive inhibitor with respect to acetyl-CoA. The K i value (0.05 pl4) is more than one order of magnitude lower than the K , of acetyl-CoA. This may indicate that ChAc contains this common structural domain, the “nucleotide fold.” This part of the ChAc molecule will be involved in the binding of the nucleotide containing coenzymes: acetyl-CoA and CoA. The importance of the nucleotide moiety in the binding of acetyl CoA by ChAc agrees with the observation that 5’-AMP was a ChAc inhibitor ( K i = 2500 pkl), and that, on the other hand, 4’phosphopantetheine did not inhibit, even slightly, ChAc at concentrations u p to 1 mM (Rossier, 1977). It has already been mentioned that the active site of ChAc contains an imidazol residue. From data presented here it seems possible to propose that acetyl-CoA will bind through its adenosine moiety in a hydrophobic nucleotide binding site close to the imidazol residue.

310

JEAN ROSSIER

F. REGULATION

Two authors (Browning, 1976; Haubrich, 1976) have recently reviewed the regulation of Ach synthesis by ChAc. Haubrich (1976) has proposed that the amount of active enzyme available for Ach synthesis may vary. Under normal conditions, part of the enzyme present in terminals would be in an inactive form. The regulation mechanism would be an increase or a decrease of the active enzyme pool in some unknown way. A more conventional mechanism would be the regulation by mass action. The reaction catalyzed by ChAc is reversible (Potter et al., 1968; Glover and Potter, 197 1 ; Pieklik and Guynn, 1975). It was suggested that in vivo the amount of Ach synthesized depends on the equilibrium constant of the reaction K = (Ach) (Coa)/(choline) (acetyl-CoA). Pieklik and Guynn have determined a value of K = 12. The low value of K shows the complete reversibility of the ChAc reaction. Therefore an increase of Ach would stimulate the back-reaction in which Ach is converted to choline and acetyl-CoA. It will be shown in the following paragraph that C1- activates ChAc and suppresses the inhibition promoted by Ach. These recent findings suggest that a more efficient mechanism than mass action regulates ChAc activity. G. EFFECTOF C1- IONS

ON

CHAC

Since rhe discovery of ChAc by Nachmansohn and Machado (1943), numerous reports have described that ChAc is inhibited by Ach (White and Wu, 1973c; Kaita and Golberg, 1969; Morris et al., 1971) and is activated by salts (Morris et al., 197 1: Potter et al., 1968; Kuczenski et al., 1975; Schuberth, 1966; McCaman and Hunt, 1965; Prince and Hide, 1971) but, as these experiments did not show large effects, their physiological importance was largely neglected. Thus it is usual to describe ChAc as a poorly regulated enzyme (Dowdall et al., 1976). On the contrary, recent studies performed by the author indicate that ChAc velocity is regulated by the concentration of anions and by its own product Ach (Spantidakis et al., 1976: Rossier et al., 1977). Figure 3 shows the effect of the ionic strength on the maximum velocity (VmaX)of the rat brain enzyme. It is clear that C1- is more effective than PO,’. Table VIII presents results obtained with a crude enzyme tested at low ionic strength (5 mM Tris buffer alone) and physiological ionic strength (buffer 145 mM NaCl). V,,, is 25 times greater in the physiological solution than in low ionic strength buffer. Also, at physiological

+

31 1

CHOLINE ACETYLTRANSFERASE

CI -

IONIC STRENGTH ( M )

FIG. 3. Effects of ionic strength on ChAc V,,,. The ionic strength was determined by using the relationship: x = Yl 2 C ; Z ; where C; is the ion concentration and 2 , is the ion valence (X-X for Tris-HCI and O - - - O for Tris-HsP04).

ionic strength, acetyl-CoA K , is increased four times and choline K , 24 times above the values at low ionic strength. 1. Regulation .f ChAc by Its Own Product Ach

Ki values for ACh are included in Table VIII. At both high and low ionic strength, ACh is a pure competitive inhibitor with respect to choline and noncompetitive with respect to acetyl-CoA. The striking difference of more than t w o orders of magnitude in K i values at low and physiological ionic strength may be important in terms of Ach regulation in vivo. At low ionic strength, the Ach K i value (0.310 mM) is well below the Ach concentration of the presynaptic terminals which has been calcuTABLE V I I I KINETICPARAMETERS OF RAT BRAIN ChAc" Assay medium (pH 7.2)

Tris acetate 5 m M V,,, (arbitrary units) K, acetyl CoA K,,, choline K , acetylcholine

(a) (a) (a)

1 0.8 22 310

Tris acetate 5 m M + NaCl 145 m M

25 3.5 540 45,000

'' Synaptosomes of rat caudate nucleus were hypoosmotically shocked. T h e dialysed 105,000 g supernatant was used as ChAc source and diluted to the final ion concentration just before assay.

312

JEAN ROSSIER

lated to be as high as 27 mM (Dunant et al., 1974). Such a high Ach concentration will completely inhibit ChAc at low ionic strength, but in the presence of 145 mM NaC1, ChAc inhibition by Ach will become negligible, as in this latter medium the Ach K i value is 45 mM.

2. Effects of Various Salts: Spec@ Efect of C1- Ions In order to determine if only the ionic strength or a specific cation or anion is involved in the phenomenon described, various salts were tested for their effects on enzyme velocity. The results are shown in Table IX, which presents the data for enzyme activation by sodium or chloride salts. The final concentration of Na+ or C1- was adjusted to 145 mM to maintain the ionic strength as constant as possible. It was found that C1more effectively activates ChAc than any other anion tested. The effectiveness of C1- ions over all other halides is particularly striking, for in most systems F- precedes C1- in order of activity. Cation concentration seems not to be involved in the activation phenomenon, as at fixed C1concentration substitution of various monovalent (K+, Tris’, Li+)or divalent (Ca”, Mg”) cations for Na+ did not change ChAc velocity. The physiological importance of the possible regulation of ChAc by C1- is presented in Sections XIV and XV. IX. Axonal Transport of ChAc

The presence of ChAc in axons was one of the experimental facts which was used by Nachmansohn (1959) to defend his unitarist theory of the axonal conduction. Soon after the presentation of this theory, Hebb, who did not believe it (Hebb, 1957), proposed that the presence of ChAc in axons was due to axonal transport (Hebb and Waites, 1956). Weiss and Hiscoe (1948) had, several years before, presented evidence of the existence of axonal transport. I t is well known that all axonal and synaptic proteins are synthesized in the perikaryon and transported by two components of axoplasmic transport, a ‘‘slow’’component with an average velocity of 1-3 m d d a y and a “fast” component with a maximum velocity of 400-500 mm/day (Lasek, 1970). ChAc is transported by the slow component (see Table X for references). Nevertheless, recent experiments by Fonnum ( 1975b) may suggest that ChAc is transported by the fast component. The recent observation by Jablecki and Brimijoin (1975) that the velocity of ChAc transport decreases with aging is in accord with the previous observations of Droz (1965) that the velocity of the slow component of axonal transport declines with increased age. Jablecki and Brimijoin (1974) had also observed that ChAc transport was markedly

313

CHOLINE ACETYLTRANSFERASE

TABLE IX EFFECTSOF VARIOUSSALTSON ACTIVATION OF ChAc VELOCITY" Sodium salt

7r Control

Chloride Salt

7( Control

NaCl (control) N aF NaBr Nal Na acetate Na, tartrate Na:{citrate Na2S04 N a2H PO1

100 64 86 22 71 61 51 59 36

NaCl Tris-CI LiCl KCI MgCI, CaCb

I00 91 91 100 86 100

,

" C r u d e enzyme was diluted in 5 m M tris acetate buffer, pH 7.2, supplemented with various salts. Values are expressed as 7r activity in the presence of 145 rnM NaCI. All salts were adjusted to a final concentration of 145 rnM Na' or CI-.

decreased in murine dystrophy. After nerve transection, the rate of ChAc transport is not affected in the axons disconnected from their cell bodies (Ekstrom and Emmelin, 1971) as well as in proximal nerve fragments (Jablecki and Brimijoin, 1975). The transport of ChAc is blocked by injection of colchicine (Fonnumet al., 1973). All the results abstracted here were obtained by nerve ligation and the rate of transport was calculated from the kinetics of the ChAc accumulation proximal to the ligature.

TABLE X AXONALTRANSPORT OF ChAc

Species

Nerve

Sciatic Sciatic Vagus H ypoglosis Rabbit Sciatic Cat Ventral roots Mouse Sciatic 2 weeks old 8 weeks old 36 weeks old Rat Rat Rabbit

Velocity (rnrn/day) 5 3 13 3 4 10 3.5 1.1

0.2

References Oesch Pt al. (1973) Wooten and Coyle (1973) Fonnum et al. (1973) Fonnum et nl. (1973) Tucek (1975) Carlsson rt 01. ( 197 1 ) Jablecki and Brimijoin (1975) Jablecki and BrimiLjoin (1975) Jablecki and Brimijoin (1975)

314

JEAN ROSSIER

Some authors have proposed the use of two ligatures. In using this protocol. Fonnum has shown that a considerable part of ChAc stays immobile between the two ligatures. Therefore, if in nonexperimental conditions a considerable part of ChAc does not move, the rates of transport previously calculated were grossly underestimated (Fonnum et al., 1973). Fonnum (1975b) has proposed that the rate of ChAc transport would be between the fast and the slow components previously described. The existence of a faster ChAc transport would be, in the mind of Fonnum (1975b), the only explanation of the observation that reserpine increased by 50% within 24 hours, ChAc levels in the preganglionic terminals of the cervical ganglion (Oesch and Thoenen, 1973; Oesch, 1974). A slow axonal transport of ChAc would not be fast enough to promote this rapid increase. A fast ChAc transport is not the only explanation of the experiments of Oesch and Thoenen, as it may be proposed that the high dose of reserpine induced the neosynthesis of ChAc in the postganglionic perikaryons. It was already shown that massive doses of reserpine induce the synthesis of a variety of proteins and enzymes in the postganglionic cells. From the data obtained in axonal transport, the half-life of ChAc was calculated to be around 12-20 days (Fonnum, 1975b). This long lifespan of ChAc almost excludes a release of this enzyme on stimulation. In conclusion, rate of axonal transport and half-life of ChAc are still to be accurately determined in avoiding such procedures as the use of ligatures. The use of a monospecific antiserum against ChAc will overcome many artifacts. After a pulse of radioactive amino acids in a cholinergic nucleus, the newly synthesized Ch Ac will become radiolabeled. The isolation of this radiolabeled ChAc will be possible by immunoprecipitation. The kinetics of the appearance and disappearance at synaptic terminals of the radiolabeled Ch Ac precipitable by its antiserum will give an accurate measure of ChAc transport and turnover.

X. Immunology

Antisera allow immunological studies on the phylogeny and the structural features of an enzyme. Moreover, a monospecific antiserum against ChAc will eventually allow studies of turnover, axonal transport (avoiding nerve ligature), and immunohistochemical localization of this enzyme. Unfortunately, the production of a monospecific antiserum, i.e., an antiserum directed only against ChAc, has not yet been achieved. Readers must bear in mind that the production of a monospecific antiserum depends on the availability of a pure antigen. Pure antigen is not yet available as ChAc has never been purified to homogeneity.

CHOLINE ACETYLTRANSFERASE

315

A. ANTIGENICITY OF ChAc For a long time it was generally believed that ChAc was a poorly immunogenic protein. Most workers involved in ChAc purification have tried to obtain antibodies. Their results were negative at first (MaltheSdrenssen et al., 1973; Rossier et al., 1973a). In addition to these negative reports demonstrating the low antigenicity of rat brain ChaC, it had also been difficult to produce the first active antiserum against ChAc (Rossier et al., 1973b). Repeated injection for 14 months, a procedure quite unusual in immunology, was necessary to elicit in a rabbit a response against rat brain ChAc. Since this early report, numerous workers have recently obtained active antisera against ChAc from various sources (Shuster and O'Toole, 1974; Singh and McGeer, 1974a; Eng et nl., 1974; Malthe-Sdrenssen, 1975; Rossier, 1976b; Polsky and Shuster, 1976b). By injecting ChAc preparation from man or beef it was relatively easy to obtain antibodies against ChAc (Singh and McGeer, 1974a; Eng et al., 1974; MaltheSdrenssen, 1975). The injection of rat and mouse enzyme was less effective. Shuster and O'Toole (1 974) required a long immuniz.ation procedure to obtain a response against ChAc from mice, and out of eight rabbits injected only two gave a positive response. I had rather similar results with rat brain ChAc. Out of 30 animals injected, only four rabbits gave a positive response (Rossier, 1976b). From these observations, it may be concluded that ChAc is not as previously stated a nonimmunogenic protein (Malthe-Sdrenssen et d., 1973; Rossier et al., 1973b). On the contrary, it must be taken into account that when investigators injected into rabbits 1 or 2 mg o f a purified ChAc preparation, they injected a crude preparation. Table XI shows the type of preparations used by various investigators. In the best cases, ChAc represented 1/50th of the injected protein if it is assumed that pure ChAc will have a specific activity around 100 pmoles/minute per mg of protein. Therefore, rabbits immunized had received a maximum of 40 pg of real ChAc protein per injection time. Such a small amount of protein was enough to induce antibodies which inhibited and precipitated ChAc. Nevertheless, a variation in mammals seems to exist: mouse and rat ChAc are less immunogenic than enzyme from other soiirces.

B. CROSS-REACTIVITY Antisera against ChAc are very similar to other enzyme antisera. They inhibit the enzymatic activity and may precipitate an enzymeantibody complex which still possesses a residual enzymatic activity. By using appropriate conditions, it is possible to precipitate all the enzymatic activity (Fig. 4).

316

JEAN R O S I E R

TABLE XI CHARACTERISTICS OF THE SERAPRODUCED AGAINSTChAc ChAc used as antigen for immunization Antisera obtained-titer"

Species Mouse cow Human

Rat

ChAc specific activity (pmolelminlmg of protein)

(PI of senim precipitating half Ch Ac activity (0.25 pmolelmin in 50 pl)

Shuster and OToole ( 1974) Eng et al. (1974) Singh and McGeer (1974a) Rossier (1976b).

5

0.07 Data not available 0.012

Half precipitation never achieved 16

0. I

0.4 ~

"

References

~

~~

~~

~~

Lowest is the value: highest is the activity of the serum.

J o ChAc activity

In

t

nmoles/mln

FIG. 4. Evidence of a precipitable enzyme-antibody complex. Variable amounts of purified ChAc were added to 20 ~l of antiserum AS 21. T h e final volume was adjusted to 100 pl with 50 mM sodlum phosphate buffer pH 7.2 containing 0.2 M NaCI, 1 mM 1-10 phenanthroline HCI, and serum albumin, 2 mglml. After incubation for 30 minutes at Y7"Cand 48 hoursat 4°C themixture wasassayed beforeandaftercentrifugation at 15,000g for 1 0 minutes. On the right of the figure, results are expressed as percent of controls performed with preimmune serum. The pellets were washed with 500 pI of the described medium and after another centrifugation, resuspended in 100 p1 of the above medium. Results of these assays are expressed as nanomoles per minute of ChAc (A-A) still present in the washed pellet (scale on the left of the figure).

317

CHOLINE ACETYLTRANSFERASE

Antisera against ChAc were used to study organ and species specificity. ChAc extracts from various origins were compared in immunoprecipitation tests (Fig. 5 ) . Reactivity of the antiserum was observed with all vertebrate ChAc extracts. However, no reactivity was seen with squid optic lobe ChAc. Also the sera produced against squid ChAc did not react with rat brain ChAc (Rossier, unpublished data). The lack of cross-reactivity does not prove that ChAc from vertebrates and ChAc from invertebrates are two totally different molecules. There may still be some identity between these two molecules that is not revealed by the antisera now available. It is possible that the antisera produced were not potent enough to demonstrate any relationship between the two molecules. For example, in a study with a less potent serum, the crossreactivity was demonstrated only in mammals (Singh and McGeer, 1974b; Rossier et al., 1973b). Now, with a more active antiserum crossreactivity is demonstrated in all vertebrates. Still it must be noted that the best serum available has a relatively low titer, 0.1 mg/ml (Rossier, 1976b). For example, a good antiserum would have a titer (milligrams of im-

1

I

1/2008

1/512

I

'11128

1/32

118

112

SERUM DILUTION

FIG.5 . Cross-reactivity of ChAc. In test tubes, various amounts of ChAc antiserum were incubated for 30 minutes at 37°C and for 48 hours at 4°C in a final volume o f 5 0 p1 with 250 pmoles/min of ChAc from a 105,000g supernatant of rat brain, chick ciliary ganglia, or torpedo electric organ crude extract. After centrifugation at 15,000 g for 15 minutes, supernatant was assayed for Ch Ac residual activity. Controls were done with preinimune serum.

318

JEAN ROSSIEH

munoglobulins per milliliter of a serum precipitating with the antigen at the equivalence point) of 2 mg/ml. The organ specificity study revealed that ChAc was identical in all rat nervous tissue extract tested, that is, brain, sympathetic ganglia, and neuromuscular junction. Also newborn rat brain ChAc and adult rat brain ChAc had the same reactivity. C. FUTUREOF THE IMMUNOHISTOCHEMICAL LOCALIZATION OF ChAc The wide cross-reactivity allows optimistic views on the future of ChAc studies using immunological methods: Antiserum produced against enzyme isolated from rich and easily available sources such as the electric organ of the torpedo might be used in all studies related to ChAc in vertebrates. In this field when workers began to produce an antiserum against ChAc they had in mind its use for immunohistochemical localization of ChAc. This project should benefit from two favorable facts. First the heat-inactivated enzyme has the same immunological reactivity as untreated ChAc (Rossier et al., 1973b). Second, ChAc treated by fixation is still active (Kasa, 1975). Therefore, the enzyme present in histological sections of fixed and paraffin-embedded material will probably react in situ with the antiserum. It is important to note that fixed and embedded materials permit better morphological studies than cryostat sections which must be used when the immunological reactivity is lost during heat embedding. However, while speculating about the prospects of ChAc immunohistochemical localization, we must recall two fundamental requirements. First, the antiserum used must have a high titer, around 1 mg/ml. Such a high titer antiserum has not yet been obtained. The best antiserum produced has a titer around 1/10 of this requirement. Second, the antiserum must be monospecific, i.e., directed only against ChAc. This seems not to be the case for all sera produced recently. Therefore, at the moment, reports claiming the immunohistochemical localization of Ch Ac seem not to be reliable (Eng et al., 1974; McGeer et al., 1974; Rossier, 1975). On the other hand, the very recent report of Cozzari and Hartman (1977) appears of very high quality. These authors have prepared a monospecific antiserum by injecting a pure preparation of ox brain ChAc. The antiserum was used for immunohistochemical localization. XI. Localization at the Cellular level

In the preceding section we have emphasized the future of- immunohistochemical localization of ChAc. Why? Because for the moment

C HOLI NE ACETY LTRANSFERASE

319

no reliable technique to histochemically visualize the enzymatic activity exists. We will abstract here some of the most interesting work done recently to localize ChAc activity. A. CLASSICAL HISTOCHEMISTRY In 1970, simultaneously and separately two workers, Kasa and Burt (Burt, 1970; Kasa et nZ., 1970), described an original method for visualizing ChAc histochemically. These two workers presented very similar techniques: the CoASH, which is released during enzymatic reaction, is precipitated in sitir as a Pb mercaptide by incubating the preparation in the presence of lead nitrate. This precipitate may be directly observed in electron microscopy. To be visualized by microscopy, the preparation is treated with ammonium sulfide and a brown precipitate of lead sulfide appears. This seems to be a good method, but in an excellent analysis Burt, one of the authors of this technique, apologized for the low reliability of the technique (Burt and Silver, 1973b). The main substrate of the reaction, acetyl-CoA, may be involved in numerous other enzymatic reactions which yield CoA as a product. To lower the acetyl-CoA deacetylation unrelated to ChAc activity, it appeared worthwhile to incubate the preparation in the presence of diisopropylphosphofluoridate (DFP) (Kasa, 1975; Burt and Silver, 1 9 7 3 ~ )This . procedure was able to reduce some of the nonspecific background. However, the entire histochemical procedure has a low specificity as little change in the staining intensity was observed when the second substrate choline was omitted. More recently, Lebbin and Waser (1 975) have, in an extensive study, compared almost all the methods proposed. They were unable to obtain. any histochemical localization of ChAc by repeating as exactly as possible the protocols originally proposed by Hebb et al. (1970), Kasa (1 975), Kim (1972), and Burt (1970). Only the method proposed by Burt and Silver (1973b) was reproducible. Nevertheless, the method of Burt and Kasa has been applied to some parts of the central nervous system. Heavy precipitates are formed in motoneurons of the spinal cord ventral horn, and few precipitates are seen in cerebellar slices. These results agree with our knowledge of the distribution of ChAc activity in the central nervous system. A new cytochemical technique was very recently proposed by Feigenson and Barrnett (1977). This technique, which may be used only at the electron microscope level, was used to study the fine structural localization of ChAc at newt myoneural junction. Fixed tissues were incubated with ChAc substrates in the presence of manganous chloride and potassium ferricyanide. CoA, the product of, the Ch Ac reaction, reduced ferricyanide to ferrocyanide, which precipitated with Mn'+ to yield an

320

JEAN ROSSIER

electron-dense marker at the site of ChAc activity. The final product was associated with membranes of synaptic vesicles and to a lesser extent with invaginations of presynaptic membranes. Schawnn process, junctional folds, muscle fibers and sarcoplasm, and the presynaptic sarcoplasm were devoid of reaction product. Nevertheless, some reaction product appeared in the absence of added substrates. This effect was also observed by Kasa and Burt, and may be due to endogenous substrate. DFP was not used in this technique, although DFP is a prerequisite in the method of Kasa and Burt. Feigenson and Barrnett have taken great care to study the specificity of their proposed method. However, some reaction product was still formed when slices were incubated with ChAc inhibitors. The low concentration of inhibitors used would be an explanation of this remaining activity. In unpublished observations Feigenson and Barrnett were able to repeat their method on rat brain. The same localization as in newt myoneural junction was found, i.e., association with membranes of vesicles and with presynaptic membranes. The association of ChAc with membranes suggested by these cytochemical experiments would agree with the hypothesis of the existence of a membrane multienzymatic complex involved in the synthesis of Ach (see Section XIII: for details). B. ChAc

AND

CHOLINERCIC PATHWAYS

Actually, to trace cholinergic pathways, neuroanatomists use old, classical techniques. They interrupt the possible cholinergic pathway by a surgical procedure and follow the disappearance of Ch Ac enzymatic acti,vity in the presumed neuronal projections. Using this technique, two well-known cholinergic pathways have been described. One pathway passes through the septum and makes synaptic contacts in the hippocampus (Lewis et al., 1967). The other one, recently described, has its origin in the habenula and projects to the interpeduncularis nucleus (Kataoka et al., 1973). We may recall here that this nucleus contains the highest specific activity of ChAc. Lesion and degeneration studies have also been used to prove the cholinergic nature of the crossed and uncrossed olivocochlear bundle (Jasser and Guth, 1973; Godfrey et al., 1976; Fex and Wenthold, 1976). The caudate-putamen nucleus contains a high concentration of ChAc and Ach. To understand from where the caudate nucleus receives its cholinergic inputs is important, as this nucleus is involved in the pathology of Parkinson’s disease. The deafferentation technique was used by McGeer et al. (1971). The ChAc activities in the caudate-putamen nucleus were assayed after lesion of several neostriatal afferent and efferent

CHOLINE ACETYLTRANSFERASE

321

pathways. The absence of alterations in ChAc activities led to the conclusion that the cholinergic neurons of the caudate-putamen nucleus may be organized exclusively within the nucleus. Therefore, the high content of ChAc in the caudate nucleus is due only to numerous cholinergic interneurons and not to a massive cholinergic projection to or from another nucleus. The existence of ChAc in short interneurons inside the caudate-putamen nucleus was confirmed recently (Butcher and Butcher, 1974).

C. ChAc

VERSUS

ACETYLCHOLINESTERASE LOCALIZATION

Histochemical detection of acetylcholinesterase (AchE), a technique discovered by Koelle, was used as far back as 1954 in order to map cholinergic pathways in the rat central nervous system (Koelle, 1954). T h e general view is that this technique has some limitations and the results obtained are open to criticism. AchE is mainly a postsynaptic enzyme and its distribution is more ubiquitous than the ChAc distribution. For example, glial cells and erythrocytes contain large amounts of AchE and do not contain any ChAc. Furthermore, the distribution of AchE does not follow the concentration of the other cholinergic synaptic markers, i.e., Ach and ChAc. This was well documented 20 years ago by Hebb and Silver (1956). T h e main example of this different distribution is the cerebellum in which a high level of AchE is found in contrast with a low level of ChAc and Ach. This was further documented recently by Brownstein et al. (1975). These authors, in numerous reports, described the value of ChAc activity in hundreds of separate nuclei and parts of the rat central nervous system. The minute amounts of tissue used were obtained by microdissection of frozen brain slices. Values of ChAc activity were compared with the intensity of the histochemical localization of AchE. Data revealed generally a good similarity between the two markers. For example, the highest AchE activity has been reported in cranial motor nuclei, which also contain a high amount of ChAc; ChAc and AchE activities are both high in the caudate-putamen. Nevertheless, certain differences between the distributions of ChAc and AchE are notable. Brownstein et al. again found low cerebellar ChAc levels in contrast with high AchE. Other regions such as the zona incerta, Fore1 field and the Raphe nuclei also present this dissimilarity. These regions stained intensely for AchE in contrast with a low level of ChAc (Kobayashi et al., 1975; Brownstein et ul., 1975; Palkovits et nl., 1974). From the enumeration of the dissociation between AchE staining and ChAc activity it may first appear that AchE is not a reliable technique to trace cholinergic pathways. Nevertheless, in the absence of any method

322

JEAN ROSSIER

to visualize ChAc, I think that still the AchE staining in combination with the method of making small stereotaxic lesions as proposed by Lewis and Shute (1963) is the best technique. Certainly we may believe the assumption of Lewis et al. (1967) that any central axon, which after interruption exhibits buildup of AchE on the cell body side and a loss of AchE staining on the synaptic side is a cholinergic axon. However, the nigroneostriatal dopamine pathway would be one of the exceptions to this rule (Butcher, 1977). This pathway, also named the dopaminelAchE pathway, is rendered highly visible by AchE histochemical staining. After lesion, a pileup of AchE is noticed on the body side. Although several interpretations have been proposed, the role of AchE on these dopaminergic axons is still unclear. In conclusion, at this time, the papers of Lewis and Shute (1967) and of Shute and Lewis (1967) are still the best sources dealing with the anatomy of the cholinergic pathways in the brain. D. ChAc ASSAYS I N SINGLE CELLSA N D

IN

DISCRETE AREAS

In the previous paragraphs I have referred to works devoted to the histochemical localization of ChAc by assaying discrete area of the brain. It must be noted that all these assays of ChAc in discrete areas were initiated as far back as 1967 by preliminary works of Giacobini. Giacobini was the first to describe a microdevice to dissect and assay ChAc in single sympathetic cells. In 1970, the Kjeller (Norway) group assayed ChAc and AchE in several layers of rat hippocampus. ChAc as well as AchE was concentrated in discrete layers both in hippocampus regio superior and in area dentata. The highest enzyme activities were found in zones close to the layers of pyramidal and granular cells (Fonnum, 1970a; Storm-Mathisen, 1970). These methods combining microdissection and microassay have been applied to the mapping of individual neurons in various invertebrate ganglia, mainly aplysia, snail, and lobster (McCaman and Dewhurst, 1970; Giller and Schwartz, 1971; Hildebrand et al., 1974; Emson and Fonnum, 1974). By a similar approach the regional distribution of ChAc in various nuclei of rat hypothalamus was recently studied. The distribution was different from that of norepinephrine, dopamine, and serotonin (Uchimura et al., 1975). Godfrey and Matschinsky (1975) have studied ChAc distribution in the pancreas. They found that ChAc activity was one order of magnitude higher in islets of Langerhans than in exocrine tissue. Therefore it may be proposed that the islets of Langerhans receive the major part of the cholinergic innervation of the pancreas. It has been proposed by Lam (1972) that photoreceptors of the retina

CHOLINE ACETYLTRANSFERASE

323

were cholinergic cells. On the contrary Ross et al. ( 1 975) recently showed that mutant mice lacking photoreceptor cells still had around the same ChAc levels in the retina as controls. Ross and MacDougal (1976) have also completed an extensive study related to the distribution of ChAc in vertebrate retina. In all the animals, the highest activity was in the inner plexiform layer, intermediate activity in the inner nuclear and ganglion cell layers, and very low activity in the photoreceptor and outer plexiform layers and optic nerve. A species difference of almost two orders of magnitude was found between the lowest (cat) and highest (goldfish) values for ChAc activity in the inner plexiform layer. Other data suggest that cholinergic neurons in retina are to be found predominantly among the amacrine cell types and that not all amacrine cells will be found to be cholinergic. E. HISTOCHEMICAL LOCALIZATION OF ChAc: FUTURE I have already discussed the future of the immunohistochemistry of ChAc. This technique will be useful only when pure antigen becomes available. Meanwhile other approaches may be tired. T h e low reliability of the histochemical method proposed by Burt and Kasa is in part due to the use of lead sulfide as a final chromophore. Lead may react with all reduced sulfide and not only with the product of the reaction, CoASH. Substitution of acetyl-CoA by acetyletheno-CoA in this method will overcome this artifact as the fluorophore visualized by fluorescent microscopy will be the direct product of the reaction, etheno-CoASH, which is highly fluorescent (Secrist et al., 1972) (Fig. 6). I have already observed that

O=P-0I

0-

Hy-OH

oc/H ,. I

fHZ

CHz

I

o//

C ‘NH-CH,--CH,--SH

FIG. 6. Etheno-coen-/yme A (etheno-CoA).

324

JEAN ROSSIER

acetylentheno-CoA was a substrate for ChAc. Unfortunately, during the chemical synthesis of acetyletheno-CoA, a contaminant compound which is an inhibitor of the ChAc reaction is also formed (Rosier, 1977). Nevertheless, once clean acetyletheno-CoA is available the method propose'd will be useful. Until now radio autographic study of the cholinergic system has never been performed. This is in contrast with the success of such methods in the catecholaminergic system. The explanation is simple: The catecholamines have a free amino group. This function may react with the fixative glutaraldehyde and therefore the radioactive catecholamines become linked to the protein matrix. Such a free amino group does not exist on the Ach molecule. Therefore, freeze drying autoradiography must be used. This technique has already given very nice results for the localization of Ach synthesis in the retina (Baughman et al., 1976). Another possibility will be to use radiolabeled choline analogs with a free amino group. For example, substitution of one methyl group by methylamine, ethylamine, or propylamine will give suitable analogs. XII. Subcellular localization of ChAc

Since the work of Fonnum it has been generally accepted that ChAc is a soluble enzyme with an even distribution throughout cholinergic nerve endings, but recent data on the high affinity choline uptake system have reopened the debate. Is ChAc really a soluble enzyme? To understand this problem more easily an historical survey will be useful. The work of Hebb and Smallman (1956), showed that homogenization of brain tissue in isoosmotic sucrose gives rise to a soluble and a particulate form of ChAc. The soluble fraction of the enzyme was recovered in the high-speed supernatant and was probably derived from cholinergic axons and perikarya. The particulate form was found associated with the mitochondria1 pellet. In fact, ChAc was not associated with mitochondria, but with particles sedimenting with mitochondria, named by Whittaker as synaptosomes (Whittakeret d.,1964). Synaptosomes are resealed nerve endings. During homogenization, nerve endings are cut from their axons, and in isoosmotic sucrose they may reseal into particles of around 0.5 pm in diameter. Synaptosomes contain all the enzymatical apparatus used to synthesize and store the neurotransmitter. The compartmentation of ChAc within synaptosomes has been a very controversial subject, especially between two famous neurobiologists, Whittaker in Cambridge and DeRobertis in Buenos Aires.

CHOLINE ACETYLTRANSFERASE

325

In 1963, DeRobertis et al. separated the constituents of rat cerebral synaptosomes. The synaptosomes were hypoosmotically disrupted and subsequently submitted to differential centrifugation. ChAc and Ach was found to be associated with fractions also containing the vesicles. In contrast, AchE was found mainly in membrane and microsomal fractions. DeRobertis concluded that ChAc was firmly bound to the vesicles, probably contributing a part of the vesicle membrane (DeRobertis et al., 1963). On the other hand, Whittaker et al. (1964) separated the constituents of disrupted synaptosomes from guinea pig cortex by density-gradient centrifugation. They found that Ach, ChAc, and AchE showed different distributions. Only Ach was found to be associated with the vesicle fractions. ChAc was soluble and AchE was present in fractions containing membrane fragments. The two groups, who were working on different animals, rat and guinea pigs, and using different centrifugation procedures, agreed in principle on the localization of Ach and AchE. They completely differed in their opinions on ChAc localization. It must be noted that the presence of ChAc in vesicles was an expected result. I f ChAc is a constituent of the vesicle membrane, synthesis and uptake of Ach by vesicles would appear to be linked mechanisms which may explain why isolated vesicles are unable to uptake preformed Ach. Whittaker et al. (1964) had data disputing this hypothesis. As an explanation of the results of DeRobertis, they claimed that DeRobertis’s procedure, i.e., differential centrifugation, led to gross contamination of the vesicular fraction with partially disrupted synaptosomes and that ChAc activity in this fraction was due to the latter. One year later, as a direct response, a research report from DeRobertis’ group claimed that insufficient unmasking of latent ChAc activity could explain the lack of ChAc activity in the vesicular fraction of Whittaker’s procedure. But this paper of DeRobertis’s group, written by McCaman et al. (1965), announced the end of the dispute. These authors made the important observation that large species differences occurred. They concluded that ChAc was firmly bound to vesicles in rat and rabbit, less strongly bound in guinea pig, and that in pigeon the enzyme was soluble. This key observation was further documented by Fonnum, working in Whittaker’s laboratory. To separate constituents of h ypoosmotical disrupted synaptosomes, Fonnum used density-gradient centrifugation, as done previously in the original Whittaker procedure, but he did not use guinea pig cortex as starting material. To prepare the synaptosomes he used rat brain, as did DeRobertis. The results showed that the fraction rich in synaptic vesicles contained a high Ach concentration but low

326

JEAN ROSSIER

ChAc activity. ChAc displayed .a bimodal distribution in the gradient, being partly soluble and partly localized to a heavy membrane fraction. The binding of ChAc to membrane was demonstrated to be a reversible process primarily dependent on pH and ionic strength. At low pH and low ionic strength, all the enzyme is bound to membrane. In increasing pH and ionic strength, ChAc lost its binding properties and became soluble. Species differences are very important. As stressed before by McCaman et al., ChAc from pigeon and guinea pig brain was more easily solubilized than the enzyme from rat, rabbit, and cat brain (Fonnum, 1967, 1968, i970b). The reversible binding of rat ChAc to membranes suggested that the enzyme has a positive charge and is therefore attracted by negatively charged membranes. It is worthwhile to note that rat ChAc may also be associated with negatively charged proteins such as tubulin (Fonnum and Malthe-Sdrenssen, 1973). Isoelectrofocusing experiments demonstrated that rat brain ChAc is a basic protein which exists in multiple forms. Isoelectrofocusing was also performed on the enzymes isolated from other species. It appeared that pigeon and guinea pig enzymes have acidic PI, whereas rabbit, rat, and cat enzymes have basic PI (Malthe-SBrenssen and Fonnum, 1972) (see Table XII). Eventually, the differences in binding of ChAc between different species were explained by these different charges found on ChAc. The conclusion of these works is that from a biochemical point of view, ChAc appears to be a soluble protein, but recent data on the uptake of choline and its transformation in Ach may reopen the debate. TABLE XI1 POINT(PI)

[SOELECTRIC

Species

PI

Cat Rat Pigeon Guinea pig Monkey Mouse Rabbit Torpedo Horseshoe crab Cockroach Squid

7.0, 7.8, 8.4 7.5, 7.9, 8.4 6.6 6.8 7.0, 7.35, 8.35 7 . 1 , 7.5, 8.4 6.9 6.6 5.3 5.0 5.0-6.3

OF

ChAc References

Malthe-Sbrenssen and Fonnum (1972) Malthe-Sbrenssen and Fonnum (1972) Malthe-Shrenssen and Fonnum (1972) Malthe-Sbrenssen and Fonnum (1972) Malthe-Sdrenssen (1976) Malthe-Sbrenssen (1976) Malthe-Sbrenssen (1976) Malthe-Sbrenssen (1976) Emson et 01. (1974) Emson et al. ( I 974) Polsky and Shuster (1976a)

CHOLINE ACETYLTFUNSFEFUSE

327

XIII. Choline Transport and ChAc

Brain tissue is unable to synthesize choline from its iinmethylated precursors, serine and monoaminethanol. Such a conclusion can be drawn from in viva and in vitro experimental studies (Ansell and Spanner, 1967: Browning and Schulman, 1968). Therefore, an external supply of choline, possibly supplied by the bloodstream, is required f o r brain choline metabolism. However, the transport of choline from blood to brain cells would then require a transport mechanism. Choline is a quaternary ammonium compound and, like other quaternary ammoniums, does not diffuse freely through membranes. Therefore, it was of no surprise when, by the end of the 1960s, a carrier system for the transport of choline into cholinergic synapses was demonstrated by several investigators (Marchbanks, 1968; Diamond and Kennedy, 1969: Potter, 1968). This carrier mechanism, demonstrated in nonphysiological medium, had a low affinity (K,,,: 50, IO-"M) for choline. In 1973 several groups simultaneously described a so-called highaffinity choline transport system. This high-affinity choline transport system is also called the Na+-dependent high-affinity choline uptake system (Simon and Kuhar, 1976). The Michaelis-Menten constant of this system, demonstrated on synaptosomes resuspended in a physiological medium, has a value around lo-'' M (Yamamura and Snyder, 1973; Haga and Noda, 1973; Guyenet et al., 1973a). While the low-affinity transport system seems to be a ubiquitous property of all cell membranes (Diamond and Milfay, 1972), the high-affinity choline transport system is specifically associated with cholinergic nerve endings. This was well demonstrated by the elegant experiments done by Kithar et al. (1973). A lesion in the medial septum of the rat was performed. I t was said in a preceding section that the medial septum projects synaptic connections in some of the layers of the hippocampiis. This lesion was followed by a loss of 90% of ChAc in the synaptosomal pellet, indicating a loss of cholinergic terminals. This loss of cholinergic terminals was accompanied by a 75% decrease o f the high-affinity choline transport system. The low-affinity choline transport system was less affected by the lesion. The association of the high-affinity choline transport system with the cholinergic terminals were also substantiated by regional studies. Sorimachi and Kataoka (1975) have shown that brain regions which contain high amounts of Ach and ChAc also contain a high amount of the high-affinity choline transport system. The low-affinity choline transport system does not show this particular distribution.

328

JEAN

ROSSIER

A. COUPLING OF CHOLINE TRANSPORT AND ChAc The possible link between the high-affinity choline transport system and ChAc was mainly substantiated by studies with synaptosomal preparations. In such preparations, the rate of choline uptake and the rate of Ach synthesis are similar. The high-affinity choline uptake system rate may be decreased by lowering the Na’ concentration in the medium. Simultaneously, the Ach synthesis rate is affected by a similar ratio (Guyenet et al., 1973b). Hemicholinium and alkylbisquaternary ammonium compounds are potent inhibitors of the high-affinity choline transport system ( K i = lo-” M). At the same low concentration, these inhibitors also block the Ach synthesis (Haga and Noda, 1973; Guyenet et al., 1973a; Holden et al., 1975). The possible link between uptake of choline and synthesis of Ach was suggested by following the specific radioactivity of choline and Ach during synaptosomal incubation with radiolabeled choline. The newly taken up radiolabeled choline is used without further isotopic dilution for the synthesis of Ach. This suggested that inside synaptosomes no choline pool involved in the synthesis of Ach exists (Guyenet et al., 1975). Similar results were obtained by using radiolabeled analogs of choline (Barker and Mittag, 1975; Zimmerman and Dowdall, 1975; Kilbinger, 1977).

TRANSPORT A N D ChAc: A POSSIBLE MULTIENZYMATIC B. CHOLINE COMPLEX The numerous experimental results abstracted here demonstrate that all choline molecules entering cholinergic terminals through the permease are directly acetylated. On the basis of these observations, Barker and Mittag (1975) and Lefresne et al. (1975) have proposed that ChAc is bound to the transport enzyme, choline permease. Our recent observation that ChAc is a hydrophobic protein (Rossier 1977) as well as the earlier observation that ChAc may be bound to membranes by ionic interactions (Fonnum, 1968) may reinforce the hypothesis that ChAc is a membrane-bound enzyme. Moreover, very recently it was shown by a fine cytochemical technique that ChAc was associated to membranes (Feigenson and Barrnett, 1977) (see Section XI). The most striking demonstration of the existence of a membranebound ChAc part of a multienzymatic complex involved in Ach synthesis was recently performed by Barker and co-workers (Barker and Mittag, 1976: Collier et al., 1977). Radiolabeled homocholine was prepared. This compound was accumulated in synaptosomes by the high-affinity choline uptake system. It has already been stated in this review that homo-

CHOLINE ACETYLTRANSFERASE

329

choline is not a substrate for ChAc (see Section VIII). Nevertheless, the homocholine accumulated by intact synaptosomes via the highaffinity uptake process was acetylated. Homocholine was also acetylated in the superior cervical ganglion of the cat, and the amount of acetylhomocholine formed was increased by preganglionic nerve stimulation. In ganglia, acetylhomocholine was available for release by preganglionic nerve impulses, and its release was Ca’+-dependent. These experiments show that homocholine can form a cholinergic false transmitter, and that the substrate specificity of ChAc in uitro might be different from that in situ. Already several investigators have shown that choline uptake and the choline acetylation process are in some way coupled. The demonstration of the acetylation of homocholine in situ indicates that the coupling of ChAc to the high-affinity choline uptake system transforms the properties of ChAc. In other words, ChAc and the choline uptake system are in close contact with each other. Therefore, it may be proposed that ChAc and the choline uptake system are part of a multienzymatic complex involved in the synthesis of Ach. The multienzymatic complex involved in Ach synthesis would also contain an acetate uptake system and an acetyl-CoA synthetase utilizing ATP. Acetate, a product of Ach hydrolysis, is reused for new Ach synthesis at the Tmpedo electric organ (Israel and Tucek, 1974; Dunant and Israel, 1975) and also at the rat neuromuscularjunction (Dreyfus, 1975). The reuptake process and the activation of acetate by ATP have not yet been characterized but it has been shown that external acetate levels regulate choline incorporation into Ach (Morel, 1975) and that presynaptic ATP and Ach levels vary rapidly and directly in phase with each other during electrical stimulation (Israel et al., 1975). Both observations are in agreement with the hypothesis that acetate uptake and activation are part of the multienzymatic complex involved in Ach synthesis.

XIV. The Role of CI- in the Regulation of Ach Synthesis by ChAe

Previous ChAc kinetic studies in 150 mM NaC1 indicated that Ach is a competitive inhibitor with respect to choline, but a poor one, as the Ki value (50 mM) was well above the Ach concentration in cholinergic terminals (about 27 mlM) (Dunant et al., 1974). Our recent findings agree only in part with these previous works. In the presence of 145 mM NaCl, Ach is a poor inhibitor(Ki = 45 mM), but in the absence of NaC1, Ach is a relatively potent inhibitor ( K ; = 0.3 1 mM). This may be significant with regard to the regulation of Ach synthesis, as in the presence of a given

330

JEAN ROSSIER

concentration of Ach, ChAc may be either completely inhibited or fully active, depending upon the ionic composition of the assay medium. There has been until now insufficient information to determine whether the increase in ChAc velocity by salts is due to the character of the ions used or to their charge (Morris et al., 1971; Potter et al., 1968; Kuczenski et al., 1975; Schuberth, 1966; McCaman and Hunt, 1965; Prince and Hide, 1971). The data presented in Fig. 3 and Table IX indicate that the increase in V,,, is promoted mainly by the increase of the anion concentration and that C1- is the most effective. I t may be concluded that an increase in C1- concentration will promote an increase in ChAc velocity by two synergistic mechanisms: ChAc will become less susceptible to Ach inhibition and ChAc Y,,, will increase. These results reveal the importance of the effect of the ionic environment on the kinetic properties of ChAc. Numerous enzymes have been shown to have ionic requirements. For example, a-amylase (Lifshitz and Levitzki, 1976) and several peptidases (Boyer, 1971) have an absolute requirement for C1- ions. The requirements of ChAc for C1- are less specific and not absolute, and ChAc will synthesize Ach in the absence of C1-. It appears that ChAc is regulated by the total anion concentration, and among anions C1- is the more effective one. It has already been shown that other enzymes involved in neurotransmission are regulated by cation or anion concentration: acetylcholinesterase is stimulated by Ca” (Marquis and Webb, 1974), glutamate decarboxylase is stimulated by K+ (Molinoff and Kravitz, 1968), and dopamine P-decarboxylase is stimulated by anions (Craine et al., 1973). The regulation of enzymes by the ionic environment has recently been emphasized by Douzou and Maurel (1976), who have proposed a model derived from the polyelectrolyte theory of Engasser and Horvath (1975). This model could be applicable to the regulation of ChAc and other enzymes involved in neurotransmission. XV. Pleiotropic Effect of Nerve Impulses on Ach Synthesis

Although C1- concentration is very low inside nerve terminals (McIlwain, 1966), it has recently been shown by Marchbanks and Campbell (1976) that C1- fluxes through synaptosomal membranes are active rather than passive. Simon and Kuhar (1976) have also recently shown that C1- concentration affects the high-affinity choline uptake system. In the absence of external C1- this system is inhibited. In view of these recent reports, we can postulate that in the multienzymatic complex involved in Ach synthesis local concentration of C1-

CHOLINE ACETYLTRANSFERASE

33 1

may rapidly increase in response to presynaptic depolarization. I t seems unrealistic to propose that the whole terminal would increase its C1concentration; however, it seems possible that in the local membrane complex C1- concentration could increase severalfold. I t is well known that depolarization of the presynaptic membrane promotes Ca‘+ entry into the axon terminal and a subsequent release of the transmitter. I t is proposed that depolarization may also promote C1entry. The resulting local increase in C1- concentration will affect ChAc velocity both by directly activating the enzyme and by suppressing Ach inhibition of enzyme activity. Therefore the same signal, depolarization of nerve terminals, would simultaneously promote synthesis and release of the transmitter. This proposed mechanism is not a classical feedback mechanism in which decreased concentration of an end product induces an increase of its synthesis. Rather, the nerve impulse would promote membrane invasion by C1- and Ca’+ such that an increase in Ach synthesis by C1- is concurrent with decrease in Ach concentration by the Ca”dependent release mechanism. Nerve impulses which alter membrane permeability to both cations and anions seem to have a pleiotropic activating effect on the mechanisms involved in Ach synthesis. Nerve impulses increase choline permease velocity (Simon et nl., 1976), ATP metabolism (Israel et nl., 1975) and may also, as suggested here, increase ChAc velocity. Therefore, I would like to propose that the multienzymatic complex involved in Ach synthesis is under direct control of the ion fluxes promoted by nerve impulses. ACKNOWLEDGMENTS The author expresses sincere thanks to Mrs. Odessa Colvin for her expert assistance in typing the manuscript. REFERENCES Alpert, A., Kisliuk, R., and Shuster, L. (1966). Biochem. Pltnrmncol. 15, 465-473. Ansell, G . B., and Spanner, S. (1967).j. Neurochtm. 14, 373-385. Aquilonius, S. M., and Eckernas, S. A. (1976).J. Neitrochem. 27, 317-318. Baker, B . R., and Gibson, R. E. ( 1 9 7 1 ) . j . Med. Chem. 14, 315-322. Baker, B. R., and Gibson, R. E. (1972).J. Med. Clim. 15, 639-642. Banns, H . E. ( 1 976).J. Neicroclzem. 26, 967-97 1 , Barker, L. A., and Mittag, T. W. ( 1 9 7 5 ) . j .Phnrmncol. Exp. Tim. 192, 86-94. Barker, L. A., and Mittag, T. W. (1976).Biochem. Phnrmcol. 25, 1931-1933. Barlow, R. B., and Dixon, R. 0. D. (1973). Bioclwm. J. 132, 15-18. Baughman, R. W.,Bader, C. R., and Schwart?.,E. A. (1976). Pror. Meet. Sac. Neiiro.sci. 6th, Torollto (Abstr. 1575). Berg, P. (1956).J. Biol. Chtm. 222, 991-1023. Berman-Reisberg, R. B. (1957). Y d e J. Biol. Mud. 29, 403-435. Berry, J. F., and Whittaker, V. P. (1959). Biochem. J. 73, 447-457.

332

JEAN ROSSIER

Birks, R. I. (1963). Can. J . Biochem. Physiol. 41, 2573-2597. Bishop, M. R., Sastry, B. V. R., Schmidt, D. E., and Harbison, R. D. (1976). Bwchem. Pharntacol. 25, 1617-1622. Boyer, P. D. (1971). “The Enzymes,” Vol. 111, 3rd ed. Academic Press, New York. Browning, E. T. (1976)Jn “Biology of Cholinergic Functions”(Goldberg, A. M., and Hanin, l., eds.), pp. 187-201. Raven, New York. Browning, E. T., and Schulman, M. P. (1968).J. Neurochem. 15, 1391-1405. Brownstein, M., Kobayashi, R., Palkovits, M., and Saavedra, J. M. (1975).]. Neurochem. 24, 35-38. Bull, G., Hebb, C., and Ratkovic, D. (1961). Nature (London) 190, 1202. Bull, G., Feinstein, A., and Morris, D. (1964).Nature (London)201, 1326. Bull, G., Hebb, C., and Morris, D. (1969). Comp. Biochem. Physiol. 28, 11-28. Bull, G., Hebb, C., and Ratkovic, D. (1970).J. Neurochem. 17, 1505-1516. Burt, A. M. (1970).J. Histochem. Cytochem. 18, 408-415. Burt, A. M., and Silver, A. (1973a).Nature (London),New Biol. 243, 157-159. Burt, A. M., and Silver, A. (1973b). Brain Res. 62, 509-516. Brain Res. 57, 518-521. Burt, A. M . and Silver, A. (1973~). Butcher, L. L. (1977). Life Sci. (in press). Butcher, S. G., and Butcher, L. L. (1974). Brain Res. 71, 167-171. Carlsson, C. A., Bolander, P., and Sjostrand, J . (1971).]. Neurol. Sci. 14, 75-93. Carson, V. G., Jenden, D. J., Cho, A. K., and Green, R. (1976).Biochem. Pharmucol. 25, 195-1 99. Chao, L. P. (1975).J. Neurochem. 25, 261-266. Chao, L. P., and Wolfgram, F. (1973).J. Netrrochem. 20, 1075-1081. Chao, L. P., and Wolfgram, F. (1974).J. Neurochem. 23, 697-701. Chmouliovsky, M., Dunant, Y., and Hojvat, S. (1974).J. Neurochem. 22, 73-75. Chuang, H. Y. K., Mohammad, S. F., and Mason, R. G. (1976). Biochem. Pharnzacol. 25, 1971-1977. Collier, B., Barker, L. A., and Mittag, T. W. (1976). Mol. Pharmacol. 12, 340-344. Collier, B., Lovat, S., llson, D., Barker, L. A , , and Mittag, T. W. (1977).J. Neurochem. 28, 331-340. Cozzari, C., and Hartman, B. K. (1977). Proc. Natl. A c d . Sci. U.S.A. (in press). Craine, J. E., Daniels, G. H., and Kaufman, S. (1973).,]. Biol. Chem. 248, 7838-7844. Crispin-Smith, J., Cavallito, C. J., and Feldes, F. F. (1967). Biochem. Phannacol. 16, 2438244 1. Currier, S. F., and Mautner, H. G. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 3355-3358. Curtier, S. F., and Mautner, H. G. (1976). Biochem. Biophys. Res. Commtrn. 69, 4 3 1 4 3 6 . DeRobertis, E., Rodriguez d e h r e s Arnaiz, G., Salganicoff, L., Pelligrino d e lraldi, A., and Zieher, L. M. (1963).J. Neurochem. 10, 225-235. Dewhurst, S. A., Croker, S. G., lkeda, K., and Mc Caman, R. E. (1972). Comp. Biochem. Physiol. B 43, 975-981. Diamond, l., and Kennedy, E. P. (1969).J. Biol. Chem. 244, 3258-3263. Diamond, I., and Milfay, D. (1972).J. Neurochem. 19, 1899-1909. Domino, E. F., Mohrman, M. E., Wilson, A. E., and Haarstad, V. B. (1973).Neurqbhamco/ogy 12, 549-561. D O L I ~ OP., U ,and Maurel, P. (1976). C. R . Acad. Sci. Park 282, 2107-2110. Dowdall, M. J., Fox, G., Wachtler, K., Whittaker, V. P., and Zimmermann, H. (1976). Cold Spring Harbor Symp. Quant. Biol. 40, 65-8 1. Dreyfus, P. (1975). C. R. Acud. Sci., Ser. D 280, 1893-1894. Droz, B. (1965).In “The Use of Radioautography in Investigating Protein Synthesis” (C. P. Leblond and K. P. Warren, eds.), pp. 159-175. Academic Press, New York.

CHOLINE ACETYLTRANSFERASE

333

Dunant, Y., and Israel, M. (1975). I n “Cholinergic Mechanisms” (P. G. Waser, ed.), pp. 161-167. Raven, New York. Dunant, Y., Gautron, J., Israel, M., Lesbats, B., and Manaranche, R. (1974).J. Neiiroclzm. 23,635-643. Ekstrom, J., and Emmelin, N. (1971).J. Physiol. (London) 216, 247-256. Emson, P. C. and Fonnurn, F. (1974).J. Neirrochem. 22, 1079-1088. Emson, P. C., Malthe-Sdrenssen, D., and Fonnum, F. (1974),J. Neurochem. 22,1089-1098. Eng, L. F., Uyeda, C. T., Chao, L. P., and Wolfgram, F. (1974). Nature (London) 250, 243-245. Engasser, J. M., and Horvath, C. (1975). Biochem. J . 145, 431-435. Feigenson, M. E., and Barrnett, R. J. (1977). Bruin Res. 119, 155-179. Fex, J., and Wenthold, R. J. (1976). Brain Res. 109, 575-585. Fitzgerald, G. G., and Cooper, J. R. (1971). Bioclzm. Pharmacol. 20, 2741-2748. Florey, E., and Winesdorfer, J. (1968).J. Neirrochem. 15, 169-177. Fonnum, F. (1967). Biochem. J. 103, 262-270. Fonnum, F. (1968). Bi0chem.J. 109,389-398. Fonnum, F. (1969). Bi0chem.J. 115, 465-472. Fonnurn, F. (1970a).J. Neitrochem. 17, 1029-1037. Fonnum, F. (1970b).J. Nertroclzm. 17, 1095-1 100. Fonnum, F. (1975a).J. Neurochem. 24, 407409. Fonnum, F. (1975b).In “Cholinergic Mechanisms” (P. G. Waser, ed.), pp. 145-159. Raven, New York. Fonnum, F., and Malthe-Sdrenssen, D. (1973).J. Neurochem. 20, 1351-1359. Fonnum. F., Frizell, M., and Sjostrand, J. (1973).J. Neirrochem. 21, 1109-1 120. Giacobini, E. (1967). Ann. N.Y. Arod. Sri. 144, 646459. Giller, E., and Schwartz, J. H. (1971).J. Neurophysiol. 34, 93-107. Glover. V. A. S., and Potter, L. T. (1971).J. Neurochem. 18, 571-580. Gnadinger, M., Walz, D., von Hahn, N. P., and Grun, F. (1967).Exp. Eye Res. 6,239-242. Gnadinger, M., Hermann, R., and Markstein, R. (1973). ExF. Eye Res. 15, 395-399. Godfrey, D. A., and Matschinsky, F. M. (1975).J. Histochem. Cytochm. 23, 645-651. Godfrey, D. A., Krzanowski, J . J.. and Matschinsky, F. M. (1976).J . Hirtochem. Cytocliem. 24,470-472. Goldberg, M. E., Salama, A. I., and Blum, S. W. (1971).J. Pharm. Pharmacol. 23,384-385. Green,A. R., Boullin, D. J., Massarelli, R., and Hanin, I. (1972).LzfeSci. 11(1),1049-1058. Green, R. W., and Mc Kay, R. H. (1969).J. Biol. Clram. 244,5034-5043. Guyenet, P., Lefresne, P., Rossier, J., Beaujouan, J. C., and Glowinski, J . (1973a). Mol. Phurmacol. 9, 630-639. Guyenet, P., Lefresne, ,.‘l Rossier, J., Beaujouan, J . C., and Glowinski, J. (1973b).Bruin Res. 62, 523-529. Guyenet, P., Lefresne, P., Beaujouan, J. C., and Glowinski, J. (1975). In “Cholinergic Mechanisms” (P. G. Waser, ed.), pp. 137-144. Raven, New York. Haga, T., and Noda, H. (1973). Biochim. Biophgs. A c k 291, 564-575. Hamprecht, B., and Amano, T. (1974). Anal. Biochm. 57, 162-172. Harbison, R. D., Olubadewo, J.. Dwivedi, G, and Sastry, B. V. R. (1975). In “Basic and Therapeutic Aspects of Perinatal Pharmacology” (L. Morsalli, S. Garattini, and F. Sereni, eds.), pp. 107-120. Raven, New York. Harbison, R. D., Dwivedi, G., and Evans, M. A. (1976). Toxi~ol.Appl. Pharmucol. 35(3), 48 1 4 9 0 . Haubrich, D. R. (1976). In “Biology of Cholinergic Function” (A. Goldberg and I . Hanin, eds.), pp. 239-268, Raven, New York. Haubrich, D. R., and Wang, P. F. L. (1976).Biochem. Pharmacol. 25(6) 669-672.

334

JEAN ROSSIER

Hebb, C. (1957). Physiol. Rev. 37, 196-220. Hebb, C. (1963). In “Handbuch der Experimentellen Pharmakologie” (G. B. Koelle, ed.), Vol. 15, pp. 55-88. Springer-Verlag, Berlin and New York. Hebb, C. (1972). Physiol. Rev. 52, 918-957. Hebb, C., and Ratkovic, D. (1962).J. Physiol. (London) 163, 307-313. Hebb, C., and Silver. A. (1956).J. Physiol. (London) 134, 718-728. Hebb, C., and Smallman, B. N. (1956).J. Physiol. (Londun) 134,385-392. Hebb, C., and Waites, G. M. H. (1956).J. Physiol. (London) 132,667-671. Hebb, C., Krnjevic, K., and Silver, A. (1964).J. Physiol. (London) 171, 504-513. Hebb, C., Kasa, P., and Mann, S. P. (1970).J. Physiol. (London) 208, 1P-2P. Hebb, C., Stephens, T. C., and Smith, M. W. (1972).Biochem.J. 129, 1013-1021. Hebb, C., Mann, S. P.,and Mead, J. (1975). Biochem. Pharmmol. 24, 1007-1011. Hemsworth, B. A.. and Smith, J. C. (197Oa).J. Neurochem. 17, 171-177. Hemsworth, A. A., and Smith, J. C. (1970b). Biochem. Pharmacol. 19, 2925-2927. Hildebrand, J. G., Townsell, J. G., and Kravitz, E. A. (1974).J, Neurochem. 23, 951-963. Holden, J. T., Rossier, J., Beaujouan, J. C., Guyenet, P., and Glowinski, J. (1975). Mol. P harmacol. 11, 19-27. Howard, R. O., Wilson, W., and Dunn, B. (1973). Invest. Ophthalmol. 12, 418-425. Howard, R. O., Zadunaisky, J. A., and Dunn, B. J. (1975).Invest. Ophthalmol. 14,592403. Husain, S. S.,and Mautner, H. G. (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 3749-3753. Ilson, D., and Collier, 8. (1975). Nature (London) 254, 618-620. Israel, M. (1970).Arch. Anal. Microsc. Murplwl. Exp. 59, 67-98. Israel, M., and Tucek, S . (1974).J. Neurochem. 22, 487-491. Israel, M., Lesbats, B., Marsal, J., and Meunier, F. M. (1975).C.R. Acad. Sci. 280,905-908. Jablecki, C., and Brimijoin, S. (1974).Nature (London) 250, 151-153. Jablecki, C., and Brimijoin, S. (1975).J. Nezirochem. 25, 583-593. Jasser, A , , and Guth, P. S. (1973).J. Neurochem. 20, 45-53. Johnson, S., and Domino, E. F. (1971).Clin. Chim. Acta 35, 421-428. Kahlson, G., and Maclntosh, F. C. (1939).J. Physiol. (London) 96,277-292. Kaita, A. A., and Goldberg, A. M. (1969).J. Neurochem. 16, 1185-1 191. Kasa, P. (1975). In “Cholinergic Mechanisms” (P. G. Waser, ed.), pp. 271-281. Raven, New I‘ork. Kasa, P., Mann, S. P., and Hebb, C. (1970). Nature (London) 226,812-816. Kataoka, K., Nakamura, Y., and Hassler, R. (1973). Brain Res. 62, 264-267. Kilbinger, H. (1977).Naunyn-Schmiedherg’sArch. Phanacol. 296, 153-158. Kim, S . U. (1972). ExperZentin 28,537-538. Kobayashi, R. M., Brownstein, M., Saavedra, J. M., and Palkovits, M. (1975).J. Neurochem. 24,637-640. Koelle, G. B. (1954).J. Comp. Neurol. 100, 21 1-228. Koelle, G. A. (1967). Neurosci. Res. f r o g . 8dl. 5 , 44. Kuczenski, R., Segal, D. S., and Mandell, A. J. (1975).J. Nevrochem. 24, 39-45. Kuhar, M. J., Sethy, V. H., Roth, R. H., and Aghajanian, G. K. (1973).J. Nezrrochzm. 20, 581-593. Kumagai, H., and Ebashi, E. (1954). Natzrre (London) 173, 871-872. Lam, D. M. K. (1972). Proc. Natl. Acad. Sce. U.S.A. 69, 1987-1991. Lasek, R. J. (1970). Int. Rev. Neurohiol. 13, 289-329. Lebbin, C., and Waser, P. G. (1975). Histochemirtry 43, 309-317. Lefresne, P., Guyenet, P.. Beaiijouan, J. C., and Glowinski, J. (1975).J. Neurochem. 25, 415-422. Lewis, P. R., and Shute, C. C. D. (1963).J. Physiol. (London) 175, 5P-7P. Lewis, P. R., and Shute, C. C. D. (1967). Bruin 90, 521-540. Lewis, P. R., Shute, C. C. D., and Silver, A. (1967).J. Physiol. (London) 191, 215-224.

CHOLINE ACETYLTRANSFERASE

335

Lifshittz, R., and Levitski, A. (1976).Biorhemi.vtly 15, 1987-1993. Lipmann, I.: (1945).J. Bid. Chem. 160, 173-190, Lipmann, F., Kaplan, N. O., Novelli, G. D., Tuttle, L. C., and Guirard, B. M. (1947).J. Biol. Clrem. 167, 869-870. McCaman. R. E., and Dewhurst, S. A. (1970).J. Nenmrhem. 17, 1421-1426. Neir~oc//em.12, 253-259. McCaman, R. E., and Hunt, J. M. (1965).J. McCaman, R. E., Rodrigties d e Lores Arnaiz, G . , and DeRobertis, E. (1965).J. NPtrrof/iem. 12,927-935. McGeer, P. L., McGeer. E. G., Fibiger, H . C., and Wickson, V. (1971). Broin Re.s. 35, 308-314. McGeer, P. L., McGeer, E. G.. Singh, V. K., and Chase, W. H . (1974). Brain Rev. 81, 373-379. Mcllwain, H . (1966). “Biochemistry and the Central Nervous System,” 3rd ed. Little, Brown, Boston. Mahoney, K., Vogel, W. H., Salvenmoser, F., and Boehrne, D. (l971).J. Ntnroc/)t?tt. 18, 3357-1359. Malthe-Sdrenssen, D. (1975).1n “Cholinergics Mechanisms“ (P. G. Waser, ed.. pp. 257-262. Raven, New York. . X61-X65. Malthe-Splrenssen. D. (1976).J. N r i i t o c h ~ m 26, Malthe-Mrenssen, D. ( 1 977).J. Nairorhtm. (in press). Malthe-Sdrenssen, D., and Fonnum, F. (1972).R i o r k m . J. 127, 229-236. Malthe-Sdrenssen, D., Eskeland. T., and Fonniini, F. (1973).Brrtin Re,\. 62, 517-522. Malthe-Sdrenssen, D., Andersen, R. A., and Fonnum, F. (1974). Bioclwm. f % W T t I / l C / J [ . 23, 577-586. Mannervik, B., and Sorbo, B. (1970). Biorhem. Pltrrrmacul. 19, 2509-25 16. Marchbanks, R. M. (1968). Bioc/tem.J. 110, 533-541. Marchbanks, R. M., and Campbell, C. W. B. (1976).J.NP?rrorhem. 26, 979-9x0. Marquis, J. K., and Webb, G. D. (3974).Bioclwm. Phcirmrrrol. 23, 3459-3465. Matsudd, T., and Yoshida, H. (1976).J. Ntitrochem. 26, 817-X22. Mehrotra, K. N. (1961).J. Imert Plrytiol. 6, 215-221. Molinoff; P. B., and Fravitz, E. A. (1968).J. Ntirmcham. 15, :391-409. Morel, N . (1975). C. R. Arrid. Sci., SPT. D.280, 999-1001. Morel, N. (1976).J . Neirrorllem. 27, 779-784. Morris, D. ( 1 966). Biochem. J. 98, 754-762. Morris, D. (1967).J. Nezrrorhem. 14, 19-27. Morris, D., and Grewaal, D. S. (1969).Life Sri. 8(11), 511-516. Morris, D., and Grewaal, D. S. (1971).E / r r . J .Biorhem. 22, 563-572. Morris, D., Maneckjee, A., a n d Hebb, C. (1971).Rior//em.J. 125, X57-X63. Nachmansohn, D. ( 1959). “Chemical and Molecular Basis of’ Nerve Activity,” 235 pp. Academic Press. New York. Nachmansotin, D. (1972).Annn. Re7l. B ~ / J C / W41, V Z . 1-28. Nachmansohn, D., and Berman, M. (1946).J. Biol. Cham. 165, 551-563. Nachmansotin, D., and Machado, A. L. (1943).5. Neziroj’hyviol. 6, 397-409. Nachniansohn, D., John, H . M., and Waelsch. H. (1043).J. R i d . C h m . 150, 485-486. Oesch, F. (lY74).J. Pharmacol. Exp. Thn.. 188(2), 439-446. Oesch, F., and Thoenen, H. (1973).Natirre (London) 242, 536-537. Oesch, F.. Otten, U., and Thoenen, H. (1973).J. Nezcrochtm. 20, 1691-17OG. Palkovits, M., Saavedra, J . M., Kobayashi, R. M., and Brownstein. M. (1974).BrrripzRe,\. 79, 443-450. Papiernik, M. (1972).J. Cell. Pl!ysio/. 80, 285-242. Persson, B. O., Larsson, L., Schuberth, J., and Sorbo. B. (1967). A c h Clrem. Srrtnd. 21, 2283-2284.

336

JEAN ROSSIER

Peterson, G. R., Webster, G. W., and Shuster, L. (1973). Dni. Eiol. 34, 119-134. Pieklik,J. R., and Guynn, R. W. (1975).J. Biol. Chem. 250,4445-4450. Polsky, R., and Shuster, L. (1976a). Biochim. Biophys. Acta 445, 25-42. Polsky, R., and Shuster, L. (l976b). Biochim. Eiophys. Acta 445,43-66. Potter, L. T. (1968). In “The lnteraction of Drugs and Subcellular Components on Animal Cells” (P. N. Campbell, ed.), pp. 293-304. Churchill, London. Potter, L. T. (1970). In “Handbook of Neurochemistry” (A. Lajtha, ed.), Vol. 4, pp. 263284. Plenum, New York. Potter, L. T., Glover, V. A. S., and Saelens, J. K. (1968).J. Eiol. C h m . 243, 3864-3870. Prince, A. K. (1971). In “Method in Enzymology” (S. P. Colowick and N. 0.Kaplan, eds.), Vol. 17B, pp. 788-797. Prince, A. K., and Hide, E. G. J. (1971). Nature (London), New B i d . 234, 222-223. Rimon, R., Puhakka, P., Venolainen, E., and Mandel, A. J . (1973). Psychiatr. Fenn. pp. 265-267. Roskoski, R. (1973). Biochemistry 12, 3709-37 14. Roskoski, R. (1974a).J. Biol. Chem. 249,2156-2159. Roskoski, R. (1974b). Biochmistry 13,2295-2298. Biochemistry 13, 5141-5144. Roskoski, R. (1974~). Roskoski, R., Mayer, H. E., and Schmid, P. G. (1974).J. Neurochem. 23, 1197-1200. Roskoski, R., Lim, C. T., and Roskoski, L. M. (1975).Biochemistry14, 5105-5110. Ross, C. D., and MacDougal, D. B. (1976).J. Neurochem. 26,521-526. Ross, C . D., Cohen, A. I., and MacDougal, D. B. (1975). Invest. Ophthalmol. 14,756-761. Rossier, J. (1975). Brain Res. 98, 6 1 9 4 2 2 . Rossier, J. (1976a).J . Neurochm. 26, 543-548. Rossier, J. (1976b).J . Neurochem. 26, 549-553. Rossier, J. (1976c).J. Neurochem. 26, 555-559. Rossier, J . (1977). Bi0chem.J. 165, 321-326. Rossier, J., Bauman, A., and Benda, P. (1973a). FEES Lett. 32, 231-234. Rossier, J., Bauman, A., and Benda, P. (1973b).FEES Lett. 36, 43-48. Rossier, J., Spantidakis, Y., and Benda, P. (1977).J. Neurochem. (submitted for publication). Rowell, P. P., and Chiou, C. Y. (1976). Biochem. Phannacol. 25, 1093-1099. Ruch, G. A., Davis, R., and Koelle, G. B. (1976).J. Neurochem. 26, 1189-1192. Ryan, R. L., and McClure, W. 0. (1976).Fed. Proc., Fed. Am. SOC.Exp. B i d . 35, 1647. Sastry, B. V. R., and Henderson, G . 1. (1972).Biochem. Phannacol. 21, 787-802. Schrier, B. K., and Klein, D. C. (1974). Brain Res. 79, 347-351. Schrier, B. K., and Shuster, L. (1967).J. Neurochem. 14, 977-985. Schuberth, J. (1966). Biochim. Eiophys. Acla 122, 470-481. Secrist, J. A., Barrio, J. R., and Leonard, N . J . (1972). Science 175,646-647. Shuster, L., and OToole, C. (1974). Lije Sci. 15, 645-656. Shute, C. C. D., and Lewis, P. R. (1967). Brain 90, 497-520. Siegel, L. M., and Monty, K. (1966). Biochim. Eiuphys. Acta 112, 346-362. Simon, J. R., and Kuhar, M. J. (1976).J. Neurochtm. 27, 93-99. Simon, J. R., Atweh, S., and Kuhar, M. J, (1976).J. Neurochem. 26, 909-922. Singh, V. K., and McGeer, P. L. (1974a).L f e Sci. 15,901-913. Singh, V. R., and McGeer, P. L. (1974b). Brain Res. 82, 356-359. Smallman, B. N. (1956).J , Physiol. (London) 132, 343-357. Sorimachi, M., and Kataoka, K. (1974a). Brain Res. 72(2), 350-353. Sorimachi, M., and Kataoka, K. (1974b). Brain Res. 70, 123-130. Sorimachi, M., and Kataoka, K. (1975).ErainRes. 94, 325-336. Spantidakis, Y., Rossier, J., and Benda, P. (1976). C.R. Acad. Scz., Ser. D.282,1621-1624. Stevenson, R. W., and Wilson, W. S. (1975).Exp. Eye Res. 21, 235-244.

CHOLINE ACETYLTRANSFERASE

337

Storm-Mathken,J. (1970).J. Neurochem. 17, 739-750. Street, G., Dunnil, P., Lilly, M. D., and Glover, V. A. S. (1973). Biotechnol. Bioeng. 15, 271-283. Strom, T. B., Sytkowski, A., Carpenter, C. B., and Merrill, J. P. ( 1 974). Proc. Natl. Acad. Sci. U.S.A. 71, 1330-1333. Tanford, C. (1961). “Physical Chemistry of‘ Macromolecules.” Wiley, New York. Thompson, S. T., Cass, K. H., and Stellwagen, E. (1975).Proc. Natl. Acad. Sci. U.S.A. 72, 669-672. Tucek, S. (1967).J. Neurochem. 14, 531-545. Tucek, S. (1973). Exp. Neurol. 40, 23-35. Tucek, S. (1975).Brain Res. 86(2), 259-270. Uchimura, H., Saito, M., and Hirano, M. (1975). Brain Res. 91, 161-164. Van Alphen, G. W. H. N. (1957). A M A Arch. Ophthalmol. 58,449-451. Weiss, P., and Hiscoe, H. B. (1948).]. Exp. Zool. 107, 315-395. Welsch, F. (1974a). Am. J . Ohstel. Gynecol. 118(6),849-856. Welsch, F. (1974b). Expmientia 30, 162-163. Wenthold, R. J., and Mahler, H. R. (1975).J. Nezirochem. 24(5), 963-968. White, H. L., and Cavallito, C. J. (1970a).J. Netrrochem. 17, 1579-1589. White, H. L., and Cavallito, C. J. (1970b). Biochim. Biophys. Acta 206, 343-358. White, H. L., and Wu, J. C. (1973a).Bioclzmnistry 12, 841-846. White, H. L., and Wu, J. C. (1973b).J. Neurochem. 21, 939-947. White, H. L., and Wu, J. C. (1973c).J. Neurochem. 20, 297-307. Whittaker, V. P., Michaelson, J. A , , and Kirkland, R. J. (1964).Biochem. J. 90, 293-303. Williams, J. D., and Cooper, J . R. (1965).Biochem. Pharmacol. 14, 1286-1289. Wooten, C . F., and Coyle, J. T. (1973).J. Neurochem. 20, 1361-1371. Yamamura, H. I., and Snyder, S. H. (1973).J. Neurochem. 21, 1355-1374. Zimmermann, H., and Dowdall, M. J . (1975). Abstr. 5th I n l . Meet. Int. Soc. Neurochem. Barcelona, p. 128.

This Page Intentionally Left Blank

SUBJECT INDEX A

Acetylcarnitine, as contaminant in choline x e t y It ran sferase assay, 28X-28!) Acetylcholine cerebral formation of, lipid-hound choline role in, 17-20 in cyclic AMP formation. 122-125 in cyclic <;IMP regulation, I4 1 synthesis of, pleiotr-opic effect of nerve impulses on, 330-59 1 Acetyl-&A. :is clioline acetvItl.aiisferase inhibitor, JOX ACTH 4-10 binding to cell niemhranes, 222-223 blood flow and, 220-221 brain uptake of, 221-222 electroph ysiological correlates of. 2 I x-220 in induction of excessive grooming. 217-21X learned beliavioi- and, 21 1-217 in humans, 214-215 in hypophysectomized rats, 21 1-21? site of action. 216 structure-actiyity studies. 215-2 1 6 morphine and. 221)-290 iieu rochemical response to, 222-2211 in neurotransmitter activity, 227-22!) in protein I'lios~~lior\latioii,22J-224 in protein synthesis. 225-227 Adenosine in cyclic A M P formation, I2 1-12? in cyclic <;MI' regulation. 141 Adenylate cyclases localimtion of. in riel-vous system, 109- I I 3 regulation of, 115-127 Affinity chromatography, ot' choliiie acetyltransferase, :100-:3o:i Amino acids in cyclic AMP limnation, 124-12.5 in cyclic G M P rehiihtion, 142 D-Amino oxidase, in brain, sr~hcellula~ localimtion of. 266-269

Anestlietics. etfects on brain ~ ~ l ~ o s ~ ~ l i o l i ~ ~ i c l s , 16-17 Avoidance heha v ior , from P-lipot ro pin , 16-9 1, 235-236 Asons 01' brain, phospholipid composition of', 5 clioline acetyltransferase trxnsport hy, 3 12-3 1'1 6

Brain ACTH 4-10 uptake o f , 221-222 D-amino osidase in, srlhcellular localimtion of', 266-261) tle\.eloping. iieii rot raiisni ission in. 6-103 ot' fetus. .)PY Fetal hlain

opiate receptors in, ACTH 4-10 ettects 011, 229-230 phospholipids of' funct ion;il niet;ilx)lisni of; 1-21) tissiie frxxionaiion of', 25:$-269 Behavior. cetitral ph;irmacolog\, cyclic nttcleotides. and. 152 Uenzodiarel,ines. ;IS pliosi'1iodiester;lse inliibitors. I J O - I :$ I Blood flow, A<:7'H 4-10 eflects on, 22022 1 Brain slices, protein a~ltopliospliol.)lation in. 137-13X C

Gilciuiii ions. i i i cyclic AhIP l'orniation, 126- I27 (:;ltecllolaininel.fiic neu i-oiis in fetal brain, 68 1) reiia tat development of, 67-82 tliymidine autoraclioglapli y of, 69-7 I (:ell nieriibraiiess A C T H 4-10 hinding 10, 222-225 Central depressants. cyclic iiiicleotide activity and. 15&155 351)

34 0

SUBJECT INDEX

Central nervous system (CNS) ACTH 4-10 effects on, 218-220 cyclic nucleotides in function of, 144-156 lipotropin and, 209-250 Central neurons, cyclic AMP in, 149-150 Central stimulants, cyclic nucleotide activity and, 1.54 C holine as choline acetyltransferase substrate, 306-308 lipid-bound, relationship to cerebral acetylcholine formation, 17-20 transport of, choline acetyltransferase in, 327-329 Choline acetyltransferase, 283-337 active site of, 305-306 affinity Chromatography of, 300-303 antigenicity of, 315 assays of, 287-291 acetylcarnitine contamination, 288-289 comparison among, 287-288 pitfalls in, 289-291 in single cells, 322-323 axonal transport of, 312-314 biophysical studies on, 303-304 chloride ion etfect on, 310-312,329-330 in choline transport, 327-329 in cholinergic pathways, 320-321 cross-reactivity of, 3 15-3 18 histochemistry of, 3 19-322 L'uture prospects, 323-324 history o f , 284-285 hydrophobic-affinity chromatography of, 302 inimunoabsorbent for, 298-300 inimunohistochernical localization of, 3 18 immunology of, 314-318 inhibitors of, 307, 308-309 isoelectric points of, 326 localimtion of, at cellular level, 3 18-324 mechanism of action of, 304-312 kinetics of, 304-305 in neuronal tissue, 294-303 in nonneuronal tissue, 291-294 purification of, 296-303 regulation of, 310 stability of, 297 stylpyridine inhibitors of, 302 subcellular localii~tionof, 324-326 substrate specificity of, 306-308 sulfliydryl reagent for, 300-302 Coenqmie A (CoA), structure of, 286

Conductance fluctuations, in fluctuation analysis, 176-180 Corneal epithelium, choline acetyltransferase in, 292-293 Cyclic AMP adaptation of systems of, 146-148 in autophosphorylation of proteins, 135-138 biochemistry of, 145-148 central neurons and, 149-150 in ganglia, 150 in nervous system. 109-138 neurophysiology of, 148-152 in neurotransmitter metabolism, 146 protein kinases dependent on, 133-1 34 transmitter release and, 151-152 Cyclic GMP in ganglia, 1.51 in nervous system, 138-144 protein kinases dependent on, 143-144 Cyclic GMP phosphodiesterases, activity of, 142-143 Cyclic nucleotides ACTH 4-10 eEects on activity of, 223-224 analogs of, activity, 153-154 central pharmacology, behavior, and, 152-156 cyclic AMP, 109-138 cyclic GMP, 138-144 in function of central and peripheral nervous system, 144-156 P-lipotropin 61-91 role in activity of, 237-238 in nervous system, 105-168 in peripheral nucleotides, 15 I D

Density gradient centrifugation, of brain fractions, 262-266 Dextran blue as choline acetyltransferase inhibitor, 302,308-309 use in choline acetyltransferase purification, 302 Dialkoxybenzyl-2-imidazolidines,as phosphodiesterase inhibitors, 132 Dialkoxyphenyl-2-pyrrolidones,as phosphodiesterase inhibitors, 132-1 33 Differential centrifugation, in brain fractionation, 257-262

34 1

SUBJECT INDEX

Dipyridamole, as phosphodiesterase inhibitor, 132 Dopamine in cyclic AMP formation, 116-1 18 in cyclic GMP regulation, 141 Drugs etfect on fetal brain catecholamines, 74-76 noise-induced, in fluctuation analysis, I 9 1-200 E

Endoplasinic reticulum. enzyme markers for, 273 Enkephalins in cyclic AMP formation, 125 in cyclic GMP regulation, 142 Enzymes, as markers, in brain fractionation, 273 Etheno-coenzyme A structure of, 323 Extrajunctional receptors, fluctuation analysis of, 196-198

Giant squid axon, fluctuation analysis of, 189-190 Glia, of brain, phospholipid composition of, 4 Glutamate receptors, fluctuation analysis of, 198-200 Grooming, excessive, ACTH 4-10 in induction of, 2 18-220 GTP, in cyclic AMP formation, 127 Guan ylate cyclases locali7ation of, 138-139 regulation of, 139-142 H Histamine, in cyclic AMP formation, 120-1 21 Hydrophobic-affinity chromatography, of choline acetyltransferase. 302 I

I mmunoabsorbent, for choline acetyltransferase, 298-300

F

Fetal brain catecholaminergic innervation in, 77-81 catecholaminergic neurons of, 67-82 regional innervation, 72-74 synatogenesis on, 76-77 catecholamines in, drug effects on, 74-76 octopamine and P-phenylethanolamine in, 71-72 Five-fraction procedure, for brain fractionation, 257-262 Fluctuation analysis in neurobiology, 169-208 conductance fluctuations. 176-1 80 of drug-induced noise, 191-200 general relationships in, 181-183 of mechanoreceptors, 203-204 methods for, 175-183 of photoreceptors, 20 1-203 results of, 183-205 of sensory systems, 200-205

Learned behavior, ACTH 4-10 and, 21 1-217 Lipotropin ACTH 4-10, 21 1-230 amino acid sequence of, 210 central nervous system and, 209-250 P-lipotropin 61-91, 232-239 0-MSH, 230-232 P-Lipotropin 47-53, see ACTH 4-10 P-Lipotropin 61-91 avoidance behavior from, 235-236 biological significance of, 238-239 cyclic nucleotide role in activity of, 237-238 effects on central nervous system, 232-239 electrophysiology of, 236-237 excess grooming activity from, 234-235 morphinelike activity of, 233-234 Lysosome, enzyme markers for, 273

G M

Ganglia cyclic AMP in, 150 cyclic GMP in, 151

Macromolecular factors, in cyclic AMP formation, 1 2 5- 126

342

SUBJECT INDEX

Mechanoreceptors, fluctuation analysis of, 203-205 Methylxanthines, as phosphodiesterase inhibitors, 130 Microtubules, protein autophosphorylation in, 136-137 Mitochondria, enzyme markers for, 273 Morphine ACTH 4-10 effects on activity of, 2 2 9-23 0 P-lipotropin 61-91 similarity to, 233-234 j.3-MSH effects on central nervous system, 230-232 Myelin of brain, phospholipid composition of, 5 enzyme markers for, 273 N

Neocortex, immature catecholaminergic innervation in, 77-81 Nerve axon, fluctuation analysis of, 183-1 90 Nerve endings of brain, phospholipid composition of, 5 enzyme markers for, 273 Nerve growth factor, in cyclic AMP formation, 125 Nervous system, cyclic nucleotides in, 105-168 Nerve tissue, choline acetyltransferase in, 294-303 Neu robiochemistry , tissue fraction in, 251-281 Neurobiology, fluctuation analysis in, 169-208 Neurons, of brain, phospholipid composition of, 4 Neurotransmission, in developing brain, 65- 103 Nerirotransmittors ACTH 4-10 effects on, 227-229 cyclic AMP in metabolism of, 146 cyclic nucleotides, behavior, and, 152-156 effects on brain phospholipids, 20-26 Nicotinic acetylcholine receptor amino acid composition of, 45 chemical and structural properties of, 44-49 identification of, 32-34

isolation and purification of, 31-63 reconstitution of, 49-58 measured by bilayer membrane conductance, 52-54 measured by Na' flux, 49-52 si7e and subunit structure of, 46-49 solubili~ationof, 39-40 purification following, 40-44 Nigrostriatal circuit postnatal development of, 82-97 dopaminergic innervation to striatium, 83-87 ontogenesis of function of, 93-97 striatal cholinergic neurons, 88-91 striatal GABAergic neurons, 91-93 Node of Ranvier, fluctuation analysis of, 183-189 Norepinephrine in cyclic AMP formation, 114-116 in cyclic GMP regulation, 140-141 0

Octopamine, in fetal brain, 71-72 Oligodendroglia, of brain, phospholipid composition of, 4 P

Papaverine, as phosphodiesterase inhibitor, 131 Peptides in cyclic AMP formation, 125 in cyclic GMP regulation. 142 Peripheral nervous system ACTH 4-10 effects on, 220 cyclic nucleotides in function of, 144-156 Peroxisome enzyme markers for, 273 Placenta, choline acetyltransferase in, 29 1-292 Plasma membrane, enzyme markers for, 273 Phenothiazines, as phosphodiesterase inhibitors, 131 /3-Phenylethanolamine, in fetal brain, 71-72 Phosphatidylinositol, metabolism of, receptor mechanisms and, 23

343

SUBJECT INDEX

P hosphodiesterases

S

inhihitors m d activators of, 130- 1:<3 in central pharmacology, 153 localimtion of. in brain tissue, 127-1 28 regulation of, 128- I30 Phospholipids of blain anesthetic effects on, 16-17 d e novo synthesis of, i - l 1 fiinctional nietaholism of. 1-29 na t II ra I1y occir rrin g niodi fica t ioii s 01'. 11-16 neurotl-;insmitter effects on, 20-26 origin a n d metabolisn~of, 6-1 1 relationship to brain function. 11-1 7 S~~IIC~La I InPd distribution of.. 2-5 in synaptic transmission, 17-26 Phosphoprotein phosphatases, cyclic A M l1 and, 134-1J.i Photoreceptors. Huctuation analysis of, 20 1-203 Pros taglari di ns in cyclic AMP formation. 123-124 in cyclic G M P regulation, 13 1 Protein (s) autopliosphor~lationof. cyclic A M P in, 13.5- I58 phosphorylation of, ACTH 4-1 0 efrects on, 223-224 synthesis of, ACTH 4-10 effects on, 225-22i Protein ki nases cyclic AhlP-dependent, 133- 134 cyclic GhlP-dependent, 143- 144 1 -H-Pyrazole[3,4b]l,yl.idines, a s phosphodiesterase inhibitors, 1.72

R

RNA metabolism, A C T H 4-10 effects on, 224-225

Sensory systems. Huctitation analysis 01, 200-205 Serotonin in cyclic A M P I'ormation, I I X - 1 2 0 in cyclic (;MP regulatioti. 1.11 Styrvlpyritline inhibitors. of rltoline ace t y It ransttr;ise , 3()2 SLlfhydi-yl reagent. tor choline acetyltransfer;ise, 300-302 Superniit;ini kicrion. of br;tin, e n ~ y i n e niarkers for-. 253 Synaptic transmission, hrain pho~pholipid role in, li-26 Synaptic vesicle en7yme markers tbr, 273 iii iissue tractionittip, 25ti-271) Syiiaptogenesis. on ktal c;itechol;rmine~-gic iieirrons, 76-7i S~naptosoines.cyclic A M P cHect on protein ~ i l i o s ~ ~ l i o ~ y l ain. t i o136 i~ T

'l'issuc fr;tctionation 01. hrain. 253-269 dilterential centi.ifugaiion. 2 5 - 2 6 2 homogeni/ation, 254-256 medium for, 2.56-257 scheme, 261 sul,fiactionatioii, 262-266 eniynie and neirrotransmitters :IS markers in, 252-275 interpretation of results of', 269-2i!) biochemical criteria. 26!)-275 rnorpliological criteria, 255-276 sy~ial)tosonies.276-259 in ne~rrobiocliemistr~. 2.5 1-2x I V

Voltage, in Iluctcrarion analysis. 1 x 0 - I X 1

This Page Intentionally Left Blank

CONTENTS OF PREVIOUS VOLUMES Volume 1

Recent Studies ot tlie Kliinenceplialon in Relation to Temporal Lobe Epilepsy and Behavioral Disorders W. R . Adey Nature of Electrocortical Potentials arid Synaptic Organimtions in Cerebral and Cerehellar Cortex Dominirk P. Ptrrpiiro Chemical Agents of the IVervous System CrrtlirYinP 0.H e l h Parasympathetic Neuroliiimors: Possible Precursors and Effect on Beha\.ior Cnrl C. Pfeifer Psychophysiology of Vision G . W. GrrrngerPhysiological and Biochemical Studies in Schi7ophrenia with Particular Emphasis on M ind-B rain Relations ti ips Robert G.HPalh Studies on tlie Role of Ceruloplasmin in Sclii/oph renia S. Mrr"rtms,S . Volllm, i o i d B . Melunder In\ estigations in Protein Metabolism in Nervous and Mental Diseases with Special Reference to tlie Metabolism of Amines I;. Gemgi, C. G . Hmegger, D . Jordan, H. P. Rieder, nnd M . RottenOerg AL'THOR INDEX-SC'BJECT

INDEX

Volume 2

Regeneration of Amphihia R . M . Gaze

the Optic Nerve in

Experimentally Induced Changes in tlie Free Selection of Ethanol Jorgr Mnrrlmze\

The Mechanism Hemiclioliniitms E: U'. S f h l / P l P ) .

of

.4ction

of

the

T h e Role of Phosphatidic Acid and Phosphoinositide in Transmemhrane 'rrat1sport Elicited t q i2cetylcholine and Other H IImot-a1 Agents Lo~oelIE . Hokin rind Mrrhrl R . Hokin Brain Neurohormones ;ind Cortical Epinephrine Pressor Responses as Aflected by Scli i/opli renic Serum Edu~nrdJ . Waln.m,k The Role of' Serotonin in Neurohiology Enninio Ciisfrr Drugs and the Conditioned Avoidance Response All>nt He?-t: Metabolic and Neurophysiological Roles of y-Aminohutyric Acid E I I ~ Robert., ~ I P and Ediinrdo Eidelbrrg OhjectiIe Psychological Tests and the Assessment of' Drug Eflects H . J . Eysench AL'THOR INDEX-SVRJECT

INDEX

Volume 3

Submicroscopic Morphology and Function ot' Glial Cells Ed~rni-diiDe Robrrti.\ and H. M. Grr.\rlwn/Pld Microelectrode Studies of' tlie Cerehral Cortex V~hE p . Amnmsin n Epilepsy Artltrir A. Ward, J I - .

Functional Organimtion of Somatic Areas of the Cerebral Cortex Hiro.clii Nakalwma 34 5

346

CONTENTS OF PREVIOUS VOLUMES

Body Fluid lndoles in Mental lllness R . Rodnight Some Aspects of' Lipid Metabolism in Nervoiis Tissue G. R . Wehstrr Convulsive EfTect of Hydrazides: Relationship to Pyridoxine Hiirty L. Williams rind Jrimrs A. Briin The Physiology of the Insect Nervous System

D.M . Vou1le.s AL'THOK INDEX-Sl'BJECT

INDEX

Volume 5 The Behavior of Adult Mammalian Brain Cells in Culture Rirth S. G r i p The Electrical Activity of a Primary Sensory Cortex: Analysis of EEC Waves Wrilter J . Freeman Mechanisms l'or the Transfer of' Inlormation along the \'iwaI Pathways Koiti M otoka 7 t n Ion Fluxes in the Central Nervous System F. J . Briiiley, J r . InterrelationshiI,s between the Endocrine System and Neuropsychiatry Richard P. Michael and J a m s L. Gihhmzs

Volume 4 The Nature of Spreading Depression in Neural Networks Sidncy Orlis Organi/ational Aspects of Some Subcortical Motor Areas Wrr-nrr P. Koellri Biochemical and Neuro1)liysiological De\,elopment of the Brain in the Neonatal Period Williriminci A . Htmulirh Substance P: A Polypeptide of. Possible Physiological Significance, Especially within the Nervous System F. Lrmbrck and G . Zeltet Ant icholinergic Psychotorn i me t ic A get1t s L . G. Ahood and J . H . Hiel Benmquinoli7ine Derivatives: A New Class of Monamine Decreasing Dnigs with Psychotropic Act ion A. Plrt~rher,A . Brossi, rind K . F. Gcy The Elt'ect of Adrenochrome and Adrenolutin on the Behavior of Animals and the Psychology of Man A . Hofrr A I ' T H O K INDEX-SL'BJECT

INDEX

Neurological Factors in the Control of' the Appetite AndrE Sotiliiirrir Some Biosynthetic Activities 01' Central Nervous Tissue R. 1,'. Coxrnz Biological Aspects of Electrocon~~~rlsive Therapy G f i nnci r H olmherg AL'THOK INDEX-SI'RJECT INDEX

Volume 6 Protein Metabolism of'the Nervous System A I d LNjfllIl Patterns of' Mrisciilar Innervation in the Lower Chordates Qlrmtifznotlr The Neural Organiiation of' the i'isiial Pa!liways in the Cat Thonui.5 H. Mvikle, J r . ( i d James M . S/irrigiie Properties of' Atierent Synapses and Sensory Neurons in the Lateral Geniculate Nucleus p. c. Bi,ShlJ/)

347

CONTENTS OF PREVIOUS VOLUMES

Regeneration in the Vertebrate Central Nervous System Carmine D . Clemente Neurobiology of Phencyclidine (Semyl), .a Drug with an Unusual Spectrum of Pliarmacological Activity Edward F. Domino

Tlie Anatomopliysical Basis of Somatosensory Discrimination David Bouisher, i t l r the rollrthorrition of Denke Alhe-Fe.wrd Drug Action on the Electrical Activity o f t h e Hippocampiis Ck. Strimpf

Free Behavior and Brain Stimulation Josi M . R . Delgado

Eflects of Drugs on Learning and Memory Jrime.~L. MrCringIi and Leii1i.s F. Petrinoi~irh

A L ' T H O R INDEX-SUBJECT

Biogenic Amines in Mental Illness Giinter G . Briine

INDEX

Volume 7

Alteration and Pathology of Cerebral Protein Metabolism A be1 Lnjtlia Micro-Iontoplioretic Studies o n Cortical Neurons K . KrnjeuiC Responses from tlie Visual Cortex of Unanesthetiied Monkeys John R . H n g l . ~ Recent Development of tlie Blood-Brain Barrier Concept Ricardo Edttrom Monoamine Oxidase Inhibitors Gordon R . Psrlieidt T h e Plienothiaiine Tranqiiiliiers: Biochemical and Biophysical Actions Pan1 S . Gath rind Mmi.c A. S1iirte.s

T h e Evolution of the Butyroplienones. Haloperidol and Trifluperidol, from Meperidine-like 4-Plienylpiperidines Port1 A. J . Janssa Amplitude Analysis of the Electroencephalogram (Review 01' tlie Information Ohtained with tlie Integrative Method) Lronide Gold~steinrind Rriymrmd A. neck A C ' T H O R INDEX-SUBJECT

INDEX

Volume 9

Development of "Organotypic" Bioelectric Activities in Central Nervoiis Tissues during Maturation in Culture Skinley M. Grain Tlie Unspecific lntralaminary Modulating System 01' the Thalamus P. Kni1)ji rind M . Mminier

Comments on tlie Selection and Use of Symptom Rating Scales for Research in Pharmacotherapy J . B . Wittenborn

Tlie Pharmacology of lmipramime and Related Antidepressants Laszlo Gyermek

Multiple Molecular Forms of Brain H ydrolases Jo.sepli Bermolin and Kevin D. Burron

Membrane Stabiliiation by Drugs: Tranquilizers, Steroids, and Anesthetics Pliilili M . S e e m n

AClTHOR INDEX-SUBJECT

INDEX

Volume 8

A Morphologic Concept o f t h e Limbic Lobe Lowell E . White, J r .

Interrelationships between I'liospliates and Calcium in Bioelectric Phenomena L. G . Abood T h e Periventricular Hypothalamus J v o m e Stitin

Stratum

01'

tlie

348

CONTENTS OF PREVIOUS VOLUMES

Neural Mechanisms of Facial Sensation I . Dorinn-Smith AL'THOR INDEX-SUBJECT

INDEX

Volume 10

A Critique of' lontophoretic Studies of Central Nervous System Neurons G . C. Sulmoiraghi and C . N . Stefiinis Extra-Blood-Brain-Barrier Brain Structures Wmner P. Koella and Jeromr Sittin Cholinesterases of the Central Nervous System with Special Reference to the Cerebellum Ann Sil7wr

Exopeptidases of the Neryous System Nniille Mark\ Biochemical Responses to Narcotic Dnrgs in the Nervous System and in Other Tissues Doris H . Clmiet Periodic Psychoses in the Light of Biological Rhythm Research F. A . Jenner Endocrine and Neurochemical Aspects of Pineal Function BPki Mexy The Biochemical Investigation of Schizophrenia in the USSR D . V. Lozoilsky Results and Trends of Conditioning Studies in Schipophrenia J . Snnrnui

Nonprimary Sensory Projections on the Cat Neocortex P. Bjiser cind K . E . Bignnll

Carbohydrate Metabolism in Schizophrenia P m S. Lingjl-mde

Drugs and Retrograde Amnesia Ahert Wei.umn

The Study of Autoimmune Processes in a Psychiatric Clinic S . F. Stmenoil

Neurobiological Action of Some Pyrimidine Analogs Harold Koenig A Comparative Histochernical Mapping of the Distribution of Acetylcholinesterase and Nicotinamide Adenine DinucleotideDiapliorase Activities in the Human Brain T. hhii und R. L . Frirde Behavioral Studies of Animal Vision and Drug Action Hiiglr Brown The Biochemistry of Dyskinesias c. Clrrzon ALITHOR INDEX-SL'BJECT

INDEX

Volume 1 1

Synaptic Transmission in the Central Nervous System and Its Relevance for Drug Action Pliilip B . Bmdley

Physiological Foundations of Mental Activity ,V. P. Heclitrreru und V. B . Gretchin AL'THOR INDEX-SLBJECT INDEX C U M U L A T I V E TOPICAL INDEX FOR VOLUMES

1-10 Volume 12

Drugs and Body Temperature Peter L o m x Pathobiology of Acute Triethyltin Intoxication R. Tvrnck, J . Gordon, und J . Pri~koji Ascending Control of' 'Thalamic and Cortical Responsiveness M . Steriude Theories of' Biological Etiology of Affecti1.e Disorders Jolrn M . Duiis

349

CONTENTS OF PREVIOUS VOLUMES

Cerebral Protein Synthesis Inhibitors Block Long-Term Memory Suniitrl H. Rarondes

iblolecular Mechanisms Processing Grorge.\ 17 "gftr

The Meclianism of Action of Hallucinogenic Drugs on a Possible Serotonin Receptor in the Brain J . R. Smyt/tir.s, F. Brriington, onri R . D. Moriri

The Effect of Increased Fiinctional Acti\ity on the Protein Metabolism ol the N e n o u s System B . Jrtkoithek rrtad R. Senitgimnjskj

Simple Peptides in Brain 1,wmti Snno

Tlie Activating Effect of Histaniine on the Central Nervous System M . Monnirr, R . Sriiier, rind A . M . Hntt Mode of Action 01' Psychomotor Stimiilant Drugs Jorqnes M. iwn Ro.atim ALTHOR INDEX-Sl'BJECT

INDEX

Volume 13

Of Pattern and Place in Dendrites Mridge E . ScheiOel ctnd Antold R . Srheibel The Fine Structural Localisation o f Biogenic Monoamines in Ner\.ous Tissue Floyd E . Rloom Brain Lesions and Amine Metabolism Robrrt Y. Moore Morphological and Functional Aspects ol Central Monoamine Neurons Kjell F i i x r , Tnnu~\ HGkfell, rind 1 'rOrin Li?gtTStPdt Lrptake and Subcellular Localimtion of Neurotransmitters in the Brain Solon~maH . Sngdet, i2lichrrt~lJ . KnItcii-, Akin I . ~ r e o i , J ( i \ e / ~T. / i C q l r , rind Edulnrd G . S/tcirkn n Chemical Mechanisms of' TransmitterReceptor Interact ion John T. Cnrlrnd and Jock DttrrII

in

Infbrmation

Pro tei t i Transport i t i Ne 11 ro ns RogmondJ . Lmrk Neurocheniical Coi-relates of' Behavior M . H. Aprison nnd J . N . Hingtgen Some Guidelines t'rom System Science tor Studying Neural Information Processing DoiariM 0.Wnlfer nnd Mortin F. Grirdinn. ALTHOK INDEX-SL'RJECT INDEX

Volume 14 The Pharmacology Geniculate Neurons J . W', Phillis

of

Thalamic

and

The Axon Reaction: A Review of the Principal Features of Perikaryal Responses to Axon InquiIy A. R . Lirhermnn

C 0 2 Fixation in the Nervous Tissue Sa-CIinh W e n g KeHections on the Role of Receptor Systems for Taste and Smell John G . Siixloir Central Cholinergic Mechanism Be hav ior S. N . Prndlton c t n d S. N . Ditttrr

and

Tlie Chemical Anatomy of' Synaptic Mechanisms: Receptors J . R. Sinsf/tie,\ AL'THOK INDEX-SL'RJECT INDEX

Volume 15 T h e Chemical Nature of the Receptor Site-A Study in the Stereochemistry of Synaptic Mechanisms J . R . Smythius

Projection of Forelimb Group I Muscle Afferents t o the Cat Cerebral Cortex InLqini

Ro\h

350

CONTENTS OF PREVIOUS VOLUMES

Physiological Pathways through the Vestibular Nuclei I'irlw J. W i h t

f'orphyria: Theories of Etiology and Treatment H. A. PrLm, D.J. C r i / ~ / ~rind s , H . H . Rrprr

Tetrodotoxin, Saxitoxin, and Kelated Substances: Their Applications in N eu rohiology Martin H . E7vin.s

SI'HJECT INDEX

The Inhibitory Action of y-Aminohutyric Acid, A Probable Synaptic Transmitter K n n ill iko Obutrr Some Aspects of Protein Metabolism of the Neuron M PSutuke ~

V o l u m e 17 Epilepsy and y-Aminobutyric Acid-Mediated Inhibition B . S. MPldrrim Peptides and Behaviol Guorgrs 1 ' n g u

Chemistry and Biology of T w o Proteins, S-100 and 14-3-2, Specific to the Nervous System Bloke W. M o m u

Acquired Biochemical Transfer 01' In for m at ion S. R . Mitcliell, J . M . Brafon, rind R . J . Bradlvy

The Genesis of the EEC Rofuul Elnl

AminotrdnSferdSe Activity in Brain M . Buncick und A . Lrrjthri

Mathematical Identification of Brain States Applied to Classitication of Drugs E. R. John, P. WalkPr, D. Couiood, M . Rtr.dr, and J . Gehrmann

The Molecular Structure of' Acetylcholine and Adreneigic Receptors: An All-Prorein Model J . R. Smythips

A I ' I H O H INDEX-SL'l3JE":T

Structural Integration 01 Neuroprotease Actit ity

INDEX

Elpriri

V o l u m e 16

Gabtiplmii

On Axoplasmic Flow Lilirinri Lnbin'sko

lModel of Molecular Mechanism Able to Schimphrenia: Perchance a Dream? Generate a Depolari~ation-Hyperi~olariJ . Cliristion Gillin ~ n r Riclrrrrd l J . Wwtl /-ation Cycle Clnru Tordri SL'BJECT INDEX Antiacetylclioline Drrigs: Chemistry. Stereochemistry. and Pharmacology T. D. Inrlr und R . W. Rrimblrcombe Kryptopyrrole and Other Monopyrroles in Moleriilar Neurobiology Dmmld G . Inline KNA Metabolism in the Brain Victur E. Shashoira A Comparison of Cortical Functions in Man and the Other Primates R . E . Pa.uingliam and G.Ettlingel-

Volume 18

Integrative Properties and Design Principles of Axons Sle/iIrm G . Wuxmun Biological Transmethylation Involving S-Adenosylmetlrionine: Development 01 Assay Methods and Implications lor Neuropsycliiatry Ross J. Bnldessarini

CONTENTS OF PREVIOUS VOLUMES

Synaptochemistry of Acetylcholine Metabolism in a Cholinergic Neuron Bertnlnn Csillik Ion and Energy Metabolism of the Brain at the Cellular Level Leif H~rtznnd Arne Sclroitshoe Aggression and Central Neurotransmitters S . N . Prndlinn A Neural Model of Attention, Reintbrcement and Discrimination Learning Ste/dirn Grossberg Marihuana, Learning, and Memory Ernest L. Abel Neiirochemical and Neiiropharmacological Aspects of Depression B. E. LPonnrd SL'BJECT INDEX

Volume 19

Do Hippocampal Lesions t'rodiice Amnesia in Animals? Siisnn D. Iiwsen

A 8 C D E F 6 H 1 1

7 8 9 O 1 2 3 4 5

35 1

Synaptosomal Transport Processes Giirlio LPvi nnd Mritirizio Rniteri Glutathione Metabolism and Some Possible Functions of Glutathione in t h e Nervous System Mnrinn Orlouaki rind Ahrnhnm Krirkow.skj Neiirochemical Consequences of Ethanol on the Nervous System Artin K . Rnwnt Octopamine and Some Related Noncatecholic Amines in Invertebrate Nervous Systems H . A . RohPrtwn nnd A . V. Jiiorio Apomorphine: Chemistry, Pharmacology, Biochemistry F. C. Col/mlert, W. F. M . Vun BpYler, rind J . E . M. F. L q w n Thymoleptic and Neiiideptic Drug Plasma Levels in Psychiatry:.Current Status Tliomcis B. Cooper, George M. Simpon, rind J . Hillrtq LPe Sl'BJECT INDEX

This Page Intentionally Left Blank