INTERNATIONAL REVIEW OF CYTOLOGY V60, Volume 60 (v. 60)

INTERNATIONAL REVIEW OF CYTOLOGY VOLUME60 ADVISORY EDITORS H. W. BEAMS ROBERT G. E. MURRAY HOWARD A. BERN GARY G. ...

Author: Author Unknown

5 downloads 731 Views 22MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

INTERNATIONAL

REVIEW OF CYTOLOGY VOLUME60

ADVISORY EDITORS H. W. BEAMS

ROBERT G. E. MURRAY

HOWARD A. BERN GARY G. BORISY

ANDREAS OKSCHE

ROBERT W. BRIGGS

W. J. PEACOCK

STANLEY COHEN

DARRYL C. REANNEY

RENE COUTEAUX

LIONEL I. REBHUN

MARIE A. DIBERARDINO

JEAN-PAUL REVEL

CHARLES J. FLICKINGER

WILFRED STEIN

M. NELLY GOLARZ DE BOURNE

ELTON STUBBLERELD

K. KUROSUMI

HEWSON SWIFT

MARIAN0 LA VIA

DENNIS L. TAYLOR

GIUSEPPE MILLONIG

TADASHI UTAKOJI

ARNOLD MITTLEMAN

ROY WIDDUS

DONALD G. MURPHY

ALEXANDER L. YUDIN

VLADIMIR R. PANTIC

INTERNATIONAL

Review of Cytology EDITED BY

G . H.BOURNE

J. F. DANIELLI

St. George's University School of Medicine St. George's, Grenada West Indies

Worcester Polytechnic Institute Worcester. Massachusetts

ASSISTANT EDITOR K. W. JEON Department of Zoology University of Tennessee Knoxville, Tennessee

VOLUME 60

ACADEMIC PRESS New York

San Francisco London

A Subsidiary of Harcouri Brace Jovonovich, Publishers

1979

COPYRIGHT @ 1979, 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.

111 Fifth Avenue, New York, New York 10003

United Kingdom Edition published by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road, London N W l I D X

LIBRARY OF CONGRESS CATALOG CARD NUMBER:52-5203 ISBN 0- 12-364360-0 PRINTED IN THE UNITED STATES OF AMERICA

79 80 81 82

987 6 5 4 3 2 1

Contents LIST OF CONTRIBUTORS . . . . . . . . . . . . . . . . . . . . . . . .

ix

Transfer RNA-like Structures in Viral Genomes TIMOTHY C . HALL I. I1. I11 . IV . V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Occurrence and Properties . . . . . . . . . . . . . . . . . . . . . . . Nucleotide Sequences of 3'-OH Ends of Aminoacylatable Viral RNAs . . . . . Biological Function of Viral tRNA-like Structures . . . . . . . . . . . . .

Summary., . . . . . . . . . . . . . . . . . . . . . . . . . . . . References

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

1 2 9 13 22 23

Cytoplasmic and Cell Surface Deoxyribonucleic Acids with Consideration of Their Origin BEVANL . REIDAND ALEXANDER J . CHARLSON

I . Introduction . . . . . . . . I1 . Cytoplasmic DNA . . . . . I11. Cell Surface DNA . . . . . IV . Some Biological Implications V . Summary . . . . . . . . . References . . . . . . . .

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

. . . . . .

. . . . . .

. . . . . .

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

27 28 38 49 49 51

Biochemistry of the Mitotic Spindle CHRISTIAN PETZELT

I. I1. I11 . IV . V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tubulin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Actin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Myosin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calcium-Dependent Regulator Protein . . . . . . . . . . . . . . . . . . VI. Dynein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . The Mitotic Ca*+-ATPase . . . . . . . . . . . . . . . . . . . . . . . VIII . Calcium in the Mitotic Cell . . . . . . . . . . . . . . . . . . . . . . IX. The Isolation of the Mitotic Spindle . . . . . . . . . . . . . . . . . . . X . Mitotic Centers . . . . . . . . . . . . . . . . . . . . . . . . . . . XI . Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

53 54 62 65 66 68 70 74

75 79 81 82

Alternatives to Classical Mitosis in Hemopoietic Tissues of Vertebrates VIBEKEE . ENGELBERT I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Imprint. Smear. Fixation. and Staining Methods . . . . . . . . . . . . . .

V

93 96

vi

CONTENTS

111. Methods Using Tritiated Thymidine . . . . . . . . . . . . . . . . . . . IV . Fluorescence Method for DNA and RNA . . . . . . . . . . . . . . . . . V . Tissue Culture in Virro of Hemopoietic Tissues . . . . . . . . . . . . . . VI. Behavior and Morphological Variations in Blast Cells and Their Nuclei . . . . VII . The Occurrence and Significance of Two Spatially Separate Nuclear Masses in Blast Cells and in Differentiating Cells . . . . . . . . . . . . . . . . . . . . VIII . Fate of the Peripheral or Shell-like Nuclear Mass, and the Inner Nuclear Mass. in Differentiating Cells of Leukemic Mice (AKR Strain) . . . . . . . . . . . . IX . Results following Injection of Tritiated Thymidine . . . . . . . . . . . . . X . Erythropoiesis in Blood of Vertebrates with Nucleated Erythrocytes. Formation of Clone Cells from Nuclei of Young Mature Erythrocytes . . . . . . . . . . . XI . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI1. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Fluidity of Cell Membranes-Current

98 98 99 100 105 107 109 114 117 117 118

Concepts and Trends

M . SHINITZKY A N D P . HENKART I . The Lipid Fluidity . . . . . . . . . . . . . . . . . . . . . . . . . . I1 The Protein Mobility . . . . . . . . . . . . . . . . . . . . . . . . . 111. Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

121 130 144 145

Macrophage-Lymphocyte Interactions in Immune Induction MARCFELDMANN. ALANROSENTHAL. AND PETERERB I. I1. 111. IV . V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Macrophage Function in Antigen-Specific T-cell Proliferation . . . . . . . . Macrophage Function in Helper T-cell Induction . . . . . . . . . . . . . Macrophage-B-Lymphocyte Interactions in Antibody Production . . . . . . . Concluding Discussion . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Note Added in Proof . . . . . . . . . . . . . . . . . . . . . . . . .

149 151 163 171 172 176 178

Immunohistochemistry of Luteinizing Hormone-Releasing HormoneProducing Neurons of the Vertebrates JULIENBARRY

I. Introduction . . . . . . . . . . . . . . . . . 11. Techniques of Study . . . . . . . . . . . . . III. Morphology of LH-RH-Reactive Perikarya . . . IV . Topography of LH-RH-Reactive Perikarya . . . V . Hypothalamohypophyseal LH-RH Tracts . . . .

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

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

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

179 180 188 194 201

vii

CONTENTS VI . Reopticoterminal LH-RH Tract . . . . . . . . . . . . . . . . . . . . . VII . Extrahypothalamic LH-RH Tracts . . . . . . . . . . . . . . . . . . . . VIII . General Discussion and Conclusions . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. .

207 209 213 214

Cell Reparation of Non-DNA Injury V . YA. ALEXANDROV I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Recovery during the Action of Injurious Agents (Reparatory Adaptation) . . . m . Repair after Elimination of Injurious Agent . . . . . . . . . . . . . . . IV . Repair of Thermal Injuries . . . . . . . . . . . . . . . . . . . . . . . V . Mechanism of Heat Injurious Action . . . . . . . . . . . . . . . . . . VI . Resynthesis or Reactivation? . . . . . . . . . . . . . . . . . . . . . . VII . Some Evidence of Protein Renativation . . . . . . . . . . . . . . . . VIII . What Happens in the Cell? . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. .

. .

223 225 229 230 252 257 259 263 264

Ultrastructure of the Carotid Body in the Mammals ALAINVERNA Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Histological Features . . . . . . . . . . . . . . . . . . . . . . . . . Ultrastructure of Type I and Type I1 Cells . . . . . . . . . . . . . . . . Type I Cell Innervation . . . . . . . . . . . . . . . . . . . . . . . . Vascular Innervation and Efferent Inhibition . . . . . . . . . . . . . . . VI. Ultrastructural Changes after Stimulation of Chemoreceptor and Pathology . . . VII . Embryology and Development . . . . . . . . . . . . . . . . . . . . . VIII . Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I. I1. 111. IV . V.

271 274 279 289 307 315 320 322 323

The Cytology and Cytochemistry of the Wool Follicle DONALDF. G . ORWIN I. I1. 111. IV . V. VI . VII . VIII .

Introduction . . Dermal Papilla . Bulb . . . . . . Medulla . . . . Cortex . . . . . Fiber Cuticle . . Inner Root Sheath Outer Root Sheath

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

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

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

331 334 337 338 342 353 360 366

...

Vlll

CONTENTS

IX. Connective Tissue Sheath . . . . . . . . . . . . . . . . . . . . . . . X . Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

SUBJECT INDEX . . . . . . . . . CONTENTS OF PREVIOUS VOLUMES.

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

369 369 369

. .

. .

375 379

List of Contributors Numbers in parentheses indicate the pages on which the authors' contributions begin.

V. YA. ALEXANDROV (223), Laboratory of Cytoecology and Cytophysiology, Komarov Botanical Institute, USSR Academy of Sciences, Leningrad, 197022 USSR JULIEN BARRY(179), U. 156 INSERM and Laboratory of Histology and Embryology, Faculty of Medicine, 59045 Lille, Cedex France

ALEXANDER J. CHARLSON (27), School of Chemistry, Macquarie University, North Ryde, New South Wales, 2113, Australia VIBEKEE. ENGELBERT* (93), The Ramsey Wright Zoological Laboratories, University of Toronto, Toronto, Canada PETERERB (149),Institute for Microbiology, University of Basel, Petersplatz 10, 4003 Basel, Switzerland MARCFELDMANN (149),ICRF Tumor Immunology Unit, Department of Zoology, University College, London WCIE 6BT, England TIMOTHY C. HALL(l), Department of Horticulture, University of Wisconsin, Madison, Wisconsin 53706 P. HENKART(121), Immunology Branch, The National Cancer Institute, Bethesda, Maryland 20014

DONALD F . G. ORWIN(331), Wool Research Organisation of New Zealand, Inc., Christchurch, New Zealand

CHRISTIAN PETZELT (53), Institute for Cell Research, German Cancer Research Center, Heidelberg, West Germany BEVANL. REID(27), Queen Elizabeth I1 Research Institute for Mothers and Infants, University of Sydney, New South Wales, 2006, Australia ALAN ROSENTHAL (149), Department of Immunology, Merck Institute for Therapeutic Research, Rahway, New Jersey 07065 M. SHINITZKY (121), The Department of Membrane Research, The Weizmann institute of Science, Rehovot, Israel ALAINVERNA(271),Laboratory of Cytology, University of Bordeaux 11, 33405 Talence. France *Resent address: Vosnaesvej 6, Loegten, 8541 Skedstrup, Denmark.

ix

This Page Intentionally Left Blank

1NTERNATlONAL REVIEW OF CYTOLOGY, VOL. 60

Transfer RNA-like Structures in Viral Genomes TIMOTHY C. HALL Department of Horticulture, University of Wisconsin, Madison, Wisconsin I. Introduction . . . . . II. Occurrence and Properties

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

A. Covalently Bound Arninoacylatable Sequences in Plant virus RNAs . . . . . . . . . . . . . . . . . . . . . B. Covalently Bound Arninoacylatable Sequences in Animal Virus RNAs . . . . . . . . . . . . . . . . . . . . . C. Noncovalent Association of tRNAs with Viral RNAs . . . . 111. Nucleotide Sequences of 3’-OH Ends of Arninoacylatable Viral RNAs . . . . . . . . . . . . . . . . . . . . . . A. Turnip Yellow Mosaic Virus RNA . . . . . . . . . . B. Brome Mosaic Virus RNA . . . . . . . . . . . . C. Tobacco Mosaic Virus RNA . . . . . . . . . . . . IV. Biological Function of Viral tRNA-like Structures . . . . . . A. Does a Biological Role Exist? . . . . . . . . . . . . B. Role as an Amino Acid Donor . . . . . . . . . . . . C. Role in Competing with Host mRNA for Ribosomes . . . . D. Role in Competing for Elongation Factors . . . . . . . . E. Role in Coat Protein Assembly . . . . . . . . . . . . F. Role in Replication . . . . . . . . . . . . . . . . V. Summary . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . ,

.

1

2 2 7

8 9 9 10 12 13 13 16 17

17 19 20 22 23

I. Introduction Most messenger RNAs (mRNAs) isolated to date have been found to contain regions that do not code for amino acid sequences. The 5’ terminus of many eukaryotic mRNAs has a 7-methylguanosine cap structure (Shatkin, 1976), as does that of most RNA viruses infecting eukaryotes. A peptide sequence covalently bound to the 5’ end of the genome RNA has been found in poliovirus (Lee et a l . , 1977), encephalomyocarditis virus (EMCV) (Hruby and Roberts, 1978), and cowpea mosaic virus (A. Van Kammen, personal communication) RNAs. Other viral RNAs, such as those of satellite tobacco necrosis virus (Lesnaw and Reichmann, 1970; L u n g et al., 1976), have neither a 7-methylguanosine cap nor a peptide sequence at their 5’ terminus. An untranslated leader sequence precedes the coding region and is variable in length; only 10 bases precede the AUG initiation codon of brome mosaic virus (BMV) RNA 4 (Dasgupta et al., 1

Copyright @ 1979 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-364360-0

2

TIMOTHY C. HALL

1975), while 36 bases precede the initiation of codon on alfalfa mosaic virus (AlMV) RNA 4 (Koper-Zwarthoff et al., 1977). Both BMV RNA 4 and AlMV RNA 4 code for the viral coat protein. Intracistronic noncoding regions have been found to be present in precursor forms of mouse P-globin mRNA (Tilghman et al., 1978) and other eukaryotic mRNAs; these are removed during processing steps yielding the mRNA molecule which is translated. The polycistronic mRNAs of procaryotic viruses such as MS2 bacteriophage contain extensive intercistronic regions that are not translated (Fiers et al., 1976). The realization that many plant and animal mRNAs have a poly(A) sequence at their 3’ end has greatly facilitated their isolation (Aviv and Leder, 1972), although the biological function of this, and of the other nontranslated structures, remains unclear. Less well known, and even more enigmatic, is the fact that several RNAs which represent viral genomes possess a 3’ sequence that can be aminoacylated. These RNAs, which also serve as messenger RNAs, are esterified by a specific amino acid through the mediation of an aminoacyl-tRNA synthetase under reaction conditions similar to those used for aminocylation of transfer RNAs (tRNAs). In this review, I describe the occurrence of tRNA-like sequences in viral RNAs and discuss the available evidence concerning their possible biological functions.

11. Occurrence and Properties

A. COVALENTLY BOUNDAMINOACYLATABLE SEQUENCES IN PLANT VIRUS RNAs The first definitive reports of aminoacylation of a viral RNA were those of Pinck et al. (1970) and Yot et al. (1970) who showed that turnip yellow mosaic virus (TYMV) RNA accepted valine. A previous article (Beljanski, 1965) had indicated that valine, and possibly other amino acids, became associated with TYMV RNA, but the nature of the association was not clearly delineated. Subsequently, the RNAs of BMV (Hall et al., 1972) and of tobacco mosaic virus (TMV) (Oberg and Philipson, 1972) were shown to be able to accept specific amino acids. Aminoacylatable sequences have now been shown to be present in the genome RNA of viruses belonging to widely differing groups of viruses (Table I). RNA from several viruses not mentioned in Table I has been tested for amino acid accepting activity, but without positive results. Plant viral RNAs that have been examined include: AlMV RNA, pea enation mosaic virus RNA, southern bean mosaic virus RNA, tobacco streak virus RNA, and tobacco rattle virus RNA. Tobacco etch virus RNA may accept histidine, but contamination with TMV RNA has not been ruled out; RNA from pepper mottle virus (like

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

3

tobacco etch virus, a member of the potato virus Y group) did not accept histidine. Bean pod mottle virus RNA may accept tyrosine after fragmentation. Poliovirus RNA has been tested unsuccessfully for amino acid accepting activity. Aminoacylation of bacteriophage RNAs (from MS2, PP7, and Qp) has been attempted unsuccessfully. It must be borne in mind that the aminoacylation reaction is sensitive to conditions such as pH, Mg2+and other ion concentrations, and source of enzyme used. Therefore, a negative result does not eliminate the possibility that future studies may demonstrate an amino acid accepting activity for any of the viral RNAs mentioned above. Some differences in the details of aminoacylation of individual viral RNAs have become apparent. Yot et al. (1970) showed that the 3' terminus of TYMV RNA was CCoH and that nucleotidyl transferase was needed to mediate the formation of a CCAoHterminus prior to the addition of valine by purified valyltRNA synthetase isolated from Escherichia coli. This has been confirmed by Litvak et al. (1973b) and by Giegt et al. (1978) who were able to obtain stoichiometric valylation of TYMV RNA bearing a C C b H terminus with purified yeast valyl-tRNA synthetase. It is now known that two RNA components of TYMV exist (Klein et al., 1976), the most abundant having a molecular weight of 2 x lo6 and the other, 2.6 x lo5; both accept valine. The genomic RNA of several viruses belonging to the tymovirus group has been found to bind valine, although RNA of belladonna mottle virus may bind alanine (Pinck et al., 1972). Native BMV RNAs terminate in C C b H and each of the four RNA components accept tyrosine (Hall et al., 1972). Preliminary data for RNA from variant 5 of BMV which grows on tobacco (Bancroft et al., 1975) indicate that it may not accept tyrosine (T. L. German, J. W. Pyne, and T. C. Hall, unpublished), although the slow growth of this variant has made it difficult to obtain adequate quantities of purified RNA. RNA from the common strain of TMV binds histidine (Oberg and Philipson, 1972), but the extent of histidylation was found to vary for several strains of TMV (Carriquiry and Litvak, 1974). The cowpea strain was found to accept valine (Beachy et al., 1976). Cucumber mosaic virus has four RNA components of sizes similar to those of BMV RNAs, and each component was found to accept tyrosine (Kohl and Hall, 1974). Recently, Gould and Symons (1978) synthesized radioactive DNA complementary to each of the four major RNAs of both cucumber mosaic virus and AlMV. The DNA was used in hybridization experiments to show that the viral coat protein cistron was present both in the small RNA (component 4) and toward the 3' end of RNA 3 for each virus. The RNAs of AlMV have not been aminoacylated in vitro, and no sequence homology between RNAs 1 , 2 , or 3 was found. Conversely, the sequences of 200 nucleotides at the 3' end of each of the four RNAs of cucumber mosaic virus were shown to be identical. A small RNA has often been found to be associated with purified cucumber mosaic virus;

TABLE I VIRALGENOMES HAVING A COVALENTLY BOUNDSEQUENCE CAPABLE OF ACCEPTING AN AMINOACID

Source of RNA Plant v i s e s Bromovirus group Broad bean mottle virus

P

Brome mosaic virus Russian strain

Cowpea chlorotic mottle virus

cucumovirus group Cucumber mosaic virus

MW (x of RNA components aminoacylated

Amino acid bound

Reference

Comments'

1 2 3 4

1.10 1.03 0.90 0.36

Tymsine Tyrosine Tyrosine Tymsine

Kohl and Hall (1974)

RNA degrades readily

1 2 3 4

1.09 0.99 0.75 0.28

Tyrosine Tyrosine Tyrosine Tyrosine

Hall et 01. (1972)

About 60% tyrosylation with synthetasefrom wheat germ or bean cotyledons. E. coli and yeast synthetases are not active in catalyzing tyrosylation. An a161 fragmentis obtained on T, RNase digestion (Dasgupta and Kaesberg, 1977); pancreatic RNase cleaves between nucleotides 65 and 66

1 2 3 4

1.15

Kohl and Hall (1974)

Over 90% tyrosylation has been obtained

0.85 0.32

Tyrosine Tyrosine Tyrosine Tymsine

1 2 3 4

1.0-1.3 0.9-1.1 0.7-0.8 0.33

Tymsine Tyrosine Tymsine Tymsine

Kohl and Hall (1974)

Associated RNA CARNA 5 does not appear to be. capable of accepting an amino acid

1.oo

Tobamovims group Tobacco mosaic virus U, (wild type or common) strain

2.00

Histidine

Oberg and Philipson (1972); Guilley er al. (1975); Beachey et al. (1976)

Serine Methionine Histidine Valine Valine Histidineb?

Sela (1972)

Valine

Lamy et al. (1975)

2.00 2.00

Histidine Histidine

Caniquiry and Litvak (1974) Caniquiry and Litvak (1974)

13% charging obtained 37% charging obtained

2.10 ?2.00 1.90

Alanine? Valine Valine

F’inck et al. (1972) F’inck et al. (1972) Pinck et al. (1972) Pinck and Hall (1978)

Preliminary result Preliminary result Early reports indicated that a 4s KNA which accepted lysine was associated with genome RNA. Although host tRNAs are found on the virion surface and are. to a small degree encapsidated, no specific association with the viral RNA wasconfirmed (Pmckand Hall, 1978). Valylation of up to 80% has been obtained

Fragments

U, strain Cowpea strain Dahlmense strain Green tomato atypical mosaic strain Holmes ribgrass (HRG or U,) Vulgare strain Tymovirus group Belladonna mottle virus Cacao yellow mosaic virus Eggplant mosaic virus

L

S

2.00 2.10 0.28 2.00 ?2.00

Carriquiry and Litvak (1974) Beachy er al. (1976) Beachy et al. (1976) Caniquiry and Litvak (1974)

Enzyme from yeast typically used for aminoacylation. About409bchargingobtained. An a 7 1 fragment is obtained on T, RNase digestion, and nucleotide 32 (from 3’ end) can be methylated (Lesiewicz and Dudock, 1978) Fragments about 55 nucleotides long obtained with a “pH 5” enzyme from tobacco 40% charging obtained 5.9% charging obtained 8.4% charging obtained 9.0% charging obtained ?

(continued)

TABLE I (continued)

Source of RNA

Okra (gombo) mosaic virus Tumip yellow mosaic virus

Animal viruses Picomavirus group Encephalomyocarditis virus

Mengovirus

MW (x lo+) of RNA components aminoacylated

H L

Amino acid bound

Reference

?2.00 2.00 0.26

Valine Valine Valine

F'inck et al. (1972) Gieg6 er al. (1978) Gieg6 et al. (1978)

2.70

Serine

Lindley and Stebbing (1977)

2.30

Histidine

Salomon and Littauer (1974)

Comments" The first definitive report of aminoacylationof viral genome RNA was valylation of TYMV RNA by Pinck er al. (1970). An a158 fragment can be obtained by TI RNase digestion (Briand et al., 1977); after enzymatic addition of a 3'-terminal bHvalylationlevels >95%for both components have been obtained with purified yeast synthetase (Giege eral., 1978). A 3' fragment of about 112 nucleotides can be obtained by RNase P digestion (Silberklang et al., 1977) K, (Val) for the H RNA = 3.7 X lO-'M, for the L RNA = 5.3 X lO-'M compared with yeast tRNAva' = 1.2 x lo-' M under the same conditions; rate constants (k,) are 20, 29, and 143 minute-' respectively (Gieg6 eral., 1978) The genome RNA contains a 3' ply(A) tract; hence, the aminoacylatable sequence is not at the 3'-OH end of the native RNA. Aminoacylation was 2.4 to 8.0%;after the reaction (using a rabbit liver synthetase) a heterogeneous population of fragments migrating on electrophoresisclose to the position of a tRNA marker was obtained The.genome RNA is fragmented in the reaction, the histidylated fragments being 0.2 to 0.35 and 1.25 X 106 M W . Aminoacylation was 18%

"Effciency of aminoacylation is given as percentage, being the number of moles of amino acid bound per 100 moles of intact genome RNA. b(?) Results not confumed or preliminary.

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

7

it has been called CARNA 5 and it has been identified as being responsible for severe and economically deleterious injury to tomatoes by this virus (Kaper and Waterworth, 1977). It appears that CARNA 5 does not accept tyrosine (L. Hirth, personal communication), and if this component represents a type of “satellite” virus, this may not be too surprising. The similar conditions and kinetics for aminoacylation of viral and tRNAs (Kohl and Hall, 1974; Giege et al., 1978) strongly suggest that viral RNAs are aminoacylated within the host. However, it has proved difficult to obtain direct evidence on this aspect. An amino acid bound to viral RNA within the virion should remain bound during acidic extraction, but would be released on alkaline extraction. Therefore, one might expect to see some difference in acceptor capacity for batches of RNA isolated under acidic or alkaline conditions. However, we found tyrosine acceptance to be identical in extent and kinetics for two lots of BMV RNA isolated in parallel from virus, one lot being extracted in the presence of Tris (pH 8) and the other in the presence of acetate buffer (pH 4.5). This suggests that BMV RNAs are not tyrosylated within the virion; if so, some deacylation step probably occurs during encapsidation. Recently, Joshi et al. (1978) microinjected TYMV RNA into Xenopus laevis oocytes in the presence of [3H]valine. Subsequently, RNA was extracted from the oocytes and [3H]valine oligonucleotides obtained by TI ribonuclease digestion were analyzed by thinlayer chromatography. The results obtained showed that the viral RNA had been valylated within the oocytes, and in fact the RNA appeared to have been cleaved, releasing a fragment similar in size to that of a tRNA. Despite these results, which strongly support the case for aminoacylation of viral RNAs in vivo, definitive data confirming the aminoacylation of viral RNA within a natural host cell have yet to be obtained. Even though trends can be seen with respect to the tRNA-like character of the plant viruses shown in Table I, there are always exceptions. For a while, it appeared that within a particular group of viruses whose RNA was capable of accepting an amino acid, the amino acid was always the same, i.e., valine for tymoviruses, tyrosine for bromoviruses, and histidine for tobamoviruses. However, if the data indicating alanine rather than valine acceptance by belladonna mottle virus RNA (Pinck et al.. 1972) are substantiated and if the cowpea strain of TMV (whose RNA binds valine) is genuinely related to the common strain (whose RNA binds histidine), these findings detract from the concept of uniformity of amino acid selection within a viral group. B. COVALENTLY BOUNDAMINOACYLATABLE SEQUENCES IN ANIMAL VIRUS RNAs Even more significant differences in the nature of amino acid binding are apparent in the cases of mengovirus and EMCV whose RNAs have been reported

8

TIMOTHY C. HALL

to bind histidine (Salomon and Littauer, 1974) and serine (Lindley and Stebbing, 1977), respectively. EMCV RNA has been reported to have a 3‘-terminal oligo(A) tract (Emtage et al., 1976). Therefore, cleavage or sequential degradation of the 3’-terminal sequence must take place before the aminoacylation character becomes functional. Some preliminary data (T. C. Hall, J. S. Semancik, and J. W. Davies, unpublished) suggested that RNA of bean pod mottle virus accepted tyrosine, but if so, the RNA of this virus must also undergo modification since it is known to be polyadenylated (Semancik, 1974), presumably at the 3‘ end. C. NONCOVALENT ASSOCIATION OF tRNAs WITH VIRALRNAs Sendai virions contain substantial amounts of 4s RNA (Barry and Bukrinsikaya, 1968). Kolakofsky (1972) reported that some of this 4s RNA serves as a substrate for tRNA nucleotidyl transferase present in Sendai virions, and hence that part of the 4s RNA is tRNA whose CCAoHend is missing. Reports of the presence of tRNA within virions of tumor viruses such as Rous sarcoma (Wang et al., 1973) and avian myoblastosis viruses (Rosenthal and Zamecnik, 1973; Stromberg and Litwack, 1973) were followed by the demonstration that a specific 4s RNA serves as a primer in the transcription of DNA from the 70s RNA of Rous sarcoma virus (Dahlberg et al., 1974). Because 4s RNA from uninfected cells can also serve as a primer for reverse transcription (Sawyer et al., 1974), it seems unlikely that this is a virally coded RNA. However, of the several tRNA species found within tumor viruses, those accepting tryptophan and, to a lesser extent, lysine were relatively difficult to dissociate from the avian myeloblastosis virus genome RNA (Waters et al., 1975). The binding site of the 4s primer was shown to be at, or close to, the 5’ end of the 35s genome RNA of avian sarcoma virus (Taylor and Illmensee, 1975). N-Formyl methionine tRNA was apparently selectively associated with the Rous sarcoma virus RNA genome (Faras, 1975), although it is now known that this observation was due to mischarging of tRNATrPwhich actually serves as primer (Harada et al., 1975). A tRNA”O that is tightly bound to the genome RNA of murine leukemia virus has been shown to serve as the major primer for DNA synthesis directed by this RNA in v i m (Peters et al., 1977). Waters (1978) found that tRNALY8 was the predominant tRNA in murine mammary tumor virus. The report (Pinck et al., 1974) that a lysine-accepting 4s RNA was specifically bound to the genome RNA of eggplant mosaic virus (EMV) was intriguing since it hinted that some function might exist in plant viral replication analogous to the primer function of 4s RNA associated with tumor viruses. Subsequent studies (Pinck and Hall, 1978) revealed that the association of lysine-accepting tRNA with EMV RNA was not highly specific. Nevertheless, tRNA molecules are found associated with EMV (both on the surface and encapsidated), and

9

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

lysine-accepting activity can be readily demonstrated. In connection with this observation, it is interesting that Waters and Mullin (1977) have stressed that a cellular origin of a viral component does not per se exclude a functional role for that component in the life cycle of the virion, and they provided extensive evidence for preferential association of host tRNAs with many RNA tumor viruses. Tight binding of host (Escherichia coli) tRNALeUto MS2 virions has been reported (Di Natale and Eilat, 1976). Changes in the spectrum, or activity, of tRNAs present in cells have been observed after infection by several viruses, for example in tRNALeU after bacteriophage T2 infection (Sueoka and Kano-Sueoka, 1970) and in tRNAPmafter QP infection (Hung and Overby, 1968). A modified (Gefter and tyrosine suppressor tRNA was found in cells infected with 80 dsu Russell, 1969). The DNA of coliphages T2, T4, T5, and T6 contains genes coding for specific tRNAs (Weiss et al., 1968; McClain et al., 1972; Paddock and Abelson, 1973). These tRNAs are functional in amino acid donation, but it appears that their functioning is not essential for multiplication of laboratory strains of bacteriophage T4 (Wilson et al., 1972; Chen et al., 1975; McClain et a l . , 1975). The association of tRNA with tumor virus RNA or coliphage DNA is probably quite unrelated to the phenomenon of tRNA-like structures in plant viruses. However, the diversity found, even within plant viruses, of the tRNA-like character does raise the question as to whether a single biological role is achieved by these structures. +

111. Nucleotide Sequences of 3'-OH Ends of Aminoacylatable Viral RNAs

A. TURNIPYELLOWMOSAICVIRUSR N A Briand et al. (1977) have established the sequence of the 3'-OH-terminal 159 nucleotides (n159) of TYMV RNA. This information is especially interesting since the sequence includes 51 nucleotides of the 3'-terminal part of the coat protein cistron; hence, there is a stretch of 108 untranslated nucleotides at the 3' extremity of TYMV RNA. A fragment of about 110 nucleotides is released by ribonuclease P (a tRNA maturation endonuclease) from the 3' terminus of TYMV RNA (Silberklang et a l . , 1977), and it is known from the work of Prochiantz and Haenni (1973) that the ribonuclease-P-released fragment can be esterified with valine. These data show that the valylation property of TYMV RNA requires no more than the 110 3'-terminal nucleotides, and probably fewer. Further, elucidation of the primary sequence of this 3' fragment has permitted construction of models for its secondary structure and comparison with the cloverleaf models for tRNAs (Fig. 1). As Briand et al. (1977) have elegantly

10

TIMOTHY C. HALL

A 3'OH end 01 TYMV RNAI

U GCCU

....

C G G G

0 C

=

uG cb Acu

U - b .G

c c

c C

D

G

C G

G D

c

c ' b c b c

FIG. 1. Sequence comparison of the 3'-OH end of TYMV RNAs (A) and of yeast tRNAVsl (B). The nucleotides of tRNAVs'common to the putative cloverleaf structure of the 3'-OH end of TYMV RNAs are in boxes. (From Giege et al., 1978.)

shown, there is surprisingly little correspondence with constant tRNA features (especially with the dihydrouridine loop) or with tRNAVa'despite similar kinetics and specificity (see Table I) for the aminoacylation reaction (Giege et al., 1978). In particular, there are no modified bases and the sequence GUUCR is absent; the equivalent pentanucleotide G T W R is present in all tRNAs, and it is thought to undergo protein-mediated base pairing with a complementary sequence of ribsomal 5 s RNA during protein synthesis (Richter et al., 1973). The models proposed by Hirth's group are also considerably larger than for authentic tRNA, although Briand et al. (1977) noted that in the structure proposed as cloverleaf A (Fig. 1 A), the excess bases are essentially limited to the dihydrouridine and T W arm analogs and might be accommodated within an L-shaped tertiary structure, such as that proposed for tRNAphe(Kim et al., 1974; Robertus et al., 1974). Particularly interesting is the close similarity between arm I1 of tRNAValand that of the postulated models for TYMV; this arm includes the CAC (valine) anticodon and may be important in recognition by valyl-tRNA synthetase. B. BROMEMOSAICVIRUSRNA

Under restrictive conditions (incubation at 0°C for 15 minutes at a substrate:T, ribonuclease enzyme ratio of 1000:1) a 3'-terminal fragment of 161 nucleotides (R161) can be cleaved from each of the RNA components of BMV (Dasgupta et al., 1975). Although these RNAs have different messenger functions (Shih and Kaesberg, 1973, 1976), Bastin et al. (1976) showed by oligonucleotide mapping that the 0161 fragments from RNAs 3 and 4 were identical. The R161 fragment

11

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

from RNA 2 differed by only one base, and the a 1 6 1 fragment from RNA 1 differed by two base substitutions from the a 1 6 1 sequence of RNAs 3 and 4. The a 1 6 1 fragments bound tyrosine as efficiently as did the intact RNA, but a 3' fragment 65 bases long obtained by pancreatic ribonuclease digestion (1: 10,000 enzymembstrate ratio) of a 1 6 1 could not be tyrosylated. However, BMV RNA fragment a 1 6 1 containing a hidden break between nucleotides 65 and 66 (obtained by S1 nuclease treatment) can be tyrosylated (Dasgupta and Kaesberg, 1977), as can some tRNAs containing a hidden break (Chambers, 1971). Figure 2 shows the secondary structure for BMV RNA a 1 6 1 proposed by Dasgupta and Kaesberg (1977). Comparison of this with the structure proposed by Briand et al. (1977) for 0,159 of TYMV RNA reveals a generally similar form. As in the case of the TYMV RNA TI fragment, the BMV RNA a 1 6 1 differs from tRNA in having no modified bases, but the aminoacylatable terminus and anticodon region are analagous to those of the corresponding yeast tRNA. The anticodon loops of the viral RNAs are greatly displaced when compared with those of tRNAs; the viral anticodon sequence is at nucleotides 65-67 (from the 3' terminus) for BMV RNA and at nucleotides 55-57 for TYMV compared with nucleotides 41 -43 for both tRNATyrand tRNAValfrom yeast.

R'

115 UC

Aclh5

A' C

U A UA CG UA AU GC

\ C

,*U

,

U

I45

UG GC AU GC AU

AC U A UA CG UA

U

GC AU A' GC C AU AU C GC AU G UUGC A C 'CG - .AGGUGCCUUU -. . .. ... UA G CG C

u::

AUC GC AGGUGCCUUU U CG \

iIa U A

__

GC

C

161

A C

A

161 A AUGUCA U A UACAGU 90U C C C M G G

UAAG G

A

A55 GCUU \

GU CG UA UA GC C G

A A

\

B UA _..

G C C 6 'UA65

FIG.2. Possible secondary structures for 161-nucleotide 3'-OH fragment obtained from BMV RNA component 4. (A) Drawn to show maximum base pairing. Preferred sites of action of T, and pancreatic ribonucleases are indicated by solid and broken arrows, respectively. (B) Drawn to illustrate a similarity to the cloverleaf structure of tRNA. (From Dasgupta and Kaesberg, 1977.)

12

TIMOTHY C. HALL

C. TOBACCO MOSAICVIRUSRNA

An 0 7 1 fragment cleaved from TMV RNA under restrictive conditions by TI ribonuclease digestion has been sequenced by Guilley et al. (1975). As may be seen from Fig. 3, there is little sequence homology with tRNAHis. It is not presently known if this TMV RNA 0 7 1 can accept histidine, as does the intact genome RNA. Sela (1972) reported the arninoacylation of fragments of TMV RNA with serine and methionine using a pH 5 enzyme from tobacco. This observation has not been confirmed. Analysis of the sequence of an 074 fragment obtained from the RNA of green tomato atypical mosaic virus (GTAMV) strain of TMV by TIribonuclease digestion revealed considerable differences from that of the wild type (Lamy et al., 1975). However, the sequences in the loops are very similar (compare structures B and C, Fig. 3) and the 3’-terminal 18 nucleotides differ only by a G-C to C-G inversion and by the addition of an extra guanosine between positions 10 and 13 in GTAMV RNA. The GTAMV RNA contains the oligonucleotide sequence UUCG which is characteristic of the T q C loop of tRNA.

AOH C C

AOH C C

X

x-x x-x x-x x-x x-x x-XlO

X

x30

X X

A. NORMAL tRNA

C

C

C CCCCGG A

U U

....

GGCGqTto G ACC GCGG A

ccccc ....#A

G20

G2o XXXXGTVrC x T

AOH

c 10 ccucucccu ....... do GAFGGGU C

G C G40 A B. TMV RNA

U

U

A C A

C. GTAMV RNA

FIG.3 . Comparison of a partial sequence of a normal tRNA (A), the nucleotide sequence of the 3‘-terminal 41 nucleotides of TMV RNA arranged in a tRNA-like structure (B), and the 3’-terminal 47 nucleotides of the green tomato atypical mosaic strain of TMV RNA (C). The sites of ribothymidine formation in normal tRNA and in TMV RNA are indicated by the arrows. The structures shown in (A) and (B) are from Lesiewicz and Dudock (1978). and the sequence shown in (C) is redrawn from Lamy et a / . (1975).

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

13

Although TMV RNA contains no modified bases other than that of the 5 ’ cap, Lesiewicz and Dudock (1978) have found that a ribothymidine (rT)-forming tRNA methyltransferase preparation from Escherichiu coli quantitatively methylates TMV RNA, 1 mole of rT being formed per mole of viral RNA. The uridine residue methylated is the 32nd base from the 3’ end, and this site resembles the normal rT position in tRNA (Fig. 3). This evidence, together with the conservation of the hairpin loops, supports the notion that tRNA-like properties are important to the function of the viral RNA. It is not clear if the similarities represent divergence of the viral RNA from an ancestral pathogen which contained a typical tRNA sequence, or if they are the result of a convergent evolution toward a form that, by mimicking a tRNA, helps the viral RNA to establish infection.

IV. Biological Function of Viral tRNA-like Structures A. DOESA BIOLOGICAL ROLEEXIST? When TYMV RNA was the single known example of a viral RNA having an amino acid acceptor capability, the possibility that this structure was fortuitous could not be discounted. However, such a possibility appears very unlikely with the discovery of other, unrelated viruses possessing this character (see Table I; it is likely that other viral RNAs will subsequently be added to this list). Despite extensive studies on tRNAs, the sequence and structural requirements for aminoacylation remain poorly defined (Rich and RajBhandary, 1976). Nevertheless, the fidelity of translation of the genetic code depends upon the specificity of aminoacylation because of the adaptor role of tRNAs between amino acid insertion into nascent protein and the messenger template. Hence, recognition of a tRNA structure by aminoacyl-tRNA synthetase must meet rigorously defined conditions. The high degree of sequence conservation exhibited by the four RNA components of BMV is in accordance with this situation. It will be interesting to learn the 3’ sequence of CCMV RNA components, since preliminary oligonucleotide maps for these RNAs suggested a different and more complex situation than for the BMV RNAs (Bastin et al., 1976). The possibility cannot be discounted that nuclease-cleaved fragments additional to those from the 3’ ends of CCMV RNAs were present in the fraction used for these initial CCMV mapping experiments. Fiers et al. (1976) have commented on the high degree of sequence conservation for the untranslated 3’-teminal 183 nucleotides of bacteriophage MS2. Since no amino acid acceptance has been found for MS2 RNA (despite attempts by several groups), the evolutionary constraint against modification of 3‘ termini may be unrelated (or additional) to the need for recognition by synthetase enzymes. Conceivably, aminoacylation represents a primitive condition

14

TIMOTHY C. HALL

and nonaminoacylatableviral RNAs have undergone modifications that serve the same role but do not need aminoacylation as an intermediate step. Certainly, the 3' terminus of MS2 phage RNA possesses a high degree of secondary structure (Fig. 4). The recognition of the 3' terminus of aminoacylatable viral RNAs by tRNArelated enzymes other than aminoacyl-tRNA synthetase provides further support for a tRNA-like function for these structures. As noted earlier, nucleotidyl transferase reacts efficiently with TYMV RNA (Yot et al., 1970; Litvak et al., 1970; Briand et al., 1977) to yield the aminoacylatable C C b H terminus from the nonchargeable CCoH terminus of virion RNA, and ribonuclease P releases an aminoacylatable 5 s fragment from the 3' end of TYMV RNA (Silberklang et al., 1977). Although no modified bases have been observed in tRNA-like fragments of native viral RNAs, the report of Lesiewicz and Dudock (1978) showing methylation of TMV RNA further demonstrates ability of the viral tRNA structure to be recognized by tRNA-related enzymes. Native BMV RNA has a 3'CCAoH sequence, but it is recognized by nucleotidyl transferase since the CCAoH terminus is reestablished after removal of the terminal adenosine by periodate oxidation and aniline cleavage. The aminoacylated, but not the nonaminoacylated, forms of TYMV, TMV, and BMV RNAs will interact with elongation factors (Litvak et al., 1973a; Bastin and Hall, 1976) as do aminoacylated tRNAs. BMV RNA, like TMV RNA, can be methylated but the position of methylation is not yet established (B. Dudock, personal communication).

FIG.4. 3'-OH end of MS2 RNA. Oriented to show vague resemblence to structures proposed for 3' termini of aminoacylatable viral RNAs. (From Fiers et al.. 1976.)

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

15

The above considerations argue strongly for the importance of the tRNA-like structure to the infectivity of aminoacylatable viral RNAs. Direct evidence for the need for a functional 3’ terminus has been obtained in the case of BMV RNA, where RNA that was modified by acetylation after tyrosylation was only 27% as infective as a control sample (Kohl and Hall, 1977). Reduction of infectivity was also evident for BMV RNA that has been periodate oxidized; however, the 5’ cap structure (Dasgupta el al., 1976) contains a free hydroxyl and hence is also susceptible to this modification. Indeed, BMV RNA that has been tyrosylated (to protect the 3’ terminus) is reduced in infectivity after 5’ modification by periodate oxidation, showing that both the 5‘ and 3’ ends of the viral RNA are important in infective processes. Aminoacylated BMV RNA is as good a messenger for protein synthesis in vitro as is virion RNA. The oxidized (Shih et al., 1974) and acetylated aminoacylated forms of BMV RNA are less efficient messengers, but no difference in the polypeptides synthesized in vitro (compared with the native forms) has been detected using these modified RNAs as messengers (J. W. Pyne and T. C. Hall, unpublished observations; T. C. Hall and M. Pinck, unpublished observations). The acetylated, aminoacylated derivative of BMV RNA (Fig. 5 ) represents a form in which only the 3’ terminus is modified, yet a drastic loss is infectivity results. Attempts to recover infectivity by removal of the acetylated terminus and repair of the 3’ end have been rewarded with partial success (P. A. Kiberstis and T. C. Hall, unpublished observations), but complications arise in these experiments from the need for essentially 100%of the RNA molecules to be tyrosylated prior to the acetylation reaction. Otherwise, infectivity resulting from the proportion of unmodified molecules (only aminoacylated molecules can be acetylated) clouds the results. Nevertheless, the data currently available strongly suggest that a functional 3’ terminus is necessary for the establishment of infectivity by viruses whose RNA is capable of aminoacylation. Although MS2 and Qp RNAs have not been aminoacylated, as noted above their 3’ ends do have some structural resemblance to those of aminoacylatable viral RNAs. Sabo et al. (1977) reported that a guanosine to adenosine transition at position 16 from the 3‘ end of Qp RNA caused loss of infectivity. Reversion of this change by site-directed mutagenesis resulted in a significant increase in specific infectivity. Salomon et a f . (1976) found that removal of 5 to 10 nucleotide residues from the 3‘ terminus of TMV RNA with polynucleotide phosphorylase from Escherichia coli eliminated the histidine-accepting capacity and infectivity of this viral RNA. They also found that periodate oxidation of TMV RNA eliminated infectivity. However, the 5’ end of TMV RNA is capped (Zimmern, 1975), and the possibility was not ruled out in their experiments that loss of infectivity resulted from modification of the free hydroxyl at the 5’ end rather than from a change in 3‘-terminal functions.

16

TIMOTHY C. HALL

Adenosine

0 .OH

+

€I$-

II

C-

0

0-P-r \,C-CH,

0 ' Tyrosylated BMV RNA

N-Acetyl succinimic acid dimethyl form am ide pH 5.6

Acetylated tyrosylated BMV RNA

FIG.5. Acetylation of tyrosylated BMV RNA

B. ROLEA S

AN

AMINOACID DONOR

It has been suggested that the tRNAs coded by DNA or bacteriophages T4 and T5 are important toward their ability to infect certain strains of Escherichia coli. These virally coded tRNAs are known to be capable of amino acid donation, although they have not been shown to be essential for translation of viral peptides coded by laboratory strains of these viruses (Chen et al., 1975). The ability of TYMV RNA to donate bound valine during protein synthesis was studied by Haenni et al. (1973), and the data were interpreted as showing a small but significant level of donation. In these experiments, valinol (an arninoalkyladenylate that binds specifically to valyl-tRNA synthetase, preventing its ability to catalyze aminoacylation)was used to prevent the cycling of radioactive valine into protein via deacylation from viral RNA, charging to tRNA, and subsequent donation to nascent polypeptides. Although a valuable control, it is less direct than is the addition of excess amounts of unlabeled amino acid substrate. Donation of radioactive amino acid from precharged tRNA added to cell-free protein

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

17

synthesis reactions is very efficient and is not adversely affected by the addition of unlabeled substrate; Haenni et al. (1973) noted that the addition of unlabeled valine essentially eliminated any trace of donation of radioactive valine from viral RNA to nascent polypeptides, suggesting that the incorporation seen in their experiments was, in fact, due to recycling rather than true donation activity of the viral RNA. In similar experiments with tyrosylated BMV RNA, it was concluded that no donation occurred (Chen and Hall, 1973). The very low levels of tyrosine transfer observed could be accounted for by traces of contamination of the charged viral RNA with Tyr-tRNA derived from the enzyme used for the preparative tyrosylation reaction. Experiments with valylated EMV RNA (EMV is a member of the tymovirus group) have shown that no donation of the bound valine occurs (Hall et al., 1978b). It may be conjectured that in vivo a 3’ fragment from the viral RNA which is more nearly the size of a true tRNA may function as a viral-specific tRNA. However, tyrosylated 3’ fragments of BMV RNAs (similar to a 1 6 1 discussed above) showed no donation activity in vitro (R. K. Wepprich and T. C. Hall, unpublished observations). The 3’-terminal fragment released from TYMV RNA by ribonuclease P can be esterified with valine (Prochiantz and Haenni, 1973), but no reports of any ability to donate the bound valine have appeared. As noted above, the pentanucleotide GTVCR believed to be important in base pairing of tRNA with 5 s RNA during protein synthesis is missing from a 1 5 9 of TYMV RNA. This, together with the other data presently available, leads to the conclusion that viral RNAs do not exhibit tRNA-like amino acid donor functions. C. ROLEIN COMPETING WITH HOSTmRNA FOR RIBOSOMES A scheme has been suggested (Hall and Wepprich, 1976) whereby the viral RNA might be able to dislodge host mRNA from ribosomes (Fig. 6). As part of the postulated mechanism, it is thought that the viral RNA might be beneficially positioned for binding to the newly released 40s ribosomal subunits (Fig. 6, Step 3). In this regard, it is interesting that Fiers et al. (1976) have presented a model for long-distance interactions in MS2 RNA in which regions toward the 5’ end of this RNA could base-pair with a region toward the 3’ end of the RNA (Fig. 7). However, to date, no experimental support for the scheme shown in Fig. 6 has been obtained. To the contrary, we have found that a plant eucaryotic mRNA (Hall et al., 1978a) is able to outcompete BMV RNA component 3 for translation in a wheat germ cell-free system (J. W. Pyne and T. C. Hall, unpublished observations).

D. ROLEIN COMPETING FOR ELONGATION FACTORS Aminoacylated viral RNAs can bind EFl and thus compete for this factor with cellular aminoacyl tRNAs. The reaction differs from that for tRNA since GTP is

T

FIG.6 . Proposed model for translational role for aminoacylation of viral RNA. (1) Host pepride synthesis. Normal host peptide synthesis, host mRNA being translated on host ribosomes to yield a host peptide. The next triplet to be read is a tyrosine codon (UAU or UAG). (2) Tyrosyl-BMV RNA enrers A sire, showing transfer funcrion. Tyrosine charged to 3' end of B M V RNA mimics tyr-tRNA and enters the aminoacyl site on the host 60s subunit. Entry of further charged tRNA molecules is blocked, inhibiting host peptide synthesis. The 5' end of the B M V RNA is now situated near the messenger binding site of the host 40s subunit. (3) Premature release of host pepride. Since the viral RNA blocks entry of more tRNAs, possibly by an inability to transfer to the "P" site, the host peptide is prematurely released, as is the host mRNA. The ribosomal subunits may dissociate, the 5' end of the viral RNA is in a favorable position to bind to the 40s subunit. The 3' end deacylates and is released from the 60s subunit and charged: tRNAs can again enter the "A" site. (4) Viralpeptide synthesis, BMV RNA as the messenger. Synthesis of viral (coat) peptide on the host ribosome can now take place, the B M V RNA Functioning as the messenger template. (From Hall and Wepprich, 1976.)

19

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

r-r!

F -G 5-4

-- G6 . A . -G.u.*

-F

- A:tlLl c I

G

- +'llSl

8

-- - - --- __ -

.-.

'--*

\

,I

-L! -2.

-c.*

,G 1

- c.c - G;Zl89

----------m----_---------" c-----------_------A 1LS:i)

130

3396

0"

3369

FIG.7. Model for long-distance interactions in MS2 RNA. The outline of the complete chain can be followed by the numbering and by the initiation-termination signals (in bold type) of the three genes. Regions which remained single stranded in the models for the other parts of the molecule were screened for ability to form theoretically stable, complementary interactions. They are not supported by direct evidence, except that some are rather nuclease resistant and hence somehow protected in the whole molecule. Long-distance interaction refers only to the fact that the segments involved are far apart in the primary sequence. (From Fiers er al., 1976.)

released on interaction with aminoacylated BMV RNA rather than forming a ternary aminoacyl RNA-EF1-GTP complex (Bastin and Hall, 1976). The release of GTP further reduces the possibility of any direct amino acid donor function for the viral RNA. However, binding of EF1 may conceivably permit the viral RNA to have a competitive advantage for translation by reducing the levels of EF1 available in the cell. The viral RNA may then be able to cycle EF1 between its use in peptide chain elongation and its binding to the 3' terminus; indeed, it has been suggested that a single EFl molecule is involved in peptide chain elongation (Grasmuk er al., 1976). However, no supportive evidence has been obtained for this concept in relation to viral RNAs. E. ROLEI N COATPROTEIN ASSEMBLY Since RNA-protein interactions are important toward encapsidation of viral RNAs, the aminoacylation function might be thought to be involved in this process. This possibility would be supported if the amino acid residue esterified to the viral RNA were the same as that at the carboxy terminus of the viral coat protein. In the case of BMV, this is not so (although tyrosine is the penultimate

20

TIMOTHY C. HALL

carboxyterminal residue). TMV is known to be encapsidated from a specific attachment site on the viral RNA (Butler and Klug, 1971; Guilley et al., 1974), but Guilley e f al. (1975) found that the a 7 1 fragment of TMV RNA neither contains codons for nor interacts with TMV coat protein. This greatly detracts from any suggestion that aminoacylation is related to assembly processes. F. ROLEIN REPLICATION From cytological and biochemical analyses, it is very clear that active replication of viral RNA takes place following infection of the host, and the replicative forms of TYMV and EMV RNAs have been shown to be capable of accepting valine (Pinck et al., 1975). Toward understanding the biochemical mechanisms through which the viral RNA is replicated, isolation of RNA-dependent RNA replicase (polymerase) has been attempted by many research groups. To date, enzymes associated with infection of bacteria by RNA phages have provided the most definitive results; the best characterized is replicase obtained from E. coli after infection by phage Qp. This replicase is usually highly template specific (Haruna and Spiegelman, 1965), but the addition of Mn2+to the reaction medium enables it to transcribe many RNA templates, including BMV RNA (Palmenberg and Kaesberg, 1974). Qp replicase is especially interesting in that only one of its four polypeptide subunits is virally coded (Landers et al., 1974). Subunit I, a translation interference factor (Groner et al., 1972), is identical with ribosomal subunit S1 (Wahba et al., 1974); subunits I11 and IV were found to be the same as bacterial elongation factors EF.Tu and EF.Ts (Blumenthal et al., 1972). Because of the dual role of elongation factors in protein synthesis and (at least in the case of Qp replicase) in transcription, it is tempting to think that the tRNA-like structure of viral RNAs could be related to a transcription event through the ability of the aminoacylated forms to bind EF1 (Litvak et al., 1973a). Figure 8 details a possible mechanism for such an interaction wherein aminoacylated RNA could be more efficiently replicated than nonaminoacylated RNA. Certainly, such a system would be highly conservative of genetic information. It would also result in the production of only small amounts of foreign protein within the host cell, thereby reducing the possibilities for any type of antiviral, reaction. However, using Qp replicase (Hall and Wepprich, 1976) no difference in template efficiency between aminoacylated or nonaminoacylated BMV RNAs was observed. The replicase used for these experiments were not deficient in elongation factors, and it is likely that conditions in which elongation factors are absent (or present in very limited amounts) must be established before viral templates complexed with elongation factor can be shown to be superior to nonaminoacylated forms. This imposes the need for purified replicase which can be dissociated into its component subunits. Unfortunately, no purified enzyme capable of mediating RNA-dependent RNA synthesis has been obtained from a

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

21

AMINO ACID

+ig

VIRAL RNA

A&J

ELONGATION FACTOR 9

+ Q

VIRAL RNA

REPLICASE SUBUNITS

/

(3)

VIRAL RMA

-

TRANSCRIPTION

v VIRAL RNA-a..

REPLICASE COMPLEX

FIG.8. Proposed model for role of minocylation of viral RNA in transcription. (1) Viral RNA is arninocylared. ( 2 ) Aminocylated viral RNA binds EF I . Charged viral RNA can compete equally with charged tRNA for EF 1 (Since no binding of EF 1 to uncharged viral RNA can occur, charged tRNA then has a competitive advantage). (3) Formation of replication complex. Replicase subunits I 11 can now bind to the BMV RNA-Tyr EF 1 complex to form a replication complex. Possibly presence of subunits I I1 (especially I) prevents the further attachment of ribosomes to viral RNA, stopping initiation of translation. (From Hall and Wepprich, 1976.)

+

+

plant source. Thus far, only relatively crude enzymes whose activities are stimulated by the addition of viral RNA template have been obtained, examples being extracts from cowpea mosaic virus-infected cowpeas (Zabel et al., 1974, 1976), TMV-infected barley (Hadidi and Fraenkel-Conrat, 1973; Kummert and Semal, 1977), and TMV-infected tobacco (Brishammer and Juntti, 1974). Romaine and Zaitlin ( 1978) concluded that similar RNA-dependent RNA polymerases were obtained from healthy and TMV-infected plants. Even in the case of enzyme prepared from cowpeas (Zabel, 1978), which probably represents the best characterized plant-derived replicase, stringent template specificity has not been

22

TIMOTHY C. HALL

obtained. Possibly this is not a major requirement, although QP replicase is highly template specific except under conditions which are probably not physiological (Palmenberg and Kaesberg, 1974). Additionally, some template specificity is to be expected, otherwise the enzyme would replicate cellular RNAs such as tRNAs and mRNAs as well as the viral RNA. For some while it was thought that viroids had a tRNA-like structure (Semancik et a l . , 1973, 1975). These infectious agents have molecular weights of about 100,000 and appear to exist only as free RNA, no coat protein having been detected; indeed they possibly are completely unrelated to viruses. Despite their tRNA-like size, no amino acid acceptor activity could be detected for citrus exocortis viroid (Hall etal., 1974), and neither citrus exocortis viroid nor spindle tuber viroid showed any messenger function when added to cell-free protein synthesis systems (Hall et a l . , 1974; Davies et a l . , 1974). More recent studies have implied that viroid pathogenicity derives from an association with host DNA replication (Semancik and Geelen, 1975), but the nucleotide sequence of potato spindle tuber viroid (Gross et a l . , 1978) does not permit any apparent tRNA-like analogy. Lindley and Stebbing (1977) suggested that the suppression of EMCV infection in mice by the addition of eucaryotic tRNA (Stebbing et a l . , 1976) might result from interference of a replicase-associated function of the tRNA-like structure present in EMCV genome RNA by the added tRNA. Sela et al. (1976) reported that interferon catalyzed the deacylation of TMV RNA and EMC RNA. A plant antiviral preparation was also reported to have the ability to discharge histidine from TMV RNA. Under their experimental conditions, histidinyl-tRNA and seryl-tRNA were resistant to the deacylating effect of interferon. Addition of a specific tRNA fraction to extracts of interferon-treated mouse cells reversed the inhibition of exogenous messenger translation (Content et al., 1974; Gupta et a l . , 1974). Gallwitz et al. (1977) found that host mRNA was not inactivated or degraded on infection by mengovirus RNA, and conjectured that the abrupt cessation of host protein synthesis on infection by picomaviruses (Lawrence and Thach, 1974) might arise from the ability of the viral RNA to inhibit host DNA replication. At present, it is difficult to relate these observations on animal infection by picomaviruses to a function of their genome-associated tRNA-like structures, but the implications for a biological role are interesting.

V. Summary The occurrence of aminoacylatable tRNA-like structures in several groups of plant viruses strongly suggests that they have a biological function. Indeed, in the case of BMV, a relatively simple 3’-specific modification of the RNA results in loss of infectivity. However, to date, no metabolic role for the tRNA-like struc-

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

23

tures has been demonstrated, and the likelihood of functions related to translation or viral assembly processes seem to be remote. Thus, a role in transcription events presently appears to be the most attractive suggestion. This is despite the fact that other RNA viruses such as pea enation mosaic virus, which lacks a functional tRNA-like 3‘ structure or even a 3’ poly(A) sequence, (German et al., 1978) are efficiently transcribed in their hosts. Because of the analogous levels of specificity for aminoacylation despite their very different structures, it is possible that comparative studies of viral and tRNAs may yield an insight to the features of RNA which permit aminoacylation to occur. Comparisons of infective properties of native and chemically modified forms of aminoacylatable viral RNAs may reveal if the tRNA structure is related to transcription or translation events; alternatively, they may reveai a mvet metabolic process.

ACKNOWLEDGMENTS I gratefully acknowledge the encouragement of Professor Paul Kaesberg toward studies in this area and helpful comments and information from Dr. L. Hirth and Dr. B. Dudock. I appreciate the constructive suggestions of Dr. J. Pyne, Dr. S. Loesch-Fries, Dr. T. German, and Ms. P. Kiberstis regarding the manuscript. Part of the studies reviewed here were supported by NIH Grant A1 11572 and NSF Grant PCM 74-21675.

REFERENCES Aviv, H., and Leder, P. (1972). Proc. Natl. Acud. Sci. U.S.A. 60, 1408. Bancroft, J. B., Motoyoshi, F., Watts, J. W., and Dawson, J. R. 0. (1975). In “Modification of the Information Content of Plant Cells” (R. Markham et al., eds.), p. 135. Am. Elsevier, New York. Barry, R. D., and Bukrinskaya, A. G . (1968). J. Gen. Virol. 2, 71. Bastin, M., and Hall, T. C. (1976). J. Virol. 20, 117. Bastin, M., Dasgupta, R., Hall, T. C., and Kaesberg, P. (1976). J . Mol. Biol. 103, 737. Beachy, R. N., Zaitlin, M., Bruening, G., and Israel, H. W. (1976). Virology 73, 498. Beljanski, M. (1965). Bull. SOC.Chim. Biol. 47, 1645. Blumenthal, T., Landers, T. A,, and Weber, K. (1972). Proc. Narl. Acad. Sci. U . S . A . 69, 1313. Briand, J.-P., Jonard, G., Guilley, H., Richards, K., and Hirth, L. (1977). Eur. J. Biochem. 72, 453. Brishammar, S., and Juntti, N. (1974). Virology 59, 245. Butler, P. 1. A., and Klug, A. (1971). Nature (London). New Biol. 229, 47. Carriquiry, E., and Litvak, S. (1974). FEBS Lett. 38, 287. Chambers, R. W. (1971). Prog. Nucleic Acid Res. Mol. Biol. 11, 489. Chen, J. M., and Hall, T. C. (1973). Biochemistry 12, 4570. Shiau, R. P., Hwang, L.-T., Vaughan, J., and Weiss, S. B. (1975). Proc. Natl. Acad. Chen, M.-J., Sci. U.S.A. 72, 558. Content, J., Lebleu, B., Zilberstein, A., Berissi, H., and Revel, M. (1974). FEBS Lerr. 41, 125.

24

TlMOTHY C. HALL

Dahlberg, J. E., Sawyer, R. C., Taylor, J. M., Faras, A. J., Levinson, W. E., Goodman, H. M., and Bishop, J. M. (1974). J. Virol. 13, 1126. Dasgupta, R., and Kaesberg, P. (1977). Proc. Narl. Acad. Sci. U.S.A. 74, 4900. Dasgupta, R., Shih, D. S., Saris, C., and Kaesberg, P. (1975). Nature (London) 256, 624. Dasgupta, R., Harada, F., and Kaesberg, P. (1976). J. Virol. 18, 260. Davies, J. W., Kaesberg, P., and Diener, T. 0. (1974). Virology 61, 281. Di Natale, P., and Eilat, D. (1976). Nucleic Acids Res. 3, 917. Emtage, J. S., Carey, N. H., and Stebbing, N. (1976). Eur. J . Biochem. 69, 69. Faras, A. J. (1975). Virology 63, 583. Fiers, W., Contreras, R., Duerinck, F., Haegeman, G., Iserentant, D., Merregaert, J . , Min Jou, W., Molemans, F., Raeymaekers, A , , Van den Berghe, A., Volckaert, G., and Ysebaert, M. (1976). Nature (London) 260, 500. Gallwitz, D., Traub, V., and Traub, P. (1977). Eur. J . Biochem. 81, 387. Gefter, M. L., and Russell, R. L. (1969). J . Mol. Biol. 39, 145. German, T. L., de Zoeten, G. A,, and Hall, T. C. (1978). Intervirology 9, 226. Giege, R., Briand, J.-P., Mengual, R., Ebel, J.-P., and Hirth, L. (1978). Eur. J. Biochem. 84,251. Gould, A. R., and Symons, R. H. (1978). Eur. J. Biochem. 91, 269. Grasmuk, H., Nolan, R. D., and Drews, J. (1976). Eur. J . Biochem. 71, 271. Groner, Y., Scheps, R., Kamen, R., Kolakofsky, D., and Revel, M. (1972). Narure (London), New Biol. 239, 19. Gross, H. J., Domdey, H., Lossow, C., Jank, P., Raba, M., Alberty, H., and Sanger, H. L. (1978). Nature (London) 213, 203. Guilley, H., Jonard, G., and Hirth, L. (1974). Biochimie 56, 181. Guilley, H., Jonard, G., and Hirth, L. (1975). Proc. Nail. Acad. Sci. U.S.A. 72, 864. Gupta, S., Sopori, M. L., and Lengyel, P. (1974). Biochem. Biophys. Res. Commun. 57, 763. Hadidi, A., and Fraenkel-Conrat, H. (1973). Virology 52, 363. Haenni, A. L., Prochiantz, A., Bernard, O., and Chapeville, F. (1973). Nature (London), New Biol. 241, 166.

Hall, T. C., and Wepprich, R. K. (1976). Ann. Microbiol. (Paris) 127a, 143. Hall, T. C., Shih, D. S., and Kaesberg, P. (1972). Biochem. J . 129, 969. Hall, T. C., Wepprich, R. K., Davies, J. W., Weathers, L. G., and Semancik, J. S. (1974). Virology 61, 486. Hall, T. C., Ma, Y.,Buchbinder, B. U., Pyne, J. W., Sun, S. M., and Bliss, F. A. (1978a). Proc. Nail. Acad. Sci. U.S.A. 75, 3196. Hall, T. C., Pinck, M., Duranton, H. M., and German, T. L. (1978b). Virology (in press). Harada, F., Sawyer, R. C., and Dahlberg, J. E. (1975). J. Biol. Chem. 250, 3487. Haruna, I., and Spiegelman, S. (1965). Proc. Narl. Acad. Sci. U.S.A. 54, 579. Hruby, D. E., and Roberts, W.K. (1978). J. Virol. 25, 413. Hung, P. P., and Overby, L. R. (1968). J. Biol. Chem. 243, 5525. Joshi, S., Haenni, A. L., Hubert, E., Huez, G., and Marbaix, G. (1978). Nature (London) 275,339. Kapex, I. M., and Waterworth, H. E. (1977). Science 196, 429. Kim, S. H., Suddath, F. L., Quigley, G. J., McPherson, A,, Sussman, J. L., Wang, A. H. J., Seeman, N. C., and Rich, A. (1974). Science 185, 435. Klein, C., Fritsch, C., Briand, J. P., Richards, K. E., Jonard, G., and Hirth, L. (1976). Nucleic Acids Res. 3, 3043. Kohl, R. J., and Hall, T. C. (1974). J. Gen. Virol. 25, 257. Kohl, R. J., and Hall, T. C. (1977). Proc. Nail. Acad. Sci. U.S.A. 74, 2682. Kolakofsky, D. (1972). J . Virol. 10, 555. Koper-Zwarthoff, E. C., Lockard, R. E., Alzner-DeWeerd, RajBhandary, U. L., and Bol, J. F. (1977). Proc. Narl. Acad. Sci. U.S.A. 74, 5504.

tRNA-LIKE STRUCTURES IN VIRAL GENOMES

25

Kummert, J., and Semal, J. (1977). Virology 77, 212. Lamy, D., Jonard, G., Guilley, H., and Hirth, L. (1975). FEBS Lett. 60, 202. Landers, T. A., Blumenthal, T., and Weber, K. (1974). J. Biol. Chem. 249, 5801. Lawrence, C., and Thach, R. (1974). J. Virol. 14, 598. Lee, Y. F., Nomoto, A., Detjen, B. M., and Wirnmer, E. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 59. Lesiewicz, J., and Dudock, B. (1978). Biochim. Biophys. Acta 520, 411. Lesnaw, J. A., and Reichmann, M. E. (1970). Proc. Nail. Acad. Sci. U.S.A. 66, 140. Leung, D. W., Gilbert, C. W., Smith, R. E., Sasavage, N. L., and Clark, J. M., Jr. (1976). Biochemistry 15, 4943. Lindley, 1. J. D., and Stebbing, N. (1977). J . Gen. Virol. 34, 177. Litvak, S., Carrk, D. S., and Chapeville, F. (1970). FEBS Lett. 11, 316. Litvak, S., Tarrag6, A., Tarrago-Litvak, L., and Allende. J. E. (1973a). Nature (London), New Biol. 241, 88. Litvak, S., Tarrag6-Litvak, L., and Chapeville, F. (1973b). J. Virol. 11, 238. McClain, W. H., Guthrie, C., and Barrell, B. G. (1972). Proc. Natl. Acad. Sci. U.S.A. 69, 3703. McClain, W. H., Marchin, G . L., Neidhardt, F. C., Chace, K. V., Rementer, M. L., and Hall, D. H. (1975). Virology 67, 385. Oberg, B., and Philipson, L. (1972). Biochem. Biophys. Res. Commun. 48, 927. Paddock, G., and Abelson, J. (1973). Nature (London), New Biol. 246, 2. Palmenberg, A . , and Kaesberg, P. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 1371. Peters, G., Harada, F., Dahlberg, J. E., Panet, A., Haseltine, W. A,, and Baltimore, D. (1977). J. Virol. 21, 1031. Pinck, L., Genevaux, M., Bouley, J. P., and Pinck, M. (1975). Virology 63, 589. Pinck, M., and Hall, T. C. (1978). Virology 88, 281. Pinck, M., Yot, P., Chapeville, F., and Duranton, H. M. (1970). Nature (London) 226, 954. Pinck, M., Chan, S-K., Genevaux, M., Hirth, L., and Duranton, H. (1972). Biochimie 54, 1093. Pinck, M., Genevaux, M., and Duranton, H. (1974). Biochimie 56, 423. Prochiantz, A . , and Haenni, A. L. (1973). Nature (London), New B i d . 241, 168. Rich, A , , and RajBhandary, U. L. (1976). Annu. Rev. Biochem. 45, 805. Richter, D., Erdmann, V. A,, and Sprinzl, M. (1973). Nature (London), New Biol. 246, 132. Robertus, J. D., Ladner, J. E., Finch, J. T., Rhodes, D., Brown, R. S., Clark, B. F. C., and Klug, A. (1974). Nature (London) 250, 546. Romaine, C. P., and Zaitlin, M. (1978). Virology 86, 241. Rosenthal, L. J., and Zamecnik, P. C. (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 1184. Sabo, D. L., Domingo, E., Bandle, E. F., Flavell, R. A,, and Weissmann, C. (1977). J. Mol. Biol. 111, 235. Salomon, R., and Littauer, U. 2. (1974). Nature (London) 249, 32. Salomon, R., Sela, I., Soreq, H., Giveon, D., and Littauer, U. 2. (1976). Virology 71, 74. Sawyer, R. C., Harada, F.. and Dahlberg, J. E. (1974). J . Virol. 13, 1302. Sela, 1. (1972). Virology 49, 90. Sela, I . , Grossberg, S. E., Sedmak, J. J., and Mehler, A. H. (1976). Science 194, 527. Semancik, J. S. (1974). Virology 62, 288. Semancik, J. S . , and Geelen, J. L. M. C. (1975). Nature (London) 256, 753. Semancik, J. S . , Moms, T. J., and Weathers, L. G. (1973). Virology 53, 448. Semancik, J. S., Moms, T. J., Weathers, L. G., Rordorf, F. B., and Kearns, D. R. (1975). Virology 63, 160. Shatkin, A. J. (1976). Cell 9, 645. Shih, D. S . , and Kaesberg, P. (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 1799. Shih, D. S., and Kaesberg, P. (1976). J . Mol. Biol. 103, 77.

26

TIMOTHY C. HALL

Shih, D. S., Kaesberg, P., and Hall, T. C. (1974). Nature (London) 249, 353. Silberklang, M., Prochiantz, A., Haenni, A.-L., and RajBhandary, U. L. (1977). Eur. J. Eiochem. 72, 465.

Stebbing, N., Grantham, C. A., Kaminski, F., and Lindley, I. J. D. (1976). J. Gen. Virol. 34, 73. Stromberg, K.,and Litwack, M. D. (1973). Eiochim. Eiophys. Acra 319, 140. Sueoka, N., and Kano-Sueoka, T. (1970). Prog. Nucleic Acids Res. Mol. Eiol. 10, 23. Taylor, J. M., and Illmensee, R. (1975). J . Virol. 16, 553. Tilghman, S. M., Curtis, P. J., Tiemeier, D. C., Leder, P., and Weissman, C. (1978). Proc. Nufl. Acad. Sci. U.S.A. 75, 1309. Wahba, J. J., Miller, M.-J., Niveleau, A., Landers, T . A., Carmichael, G. G., Weber, K., Hawley, D. A., and Slobin, L. I. (1974). J. Eiol. Chem. 249, 3314. Wang, S., Kothari, R.M., Taylor, M., and Hung, P. (1973). Nurure (London), New Eiol. 242, 133. Waters, L. C. (1978). Eiochem. Eiophys. Res. Commun. 81, 822. Waters, L. C., and MuUin, B. C. (1977). Prog. Nucleic Acids Res. Mol. Eiol. 20, 131. Waters, L. C., Mullin, B. C., Bailiff, E. G., and Popp, R. A. (1975). J . Virol. 16, 1608. Weiss, S. B., Hsu, W.-T., Foft, J. W., and Scherberg, N. H. (1968). Proc. Narl. Acad. Sci. U 3 . A. 61, 114. Wilson, I . H., Kim, I . S., and Abelson, J. N. (1972). J. Mol. Biol. 71, 547. Yot, P., Pinck, M., Haenni, A. L., Duranton, H. M., and Chapeville, F. (1970). Proc. Nutl. Acad. Sci. U.S.A. 67, 1345. Zabel, P. (1978). Ph.D. Thesis, University of Wageningen, Holland. Zabel, P., Weenen-Swaans, H., and Van Kammen, A. (1974). J. Virol. 14, 1049. Zabel, P., Jongen-Neven, I., and Van Kammen, A. (1976). J. Virol. 17, 679. Zimmern, D. (1975). Nucleic Acids Res. 2, 1189.

INTERNATIONAL REVIEW OF CYTOLOGY, VOL. 60

Cytoplasmic and Cell Surface Deoxyribonucleic Acids with Consideration of Their Origin BEVANL. REID Queen Elizabeth I1 Research Institute for Mothers and Infants, University of Sydney, New South Wales, Australia

ALEXANDER J. CHARLSON School of Chemistry, Macquarie University, North Ryde, New South Wales, Australia

. . . . . . . . . . . . . . . . . . . . . . . A . Morphological Evidence . . . . . . B. Biochemical Evidence . . . . . . . Cell Surface DNA . . . . . . . . . . A . Observations . . . . . . . . . .

I. Introduction

11. Cytoplasmic DNA

111.

B. A Possible Source of Cell Surface DNA IV. Some Biological Implications . . . . . V. S u m m a r y . . . . . . . . . . . . . . References . . . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . .

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

27 28 28 33 38 38 45 49 49 51

I. Introduction A study of the cell wall and cell edge has as some of its more direct goals the discovery of that elusive factor or factors initiating or controlling cell growth and function. In pursuing this search from the cell interior toward its wall and ultimately to its boundary layer with the environment, the experimentalist may have had in mind the earlier views of Jacob et al. (1963) on the location of a proposed gene-activating apparatus at the surface of the bacterial cell. It is the aim of this review to draw attention to existing evidence for the presence of DNA in the cytoplasm and the cell surface other than that known to occur in various organelles. In view of complicity of the surface in known control mechanisms in the prokaryotes the presence of DNA in these less well-known sites may offer an insight into the more complicated mechanisms which are believed to operate in the eukaryote cell. The deoxyribonucleic acids are not as well-known components of the cell wall as are the proteins and lipids, especially glycoproteins and glycolipids. Yet there 27

Copyright @ 1979 by Academic F’ress, Inc. All rights of reproduction in any form reserved. ISBN 0-12-3643600

28

BEVAN L. REID AND ALEXANDER J . CHARLSON

is an important link between the nucleic acids and those types of sugars which characteristically occur in glycoproteins and glycolipids, a link on which there is some existing knowledge. This review takes this link into consideration in attempting to see in it the establishment of a bridge between cell and environment, a bridge which speculatively can then become an important element in the control mechanism. The evidence is conveniently divided for descriptive purposes into morphological and biochemical studies as applied to the two compartments of the cell under consideration, the cytoplasm and the cell wall including the plasma membrane and the glycocalyx.

11. Cytoplasmic DNA

A . MORPHOLOGICAL EVIDENCE 1 . Histological and Histochemical Investigations The subject of nucleocytoplasmic relationships has been a lively one among microscopists over the past 50 years. Structures were seen in the cytoplasm either in direct continuity with the nuclear membrane or in association with the nucleus sufficiently close as to warrant the term extranuclear chromatin for their description. The observation that these structures reacted with the Feulgen stain or that they absorbed light in the ultraviolet range specific for DNA led to the idea that DNA may be present in the cytoplasm as a result of its transfer from the nucleus (Clark, 1960). These observations were prominent in botanical literature. Examples are to be found in the study of Feulgen-positive cytoplasmic bodies by Sparrow and Hammond (1947) in meiotic phases of microsporocytes of Lillium sp., Trillium sp., and Allium sp. and in the presence of cytoplasmic granules which absorbed intensely at 265 nm in meristem cells of Viciafuba as reported by Chayen (1960). Chayen drew attention to the mobility of the material responsible for the absorption such that a meristem cell in an undamaged state contained less than 15% of the material in its nucleus, the remainder residing in the cytoplasm. Damaged cells had 100% of their absorption confined to the nucleus. The Feulgen reaction was used to study the egg cytoplasm of anuran amphibians by Brachet (1965). The author interpreted the accumulation of strongly Feulgen-positive spheres in the cytoplasm as an aggregation of DNA derived from the nuclear sap. Bridges often form between motile cells in culture, and Feulgen-positive material has been found in these bridges between epithelial cells (Lindholm and Britten, 1967). In a tissue culture of motile stromal cells from the human extocervix uteri, Coppleson and Reid (1969) showed the presence of filamentous Feulgen-positive material in the cytoplasm whose continuity with

CYTOPLASMIC AND CELL SURFACE DNA

29

Feulgen-positive material in the nucleus was demonstrable. Electron micrographs of one of these stromal cells which had been previously identified and observed to be actively motile showed the presence of microfilaments continuous with the heterochromatin of the nuclear margin coursing through the cytoplasm to the cell edge (Reid and Blackwell, 1971). Many studies with the electron microscope have shown that processes emanating from the nuclear membrane become disperse in the cytoplasm. This is especially documented in the case of thymus cells (Toro and Olah, 1966; Sebuwufu, 1966; Reid and Blackwell, 1970) and lymphocytes, particularly neoplastic lymphocytes (Mollo and Stramignoni, 1967). Nuclear membrane extensions have been described during the development of Lycopodium sp. spores. The extensions may be observed close to the plasma membrane where they may be integrated with the developing intine (Gullvag, 1970). A similar migration across the cytoplasm to the plasma membrane has been suggested to befall granular and fibrillar material derived from chromosome 6 and subsequently extruded through the nuclear membrane of oocytes of the cricket Acheta (Jaworska and Lima da Faria, 1973) (Fig. 1). For their investigation using motile baby rat thymus cells, Reid and Blackwell (1970) developed a technique involving the serial use of phase contrast and electron microscopy of the same cell. By this means it was shown that the emergence of pseudopodial spikes from the surface of the motile cell was accompanied by a striking change in the nuclear membrane underlying the site of the spike. The heterochromatin beneath the spike became dispersed and from its edge, microtubular elements could be resolved proceeding as a bundle into the spike. A similar arrangement of axial filaments is seen in electron micrographs of surface spikes of cultured rat embryonic cells reported by Buckley (1975), although in this case the filaments are derived from a cytoplasmic microfilament network beneath the plasma membrane. Microfilaments from heterochromatin at the nuclear margin were found coursing through the cytoplasm to the plasma membrane in ultrastructural studies of cultured human uterine cervical stromal cells (Reid and Blackwell, 1971) which were fixed during actively motile stages as seen by phase microscopy (Coppleson and Reid, 1969). 2. Autoradiographic Investigations The purely morphological evidence for the presence of DNA in the cytoplasm and cell wall is considerably strengthened by the use of autoradiography with tritiated thymidine. In addition to the histochemical observations using Feulgen stain to show the presence of positive granules in intercell bridges, Lindholm and Britten (1967) demonstrated the incorporation of label at the same site following the use of tritiated thymidine. In a study of the cytoplasmic labeling following exposure to tritiated thymidine by cells of the growing root tip of Allium cepa, Fussell (1968) showed that two-thirds of the label appeared in sites unrelated to

30

BEVAN L. REID AND ALEXANDER J. CHARLSON

CYTOPLASMIC AND CELL SURFACE DNA

31

cell organelles. The label was related to the cytoplasmic ground substance, the plasma membrane, and the cell wall. The same isotope incorporated into the amoeba Hartmanella rhysoides showed labeling in the cytoplasm as well as in the mitochondria, the plasma membrane, and the glycocalyx (Fig. 2). The latter two sites accounted for thee times as many grains as did that of the mitochondria (Ito et al., 1969). Fussell found that most of the cytoplasmic label was removed after treatment with deoxyribonuclease and Ito et al. removed all the cell membrane label with this enzyme. When complement-free serum from animals immunized with lymphocytes is added to the medium in which lymphocytes of the same type are cultured for a few hours, the cell nuclei become enlarged and pale and a bulky extracellular mucoid coat forms about the lymphocytes tending to aggregate them. Several days later the lymphocytes proceed to mitosis. If the lymphocytes have been previously labeled in vivo with tritiated thymidine, a control culture free of specific antiserum shows typical dense labeling over the nucleus. After contact with antilymphocyte serum a broad extracellular mucoid coat forms about the lymphocyte. Autoradiographs of these cells now show that the label is formed in the extracellular coat and reciprocally the nuclear labeling has decreased (Reid and Blackwell, 1971). In another study of this type, autoradiographs were prepared from lymphocytes stimulated by phytohemagglutinin and labeled for 7 hours with tritiated thymidine (Politis el al., 1975). Heavily labeled cell nuclei resulted. Sheep red blood cells added to the lectin-stimulated cell culture formed rosettes about the lymphocytes. After a few minutes of incubation the label was now found covering the red cells. The authors construed this as labeling of an excretion from the lymphocytes which proceeded to coat the added red cells. In an investigation of the DNA fractions of embryonic mouse liver cultured for brief periods, to be reported in Section II,B, Williamson (1970) used autoradiography following incorporation of tritiated thymidine. About 10% of the labeled cells showed a distribution of grains shared between nucleus and cytoplasm. Because of the lack of definition of nuclear staining in one-half of these cultured cells, the author interpreted the cytoplasmic label as due to damaged cells and therefore probably artifactual. He did not attempt to reconcile this finding with a further 7% of cells with dual labeling which, however, exhibited nuclear staining of normal intensity.

FIG. 1. A series of six light micrographs of oocyte of house cricket Acheta sp. illustrative of one mechanism for the origin of cytoplasmic DNA. In the uppermost micrograph a large nucleolusassociated sphere of DNA-RNA composition derived from a chromomere of chromosome 6 moves across the nucleus to the nuclear membrane which it breaches taking a fragment of the membrane into the cytoplasm. In the cytoplasm the particle now coated with the nuclear membrane migrates to the cell wall where it breaks up into smaller particles (lowermost micrograph). Feulgen stain. (From Jaworska and Lima da Faria, 1973.)

32

BEVAN L. REID AND ALEXANDER J . CHARLSON

FIG.2. Portion of the amoeba Harrmanella labeled with tritiated thymidine. There is a significant number of grains over the plasmalemma and the ectoplasm. Label also overlies the gfycocalyx. Electron autoradiograph. (From Ito er al., 1969.) Bar = 0.5 pm.

Cytoplasmic labeling of certain ovarian cells of the rat was studied in autoradiographs prepared with tritiated thymidine (Smith et al., 1975). Following prolactin stimulation, cells of the corpus luteum showed cytoplasmic labeling which was resistant to solubilization by trichloroacetic acid. A combined

CYTOPLASMIC AND CELL SURFACE DNA

33

biochemical study of cell fractions confirmed the presence of a cytoplasmic DNA component in the luteal cells. B. BIOCHEMICAL EVIDENCE Over the past 20 years the notion that DNA may not be restricted to the nucleus and mitochondria was growing apace in the biochemical world as well as among histologists. The presence of DNA in material obtained by ultracentrifugation of disrupted cells was not only found in nuclear and mitochondrial fractions. Postmitochondrial residues including microsomal and postmicrosomal supernatant fractions also showed the presence of measurable and consistent quantities of DNA. Examples of this type of study are to be found in the papers of Bach (1962), Bond et al. (1969), and Schneider and Kuff (1969). Bach cultured HeLa cells in the presence of tritiated thymidine. After 2 hours cells were harvested and homogenized. The cytoplasmic fraction separated on sucrose gradients was radioactive with a specific activity which showed a more rapid turnover than liver or mitochondrial fractions from which the cytoplasmic fractions were clearly separated. Bond et al. (1969) studied DNA in cells of mouse liver following separation of homogenized cells on sucrose gradients. Postmitochondrial fractions were extracted by methods using protease followed by ribonuclease and precipitation of DNA in the supernatants from these enzyme digests with ethanol. Specific radioactivity studies showed that tritiated thymidine was incorporated into cytoplasmic DNA faster than it was incorporated into mitochondrial or nuclear DNA. The absence of significant amounts of the satellite fraction, characteristic of bulk DNA of the mouse, in cytoplasmic DNA suggested that there was little contamination of the preparation with nuclear or mitochondrial DNA. Schneider and Kuff (1969) using similar material prepared by similar methods of homogenization, followed by extraction with phenol, showed that the specific activity of mitochondrial fractions was the greatest followed by cytoplasmic and nuclear DNA. The melting profiles of cytoplasmic and nuclear DNA, as well as the buoyant densities and hybridization properties with cytoplasmic RNA, were similar. Moreover the cytoplasmic DNA contained satellite DNA. This suggested to the authors that at least some of the cytoplasmic DNA may have been derived from nuclear DNA although a difference in the specific activities of labeled DNA from the two sources refuted this suggestion. An extensive study of a labeled cytoplasmic fraction of embryonic muscle cells following incorporation of tritiated thymidine was made by Bell (1969). After 2 hours of exposure the labeled material showed a sedimentation coefficient of 7s on sucrose gradient density centrifugation which increased to 16s after 4 hours of culture. Treatment of the 16 S particles with sodium dodecyl sulfate to remove the protein resulted in the recovery of 7s particles. Bell attempted to answer the criticism that the labeled particles were of accidental

34

BEVAN L. REID AND ALEXANDER J . CHARLSON

nuclear origin by adding labeled nuclei to unlabeled muscle and showing that no labeled particles were recovered from the cytoplasm. Using similar techniques Fromson and Nemer (1970) proposed that the cytoplasmic DNA which they found in the cells of the sea urchin was a result of nuclear contamination. Bell (197 1) replied that their labeling period of 3 hours was insufficient to detect the cytoplasmic fraction at 18°C stating that his own labeled material did not appear after such an interval. He went on to show that the incorporation of the label is more sensitive to hydroxyurea and less sensitive to deoxyfluorouridine than is the case with nuclear DNA. Differentiation from mitochondrial DNA in the chick muscle material was possible with the use of ethidium bromide by which only the mitochondrial DNA synthesis is specifically blocked. Additionally in these studies cytoplasmic DNA had a higher buoyant density. Williamson (1970) kept embryonic mouse liver cells in culture for 16 hours before their lysis in M magnesium chloride solution. The postmitochondrial fractions produced by ultracentrifugation were extracted by phenol and their nucleic acid content was recovered by ethanol precipitation. Final separation from cytoplasmic RNA was made by gradient centrifugation in cesium sulfate. Up to 20% of the total cell DNA was found in the cytoplasmic fractions and this fraction was labeled 50% heavier than the nuclear DNA. Cytoplasmic DNA was also found when the culture step was omitted. Note that by contrast, adult mouse liver cell cytoplasm contains a much smaller fraction of the total cell DNA (Bond et al., 1969). Williamson found the base compositions of nuclear and cytoplasmic DNA to be similar as were their buoyant densities in cesium sulfate. When liver cell nuclei were disrupted by hypotonic treatment only a small fraction of DNA was lost from the damaged nuclei, insufficient to account for the amount found in the cytoplasm. Nevertheless the finding of imprecise nuclear outlines in the autoradiographs made at the time and described in Section II,A persuaded the author that nuclear breakdown most likely resulted in cytoplasmic contamination during cell division. Disruption by osmotic or mechanical means of cultured liver cells was used by Koch and von Pfeil(l971) to produce a cytoplasmic fraction on sucrose gradients which was labeled after a 10-hour period of incorporation of tritiated thymidine. Although this cytoplasmic fraction differed from mitochondrial DNA in buoyant density, no such difference could be found between cytoplasmic and nuclear DNA. These observations were extended to kidney cell lines (Koch, 1973) following a 6-hour exposure to tritiated thymidine. Kidney cell cytoplasmic DNA also differed in buoyant density from mitochondrial DNA. The reassociation of single strands of cytoplasmic DNA was slower than that of nuclear DNA. The author noted that tumor and embryonic cells had more cytoplasmic DNA than their normal counterparts where no cytoplasmic DNA could be found. Koch favored the idea that there was a gradual migration of DNA from nucleus to cytoplasm.

CYTOPLASMIC AND CELL SURFACE DNA

35

Accounting for the well-known distribution of DNA polymerase in cytoplasmic and cell sap fractions has remained a problem for the biochemist (Baril et al., 1973; see Reid, 1974, for discussion). Novak and Elford (1973)investigated the possibility of cytoplasmic DNA acting as a template for synthesis of DNA involving locally available polymerase. Disrupted liver cells from several sources were fractionated on sucrose gradients. Deoxyribonucleic acids were extracted from the cytoplasmic fraction and used as a template primer for in vitro synthesis studies. New synthesis was monitored with the use of labeled nucleotides. An active primer source was obtained from the cytoplasm of regenerating liver cells obtained 12 hours after resection. The rate of incorporation was 6-fold that of the same reaction primed by extracts from normal liver cells. However the priming capacity of cells of hepatoma origin was about 8-fold that of regenerating liver. DNA was found with RNA in association with a sucrose gradient postmitochondrial fraction of disrupted hen ovocytes by Schjeide and 1-San Lim (1970).On electron microscopic study this fraction showed bilaminar membranes disposed as vesicles. Chemical analysis showed that the bulk of the dry weight of the fraction was made up of protein and lipid but about 3% was found to be RNA and one-tenth of this weight was DNA. No stranded material was found associated with the vesicles leading the authors to conclude that the DNA must be tightly bound to or integral with the membrane itself. The buoyant density of the membrane-vesicle DNA showed two peaks in cesium chloride which were different from the value shown by the peak from the bulk of the ovocyte DNA. From the studies of Solage and Loskov (1975)about I-2% of the DNA of mouse melanoma cells is found in the cytoplasmic fraction released from detergent-lysed cells. The authors noted the similarity of their product to that isolated by Schneider and Kuff (1969)although its isotope incorporation rate was similar to that of nuclear DNA. In a study of cytoplasmic granules aimed especially at their DNA content, Schneider er al. (1975)used livers from Amphiuma sp. and mouse melanoma tumors homogenized in sucrose. The granules were separated from the mitochondria by appropriate sucrose gradients and the absence of succinic dehydrogenase in granule fractions was considered to denote freedom from mitochondrial contamination. About 1 % of the cellular DNA was estimated to reside in the granule fraction and this DNA was different from the nuclear DNA in some respects. Although the buoyant densities were similar, nuclear DNA showed a sharp band in the gradient whereas that of granule DNA was broader. The melting curve for granule DNA was bitonal with peaks at 64"and 84°C. In the tumor tissue the specific radioactivity of granule DNA was always lower than that of nuclear DNA. In spite of several similarities between nuclear and cytoplasmic DNA in thess various investigations, the concensus of the various authors was that they were

TABLE I PROPERTIES OF NONMITOCHONDRIAL CYTOPLASMIC DNA

Source

Preparative method

Proportion of total cell DNA (%)

Molecular weight ( x 10")

1419

Units of0.135

Rate of labeling vs nuclear DNA 1.5X Faster

Buoyant density (gm cm-I)

Sensitivity to antimetabolites vs nuclear DNA

1.7OOa

-

Williamson (1970)

Reference

Embryonic mouse liver

Osmotic lysis

Adult mouse liver

Homogenization, 0.25 M sucrose

2.5X FasteI

-

Bond er al. (1969); Schneider er al. (1975)

Adult rat liver

Homogenization, 0.25 M sucrose

2.5~ Faster

-

Schneider er al. (1975)

Humm liver cell line

Osmotic lysis or mechanical disruption

-

Koch and von Pfeil(l971)

Amphiuma liver

Homogenization, 0.25 M sucrose

-

Schneider et al. (1975)

3

1

Faster or slower

Embryonic chick muscle

Osmotic lysis

0.3

1.703'

More sensitive to hydroxyurea; less sensitive to S-deoxyfluorouridine

Viral transformed

Osmotic lysis

0.3-3

I . 702a

-

Koch (1973)

-

Bach (1962)

Bell (1971)

hamster kidney cell line HeLa cells

Homogenization

Mouse myeloma

Mechanical disruption Homogenization, 0.25 M sucrose

Mouse melanoma

Faster in early stages 1-2 1

"Nuclear DNA had identical buoyant density. "Nuclear DNA had slightly lower buoyant density.

Same

Slower

1.700"

Solage and Loskov (1975) Schneider et al. (1975)

38

BEVAN L. REID AND ALEXANDER J. CHARLSON

dealing with a specific fraction of the cell DNA and not with a preparative artifact of nuclear origin. A more consistent property was the elevated specific activity of isotope-labeled cytoplasmic DNA as well as metabolic differences during synthesis as revealed by the effect of antimetabolites. The principle investigations are summarized in Table I.

111. Cell Surface DNA A. OBSERVATIONS By their nature, biochemical methods cannot discern the precise localization of the cytoplasmic DNA. Over the past several years a body of research has appeared which indicates that some part of the cytoplasmic DNA is found not only in the cytoplasm but also at the cell surface, both in association with the plasma membrane and beyond in a cell envelope termed the glycocalyx. Lerner’s group in La Jolla set out to determine if a direct link could be established between the cell wall and the nucleus using the lymphocyte as a cell type (Lerner et al., 1971; Hall et al., 1971; Meinke et al., 1973). They sought and found a DNA fraction associated with the cell wall of cultured human lymphocytes. Using synchronized cell cultures they showed that the cell wall fraction is synthesized in the nucleus during the S phase, after which it is apparently transported to the plasma membrane. Density marker studies showed turnover of the membrane fraction to be slower than that of the nuclear DNA but the authors also showed that the membrane fraction was synthesized at times in the cell cycle other than the S phase. Electron microscopy of the membranefraction DNA preparations showed that linear filaments were often associated with membrane debris. The membrane DNA component was assessed at 0.5% of the total cell DNA. Different experiments were made to minimize a likely contamination with DNA of nuclear, mitochondrial, or environmental origin, such as a microorganismal source, with no change in the results. The buoyant densities of membrane and nuclear DNA were similar. Later attempts to characterize the DNA by kinetic renaturation studies revealed wide differences between DNA from the two sources. The cell surface DNA showed up to nine times the content of more homogeneous DNA. The metabolic patterns of its synthesis were evidently different since they could be blocked by a concentration of rifampicin which left the nuclear DNA unaffected. Because of the known similarity in some of the properties of surface and nuclear DNA of lymphocytes, other investigators have been at pains to show surface DNA as a contaminant of nuclear origin. Melera and Cronin-Sheridan (1975) used the same material and the same extractive techniques as the La Jolla group save for one modification. By including 5 mM magnesium ions in the solution used for initial disruption, Melera and Cronin-Sheridan showed that the

CYTOPLASMIC AND CELL SURFACE DNA

39

nuclear membrane was stabilized to the exent that, in the nuclear fraction after isotope incorporation, counts exceeded those of controls which had no added magnesium. Counts in the cytoplasmic membranes, on the other hand, were insignificant when they were extracted in magnesium-enriched solutions but were appreciable in the controls where the membranes were magnesium-low at a level equivalent to that reported in Lerner’s work. The authors thus viewed the membrane DNA as a contaminant from the nucleus produced by leakage during preparation from a nuclear membrane which was insufficiently stabilized. It is to be noted, nevertheless, that Melera and Cronin-Sheridan ’s study also revealed an even more striking reduction, of a 10-fold order, in the labeling of mitochondria1 fractions in the presence of magnesium ions. On logical grounds they could have concluded that the normal DNA content of mitochondria found in the absence of magnesium enrichment was, in the same way, a preparative artifact produced by leakage from the nucleus. Clearly this conclusion is unwarranted and some other reason for the mobility of DNA between nuclear and nonnuclear compartments must be sought. Such mobility is referred to later in this article. Further characterization of cell membrane DNA by means of reassociation studies was made by Meinke and Goldstein (1974). They separated two fractions, a rapidly reassociating fraction comprising 70% of the sample and a more slowly reassociating fraction comprising the remainder. The former showed homology with 4% of the repeated sequence (rapidly reassociating) of fractions of nuclear DNA and the latter hybridized with 11% of the unique sequence (slowly reassociating) of DNA of nuclear origin. To the extent that these homologies reflect the base sequence cell surface DNA is a mixture of heterogeneous and homogeneous polymer. It is not only of a predominantly more homogeneous base sequence but contains novelty amid this homogeneity, a feature brought out by other studies to be described. At the same time Lerner’s investigations seeking a link between nucleus and cell wall were in progress, Reid and associates were engaged in a similar inquiry. They sought the presence of nucleic acids in the cell wall empirically as a possible factor underlying the secondary or long-term response of subsequent generations of immune cells to exhibit an enhanced response, the so-called lymphocyte “memory. ” Ultrastructural studies of baby rat thymus cells mounted whole disclosed in the glycocalyx of these cells an ornate filamentous network (Reid and Blackwell, 1970; Singer and Reid, 1970) (Figs. 3A and B). Histochemical study of these filaments revealed a nucleic acid content of both ribose and deoxyribose type (Reid and Blackwell, 1970). This coat material was shown to be liberated to the culture fluid from the parent cell without any shear or agitation. The supernatant could be easily separated from the thymus cell culture by low-speed centrifugation and a preparation of DNA obtained from it by standard biochemical methods of extraction (Reid and Blackwell, 1972a) (Fig. 4A). Under the electron microscope this DNA was found to be composed of short lengths as contrasted with the long lengths characteristic of DNA prepared from

40

BEVAN L. REID AND ALEXANDER J . CHARLSON

FIG. 3. (A) Electron micrograph of the edge of a strand of mucus from the human cervix uteri. A branched filamentous network projects into a clear channel in the mucus. Stained with ammonium molybdate. From Singer and Reid (1970). Bar = 0.5 pm. (B) Electron micrograph of the edge of a small unsectioned thymus cell from a baby rat prepared as for Fig. 3A. The cell body occupies the lower right comer of the photograph. The remainder consists of its mucoid coat (glycocalyx). At the top and bottom of the photograph the surface coat has smaller filaments bearing an overall resemblance to those of cervical mucus (Fig. 3A). In the middle of the photograph, a coarser fiber is present which characteristically breaks up into finer filaments which join the filamentous networks at the top and bottom of the photograph. Stained with ammonium molybdate. (From Reid and Blackwell, 1970.) Bar = 0.5 pm.

CYTOPLASMIC AND CELL SURFACE DNA

41

FIG. 3B

whole cells or cell nuclei (Fig. 4B). The supernatant DNA showed melting properties indicative of a more homogeneous base composition than DNA of the bulk phase. Unlike the position in human lymphocytes (Meinke et al., 1973) the buoyant density in cesium chloride was different in DNA from the two

42

BEVAN L. REID AND ALEXANDER J. CHARLSON

FIG.4. (A) Electron micrograph of the human sperm head mounted whole and stained with uranium oxide. The head is in thelower left of the photograph. The remainder of the photograph shows an extracellular network connected to the cell edge by coarser fibers. From Reid and Blackwell (1972b). Bar = 0.5 pm. A similar network appears at the edge of baby rat thymus cells. (B) Electron micrograph of a film using the Kleinschmidt technique for DNA prepared from the surface coat of the spermatozoon as shown in Fig. 4A.The molecules are characteristically short. A similar product is obtained in the same way from the extra cell network of the thymus cell. (From B. L. Reid, unpublished.) Bar = 0.1 pm.

sources. In the study of Reid and Blackwell (1972a) particular attention was paid to the problem in interpreting the contribution of dying cells to the supernatant DNA. Cells killed by freeze-thawing and then returned to standard culture conditions released one-thirtieth of the amount released by normal cells. Cells kept at

CYTOPLASMIC AND CELL SURFACE DNA

43

FIG.4B.

4°C behaved like killed cells but started to contribute to the superntant again when the temperature was raised to 37"C, at which time motile cells were noted in the culture. A study of the DNA of cell membranes using isolation procedures in which the absence of nuclear contamination could be monitored was made by Binkherd and Toliver (1974). The results showed that RNA primers were necessary for the synthesis of membrane DNA just as they are for DNA synthesis elsewhere. Detailed studies of the export of DNA from cultured lymphocytes have been made by Rogers et al. (1972) who found that phytohemagglutinin was necessary to stimulate the process. The studies were prompted by the finding of a discrepancy between the number of cells synthesizing DNA and those actually dividing following the exposure of cultures to lectin. Almost all of the label which had been incorporated was excreted within 3 days of the initial dose. Later, Rogers

44

BEVAN L. REID AND ALEXANDER J . CHARLSON

noted that the lectin was not essential for the excretion of label and he made a detailed study of the nature of the DNA involved in the excretion (Rogers, 1976). It comprised 10-15% of the total DNA in the culture, was double stranded, and sedimented at 7-8 S in alkaline sucrose gradients. The midpoint of its melting curve (T,) was substantially lower than that of whole cell DNA. The kinetics of hybridization of excreted DNA when compared with those of parent cell DNA showed that 60% of excreted DNA reassociated slowly while half of the remainder reassociated rapidly. The slower reassociating fraction had a sequence complexity equivalent to that of 10% of the cell genome DNA, equivalent approximately to the unique sequences of the genome. This fraction was found to be common to the lymphocytes of several donors. Insofar as structural genes are transcribed from unique sequences, the author reasoned that structural genes may be present in surface DNA the protein products of which may be important in the immune process. Using the same material stimulated in the same way with lectins and fractionated in the same way on sucrose gradients, Sarma and Rutman (1972) showed a similar excretion of labeled DNA which had recently been incorporated into nuclear DNA. Some 24 hours after incorporation, as much as 70% of the label was released from small DNA fragments to the medium. The phenomenon was not associated with death of the lymphocytes since their metabolic patterns were not disturbed and they could exclude dyes. Several investigations of DNA excreted by tissues and by cells have been made over the past decade by Anker and associates (1975) and Stroun et al. (1977). They showed that normal human lymphocytes could excrete DNA (Anker et al., 1975). The excretion was stimulated by transfer to a new medium where the rate reached a plateau within an hour. Extraction of the medium for nucleic acids showed material with an absorption maximum at 260 nm which was sensitive to deoxyribonuclease and sedimented at 16 and 7 S in contradistinction to cell DNA which sedimented at 18 S. Buoyant density values for cellular DNA and medium DNA were equal. Lymphocytes, in common with other bacterial and vertebrate systems studied in the past by Anker’s group, released DNA which reached a constant level in the medium. Addition of fresh medium promoted a burst of extracellular synthesis. Careful characterization of this synthesis (Anker et al., 1976) showed that it involved a true precursor requiring template and enzyme. From nearest neighbor analysis and hydrolysis using specific nucleases they excluded terminal additions to the existing molecule as well as physical adsorption of the precursor. Rather, the sensitivity of the synthesis to inhibition by deoxyribonucleases, ribonuclease, protease, and actinomycin D suggested that a true synthetic process was occurring in the medium apparently independent of the presence of the parent cell. Reassociation studies of the newly synthesized polymer showed a level of hybridization with cell DNA which indicated either that its sequences were unique or that they possessed little homology

45

CYTOPLASMIC AND CELL SURFACE DNA

or that both were occurring. Uniqueness due to synthesis by random attachment of terminal nucleotides was ruled out in view of the skewing of results confirmed in nearest neighbor nucleotide analyses. Some factor was evidently ordering the sequence of newly added nucleotides. It was reasoned that a homology with unique sequences present in the parent cell was a more plausible explanation of the reassociation data. Note that the uniqueness of the DNA synthesized extracellularly compares with the failure of cell wall membrane DNA to hybridize with the great majority of the cell DNA (Meinke and Goldstein, 1974). Perhaps the uniqeness of its base sequence is a characteristic of DNA nearest the outside of the cell, a point thought to be of possible significance in evolution and discussed elsewhere (Reid et al., 1976). Evidence of the presence of nucleic acids at the cell surface came from an entirely new direction when Rosenberg and collaborators showed that certain organoplatinum compounds, which have a known affinity for the nucleic acids; could be used as an electron stain. The platinumpyrimidine blues with a molecular weight of 1000-3000 (Lippert, 1977) (Fig. 5 ) bind rapidly to DNA in vitru. The electron micrographs of Aggarwal et al. (1975) show dense deposits of platinum both in the nucleus and at the cell surface overlying the glycocalyx. The authors describe the surface stain as peculiar to cancer cells but include a micrograph of surface staining of a normal lymphocyte. The stain was removed by deoxyribonuclease of diverse sources and by neuraminidase but not by protease. Later studies from Rosenberg’s group failed to remove the surface stain by the use of deoxyribonuclease (McAllister et al., 1977). A summary of main sources of work on cell surface DNA appears in Table 11.

B. A POSSIBLE SOURCE OF CELLSURFACE DNA A synopsis of the studies described so far suggests that DNA is found in cytoplasmic and cell surface compartments where its presence is probably not

+

2AgCl

I

OH FIG.5 . Scheme for synthesis of a platinum pyrimidine blue from uracil. The reaction of the “diaquo species” with uracil can be carried out in water or N,N-dimethyl formamide at pH 2-7.

TABLE I1 PROPERTIES OF CELLSURFACE DNA

Source

Preparative method

Proportion of total cell DNA (%)

Human diploid lymphocytes

Detergent lysis

0.5-2

Baby rat thymus cells

Extracted from culture fluid

0.016

Lectinstimulated human lymphocytes

Extracted from culture fluid

10-15

HlUTlan lymphocytes

Extracted from culture fluid

8

"Nuclear DNA had identical buoyant density.

Molecular Weight ( X 10") 4.4

Rate of labeling vs nuclear DNA Slower

0.45-0.62

70X Faster

Buoyant density (gm cm-')

Sensitivity to antimetabolites vs nuclear DNA

Reference

1.699"

Blocked by rifampicin; nuclear DNA unaffected

Lemer er al. (1971); Meinke er al. (1973)

1.707

-

Reid and Blackwell ( 1972a)

Broad band 1.700

-

Rogers (1976)

-l7

-

Anker e? al. (1975, 1976)

CYTOPLASMIC A N D CELL SURFACE D N A

47

due to nuclear contamination. Embryonic, regenerating, and neoplastic cells seem to contain more than adult cells. Many physical and physicochemical properties of surface and cytoplasmic DNA are shared with that of DNA of nuclear origin and, if the former is not an artifact of preparation resulting from contamination by the latter, there is a tacit understanding that cytoplasmic and cell surface DNA are physiological products of the nucleus. Indeed autoradiographic studies with human material show that DNA, isolated from the surface membranes of previously labeled lymphocytes and hybridized to cultured fibroblast chromosomes, labels the heterochromatic segments of all the chromosomes, more especially that of chromosome 9 (Kuo et al., 1975). The origin of cytoplasmic DNA is not discussed in the biochemical literature beyond the implied concept that its source is in the nucleus from which it is somehow transferred to the periphery. Morphological studies correlating phase contrast and electron micrographic techniques have shown that the mechanism is closely correlated with amoeboid movement in the differentiating cell (Reid, 1974). The finding by Anker and colleagues (1976) that a small fraction of the DNA results from extracellular synthesis indicates that the nucleus is not the sole source of the surface DNA. The synthesis of nuclear DNA is of course the subject of a voluminous tract of literature in contrast to the pathways of cell surface synthesis which are unknown. However, some speculation on a possible pathway is warranted inasmuch as aspects of biology of far-reaching consequence may be involved. In terms of the following discussion it is possible that the synthesis may be connected with synthesis of the carbohydrate component of surface glycoproteins and glycolipids. Glycoproteins and Glycolipids A combined biochemical and morphological approach over the past decade has produced an insight into the synthesis of glycoproteins and glycolipids by animal cells. The results of these investigations have shown that glycosylation occurs, in the main, on the Golgi apparatus (Bennett et al., 1974). Toward the end of the last decade, a paper appeared (Swenson and Kern, 1968) suggesting that terminal additions to the sugar chains of glycoproteins undergoing synthesis could be added in membrane sites outside the Golgi apparatus extending as far as the plasma membrane. In one of the mechanisms for glycosylation, a nucleotide diphosphate sugar donor transfers its sugar residue to the growing oligosaccharide chain, the reaction being catalyzed by a specific glycosyl transferase (Pazur and Aronson, 1972; Hughes, 1975). Since DNA has been isolated from cell membranes, it occurred to us that nucleotide monophosphates, produced from nucleotide diphosphate sugars, might be incorporated into DNA. Alternatively, the nucleotide diphosphate might be incorporated into DNA by the action of nucleotide phosphorylases. If this type of combined pathway could occur on the cell surface, the pcesence of DNA and complex carbohydrates in mucoid coats might be rationalized. The demonstration by Anker et al. (1976) of an

48

BEVAN L. REID A N D ALEXANDER J . CHARLSON

extracellular synthesis of DNA implies that nucleotide precursors and the necessary polymerases for the synthesis are present on or near the surface. a. Ectoglycosyl Transferases. Evidence for the existence of glycosyl transferases on cell surfaces has been accumulating. The biochemical studies which have provided this evidence have been the subject of a number of recent reviews (Roth, 1973; Keenan and Morre, 1975; Shur and Roth, 1975). In summary, the surface syntheses demonstrated in this work make use of nucleotide diphosphate sugars as substrates. The general plan of experiments to show the presence and activity of cell surface glycosyl transferases involves the addition of labeled nucleotide diphosphate sugars to cultures of whole cells followed by separation and recovery of labeled products by chromatographic means. Recovery of labeled product is taken as evidence for the presence of the glycosyl transferase with respect to four conditions. First, the nucleotide diphosphate sugar is thought not to enter the intact cell. Second, the presence of an excess of the specific unlabeled sugar does not affect the amount of labeled product recovered. Third, there is no cell death or lysis during the period of the experiment. Last, the density of cells in the culture flask is such that the cells have not yet reached confluence. An extensive discussion of these conditions and their varying fulfilment in numerous studies by different investigators is given in the review of Shur and Roth (1975). Morphological evidence employing autoradiography has also been used to study the localization of the labeled product (Porter and Bernacki, 1975). These studies corroborate the biochemical studies in showing that, in the case of L-1210 leukemia cells, the label is confined to the plasma membrane of the cell and to the surrounding extracellular coat. b. Su$ace Pyrophosphatases. Evans (1974) has shown that there is an enzyme on the outer surface of liver cells which can catalyze the degradation of a variety of nucleotides, such as adenosine triphosphate and uridine diphosphate galactose, to nucleotide monophosphates. This enzyme is a sialoglycoprotein. Evans also pointed out that nucleotide pyrophosphatases have been detected on a number of mammalian cell surface membranes, and that, in the case of cultured hamster cells, there was indirect evidence for the enzyme being on the outer surface of the membrane. Deppert et al. (1974) suggest that pyrophosphatases on the surface of hamster kidney cells grown in culture can catalyze the degradation of uridine diphosphate galactose, guanosine diphosphate fructose, and cytidine diphosphate N-acetylneuraminic acid to yield the 1 -0-phosphates of the corresponding monosaccharides. Presumably then the other products are the nucleotide monophosphates. The demonstration of glycosyl transferases and pyrophosphatases on cell surfaces makes our suggestion, of nucleotide diphosphate sugars acting as substrates for synthesis of both nucleic acids and complex carbohydrates, plausible. A major problem, however, is that there appears to be no information on the presence of endogenous sugar nucleotides occurring as a normal component of

CYTOPLASMIC AND CELL SURFACE DNA

49

the external cell membrane or its mucoid coat. It seems most unlikely that there are no substrates for the ectoglycosyl transferases, and it would appear that the sugar nucleotides have not been detected either because they occur in very small amounts or because they are transient metabolites. Detailed chemical investigations in this area are necessary before the biological function of the ectoglycosyl transferases can be understood. It would also be valuable to conduct experiments with nucleotide diphosphate sugars bearing a radioactive label in the heterocyclic base to ascertain whether the nucleotide portion of the molecule can be incorporated into cell surface DNA.

IV. Some Biological Implications The evidence from biochemistry and morphology reviewed in this article indicates that three separate compartments exist for accommodation of cellular DNA; nuclear, cytoplasmic including the plasma membrane, and extracellular or excreted DNA. If these three compartments are part of a covalently linked polymer system, as seems not unlikely, then an interchange possibly of great velocity between the compartments is equally as likely. It is possible in such a close-knit system that a rearrangement of the carbohydrate chains of surface glycoprotein would be inevitably accompanied by a rearrangement of the sequence of a nucleotide polymer formed on the surface at the time. Some such mechanism could account for the uniqueness of the base sequence of the surface DNA which is such a striking feature of every study devoted to the topic. If the carbohydrates are the most outward of the macromolecular groups at the cell wall, the oligosaccharide chains of surface glycoproteins and glycolipids, which consist of a spectrum of monosaccharide residues, may be useful in the steric matching of environmentally situated shapes, more especially those induced in water. A record of the heterogeneity of the environment would now be available as a sequence of monosaccharides and so of nucleotides in polynucleotide form. Such matching may be of overall significance in adapting the cell to its external environment. Continuity of the DNA concerned with cytoplasmic and, more remotely, nuclear compartments by interpolymeric valency or protein links coupled with the demonstrable ion-assisted movement of DNA between the compartments then represents a possible mechanism for more permanent and central storage of such a record.

V. Summary Nucleocytoplasmic relationships have been the subject of numerous essays in the past and the more recent papers have drawn attention to the existence of DNA

50

BEVAN L. REID AND ALEXANDER J . CHARLSON

not only in mitochondria and plastids but also disseminated throughout the membrane system of the cytoplasm, the cell wall, and beyond into the mucoid coat. Advanced biochemical and biophysical techniques have been used to differentiate cytoplasmic and cell surface DNA from nuclear DNA and to counter the trivial argument that such DNA is a contaminant produced during the preparation. Evidence is presented for the presence of cytoplasmic DNA other than that found in mitochondria using the disciplines of morphology and biochemistry. The morphological evidence is subdivided into that derived from histochemical, ultrastructural, and autoradiographic approaches. Histochemical studies on the topic are of greater antiquity in suggesting a passage of material from nucleus to cytoplasm especially in plants and invertebrate animals. Many of them rely on Feulgen methods which have sometimes been criticized. Ultrastructural studies of unusual nuclear membrane protuberances have been interpreted as evidence for an active nuclear contribution to the cytoplasm. Their origins in the peripheral nuclear heterochromation is a description common to several observations. Several autoradiographic studies attest to the presence of specific DNA label sensitive to deoxyribonuclease both over the cytoplasm unrelated to mitochondria and over the cell wall and associated mucoid coat. Biochemical methods are concerned with cell fractionation studies using centrifugation in various media. They reveal that a varying but small fraction of the DNA is found in postmitochondrial fractions which in different studies may or may not vary from the nuclear DNA in several respects, including molecular weight, specific activity, sensitivity to antimetabolites, melting profile, and buoyant density. Most authors consider that there is sufficient similarity in these properties as to warrant the conclusion that the origin of this DNA is in the nucleus. They differ in their interpretation of this conclusion. A minority feel that this indicates nuclear contamination during preparation. The majority of investigators see in the disparity in the values obtained between nuclear and cytoplasmic DNA, clear evidence that the latter is a fraction whose specificity is inconsistent with any origin through contamination and propose some as yet unspecified physiological role. Evidence is presented for the existence of DNA both in the plasma membrane and beyond it into the mucoid coat of mammalian cells in culture using the same techniques in principle as for DNA in the cytoplasmic location. Parent materials used were those from separation of wall membranes and the supernatant from cultures of growing cells. Among the results of the latter studies is shown an actual synthesis of DNA occurring extracellularly . The newly synthesized polymer differs from the nuclear DNA in several biochemical parameters. A persistent and striking feature of these studies is the dissimilarity between surface and nuclear DNA as measured by their capacity to hybridize. Perhaps more so

CYTOPLASMIC AND CELL SURFACE DNA

51

than those investigating cytoplasmic nucleic acid, researchers in the field of surface DNA aver that they seem to be dealing with a unique species of DNA which is actually excreted by the cell sometimes in striking quantities. A synopsis of aspects of recent research on glycoprotein and glycolipid synthesis at the cell surface is given to draw attention to the fact that the substrates used by investigators are the sugar nucleotides. Should an endogenous sugar nucleotide be discovered on the cell surface, which is not unlikely, a metabolic path may be available for the polymerization of nucleotides extracellularly into a chain which bears strict correspondence with sugar chains being assembled at the same time. The use of such a polynucleotide when stored for adaptation of the cell to its environment presently or in the future is mentioned.

REFERENCES Agganval, S. K., Watner, R. W., McAllister, P. K., and Rosenberg, B. (1975). Proc. Natl. Acad. Sci. U.S.A. 12, 928. Anker, P., Stroun, M., and Maurice, P. A. (1975). Cancer Res. 35, 2375. Anker, P., Stroun, M., and Maurice, P. A. (1976). Cancer Res. 36, 2832. Bach, M. K. (1962). Proc. Natl. Acad. Sci. U.S.A. 48, 1031. Baril, E. F., Jenkins, M. D., Brown, 0. E., Laszlo, J., and Moms, H. P. (1973). Cancer Res. 33, 1187. Bell, E. (1969). Nature (London) 224, 326. Bell, E. (1971). Science 174, 603. Bennett, G., Leblond, C. P:, and Haddad, A. (1974). J. Cell Eiol. 60, 258. Binkerd, P., and Toliva, A. (1974). Mol. Cell. Biochem. 5 , 177. Bond, H. E., Cooper, J. A., 11, Conington, D. P., and Wood, J. S. (1969). Science 165, 705. Brachet, J. (1965). Nature (London) 208, 596. Buckley, I. K. (1975). Electron. Microsc., Proc. Int. Congr., 8th, 1974 Abstract, Vol. 2, p. 338. Chayen, J. (1960). Exp. Cell Res. 20, 150. Clark, W. H. (1960). J. Eiophys. Eiochem. Cytol. 1, 345. Coppleson, M., and Reid, B. L. (1969). Lancet 2, 216. Deppert, W., Werchau, H., and Walter, G. (1974). Proc. Natl. Acad. Sci. U.S.A. 71. Evans, W. H. (1974). Nature (London) 250, 391. Fromson, D., and Nemer, M. (1970). Science 168, 266. Fussell, C. P. (1968). J . Cell Eiol. 39, 264. Gulvag, B . M. (1970). Grana 10, 31. Hall, M. R., Meinke, W., Goldstein, D. A,, and Lerner, R. E. (1971). Nature (London), New Eiol. 234, 227. Hughes, R. C. (1975). Essays Eiochem. 11, 1 . Ito, S., Chang, S . , and Pollard, T. D. (1969). J . Protozool. 16, 638. Jacob, F., Brenner, S., and Cuzin, F. (1963). Cold Spring Harbor Symp. Quant. Eiol. 28, 329. Jaworska, H . , and Lima da Faria, A. (1973). Hereditas 74, 187. Keenan, T. W., and Moore, D. J. (1975). FEBS Lett. 55, 8. Koch, J. (1973). FEES Lett. 32, 22. Koch, J., and von FYeil, H. (1971). FEES Len. 17, 312. Kuo, M. T., Meinke, W., and Saunders, G. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 2004.

52

BEVAN L. REID AND ALEXANDER J. CHARLSON

Lerner, R. E.,Meinke, W., and Golstein, D. A., (1971). Proc. Nut/. Acad. Sci. U.S.A. 68,1212. Lindholm, L., and Britten, S. (1967). Exp. Cell Res. 48, 660. Lippert, B. (1977). J. Clin. Hematol. Oncol. 7, 26. McAllister, P. K., Rosenberg, B., Aggarwal, S. K., and Wagner, R. W. (1977). J. Clin. Hematol. Oncol. 7, 717. Meinke, W., and Goldstein, D. A. (1974). J. Mol. Biol. 86, 757. Meinke, W., Hall, M. R., Goldstein, D. A., Kohne, D. E., and Lerner, R. E. (1973). J. Mol. Biol. 78, 43. Melera, P. W., and Cronin-Sheridan, A. P. (1976). Biochim. Biophys. Acta 432, 300. Mollo, F., and Stramignoni, A. (1967). Br. J. Cancer 21, 519. Novak, B., and Elford, H. (1973). Biochem. Biophys. Res. Commun. 54, 633. Pazur, J. H., and Aronson, N. N., Jr. (1972). Adv. Carbohydr. Chem. Biochem. 27, 301. Politis, G., Plassara, M. G., and Thomou-Politi, H. (1975). Nature (London) 257, 485. Porter, C. W., and Bernacki, R. J. (1975). Nature (London) 256, 648. Reid, B. L. (1974). BioSystems 5, 207. Reid, B. L., and Blackwell, P. M. (1970). Ausr. J . Med. Technol. 1, 2 . Reid, B. L., and Blackwell, P. M. (1971). In “Informative Molecules in Biological Systems” (L. Ledoux, ed.), p. 285. North-Holland Publ., Amsterdam. Reid, B. L., and Blackwell, P. M. (1972a). Aust. J . Med. Technol. 3, 121. Reid, B. L., and Blackwell, P. M. (1972b). Acra Fertil. Sreril. 3, 193. Reid, B. L., Hagan, B., and Kaye, M. (1976). Aust. Vet. Practitioner 6, 235. Rogers, J. C. (1976). J. Exp. Med. 143, 1249. Rogers, J. C., Boldt, D., Kornfeld, S., Sr., Skinner, A,, and Valeri, C. R. (1972). Proc. Narl. Acad. Sci. U.S.A. 69, 1685. Roth, S. (1973). Q. Rev. Biol. 48, 541. Sarma, D. S. R., and Rutrnan, J. (1972). Fed. Proc., Fed. Am. SOC. Exp. Biol. 31, 607. Schjeide, 0. A., and I-San Lin, R. (1970). In “Cell Differentiation” (0. A. Schjeide and J. de Vellis, eds.), p. 224. Van Nostrand-Reinhold, Princeton, New Jersey. Schneider, W. C., and Kuff, E. L. (1969). J. Biol. Chcm. 244, 4843. Schneider, W. C., Shelton, E., and Kuff, E. L. (1975). J. Natl. Cancer Inst. 55, 665. Sebuwufu, P. H. (1966). Nature (London) 212, 1382. Shur, B. D., and Roth, S. (1975). Biochim. Biophys. Acta 415, 473. Singer, A., and Reid, B. L. (1970). J . Reprod. Fertil. 23, 249. Smith, G. H., Kiddwell, W. R., and Schneider, W. C. (1975). Exp. Cell Res. 96, 321. Solage, A., and Loskov, R. (1975). Eur. J. Biochem. 60, 23. Sparrow, A. H., and Hammond, M. R. (1947). Am. J. Bot. 34, 439. Stroun, M., Anker, P., Maurice, P., and Gahan, P. B. (1977). h i . Rev. Cytol. 51, 1. Swenson, R. M., and Kern, M. (1968). Proc. Natl. Acad. Sci. U.S.A. 59, 546. Toro, I . , and Olah, I. (1966). Nature (London) 212, 315. Williamson, R. (1970). J . Mol. Biol. 51, 157.

INTERNATIONAL REVIEW OF CYTOLOGY. VOL . 60

Biochemistry of the Mitotic Spindle CHRISTIAN PETZELT Institutefor Cell Research. German Cancer Research Center. Heidelberg. West Germany I . Introduction . . . . . . . . . . . . . . . . . . I1. Tubulin . . . . . . . . . . . . . . . . . . . A . Isolation . . . . . . . . . . . . . . . . . . B . Properties . . . . . . . . . . . . . . . . . C . Tubulin in the Spindle . . . . . . . . . . . . . 111. Actin . . . . . . . . . . . . . . . . . . . . A . Characteristics . . . . . . . . . . . . . . . . B . Actin in the Spindle . . . . . . . . . . . . . . IV . Myosin . . . . . . . . . . . . . . . . . . . . A . Myosin in Cells . . . . . . . . . . . . . . . B . Myosin in the Spindle . . . . . . . . . . . . . V . Calcium-Dependent Regulator Protein . . . . . . . . A . Characteristics . . . . . . . . . . . . . . . . B . Calcium-Dependent Regulator Protein in the Spindle . . VI . Dynein . . . . . . . . . . . . . . . . . . . . A . Characteristics . . . . . . . . . . . . . . . . B . Dynein in the Spindle . . . . . . . . . . . . . VII . The Mitotic Ca2+-ATPase . . . . . . . . . . . . . A . The Enzyme in the Cell Cycle . . . . . . . . . . B . Characteristics . . . . . . . . . . . . . . . . C . Functions . . . . . . . . . . . . . . . . . VIII . Calcium in the Mitotic Cell . . . . . . . . . . . . IX . The Isolation of the Mitotic Spindle . . . . . . . . . A . Isolation Procedures . . . . . . . . . . . . . . B . Spindle Models . . . . . . . . . . . . . . . X . Mitotic Centers . . . . . . . . . . . . . . . . . XI . Concluding Remarks . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . .

. .

53

. . . .

54 54

. . . .

55 59 62 62 63 65 65 65 66 66 68 68 68 69 70 70 71 71 74 75 75 78 79 81 82

. .

. .

. . . .

. . . .

. .

. .

. . . .

. . . .

. . . .

. . . .

. . . . . .

. .

. . . . . .

.

I Introduction The mitotic spindle has been described by Flemming (1880) to consist of the chromatic and the achromatic figure . This description is still quite useful today . The chromatic figure consists of the chromosomes and includes their attachment points to the spindle. the kinetochores . The achromatic figure is more difficult to define . It consists of the transient spindle-like structure. and. where present. the asters and the centrioles . In this paper only the biochemistry of the achromatic 53

Copyright @ 1979 by Academic Press. Inc . All rights of reproduction in any form reserved . ISBN 0-12-364360-0

54

CHRISTIAN PETZELT

figure will be discussed since on the composition of the mitotic chromosomes not very much is known which bears a direct relationship to mitosis. This review deals mainly with results from the last decade. For a detailed review of our knowledge of mitosis before that the reader is referred to the excellent articles by Mazia (1961) and Luykx (1970). The evolution of the mitotic spindle has been described by Kubai (1975) and the ultrastructure of the mitotic spindle has been reviewed most recently by Fuge (1978). Any discussion on the biochemical properties of such an ill-defined and labile structure as the mitotic spindle has to center on the question whether the fact described is either a methodological artifact or a cytoplasmic contamination unrelated to the mitotic apparatus or a specific element of this structure. To follow this strategy as much as possible, it is assumed that biochemical facts and observations obtained for the spindle in the living cell will provide the least distorted image of reality, open only for misinterpretations by the observer. The second step in checking facts will deal with informations obtained from the fixed cell. In discussing the results obtained by the various immunofluorescent techniques and by some physical methods the problem of artifacts versus true and accepted properties of the spindle will have to draw most of our attention. The third level of discussion describes the molecules in the isolated mitotic apparatus. After discussing the characteristics of the various isolation procedures and the criteria used for evaluating them, the recent literature on the composition of the isolated spindle will be reviewed. Finally, we report on the state of the art of the ultimate goal of every student of mitosis, the repetition of the mitotic process in vitro using spindle proteins in spindle models.

11. Tubulin

Microtubules are one of the prominent structures found in the spindle. They are composed of one major protein which was named tubulin (Mohri, 1968). According to Bryan (1974a) this name should be restricted “to microtubule protein(s) in a native configuration as judged by the ability to repolymerize or bind any of several antimitotic drugs. The two protomeric subunits are termed a-tubulin and P-tubulin (Bryan, 1974a). ”

A. ISOLATION The protein was identified first by Taylor (1965), Borisy and Taylor (1967a,b), and Wilson and Friedkin (1967) on the basis of its colchicine-binding activity. Weisenberg et al. (1968) described the protein as a dimer with a molecular weight of 55,000 for the subunit. One of its main characteristics is its capacity to bind specifically colchicine (Taylor, 1965). On the basis of this

BIOCHEMISTRY OF THE MITOTIC SPINDLE

55

property it can be purified by ammonium sulfate precipitation and ion-exchange chromatography on DEAE-Sephadex as a colchicine-tubulin complex (Eipper, 1972). Another characteristic property was used by Bryan to purify tubulin from cells by inducing the formation of paracrystals in sea urchin eggs by treatment with vinblastine (Bryan, 1971, 1972a). Using nonionic detergents the crystals formed in the cell can be isolated. They are birefringent and labile in solutions of low ionic strength (Bryan, 1971). It has been shown that they consist mainly of tubulin and that the protein has retained its colchicine-binding capacity (Bryan, 1972a). In the following a short description of the isolation technique and the chemistry of tubulin is given. For a complete survey the reader is referred to the reviews by Bardele (1973), Roberts (1974), Wilson and Bryan (1975), Snyder and McIntosh (1976a), Stephens and Edds (1976), Mohri (1976), and, especially, to the recently published extensive review of the field by Dustin (1978). The most natural approach to isolate tubulin appears to be the isolation of microtubules and their subsequent dissociation. This has been done by a number of workers who used the stability of a certain class of microtubules, the so-called “stable” microtubules, for isolation (Stephens, 1968; Renaud et al., 1968, Shelanski and Taylor, 1968; Jacobs and McVittie, 1970; Everhardt, 1971; Witman e t a l . , 1972; Meza et a f . , 1972). Microtubules are called “stable” when they resist a variety of treatments like low temperature, antimitotic agents, or pressure (Behnke, 1970). They form the microtubular structure in cilia, flagella, centrioles, and axostyles. B. PROPERTIES The characteristics described for tubulin from these sources are nearly identical to those for tubulin of other origins like the brain. Tubulin is a heterodimer consisting of equal amounts of a-tubulin and /3-tubulin (Bryan and Wilson, 1971; Wilson and Bryan, 1975; Ludueiia et al., 1975). It binds 2 moles of guanine nucleotide per mole of tubulin dimer. One of these is tightly bound, whereas the other is rapidly exchangeable (Berry and Shelanski, 1972; Weisenberg, 1975). Additionally, it has special binding sites for a variety of antimitotic drugs. Colchicine and colcemid (Taylor, 1965), rotenone (Brinkley et a f . , 1974; Barham and Brinkley , 1976a,b), podophyllotoxine (Bryan, 1972b), and steganacine (Wang et a f . , 1977) share a binding site on the protein molecule, whereas vinblastine and related compounds (Bryan, 1971, 1972a), chlorpromazine (McGuire et al., 1974), mescaline (Harrison et al., 1976), oncodazole (Hoebeke et af., 1976), maytansine (Remillard et a f . , 1975), and griseofulvine (Roobol et al., 1976; Weber et al., 1976a; Schatten, 1977) bind to different sites. The interaction of tubulin with these drugs is in most cases quite

56

CHRISTIAN PETZELT

specific thus allowing the molecular understanding of their antimitotic effects. Colchicine, especially, has been used for many years as an antimitotic drug (cf. Eigsti and Dustin, 1955), and we know now that by its binding to tubulin the polymerization to microtubules is inhibited. The antimitotic drugs do not have a direct effect on existing microtubules. It seems that their binding site is buried after the tubulin is polymerized. The effect of colchicine and other drugs on mitotic spindles may be understood if the state of the polymerized microtubules is considered a labile one, i.e., that an equilibrium exists between tubulin subunits and the polymerized microtubules. An inactivation of the subunits by the antimitotic drug would thus cause the polymerized microtubules to disintegrate in order to maintain the equilibrium (cf. Margolis and Wilson, 1978). 1. In Vitro Polymerization of Brain Tubulin

Major progress was made in tubulin chemistry when Weisenberg (1972) and Borisy and Olmstedt (1972) succeeded in polymerizing tubulin to microtubules in vitro using a high-speed supernatant of a mammalian brain homogenate. The microtubules obtained were virtually identical to the microtubules observed in vivo as judged by electron microscopy. They were cold labile and the polymerization process could be inhibited by the antimitotic drugs described above. Since then, several facts on the polymerization of tubulin to microtubules in vitro have been established: It is necessary to keep a very low Ca2+ concentration in the tubulin solution to obtain polymerization. More than 1 mM Ca2+ will block the process. Therefore, a Ca2+ chelator, EGTA, is commonly used (concentration ca. 1 mM). Mg2+ions have an inhibiting effect at concentrations above 10 mM; low Mg2+ (ca. 0.5 mM) increases the polymerization rate. Addition of GTP is necessary for the polymerization; other nucleotides like ATP result in a decreased polymerization rate but microtubules may be obtained. However, a transphosphorylation between ATP and GTP may allow the polymerization to proceed in spite of a strong specificity of the tubulin for GTP (Penningroth and Kirschner, 1977). The pH range for polymerization is rather broad (pH 5.5-7.5), the optimum being about pH 6.6. For a buffer, normally either 0.1 M Mes or Pipes is used. The addition of glycerol in concentrations up to 4 M enhances the polymerization rate although it is not yet clear if the addition of glycerol to the cytosol does not favor a biased selection of tubulins. The polymerization of tubulin to microtubules can be followed by measuring either the increasing viscosity or the increasing turbidity of the solution and can be most easily verified by negative staining of the polymerized microtubules with uranyl acetate and visualization of them in the electron microscope (Olmsted and Borisy, 1973). 2. In Vitro Polymerization of Tubulin from Nonneuronal Sources All the results reported so far were obtained using brain tubulin of a variety of species. Only recently, however, was it possible to purify tubulin from non-

BIOCHEMISTRY OF THE MITOTIC SPINDLE

57

neuronal cells and polymerize it to microtubules. Green et al. (1975) succeeded first using embryos of Drosophila melanogaster. Subsequently, Barnes et al. (1975) with renal medullary cytosol, Wiche and Cole (1976) with homogenates of C,-glioma cells, Fuller et a l . (1975b) with transformed and untransformed 3T3 cells, Kuriyama (1977) with sea urchin and starfish eggs, Doenges et al. (1977) with Ehrlich ascites cells, Nagle et al. (1977) with neuroblastoma cells, Farell and Wilson (1978) with outer doublet microtubules of sea urchin sperm tails, Weatherbee et al. (1978) with HeLa cells, and Maekawa and Sakai (1978) with Tetrahymena obtained a reversible assembly-disassembly of microtubules with tubulin from these sources. If one looks for a common property of all the assembly mixtures, one finds as the main prerequisite for the assembly of tubulin from nonneuronal cells a high ratio of cells to buffer at homogenization. If the protein concentration and, concomitantly, the tubulin concentration is too low, one does not obtain assembly. This phenomenon of a critical concentration for the tubulin assembly mixture has been known for a few years (Johnson and Borisy, 1974). The data, however, were obtained with brain tubulin. It took some time until it was realized that the tubulin content in the cell is much lower than in the brain and that, therefore, a much higher cell to buffer ratio was needed to obtain a tubulin concentration in the homogenization mixture above the critical concentration. All other conditions for assembly were comparable to the conditions described for the brain system. If the protein concentration is high enough, the presence of 1 mM GTP, 1 mM EGTA, and the Mes buffer system at a pH of 6.4 is sufficient to allow polymerization. The microtubules obtained in this way all look alike provided the same preparation procedures are used. Bryan et al. (1975) described an assembly-inhibiting factor which was identified as RNA. Recently, Naruse and Sakai (1978) isolated an assemblyinhibiting factor from the cortex of sea urchin eggs. It is not clear yet if these factors are actually used in vivo or if the assembly-inhibiting property of the isolated component is just a concomitant effect of a molecule otherwise unrelated to the microtubule system. 3. Microtubule-Associated Proteins It was thought that the way to purify tubulin to homogeneity was to assemble-disassemble microtubules in vitro and that the result would be pure tubulin. However, soon after the in vitro model of tubulin was described (Weisenberg, 1972; Borisy and Olmsted, 1972), it was found that several proteins copolymerized with tubulin. After several assembly cycles the relative amounts of these proteins remained the same compared to tubulin. They were described as proteins with a high molecular weight of 271,000 to 345,000 by Murphy and Borisy (1974, 1975) and Borisy et al. (1975) and were called, therefore, high-molecular-weight proteins (HMW proteins). They can be separated by ion-exchange chromatography on DEAE-cellulose or phosphocellulose from tubulin. The latter, being then electrophoretically pure, is unable to

58

CHRISTIAN PETZELT

polymerize. Only a high increase in its concentration (>lo mg/ml) or in the Mg2+concentration (>16 mM) or the presence of DMSO (Himes et al., 1977) will result in microtubule formation (Himes et al., 1976a; Lee and Timasheff, 1977; Herzog and Weber, 1977). Weingarten et al. (1975) described a different class of microtubule-associated proteins (MAPs) which they called the Tau factor. Electrophoretically they consist of four closely spaced proteins with a molecular weight of 55,000 to 62,000 (Cleveland et al., 1977a,b). Like at least some of the HMW proteins, the Tau proteins not only promote the initiation of microtubule assembly but are also required in the elongation process (Witman et al., 1976). Like the HMW proteins the Tau proteins can be separated from tubulin by chromatography on phosphocellulose. Their exact function in microtubule assembly is still unclear. In the brain tubulin system it seems that only in the presence of MAPs can ring-like intermediate structures be formed (Vallee and Borisy, 1978). These rings have never been found during assembly of tubulin from nonneuronal cells. However, if the MAPs from the latter cells are removed by ion-exchange chromatography and the resulting pure tubulin is combined with MAPs derived from brain, rings are formed in the course of tubulin assembly (Nagle et al., 1977). Several hypotheses have been put forward to explain the function of the MAPs. They seem to be an integral part of the polymerized microtubule (Dentler et al., 1975; Murphy and Borisy, 1975); their role in the assembly reaction seems to be equally important since they are at the low tubulin concentrations (Cleveland et al., 1977a) required for polymerization (Murphy and Borisy, 1974). They are specific for a given tubulin system like brain tubulin versus nonneuronal tubulin. All MAPs described so far have a molecular weight higher than that of tubulin. Klein et al. (1978) describe a HMW protein of 290,000 from SV40transformed mouse fibroblasts which copolymerizes with rat brain tubulin. Interestingly, about 20% of it is in an insoluble form bound to the plasma membrane. Recently however, a protein of molecular weight 49,000 has been reported to copolymerize with tubulin (Nagle ef al., 1977). Since we are just beginning to investigate the different microtubular systems, we will probably discover more MAPs which will add new possibilities to the diversity of the microtubular system (e.g., Gaskin et al., 1974). In addition to the somewhat artificial conditions used to assemble tubulin without MAPs to microtubules described by Lee and Timasheff (1977), Himes et al. (1976b), Herzog and Weber (1977), and Himes et al. (1977), a recent report by Farell and Wilson (1978) describes a polymerization-competent tubulin system derived from the outer doublets of sea urchin flagella where tubulin can be purified by ionexchange chromatography on phosphocellulose and still retain its full polymerizability. An additional remarkable fact is that this tubulin derives from so-called “stable” microtubules of sperm flagella. Once it is solubilized by sonication, it is sensitive to cold, colchicine, and calcium as are the other tubulin systems.

BIOCHEMISTRY OF THE MITOTIC SPINDLE

C. TUBULIN IN

THE

59

SPINDLE

Up to now no mitotic spindle in the eukaryotic cell has been found without microtubules (cf. Kubai, 1975; Fuge, 1978). Tubulin is an integral part of the mitotic spindle (Cohen and Rebhun, 1970). Burnside et al. (1973) have estimated that in Spisula eggs the tubulin content is 2.0 to 2.4 x pg/egg or 3.3%of total cell protein. The presence of microtubules in the spindle of a living cell can be demonstrated by its property to show birefringence in polarized light (Inoue and Dan, 1951) and a correlation of the amount of microtubules to the amount of birefringence has been established (Inoue and Sato, 1967; Rebhun and Sander, 1967; Stephens 1972a, 1973; Inoue et al., 1975; Sat0 et al., 1975; Fuseler, 1975; Salmon, 1975a,b,c; Sluder, 1976). Sat0 et al. (1975) have calculated that “in the living Pisaster spindle the microtubular contribution to the dry mass is a minimum of 24%. In the fixed cell, microtubules can be shown to be present in the spindle by electron microscopy or by immunofluorescent techniques. Immunologically, it was shown that sera containing antibodies to tubulin from flagellar and mitotic microtubules cross-react with tubulin from both sources (Fulton et al., 1971; Dales et al., 1973). Ferritin-labeled antibodies to tubulin from vinblastine paracrystals stain the microtubules in the spindle (Dales, 1972). By using a monospecific antibody to brain tubulin it was demonstrated that cytoplasmic and spindle microtubules react with antibodies against bovine brain tubulin (Fuller et al., 1975a). This technique, using mostly a monospecific antibody to tubulin and indirect immunofluorescence, has been applied by a number of workers to establish the microtubule pattern in the spindle and to show that tubulins from nearly all cells and sources cross-react with each other (Franke et al., 1977; Fujiwara and Pollard, 1978; Fuseler et al., 1976; Pepper and Brinkley, 1977; Osborn and Weber, 1976; DeBrabander et al., 1977; Weber et al., 1975, 1976b). Recently, however, Morgan et al. (1978), using 1251-labeled tubulins from different species, observed immunological differences in its binding capacity indicating different densities of shared antigenic determinants. By a careful comparative analysis using electron microscopic and immunofluorescent methods on the same preparation, Osborn et al. (1978) proved that antibodies to tubulin stain polymerized microtubules (cf. Weber et al., 1978). As has been shown for mammalian cells, the spindle tubulin is not synthesized during a distinct period of the cell cycle but is made continuously with only a slight increase in the rate of synthesis toward the end of the cycle (Lawrence and Wheatley, 1975). The tubulin pool in the cell does not even change dramatically during the morphological differentiation of mouse neuroblastoma cells (Morgan and Seeds, 1975). During the early embryogenesis tubulin is used for the spindle after having been synthesized earlier during oogenesis (Bibring and Baxandall, 1977). So about 12% of the total protein synthesis during the oogenesis of sea urchins is concerned with the synthesis of tubulin (Cognetti et al., 1977). Tubulin can also be made on preformed mRNA (Raff et al., 1971, 1975; Raff and ”

60

CHRISTIAN PETZELT

Kaumeyer, 1973). Up to now it has been difficult to decide if only a given tubulin can fulfill a certain task or if the different tubulins described in a cell are interchangeable. Some time ago different classes of microtubules were suggested, based on different solubility properties. Behnke and Forer (1967) presented evidence that in sea urchins four classes of tubulins may exist: cytoplasmic and spindle microtubules, the central pair tubules of flagella, the Bsubfibers of the outer fibers, and the A-subfiber (cf. Stephens, 1970, 1976). Tilney and Gibbins (1968) explained the stability differences by modification of the tubulin itself, not by assuming different tubulins. Brinkley and Cartwright (1970) showed that even in a single spindle, chromosomal and nonchromosomal fibers react differently to cold, thereby suggesting differences among tubulins. A similar result was obtained by Lambert and Bajer (1977) who found a differential sensitivity of plant microtubules to cold. Salmon et al. (1976) described a differential stability of spindle microtubules to high pressure applied on cells. The interpolar and astral microtubules appear to be the most labile ones, the kinetochore microtubules are less sensitive, and the microtubules of the midbody of telophase cells are not afflicted at all by the treatment. Fuseler et al. (1976) showed that spindle microtubules are more sensitive to colchicine than cytoplasmic microtubules. Although it reacts with antibody to brain tubulin (Franke et a l . , 1977), plant tubulin is much less susceptible to the action of colchicine than tubulin from animal cells (Hart and Sabnis, 1967a,b; Rubin and Cousins, 1976). The same holds true for tubulin from yeast (Haber et a l . , 1972) and fungi (Heath, 1975a,b). Microtubular structures like centrioles are formed despite the presence of colchicine at concentrations which completely inhibit spindle formation (Dustin et a l . , 1976; Flament-Durant et ul., 1976). Bamburg et al. (1973) demonstrated solubility and colchicine-binding changes of the tubulin from chick brain during development. Kowitt and Fulton (1974) studied the synthesis of tubulin for the flagella that develop during cell differentiation in Nuegleria gruberi and found no interconversion between existing tubulin and new flagellar tubulin. On the other hand, Auclair and Siege1 (1966) and Stephens (1972b) pointed out the possibility that during sea urchin embryogenesis tubulin from the spindle is used for the formation of cilia. More evidence for the existence of different tubulins was given by Feit et al. (1971, 1977). By isoelectric focusing of tubulin they obtained up to six a- and two @-subunits.Tubulin can be extracted from isolated mitotic apparatus of sea urchin eggs by treatment with meralluride sodium (Bibring and Baxandall, 1968, 1971). If this tubulin is compared with tubulin derived from cilia and sperm flagella, the a-tubulin of the mitotic apparatus and of the A-tubule of the ciliary doublets is resolved electrophoretically into two bands, while the a-subunit of the flagellar doublet tubulin gives a single band. Additionally, the mitotic and the ciliary tubulins differ in the mobilities of their two a-species (Bibring et a l . , 1976). A decisive answer to the question of different tubulins will only be possible by analysis of the primary structure. Ludueiia and Woodward (1973)

BIOCHEMISTRY OF THE MITOTIC SPINDLE

61

compared the sequence of the first 25 NH,-terminal amino acids of a- and P-tubulin from outer doublets of sea urcin sperm and chick embryo brain and found a close resemblance between the corresponding subunits of the two species. Recent work by Ponstingl et al. (1978) indicates several amino acid exchanges in the a-and P-chains of tubulin derived from pig brain, proving the existence of different a-and P-tubulins. Limited proteolysis of a-and P-tubulins from baby hamster kidney cells and chicken brain followed by separation in SDS-gels according to Cleveland et al. (1977~)was used by Starger et al. (1978) to show clear species differences between a-and P-tubulin. Summarizing, one can say that there is strong evidence that several tubulins exist which are closely related to each other but may be responsible for different structures or functions in a cell. 1. Microtubule-Associated Proteins in the Spindle Studies on the microtubule-associated proteins add more complexities to the originally simple scheme of the microtubule system. Not only are there different tubulins within a cell but microtubule-associated proteins also seem to differ from one cell system to another. It was shown by Nagle et al. (1977) that an assembly-competent tubulin solution from nonneuronal cells lacks the HMW proteins described by Borisy et al. (1974) for the brain system, whereas Wiche and Cole (1976) found those proteins present in rat glial cells and copurifying with tubulin. Sherline and Schiavone (1978) obtained good immunofluorescent staining of the mitotic spindle of 3T3 cells with antibodies prepared to the HMW proteins from rat brain. A similar result was obtained by Connolly et al. (1978). By immunofluorescent techniques they could stain rat glial cells with antibodies to the HMW protein and obtain images indistinguishable from that seen when cells are treated with antitubulin serum. However, the same authors state that antibodies to Tau protein do not show immunofluorescent staining in rat glial cells. Anti-Tau proteins stain mouse fibroblasts, however, and this staining is also like the image obtained by antitubulin treatment (Connolly et a l . , 1977). Lockwood (1978) purified further the assembly-promoting MAPs of the Tau protein group and identified the most effective fraction as a protein of molecular weight 67,000. He named it tubulin-associated protein (TAP). Antibodies to this protein stain a variety of cells and again the image obtained is identical to that produced by antitubulin staining. A preliminary conclusion would be that as microtubules and, therefore, tubulin are an integral part of the mitotic spindle, microtubule-associated proteins are likewise a natural part of the spindle. Until now, it has been completely unclear if and how these MAPs are used by the cell although a number of authors suggest a control function for them in the tubulin assembly-disassembly system (Connolly et al., 1977; Murphy et al., 1977; Lockwood, 1978). Another possibility is discussed by Wiche et al. (1978). They describe a preferential binding of MAPs to mouse satellite DNA and, assuming that kinetochore DNA corresponds to

62

CHRISTIAN PETZELT

satellite DNA, they see the MAPS as linkers between the DNA and the microtubules.

111. Actin

Proteins typical of muscle have their counterpart in other cells. Actin has been described to occur in all types of cells ranging from vertebrates to invertebrates, plants, and protozoa, e.g., human thrombocytes (Bettex-Galland and Liischer, 1959), erythrocytes (Tilney and Detmers, 1975), macrophages (Stossel and Hartwig, 1975; Hartwig and Stossel, 1973, leukocytes (Boxer and Stossel, 1976), blood platelets (Elzinga et a l . , 1976), brain (Storti et al., 1976), liver (Govindan and Wieland, 1975), cultured cells (Bray and Thomas, 1975), echinoderm eggs (Kane, 1975, 1976), echinoderm sperm (Tilney et a l . , 1973), Limulus sperm (Tilney, 1975), plants (Forer and Jackson, 1976; Jackson and Doyle, 19771, Physarum (Hatano and Oosawa, 1966a,b; Loew, 1952), and Acanthamoeba (Weihing and Korn, 1971, 1972). For an extensive review on the occurrence of actin in nonmuscle cells the reader is referred to Pollard and Weihing (1974), Goldman et al. (1975), and Clarke and Spudich (1977). A. CHARACTERISTICS All the proteins described as actins share very similar properties. The molecular weight is 42,000 as determined by sequence analysis. Cellular actin like muscle actin polymerizes, forming thin filaments with a diameter of 40 to 60 %i or filament bundles which are then birefringent in the polarizing microscope. It can activate the Mg2+-ATPaseof myosin as has been demonstrated by Gordon et al. (1976) for the activation of heavy meromyosin from muscle by Acanthamoeba actin. Upon extraction it will form a gel under suitable conditions and-in certain cases-undergo contraction. This process has been demonstrated for actin extracts from many cell types like sea urchin eggs (Kane, 1975, 1976), mouse fibroblasts (Weber et a l . , 1977), Acanthamoeba (Pollard, 1976), Ehrlich ascites tumor cells (Kane et a l . , 1977; Moore and Caraway, 1978), Amoeba (Condeelis and Taylor, 1977), HeLa cells (Weihing, 1977), and macrophages (Hartwig and Stossel, 1977). The components necessary for gelation and contraction have been identified. In addition to actin, a MW 58,000 protein and a HMW protein of 220,000 (actin-binding protein) (Bryan and Kane, 1977) or, in mammalian cells, filamin (MW 250,000) (Wang et a l . , 1975; Davies et al., 1977), are required for gelation. Contraction occurs only in the presence of myosin. The gelation process is Ca2+ sensitive; more than micromolar amounts of Ca2+ are inhibitory. This gelation-contraction model provides experimental approaches to study the function of actin in the cell. This seems to be highly desirable since by no means is the cellular role of actin understood. It is part of the cytoskeletal

BIOCHEMISTRY OF THE MITOTIC SPINDLE

63

structure and essential for dynamic cellular processes such as transport and cytokinesis. But are these functions enough to account for the roughly 10% of total cell protein which actin represents? Actin itself is different in muscle and nonmuscle cells. Messenger RNA from chicken brain and chicken muscle produce different actins in cell-free protein synthesis in wheat germ extracts (Storti and Rich, 1976). Gruenstein and Rich (1975) describe changes in the tryptic peptide pattern of brain actin and muscle actin from chicken. By sequencing muscle actin and comparing it to some actin fragments from nonmuscle cells (blood platelets) Elzinga et al. (1976) showed the existence of different actin genes within the same organism. Further indications for these differences come from the work of Whalen ef a/. (1976), Carrels and Gibson (1976), and Rubenstein and Spudich (1977), who demonstrated by isoelectrofocusing the existence of three main actins. They named them a-,p-, and y-actin. a-Actin occurs only in muscle whereas p- and y-actin are found in nonmuscle cells. p- and y-actin can be further distinguished by different properties. If neurotubulin is assembled to microtubules in the presence of actin derived from CHO cells, y-actin associates predominately with the microtubules (Izant and McIntosh, 1977). A definite proof for the existence of several genetically different actins in a cell was recently given by Vandekerckove and Weber (1978). They compared muscle and cytoplasmic actins by fingerprint analysis and partial amino acid sequence determination and found that cytoplasmic actins differ from muscle actin by at least 25 amino acids and that cytoplasmic actins are the products of at least two genes, Furthermore, they show that mammalian cytoplasmic actins of the same type are very similar if not identical. B. ACTINI N THE SPINDLE As early as 1965 Aronson showed that isolated spindles from sea urchin eggs bind fluorescent heavy meromyosin. His pioneering work was followed up years later after Ishikawa et a / . (1969) described extensively the interaction between heavy meromyosin and actin, demonstrating the formation of arrowheaddecorated filaments in nonmuscle cells. These decorated filaments have been shown to be actin in the muscle system (Huxley, 1963). Two years after Ishikawa’s paper Behnke et a / . (1971) and Gawadi (1971) described filaments in the spindle which could be decorated by heavy meromyosin and which showed the typical arrowhead structure. Subsequently, Forer and Behnke (1972) demonstrated that thick filaments can be decorated by heavy meromyosin in the meiotic spindle of the crane fly Nephrotoma, Hinkley and Telser (1974) showed similar filaments in spindles of neuroblastoma cells, Gawadi (1974) gave an extended description of her original observations of heavy-meromyosin-decorated filaments in the mitotic spindle of locust spermatogonia, and Forer and Jackson (1976) showed that heavy-meromyosin-decorated filaments can also be found in plant spindles after they had studied mitoses in the endosperm of Haemanthus

64

CHRISTIAN PETZELT

katherinae. Schloss et a!. (1977) demonstrated actin filaments in the spindle of cultured cells by their binding of myosin subfragment. The work of Aronson (1965) was taken up by Sanger (1975a,b,c). He used fluorescent heavy meromyosin to study the distribution of actin in the spindle of rat kangaroo cells using the light microscope. He found fluorescent fibers between the chromosomes and the poles, a relatively strong staining of the kinetochores, and some fluorescence in the cleavage furrow at cytokinesis. He extended his work by modifying fixation and staining procedures and still obtained the same results (Sanger and Sanger, 1976; Sanger, 1977). A different approach was followed by Cande et al. (1977) who used the indirect fluorescent antibody technique with an actin antibody (Lazarides and Weber, 1974) to study, in rat kangaroo mitotic cells, the distribution of actin and to compare it to the tubulin distribution in the spindle. They also found actin predominately along the chromosomal fibers and obtained a distinct pattern of fluorescence for actin and tubulin. A recent paper by Herman and Pollard (1978) confirms most of the results mentioned insofar as actin is shown to be present in the spindle along the chromosomal fibers. The authors used fluorescent heavy meromyosin and tried to control fixation and staining procedures. However, in contrast to Sanger (1975a), they found no staining of kinetochores and no distinct staining of the cleavage furrow, but they did find a concentration of heavy meromyosin-actin in the interzone at anaphase, a part of the spindle which showed no fluorescence in Sanger’s experiments. All these studies point to the existence of actin in the spindle. However, since for the heavy meromyosin-actin reaction and for the indirect fluorescent antibody technique the cells have to be glycerinated in order to allow the formation of the complex, critics argue either that during glycerination actin fibers are translocated within the cell and their occurrence in the spindle is an artifact, or that the glycerol or the heavy meromyosin treatment itself induces polymerization of actin with the same result of an artifactual distribution in the spindle. Indeed, surprisingly few electron microscopic reports exist which show thin filaments of 40 to 60 8, in the spindle (Bajer and Mole-Bajer, 1969; Miiller, 1972; McIntosh e f al., 1975; Blecher, 1975; Sanger and Sanger, 1975; Schroeder, 1976; Euteneuer et a f . , 1977). LaFountain (1974, 1975) and LaFountain and Zobel (1976) describe thin filaments in the contractile ring of Nephrotoma at cytokinesis but are unable to see similar filaments in the spindle. They argue that the preparation for the electron microscopy (fixation and staining) may destroy these filaments. This view has been reinforced by the recent work of Maupin-Szamier and Pollard ( 1 978) who demonstrated the destruction of actin filaments with osmium tetroxide. That actin occurs in the spindle is not surprising. In fact, it would be difficult for the spindle to avoid trapping the actin in whatever state it might be. The distribution of actin between the nucleus and the cytoplasm may serve as an

BIOCHEMISTRY OF THE MITOTIC SPINDLE

65

example for a similar situation. “Actin may associate with nuclei for the trivial reason that the nuclear envelope is no barrier to free movement of that protein between the two compartments” (Goldstein et al., 1975a,b). Since the compartmentalization of the spindle is certainly less developed than it is in the nucleus, the presence of actin in the spindle is even less surprising. In 1966 Sakai had already obtained evidence for actin in the isolated spindle. Only a functional test may determine whether actin is an essential part of the spindle. The first experiments reported so far are difficult to interpret. Neither antibodies to actin, nor DNase 1, which binds strongly to actin (Lazarides and Lindberg, 1974), nor phalloidin, which stabilizes actin efficiently (Wieland and Govindan, 1974), has any effects on chromosome movement in cell lysates. However, all the agents used do not inhibit the contraction of glycerinated muscle fibers (McIntosh, 1974), and, therefore, any conclusions would be premature now.

IV. Myosin A. MYOSININ CELLS Since actin has been described as possibly a part of the mitotic apparatus, a search for its counterpart myosin and other contractile proteins was undertaken. It is difficult if not impossible to demonstrate the presence of cytoplasmic myosin by ultrastructural analysis (cf. Schroeder, 1973). Myosin had been isolated from nonmuscle cells, e.g., platelets (Pollard et al., 1974; Niederman and Pollard, 1975), echinoderm gametes (Mabuchi, 1973, 1974, 1976a,b), mouse fibroblasts (Adelstein et al., 1972), brain tissue (Burridge and Bray, 1975), leukocytes (Stossel and Pollard, 1974), baby hamster kidney cells (Yerna et al., 1977), and intestinal epithelial cells of the brush border (Mooseker et al., 1977), and was shown to be different from muscle myosin. Pollard et al. (1977) even gave preliminary evidence that in Acantharnoeba two forms of myosin coexist, one resembling muscle myosin and the other cytoplasmic myosin. Weber and Groeschel-Stewart (1974) used an antibody to chicken gizzard myosin to stain specifically microfilament bundles in a variety of fibroblasts. Those microfilaments had been shown previously to bind antiactin (Lazarides and Weber, 1974). It was assumed, therefore, that the myosin distribution closely follows the actin distribution in the cell. Similar results were obtained by Otto et al. (1977), Fujiwara et al. (1977), and Fujiwara and Pollard (1976, 1978). B. MYOSINI N

THE

SPINDLE

Fujiwara and Pollard (1976) used a myosin-specific antibody against human platelet myosin coupled to fluorescent dyes to study the distribution of myosin in

66

CHRISTIAN PETZELT

the cytoplasm and the mitotic spindle of human cells. They found a diffuse staining of the spindle; the region between the chromosomes and the poles particularly showed a higher fluorescence. No fibrous structure was detectable. Cells in cytokinesis showed a clear staining of the cleavage furrow. These results were extended in a second paper in which the authors compared simultaneously the distribution of myosin and tubulin in a given cell using antibodies to these two proteins labeled with contrasting fluorochromes (Fujiwara and Pollard, 1978). In addition to the characteristic staining pattern of the spindle by antitubulin, antimyosin gave a diffuse staining of the region between the chromosomes and the poles similar to that described above. Although these results apparently demonstrate the presence of myosin in the spindle, additional evidence seems to be necessary to establish unambiguously the preferential localization of myosin in the spindle. As in the case of actin the mere presence of a protein in the spindle is just not enough to accept it as an integral part of this structure required for the structure and function of the mitotic spindle. The injection of a myosin antibody into the living cell provides a more direct approach to this problem. Mabuchi and Okuno (1977) injected antimyosin into starfish blastomeres. There was no influence on nuclear division although in some experiments the size of the spindle was somewhat reduced. Instead, cytokinesis was completely inhibited. These results were confirmed by Kiehart et al. (1977), who used the same system. Sakai et al. (1976) prepared a spindle isolate in which they described chromosome movements similar to the in vivo behavior. These movements could not be blocked by incubating the isolated spindles in antimyosin serum. Using the same antimyosin serum, Mohri et al. (1976) obtained in isolated spindles from sea urchin zygotes a weak staining at the poles indicating the presence of myosin there. Summarizing, one can conclude that myosin is most probably part of the contractile machinery which at cytokinesis cleaves one cell into two. Whether myosin is an integral part of the spindle structure or is required for spindle function remains to be proven.

V. Calcium-Dependent Regulator Protein A. CHARACTERISTICS A few years ago a low-molecular-weight, thermostable protein was described to activate brain cyclic nucleotide phosphodiesterase (Kakiuchi et al., 1970; Cheung, 1970, 1971; Cheung et al., 1975) and adenyl cyclase (Brostrom et al., 1975). Since then, the protein has been found in all tissues studied (Smoake et al., 1974; Kakiuchi et al., 1975; Waisman et al., 1975), and it was shown by Teo and Wang (1973) that the activation of the enzyme requires micromolar levels of calcium. Therefore, it was named “calcium-dependent regulator protein

BIOCHEMISTRY OF THE MITOTIC SPINDLE

67

(CDR protein). " Recently, Schulman and Greengard (1978) presented evidence that the CDR protein activates specific protein kinases in brain which phosphorylate membrane proteins. Several purification procedures for the protein have been published (Lin et al., 1974; Watterson et al., 1976; Dedman et al., 1977a). Dedman et al. (1978) determined the primary structure of the CDR protein and

FIG. 1. Comparison of HeLa cells during mitosis stained with antibodies to the calciumdependent regulator protein (left four panels) or with antibody to tubulin (right four panels). (From Andersen et al., 1978.)

68

CHRISTIAN PETZELT

found in the amino acid sequence a 50% direct homology with troponin-C. CDR protein from rat testis shows biological cross-reactivity to troponin-C from rabbit skeletal muscle (Dedman et al., 1977b). Although both troponin-C and CDR protein bind 4 moles of calcium per mole of protein, the two proteins have different binding sites for it (Dedman et al., 1977b). However, CDR protein can substitute for troponin-C in the Ca2+regulation of an in vitro actomyosin system (Amphlett et al., 1976). B. CALCIUM-DEPENDENT REGULATOR PROTEIN

IN THE

SPINDLE

Welsh et al. (1977, 1978) prepared antibodies to CDR protein from rat testis and used indirect immunofluorescence to study the distribution of this protein in a variety of vertebrate cells. In interphase, mainly the stress fibers are stained. At prometaphase, the protein begins to appear associated with the spindle and remains localized between the chromosomes and the poles until the end of anaphase. Then it is concentrated in two small regions, one on each side of the midbody. In no case was a staining of the cleavage furrow obtained. The authors speculate that, in analogy to troponin-C in the muscle system, the CDR protein is part of the Ca2+-regulatingsystem of the spindle. The presence of CDR protein in the spindle was confirmed by Andersen et al. (1978). Antibodies to CDR protein from bovine brain gave only a diffuse staining pattern in the interphase cell with the indirect immunofluorescence technique. In the mitotic spindle, however, the CDR protein is strongly concentrated in the two polar parts of each half spindle. As described above, at the end of mitosis the protein is localized in two small regions of the midbody. The staining pattern is distinct from that obtained with antitubulin (Fig. 1). It is difficult at present to assess the presence of CDR protein in the spindle. Too many functions have been attributed to it, like the regulation of Ca2+ions or the involvement in the metabolism of cyclic nucleotides. Exciting first results on its role in mitosis were recently presented by Brinkley et al. (1978). They showed that CDR protein sensitizes tubulin such that its polymerizability becomes susceptible to inhibition by micromolar concentrations of Ca2+ ions.

VI. Dynein A. CHARACTERISTICS Dynein represents a protein fraction with ATPase activities and was isolated first from Tetrahymena cilia (Gibbons and Rowe, 1965). Subsequently, it was localized at the arms of the external microtubule doublets (Gibbons, 1965). It is highly specific for ATP and is Mg2+ dependent (Gibbons and Fronk, 1972). It

BIOCHEMISTRY OF THE MITOTIC SPINDLE

69

has in its soluble form a sedimentation coefficient of 13s. Electrophoretically, it can be separated on SDS-gels into several high-molecular-weight proteins (Fronk et al., 1975). The dynein ATPase from sea urchin sperm flagella occurs in two isoenzymic forms (Ogawa and Gibbons, 1976). Attempts have been made recently to attribute the different proteins to different parts of the arms of the outer doublets (Baccetti et al., 1977). The function of dynein in the sliding of microtubules during flagella movement is by now well established (cf. Gibbons, 1975). B. DYNEINI N THE SPINDLE The identification of the dynein-microtubule system as the basis for beating of flagella and cilia makes it tempting to draw analogies to the spindle, a structure, where movement of chromosomes is somehow connected with microtubules. Dynein-like arms were postulated by McIntosh et al. (1969) when they formulated the sliding filament hypothesis for mitosis. Arm-like structures on spindle microtubules have been repeatedly described (e.g., McIntosh, 1974). For a critical discussion on the ultrastructural observations of arms and bridges in the mitotic spindle the reader is referred to Fuge (1978). Biochemically, the evidence for a functional role of dynein in the spindle is even more circumstantial. Weisenberg and Taylor (1968) isolated a protein fraction with ATPase activity from whole sea urchin eggs and isolated mitotic apparatus which had a sedimentation coefficient of 13S, similar to that of dynein. However, the possibility of cytoplasmic contaminations in the isolated spindle preparation could not be excluded. Ogawa (1973) prepared a tryptic fragment of dynein, fragment A. This peptide shows the same ATPase activity as dynein. Antibodies to it inhibit the ATPase activity (Ogawa and Mohri, 1975) and stop the movement of glycerinated or Triton-treated sperms (Okuno et al., 1976). If these antibodies are applied to sea urchin eggs in mitosis or isolated mitotic apparatus, the use of the indirect immunofluorescence technique results in a strong staining of the mitotic spindle (Mohri et al., 1976). Although these results can be taken as evidence for the presence of dynein in the spindle, the apparently preferential localization of the enzyme in the spindle might also be interpretable by the special composition of the cell used. The sea urchin egg is filled with yolk platelets and other metabolically inert material stored for the early embryogenesis. At mitosis all these optically dense elements are pushed away by the spindle and the cytoplasm becomes visible. Additionally, this cytoplasm contains a huge amount of proteins synthesized in advance for the ciliogenesis of the blastula. Dynein being an essential part of the ciliary structure is already present in the sea urchin egg. Therefore, the strong fluorescence of mitotic sea urchin eggs treated with antidynein may be a concomitant observation of a protein localization completely unrelated to mitosis. Sakai et al. (1976) have

70

CHRISTIAN PETZELT

described spindle isolates in which chromosomes reportedly continue to move. As mentioned earlier the authors were unable to block chromosome movement with antimyosin serum or other sera; only the serum containing antibodies to fragment A inhibited the movement. McIntosh et al. (1975) describe cell lysates from rat kangaroo cells in which the chromosomes are also still moving after lysis. They stop in a high-salt solution and can be restarted by the addition of a dynein extract from starfish sperm flagella. However, as stated by McIntosh (1977), these results are difficult to repeat and await further confirmation.

VII. The Mitotic Ca2+-ATPase A. THE ENZYMEI N

THE

CELLCYCLE

A calcium-activated ATPase activity was described in sea urchin eggs showing cyclic variations of activity during the cell cycle with one peak of activity at mitosis and the other peak in the first half of the interphase (Petzelt, 1972a). Mitotic apparatus from sea urchins isolated by a method which preserves enzymatic activities show enzymatic activity roughly three times higher than that of the surrounding cytoplasm (Mazia et al., 1972). The apparent connection of the enzyme with mitosis was further underlined by the finding that by changing the length of the cell cycle the course of the enzymatic activity is also altered (Petzelt, 1972b). Since the increase in enzymatic activity was assumed to be connected with the establishment or maintenance of the mitotic apparatus, a system was studied in which different spindle-like structures are found according to variable experimental conditions. During the parthenogenetic development of sea urchins either cytasters, monasters, or normal spindles are formed, and whenever such a structure appears, the enzymatic activity increases (Petzelt and von Ledebur-Villiger, 1973). As described by Rustad (1959), UV irradiation extends the cell cycle of fertilized sea urchin eggs, if they are irradiated at a certain time. If irradiated later in the cycle, not the upcoming mitosis but the second one is delayed. This UV-sensitive period is marked by an increase in the Ca2+ATPase activity although the enzyme itself appears to be UV-insensitive (Petzelt, 1974a). The Ca2+-ATPase was shown to occur in nearly all cycling cells studied. The course of the enzymatic activity showed the same cyclic increase of activity at mitosis as in the sea urchin eggs. This was demonstrated in mouse fibroblasts (Petzelt, 1974b), surf clam eggs (Rebhun, 1976), mouse mastocytoma cells (Petzelt and Auel, 19771, and, most recently, even in Physarum polycephalum (Petzelt et a / . , 1979). If only parts of the cell cycle are turned on, as can be done in unfertilized sea urchin eggs by treatment with ammonia (Mazia, 1974), the course of the Ca2+-ATPase follows closely the activated chromosome cycle (Petzelt, 1976). That the activity changes during the cell cycle (in mouse mas-

BIOCHEMISTRY OF THE MITOTIC SPINDLE

71

tocytoma cells) are activations of protein molecules already present and not the result of an increased synthesis was shown by Petzelt and Auel (1977). They presented evidence that the enzyme is synthesized in the first part of the interphase, around the beginning of the S-phase, in a period when the enzymatic activity is at its lowest.

B . CHARACTERISTICS Recently, the enzyme has been purified and many of its properties characterized (Auel and Petzelt, 1978; Petzelt and Auel, 1978). It is a membrane-bound protein and has in its native configuration a molecular weight of 260,000 (D. Auel, unpublished). It consists of several subunits, each being a dimer. The subunits have molecular weights of 54,000 and 55,000. It has been found in all cells capable of proliferation, ranging from slime molds, plants, and echinoderms to mammals. It is different from other ATPases in the cell since none of the known specific ATPase inhibitors shows any effect on the mitotic CaZ+-ATPase. We tested methylene blue [inhibits the Caz+-ATPase of the sarcoplasmic reticulum according to Yamada and Tonomura (1972) and Yu et al. (1974)], ruthenium red [also inhibits the Caz+-ATPase from sarcoplasmic reticulum and the Ca2+-ATPasefrom erythrocyte membranes according to Watson et al. (1971) and Vale and Arselio (1973)], oligomycin [inhibits the Ca2+ATPase of mitochondria according to Brierley et al. (1964)], ouabain [inhibits the Na+-K+-ATPase of plasma membranes according to Dahl and Hokin (1974)], and vanadate [inhibits the Na+-K+-ATPase according to Cantley et al. (1977) and is a potent inhibitor of dynein according to Gibbons et al. (1978)l. None of these agents had any inhibiting effect on the mitotic Ca2+-ATPase. Nath and Rebhun (1976a) describe an inhibition of the enzyme by caffeine and show that diamide inhibits the Ca2+-ATPaseboth in vivo and in vitro causing a concomitant mitotic block. The inhibition can be overcome by addition of dithiothreitol or mercaptoethanol (Rebhun et a l . , 1975, 1976; Nath and Rebhun, 1976b). These results demonstrate the functional importance of SH groups for the mitotic Ca2+-ATPase, and are confirmed by the fact that p-chloromercuribenzoate inhibits the enzymatic activity completely (Petzelt and Auel, 1978). C. FUNCTIONS The facts that, in all cells studied, the enzyme is most active at mitosis and is concentrated in the mitotic apparatus point to its involvement in the mitotic process. Soon after its discovery, the enzyme was thought to be connected with the membranous components known to occur in or around the mitotic spindle, and to possibly have a function analogous to that of the Ca2+ pump of the sarcoplasmic reticulum (Mazia et al., 1972; Petzelt and von Ledebur-Villiger,

72

CHRISTIAN PETZELT

1973). Membranes in the spindle have been pointed out by many authors (e.g., Porter and Machado, 1960; Robbins and Jentzsch, 1970; Friedlander and Wahrman, 1970; for review, see Fuge, 1978). Ito (1960), Hams (1962), Harris and Mazia (1962), and, using improved techniques, Harris (1975, 1976) and Hepler (1976a,b) described vesicles in the mitotic spindle and discussed their relation to the regulation of divalent cations at mitosis. Similar observations have been made by Longo (1976a,b) who found a close association of vesicles (dense aggregation of cisternae of smooth endoplasmic reticulum) and microtubules in the centrosphere of the sperm aster in rabbit zygotes. Ca2+ ions are the obvious candidates as regulating ions and a role for them in the formation of the spindle had been implied long ago by Heilbrunn (1921). Subsequently, they were always thought to be important, but not enough facts were known on possible sites of spindle components affected by Ca2+ ions. The situation changed somewhat when Weisenberg ( 1972) discovered that more than micromolar concentrations of Ca2+ ions had an inhibitory effect on microtubule polymerization in vitro. Rosenfeld et al. (1976) found a dependence of the Ca2+ sensitivity of tubulin on the Mg2+ concentration. Weisenburg’s results were in disagreement with those of Borisy and co-workers (Olmsted et al., 1974; Olmsted and Borisy, 1975) who found an insensitivity in tubulin polymerizability up to millimolar concentrations of Ca2+ions. An even lower sensitivity to Ca2+ions is reported by Doenges (1978) for the assembly of tubulin from ascites cells. Similar results on a differential Ca2+ sensitivity of the microtubule reassembly system depending on the use of crude brain extracts or purified microtubular proteins were described by Nishida and Sakai (1977). Schliwa (1976), using the ionophore A 23 187, found clear evidence for the regulation of microtubule assembly in the heliozoan axopodium by Ca2+ions. In the same system he also obtained evidence that disruption of microtubule links by Ca2+ ions precedes the depolymerization of microtubules (Schliwa, 1977). The recent discovery of the calciumsensitizing effect of the calcium-dependent regulator protein on tubulin (Brinkley et al., 1978) and its preferential localization in the mitotic spindle (Welsh et al., 1978; Andersen et al., 1978) may have solved the dispute on the differential sensitivity of tubulin to Ca2+ ions. The elegant experiments by Kiehart and Inoue (1975, 1976) provide probably the most convincing support for the idea that Ca2+ ions regulate the integrity of the spindle and that a Ca2+control system exists at the site of the spindle. They microinjected Ca2+ions (up to 10 mM) onto a restricted area of the spindle in an echinoderm egg and thereby caused a local dissolution of the spindle seen as a reduction in the birefringence of spindle fibers. The birefringence of the rest of the spindle remained unaltered. This effect of calcium is reversed in a matter of minutes and the normal birefringence is restored. If Ca2+ ions are injected not onto the spindle but into the cytoplasm, the morphology of the spindle and the

BIOCHEMISTRY OF THE MITOTIC SPINDLE

73

course of mitosis are not affected. Injections of Mg2+ions or distilled water also have no effect on the spindle. The sensitivity of the mitotic spindle to Ca2+ ions has also been demonstrated by the recent development of gentle spindle isolation methods (for details, see Section IX). It was shown that micromolar quantities of Ca2+ions will cause the spindle to lose most of its birefringence (Goode and Roth, 1969; Smith and Rebhun, 1974; Cande et al., 1974; Rebhun et al., 1974; Salmon and Jenkins, 1977). Therefore, most of the present isolation procedures now include EGTA to chelate Ca2+ ions. How and where in the cell is the concentration of Ca2+ions regulated? If we look at the muscle system, the Ca2+-regulating pump is located in specialized vesicles of the endoplasmic reticulum. It is a Ca2+-activated ATPase and has been well characterized (cf. Hasselbach, 1972). One can assume afunctional resemblance of the mitotic CaZ+-ATPaseto the enzyme although biochemically the two enzymes are quite different (e.g., molecular weight, sensitivity to inhibitors). The Ca2+-ATPaseof the sarcoplasmic reticulum starts to appear in the later stages of myotube formation (Jorgensen et a l . , 1976). The mitotic Ca2+-ATPase is thought to be part of the vesicles described in the spindle (see above). Vesicles active in Ca2+transport have been described by Kinoshita and Yazaki (1967) in sea urchin eggs. These vesicles are still able to accumulate calcium and seem to be localized in the asters and around the central spindle. Preliminary evidence by Wick and Hepler (1976) shows by antimonate staining of dividing cells of Marsilea and Hordeum that Ca2+ ions are localized in the mitochondria, the nuclear envelope, and the endoplasmic reticulum. Kato and Tonomura (1977) also found an uptake of calcium into vesicles isolated from Physarum. This Ca2+-regulatingmembrane system is like the mitotic Ca2+-ATPase;it is caffeine sensitive and thought to be involved in protoplasmic motility (Matthews, 1977). Also similar to the mitotic Ca2+-ATPase is the vesicular ATP-dependent Ca2+uptake system described in plants by Gross and M m e (1978) and in mouse fibroblasts by Moore and Pastan (1977). Gallin and Rosenthal (1974) exposed granulocytes to chemotactic stimuli thereby causing Ca2+ release, decreased Ca2+ uptake, and an associated shift of cellular calcium from cytoplasmic to granular fractions. In a preliminary report Baugh et al. (1976) described a membrane-bound Ca2+-ATPaseat the base of the cilia of Tetrahymena which is supposed to be involved in the Ca2+ regulation in the cell. By no means is it clear, however, that the only role of Ca2+ ions for the mitotic spindle is their regulation of microtubule assembly-disassembly. Our failure to answer this question may just be a sign of our ignorance of other processes in the mitotic spindle which are also regulated by Ca2+ions but about which we know nothing. There are reports in the literature on interactions of membranes with tubulin or microtubules. Smith et al. (1970) describe a close relationship between vesicles

74

CHRISTIAN PETZELT

and microtubules in the axons of the spinal cord of Petromyzon marinus. A linkage by bridges between microtubules and membrane-bound vesicles has been found by Allen (1975). Bhattacharya and Wolff (1976) describe a membraneattached tubulin which copolymerizes with brain tubulin only after solubilization with detergents. Similar reports had been published earlier by Stadler and Franke (1972, 1974) on colchicine-binding activities of membrane fractions from rat and mouse liver. Cordier (1975, 1976) found a connection between the ciliary rootlets of the basal body and the smooth endoplasmic reticulum and vesicles of thymic cysts. Paulson and McClure (1975) found a close proximity of vesicles and microtubules in nerves and discussed their relationship to the fast axoplasmic transport. Thyberg and Hinek (1977) and Hinek et al. (1977) describe in vitro effects of antimicrotubular agents on the Golgi complex of embryonic chick spinal ganglion cells. If one summarizes all the facts described so far, it is evident that a Ca2+regulating system exists in most cells which seems to be localized in vesicles. These vesicles are found in or around the mitotic spindle (cf. Rebhun, 1977). The mitotic Ca2+-ATPase appears to function there as part of the Ca2+ pump. The prime candidate for a regulation by Ca2+ions is the microtubule assembly system (Sawada and Rebhun, 1969) although other Ca2+ targets may be important at mitosis (e.g., Rubin et al., 1978).

VIII. Calcium in the Mitotic Cell Fluctuations in the calcium distribution in the cell at mitosis have been analyzed by Timourian et al. (1974) using electron microprobe analysis. They found a higher concentration of total calcium in the spindle region than in the surrounding cytoplasm. Recently, a highly sensitive proton microprobe has been described (Bosch et a l . , 1978). Observations on the distribution of calcium in the mitotic sea urchin egg show a preferential concentration of it in the spindle (C. Petzelt and K. Traxel, unpublished). However, up to now these studies have had to be done on fixed cells. Work is in progress to determine the distribution of free and bound calcium using this new technique and preparing cells in a more native state by shock-freezing. Ridgway and Durham (1976) and Ridgway et al. (1976) injected aequorin into Medaka eggs. Aequorin is a photoprotein which emits light in response to changes in free calcium concentration. Aequorininjected eggs show a strong increase in luminescence at activation and a weaker one at the first and second mitosis indicating a rise in the Ca2+concentration in the cell. Whereas the increase in free calcium at activation, respectively, at fertilization, was described long ago by Mazia (1937) and recently confirmed by Johnston and Paul (1977) and Paul and Johnston (1978), the Ca2+ increase at

BIOCHEMISTRY OF THE MITOTIC SPINDLE

75

mitosis was worked out only recently. A cyclic increase in Ca2+uptake at mitosis was demonstrated for Physarum polycephalum by Holmes and Stewart (1977). For this organism Matthews (1977) had shown the existence of a vesicular Ca2+ pump. Similar results on Ca2+ uptake had been obtained earlier by Clothier and Timourian (1972) for the sea urchin egg at mitosis for which they demonstrated a cyclic increase in the rate of Ca2+uptake.

IX. The Isolation of the Mitotic Spindle A. ISOLATIONPROCEDURES Ultimately, there is only one criterion for the successful isolation of the mitotic spindle, that is, it has to retain its functionality and allow the process of mitosis to occur in vitro. Since we have not reached this point yet, other criteria have to be applied to judge the different isolation procedures. Historically, the first mass isolation of mitotic apparatus was performed by Mazia and Dan (1952). They were confronted with the two now classic difficulties one encounters at isolation. Being a labile and transient structure the mitotic apparatus has to be stabilized (reversibly if possible) and then separated from the rest of the cell. Mazia and Dan stabilized the mitotic apparatus by immersing sea urchin eggs at metaphase in cold 30% ethanol. Thereafter, the cells were dispersed and the mitotic apparatus set free by detergents. In principle, the alcohol-detergent method is still in use today (Mazia et al., 1972) since it gives excellent preservation of proteins. However, the ultrastructure of the spindle cannot be compared with that of the spindle in cells fixed as a whole. Additionally, the birefringence of the spindle and its sensitivity to cold, KCI, and calcium solubilization are lost. This means that a permanent stabilization, almost a fixation, has been obtained in cold ethanol. Mazia et al. (1961a) then introduced dithiodiglycol as a new isolation medium. It gave a more reversible stabilization and good preservation of the ultrastructure. It was especially possible now to isolate the mitotic apparatus directly from the living cell without prior stabilization (Mazia et al. 1961b). Kane (1962), starting from the dithiodiglycol procedure, generalized the method by showing that many glycols with longer chains could substitute for dithiodiglycol and subsequently described a variety of conditions to obtain pure spindles in high yield (Kane, 1965). He observed the importance of pH, the isolation with the glycols being effective only at slightly acid pH. Kane and Forer (1965) studied changes of the sea urchin spindle in birefringence and solubility after isolation with Kane 's standard procedure using 1 M hexylene glycol (2-methyl 2,4-pentanediol), 0.01 M KH,P04, pH 6.4. They found a rapid loss of solubility after isolation correlated with a decrease in the birefringence of the mitotic apparatus. Parallel to these phenomena the microtubule-like structures in the spindle disappeared.

76

CHRISTIAN PETZELT

Kane (1967) and Stephens (1967) described in detail many properties of the hexylene glycol-isolated mitotic apparatus and gave a first analysis of the proteins solubilized by 0.6 M KCl treatment. By analytical ultracentrifugation the solubilized proteins were separated into two groups, one with a sedimentation constant of 22 S and the other, 4 to 6 S. In later studies it was shown that the 22 S protein was not a spindle-specific protein (Bibring and Baxandall, 1969). Rebhun and Sander (1967) and Goldman and Rebhun (1969) correlated the birefringence of the isolated mitotic apparatus to the amount of microtubules in the spindle although they assumed that small ribosome-like particles also contribute to the form birefringence. Hartmann and Zimmerman (1968) prepared ribosomes and nucleoproteins from hexylene glycol-isolated sea urchin spindles. Using interference microscopy, Forer and Goldman (1969) studied the preservation of dry matter when mitotic apparatus are isolated. They found that the solubility, the total amount of dry matter, and the birefringence of isolated mitotic apparatus all depend on the pH of the isolation medium and that the specific pH values of the properties in question are not necessarily the same for different species. Forer and Goldman ( 1 972) demonstrated that use of the hexylene glycol isolation medium causes the mitotic apparatus to loose up to 80% of the dry mass of the in vivo concentration, depending on the pH of the isolation medium. Subsequently, efforts were made to extract the microtubular protein from the mitotic apparatus. Bibring and Baxandall(l968) isolated the mitotic apparatus with hexylene glycol and treated it with mild (pH 3) hydrochloric acid. They found that microtubules disappeared upon this treatment. Later (Bibring and Baxandall, 1971), they used meralluride sodium, an organic mercurial, for extraction and obtained morphologically a breakdown of the microtubules in the isolated spindle. The extracted protein was identified as tubulin on the basis of its precipitation by calcium and vinblastine and by its electrophoretic mobility. An antiserum against a preparation of sperm tail outer doublet microtubules cross-reacts with the extract from mitotic apparatus. The by now widely used hexylene glycol procedure was applied to a variety of species. Bryan and Sato (1970) isolated meiotic spindles from the starfish Pisaster ochraceus. Sisken et al. (1967) and Chu and Sisken (1977) adapted the method to isolate spindles from mammalian cells by omitting the phosphate buffer and adding 1 x M Ca2+to the isolation medium to stabilize the spindle. Wray and Stubblefield (1970) and Wray ( 1 973) increased in a mammalian cell culture (naturally asynchronous) the yield of mitotic cells by pretreatment with colcemide and used 1 M hexylene glycol, 5 x M Ca2+,5 x M Pipes buffer, pH 6.5, for isolation. Milsted and Cohen (1973) isolated spindles from Drosophia melanogaster embryos and Muller (1972) also used hexylene glycol to isolate meiotic spindles from the crane fly Pales ferruginea. Several criteria for a good isolation procedure are fulfilled by Kane’s method: It lyses all cells very efficiently, invertebrate cells (mostly echinoderm eggs) as well as mammalian cells. The preservation of the birefringence and of the ultrastructure is compar-

BIOCHEMISTRY OF THE MITOTIC SPINDLE

77

able to that in the unlysed cell. The mitotic apparatus are soluble in high-salt solutions. However, the method also has some drawbacks. The enzymatic activity of the mitotic Ca”-ATPase is inhibited (Petzelt, 1972a). The sensitivity of the isolated spindles to cold, calcium, or pressure treatment is lost. The birefringence decays. The spindles are insensitive to tubulin incorporation. In a pioneering study Goode and Roth (1969) isolated the mitotic apparatus from the giant amoeba Chaos carolinensis using only a dilute buffer. They obtained a spindle which reacted to changes in Ca2+ and K+, the ionic strength, and the pH. The authors even observed an elongation of the isolated spindle similar to the anaphase movement in vivo. Great progress was made in 1974 when several laboratories published new isolation procedures. Forer and Zimmerman ( 1974) introduced a glyceroldimethylsulfoxide medium which allowed the isolation of spindles while retaining many of their in vivo properties. Pressure treatment of the spindles reduced their birefringence (Forer and Zimmerman, 1976a). They were susceptible to cold and stayed soluble for weeks (Forer and Zimmerman, 1976b). By transferring hexylene glycol-isolated spindles immediately after isolation to a glyceroldimethylsulfoxide medium, their birefringence and sensitivity against 0.6 M KCl treatment could be preserved (Forer et al. , 1976). Since by KCl treatment the authors obtain a 45% loss of the birefringence, with the microtubules apparently remaining intact, they asume that a nontubulin component (called substance “ y ” ) which is extracted by the KCl contributes 45% of the spindle birefringence. Sakai and Kuriyama (1974) used 1 M glycerol, 1 mM EGTA, 5 mM Mes, pH 6.15, for spindle isolation. This is essentially a microtubule-polymerizing solution (cf. Section 11). Indeed, the isolated spindles were extractable by calcium or-in the presence of GTP-by cold and the extracted protein proved to be tubulin which had even retained its colchicine-binding properties. An important contribution was made when Rebhun et al. (1974) described the isolation of mitotic apparatus of surf clam eggs which had retained much of their in vivo characteristics. The authors used 0.1 M Mes or Pipes buffer, pH 6.85,0.25-1 mM MgC12, 1-5 mM EGTA, a 10 mM concentration of a proteolytic enzyme inhibitor p-tosyl arginine methylester HCl (TAME), 0.2-1% Triton X-100 for isolation. The isolated spindles can incorporate heterologous tubulin (from chick brain) and, after removal of homologous tubulin, can assemble heterologous tubulin into birefringent fibers similar in distribution to those of spindles in living cells. However, the incorporation of tubulin achieves only an increase in length; the spindles never shorten. Or in the terms of tubulin chemistry, the TAME spindles take part in the microtubule assembly as in the living cell; however, disassembly fails to occur. Rebhun’s method was also used by Milsted et al, (1977) to isolate spindles from Drosophila embryos. They observed the isolated spindles with the scanning electron microscope after they had glued them with polylysine onto small glass or plastic pieces as described by Mazia et al. (1975).

78

CHRISTIAN PETZELT

B. SPINDLEMODELS From 1974 on, the ability of the isolated spindle to interact with exogenous tubulin was taken as one of the main criteria for the quality of an isolation procedure. Cande et al. (1974) lysed mitotic rat kangaroo cells individually with a nonionic detergent into tubulin assembly buffers (0.1 M Pipes, 1 mM EGTA, 0.1 mM GTP) containing various amounts of brain tubulin and obtained a spindle which lost and gained birefringence when cooled and warmed, respectively. Additionally the authors reported that cells lysed in early anaphase show continuous chromosome movement to the poles. A difficulty becomes apparent here. Are those spindles already isolated or do we still look at a whole cell being able to move chromosomes despite the wounds inflicted? The range of the effective Triton concentration is so small that a decision between observation of a “spindle in vitro ’’ and a living cell is not always unambiguously possible. InouC et al. (1974) have used oocytes of the marine worm Chaetopterus to break cells in hypotonic medium and to study the growth and ability of metaphase spindles in varying concentrations of porcine neurotubulin. The spindles will not only increase their birefringence but, if the concentration of the exogenous tubulin is high enough, they will even increase in length. The spindles are similar to the spindles described by Cande et al. (1974), Ca2+ sensitive. Colchicine does not show any effect. Sakai et al. (1975, 1976) developed an isolation medium for mitotic apparatus from echinoderm eggs which may give the most lifelike spindles as yet. The medium was composed of a mixture of glycerol, EGTA, Ca2+, Mg2+, GTP, CAMP, ascorbic acid, glutathione, Mes buffer, and heterologous tubulin. A chromosome-movement-inducing solution consisting of sucrose or mannitol, K-acetate, EGTA, Ca2+,Mg2+, GTP, ATP, Mes buffer, and heterologous tubulin must be added after isolation. Chromosome motion was followed by taking photographs using mostly the phase microscope and measuring the distance to the poles. The authors describe anaphase-like spindle elongation and movement of chromosomes to the poles; the rate of motion, however, is about one-tenth of the in vivo rate. The movement can be blocked by the addition of an antibody to fragment A of flagellar dynein; antimyosin does not show any effect. Colchicine suppresses completely the chromosome motion in vitro. The authors conclude that a microtubule-dynein system, but not a myosin-actin system, is involved in chromosome motion. It does not seem to be clear yet if the interpretation of Sakai el al. can be unambiguously accepted in toto. There is confirmatory evidence that actin-myosin may not be a functional part of the spindle (cf. Sections I11 and IV). That microtubules are involved in chromosome movement is a generally accepted fact. However, the effect of antidynein serum on the chromosome movement could also be explained by a simple precipitation of the dynein in the spindle by the antibody, the resulting complex interfering with the

BIOCHEMISTRY OF THE MITOTIC SPINDLE

79

spindle’s functional structure. In the section on dynein we saw that the presence of dynein in a spindle deriving from an egg cell is no surprise. Further experiments, e.g., on other spindles or by using specific inhibitors, are needed to obtain detailed information on this fascinating model.

X. Mitotic Centers The spatial arrangement of the components of the mitotic spindle only recently became open for experimental studies. Generally it is assumed that all the components for the mitotic apparatus are synthesized before spindle formation (Wilt et al., 1967) although a low level of protein synthesis can be demonstrated in the mitotic cell (Prescott and Bender, 1962; Bibring and Cousineau, 1964; Parchman and Stem, 1969; Petzelt, 1970). The cell must be able to organize all the necessary components into the mitotic spindle. This process is being studied on the most prominent spindle component, the microtubules. Porter (1966) postulated the existence of a complex of tubule-initiating sites where microtubules assembly could occur and where microtubules could be oriented. Pickett-Heaps (1969, 1974) defined these entities as microtubule-organizing centers (MTOC) and their ultrastructure has been described in detail (cf. Pickett-Heaps, 1969; Tucker, 1977; Fuge, 1978). In higher eukaryotes the kinetochores and-if present-the centriolar complex are the principal structures involved in the formation of the mitotic spindle. Tubulin will polymerize in vitro, forming randomly distributed microtubules. Only if structures with an orienting capacity are present, will microtubules show a directionality of growth (Snell et al., 1974). These structures are the kinetochores, the centriolar complex and, to a certain extent, even the isolated spindle (Fig. 2). The centriolar complex consists of a pair of centrioles with electrondense pericentriolar material round it. Upon isolation and incubation in purified microtubule protein it shows preferential assembly of microtubules (Weisenberg and Rosenfeld, 1975; McGill and Brinkley, 1975; Binder et al., 1975; Steams et al., 1976; Snyder and McIntosh, 1975, 1976b; Gould and Borisy, 1976; Borisy and Gould, 1977). The microtubule-organizing capacity of kinetochores was demonstrated by Telzer et al. (1975), McGill and Brinkley (1975), and Borisy and Gould (1977). Chromosomes were isolated from cell lysates and incubated with heterologous tubulin under polymerizing conditions. Microtubules were found to assemble specifically at the kinetochores, confirming that these sites serve to nucleate microtubles. Not very much is known of the biochemistry of the centriolar complex. From the ultrastructural analysis of the centriole the presence of microtubules is evident. However, their reaction to colchicine and other tubulin poisons shows such drastic differences in sensitivity that a special composition of the centriolar microtubules is highly probable (cf. Dustin, 1978). A Mg2+-activated ATPase

80

CHRISTIAN PETZELT

FIG. 2. Isolated spindle of an L-cell (mouse fibroblast) in late anaphase. Upon incubation in porcine brain tubulin, microtubules will polymerize onto the spindle if the mitotic Ca2+-ATPase is present in the solution (cf. Petzelt, 1974b).

has been demonstrated by Abel et al. ( 1972). Anderson and Floyd ( 1977) gave a preliminary report on an ATPase in isolated basal bodies which shows properties similar to those of the mitotic Ca2+-ATPase. RNA has been shown to be present in basal bodies (Hartman et a l . , 1974; Dipple, 1976), structures which can be compared with centrioles or which can even be changed into them as the following experiments have shown. Heidemann and Kirschner (1975) and Heidemann et al. (1977) isolated basal bodies from Tetrahymena and Chlamydomonas and injected them into Xenopus oocytes. In those eggs which had been activated before injection such that the nuclear envelope had already disappeared many asters formed about 1 hour after injection. RNase treatment of the basal bodies after isolation but before injection completely inhibited the capacity of the egg to form asters after basal body injection. Several conclusions can be drawn from these experiments. A basal body can form a mitotic center but only in a competent cell (when nonactivated eggs were injected, no asters were formed upon the

BIOCHEMISTRY OF THE MITOTIC SPINDLE

81

injection of the basal bodies). It takes about 1 hour until the asters appear; the mitotic center has to undergo a maturation process (Maller et a l . , 1976). Weisenberg and Rosenfeld (1 975) observed this first on lysates of Spisula eggs. For the function of the mitotic center an RNA specific for centrioles is a necessary component. This has been confirmed by Zackroff et al. (1975). They showed that after RNase treatment mitotic centers from Spisula eggs initiate smaller asters than untreated centers. Additional support for the presence of RNA in the centriolar complex comes from the work of Berns et al. (1977) who extended considerably our knowledge of the essential components of the centriolar complex (Berns et a l . , 1977; Peterson and Berns, 1977; Berns and Meredith, 1977). They treated rat kangaroo (PtK,) cells with nucleic acid intercalating and photoreacting drugs, acridine orange, or a psoralen (4’-aminomethyl-4,5’ ,8-trimethyl-psoralen) and irradiated them with an argon laser microbeam. Only if the laser damages the pericentriolar material around the centrioles, can an effect be seen such that after the formation of a normal looking metaphase no chromosome separation and no anaphase movement occur. The centrioles appear undamaged by the irradiation. It was concluded, then, that not the centrioles but the pericentriolar material is the causative agent for the capacity of the centriolar complex to function as mitotic center. This view is supported by the results from Brenner et al. (1977). They observed in tetraploid PtK, cells meiosis-like reduction divisions where two spindles separated the chromosomes to four poles. In the electron microscope these cells revealed two duplices of centrioles (one at each of the two spindle poles); the other two spindle poles did not have centrioles but the microtubules of the spindle terminated in the pericentriolar material. The authors assume, therefore, that “the centriole is not essential for spindle pole formation and division” and that the pericentriolar material is a necessary component of the spindle apparatus. The above-mentioned experiments by Berns et al. give the first hints on the biochemical nature of the pericentriolar material insofar as RNA seems to be involved.

XI. Concluding Remarks Any survey of such a diverse and fast-moving field as biochemistry of the mitotic spindle should probably give an impression of how firm the land upon which we presently are is and of where and how far one will have to go in the future. There is no reason whatsoever for pessimism or even despair that mitosis is too complicated to study and that the molecular background of it will elude US for many more years to come. On the contrary, if one looks at the last decade and the nearly incredible rate of growth of our knowledge on the composition of the spindle, one should expect some exciting progress in the next few years. There are groups working on spindle models who should be able to define in a more

82

CHRISTIAN PETZELT

precise way the necessary ingredients in the brew in which chromosome move in the isolated spindle. There are the purists, biochemists working on isolated and purified proteins (deriving not always from the spindle but sometimes thought into it); most of them realize that components which are very conspicuous are not always the most important ones. And for the student of biochemistry of the mitotic spindle, there is a new field, the analysis of mutants in mitosis. Great progress has been made on the genetic side; many mutants have been isolated in yeast (cf. Hartwell, 1978) as well as in mammalian cells (cf. Wang, 1974). It is now up to the biochemist to make use of this promising material.

ACKNOWLEDGMENTS I am grateful to Dr. M. Osbom and Dr. K. Weber for providing Fig. 1, N. Sautter for continuous encouragement, and U. Joa for her secretarial help. Part of my work was supported by the Deutsche Forschungsgemeinschaft Grant Pe 164/1-6.

REFERENCES Abel, H. H., Jokelainen, P. T., and Verhage, H. G. (1972). Cytobiologie 6, 468. Adelstein, R. S . , Conbe, M. A,, Johnson, G., Pastan, I., and Pollard, T. (19721. Proc. Natl. Acad. Sci. U.S.A. 69, 3693. Allen, R. D. (1975). J. Cell Biol. 69, 497. Amphlett, G. W., Vanaman, T. C., and Perry, S. V. (1976). FEES Lerr. 72, 163. Andersen, B., Osbom, M., and Weber, K. (1978). Cyfobiologie 17, 354. Anderson, R. G. W., and Floyd, A. K. (1977). J. Cell Biol. 75, 273a. Aronson, J. F. (1965). J . Cell Biol. 26, 293. Auclair, W., and Siegel, B. W. (1966). Science 154, 913. Auel, D., and Petzelt, C. (1978). Cyfobiologie 18, 132. Baccetti, B., Bunini, A. G., Dallai, R., and Pallini, V. (1977). J. Cell Biol. 75, 300a. Bajer, A. S., and Mole-Bajer, J. (1969). Chromosoma 27, 448. Bamburg, J. R., Shooter, E. M., and Wilson, L. (1973). Biochemistry 12, 1476. Bardele, C. F. (1973). Cytobiologie 7, 442. Barham, S. S . , and Brinkley, B. R. (1976a). Cyrobios 15, 85. Barham, S. S., and Brinkley, B. R. (1976b). Cyrobios 15, 97. Barnes, L. D., Engel, A. G., and Dousa, T. P. (1975). Biochim. Biophys. Acra 405, 422. Baugh, L. C., Satir, P., and Satir, B. (1976). J. Cell Biol. 70, 66a. Behnke, 0. (1970). Inr. Rev. Exp. Parhol. 9, 1. Behnke, O., and Forer, A. (1967). J. Cell Sci. 2, 169. Behnke, O., Forer, A., and Emmersen, J. (1971). Nafure (London) 238, 408. Berns, M. W., and Meredith, S. (1977). J. Cell Biol. 75, 282a. Berns, M. W., Rattner, J. B., Brenner, S., and Meredith, S. (1977). J. Cell Biol. 72, 351. Berry, R. W . , and Shelanski, M. L. (1972). J. Mol. Biol. 71, 71. Bettex-Galland, M., and Liischer, E. F. (1959). Narure (London) 184, 276. Bhattacharya, B., and Wolff, J. (1976). Nature (London) 264, 576.

BIOCHEMISTRY OF THE MITOTIC SPINDLE

83

Bibring, T., and Baxandall, J. (1968). Science 161, 377. Bibring, T., and Baxandall, J. (1969). J . Cell Biol. 41, 577. Bibring, T., and Baxandall, J. (1971). J. Cell Biol. 48, 324. Bibring, T., and Cousineau, G. H. (1964). Nurure (London) 209, 805. Bibring, T., and Baxandall, J. (1977). Dev.Biol. 55, 191. Bibring, T., Baxandall, J., Denslow, S., and Walker, B. (1976). J. Cell Biol. 69, 301. Binder, L. I., Dentler, W. L., and Rosenbaum, J. L. (1975). Proc. Narl. Acad. Sci. U.S.A. 72, 1122. Blecher, S. R. (1975). Cyrobiologie 11, 190. Borisy, G. G., and Gould, R. R. (1977). In “Mitosis, Facts and Questions” (M. Little erul., eds.), p. 78. Springer-Verlag, Berlin and New York. Borisy, G. G., and Olmsted, J. B. (1972). Science 177, 1196. Borisy, G. G., and Taylor, E. W. (1967a). J. Cell Biol. 34, 525. Borisy, G. G., and Taylor, E. W. (1967b). J. Cell Biol. 34, 535. Borisy, G. G., Olmsted, J. B., Marcum, J. M., and Allen, C. (1974). Fed. Proc., Fed. Am. SOC. Exp. Biol. 33, 167. Borisy, G. G., Marcum, J. M., Olmsted, J. B., Murphy, D. B., and Johnson, K. A. (1975). Ann. N . Y . Acad. Sci. 253, 107. Bosch, F., El Goresy, A., Martin, B., Pouh, B., Nobiling, R., Schwalm, D., and Traxel, K.(1978). Science 199, 765. Boxer, L. A. and Stossel, T. P. (1976). J. Clin. Invesr. 57, 964. Bray, D., and Thomas, C. (1975). Biochem. J. 147, 221. Brenner, S., Branch, A., Meredith, S., and Bems, M. W. (1977). J. Cell Biol. 72, 368. Brierley, G., Murer, E., and Bachmann, E. (1964). J. Biol. Chem. 239, 2706. Brinkley, B. R., and Cartwright, J., Jr. (1970). J. Cell Biol. 57, 25a. Brinkley, B. R., Barham, S. S., Bmanco, S. C., and Fuller, G. M. (1974). Exp. Cell Res. 85,41. Brinkley, B. R., Marcum, J. M., Welsh, M. J., Dedmon, J. R., and Means, A. R. (1978). J . Supramol. Srrucr., Suppl 2, 290. Brostrom, C. 0..Huang, Y. C., Breckenridge, B. M., and Wolff, D. J. (1975). Proc. Nurl. Acad. Sci. U S . A . 72, 64. Bryan, J. (1971). Exp. Cell Res. 66, 129. Bryan, J. (1972a). J . Mol. Biol. 66, 157. Bryan, J. (1972b). Biochemistry 11, 261 1. Bryan, J. (1974a). Fed. Proc., Fed. Am. SOC.Exp. Biol. 33, 152. Bryan, J., and Kane, R. E. (1977). J . Cell Biol. 75, 268a. Bryan, J . , and Sato, H. (1970). Exp. Cell Res. 59, 371. Bryan, J., and Wilson, L. (1971). Proc. Narl. Acad. Sci. U.S.A. 68, 1762. Bryan, J., Nagle, B. W., and Doenges, K. H. (1975). Proc. Narl. Acad. Sci. U.S.A. 72, 3570. Burnside, B., Kozak, C., and Kafatos, F. C. (1973). J . Cell Biol. 59, 755. Bunidge, K., and Bray, D. (1975). J . Mol. Biol. 99, 1. Cande, W. Z., Snyder, J., Smith, D., Summers, K.,and McIntosh, J. R. (1974). Proc. Narl. Acad. Sci. U.S.A. 71, 1559. Cande, W. Z., Lazarides, E., and McIntosh, J. R. (1977). J. Cell Biol. 72, 552. Cantley, L. C., Josephson, L., and Warner, R. (1977). J. Biol. Chem. 252, 7421. Cheung, W. Y. (1970). Biochem. Biophys. Res. Commun. 33, 533. Cheung, W. Y . (1971). J. Biol. Chem. 246, 2859. Cheung, W. Y., Bradham, L. S., Lynch, T. J., Tin, Y. M., and Tallant, E. A. (1975). Biochem. Biophys. Res. Commun. 66, 1055. Chu, L. K., and Sisken, J. E. (1977). Exp. Cell Res. 107, 71. Clarke, M., and Spudich, J. A. (1977). Annu. Rev. Biochem. 46, 797.

84

CHRISTIAN PETZELT

Cleveland, D. W., Hwo, S. Y., and Kirschner, M. W. (1977a). J. Mol. B i d . 116, 207. Cleveland, D. W., Hwo, S. Y., and Kirschner, M. W. (1977b). J. Mol. B i d . 116, 227. Cleveland, D. W., Fischer, S., Kirschner, M. W., and Laemmli, K. (1977~).J. B i d . Chem. 252, 1102. Clothier, F., and Timourian, H. (1972). Exp. Cell Res. 76, 105. Cognetti, G., DiLiegro, T., and Cavarretta, F. (1977). Cell Difer. 6, 159. Cohen, W. D., and Rebhun, L. I. (1970). J. Cell Sci. 6 , 159. Condeelis, J. S., and Taylor, D. L. (1971). J . Cell Biol. 74, 901. Connolly, J. A,, Kalnins, V. I., Cleveland, D. W., and Kirschner, M. W. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 2437. Connolly, J. A., Kalnins, V. I., Cleveland, D. W., and Kirschner, M. W. (1978). J. Cell B i d . 76, 781. Cordier, A. C. (1975). Anar. Rec. 181, 227. Cordier, A. C. (1976). Cell Tissue Res. 166, 315. Dahl, J . L., and Hokin, L. E. (1974). Annu. Rev. Biochem. 43, 327. Dales, S. (1972). J . Cell Biol. 52, 748. Dales, S., Hsu, K., and Nagayama, A. (1973). J . Cell Biol. 59, 643. Davies, P., Wallach, D., Willingham, M., and Pastan, I. (1977). J . Cell B i d . 75, 263a. DeBrabander, M., DeMey, J., Joniau, M., and Geuens, G. (1977). Cell Biol. Int. Rep. 1, 177. Dedman, J. R., Potter, J. D., and Means, A. R. (1977a). J. Biol. Chem. 252, 2437. Dedman, J. R., Potter, J. D. Jackson, R. L., Johnson, J . D., and Means, A. R. (1977b). J . B i d . Chem. 252, 8415. Dedman, J. R., Jackson, R. L., Schreiber, W. E., and Means, A. R. (1978). J. B i d . Chem. 253, 343. Dentler, W. L., Granett, S., and Rosenbaum, J. L. (1975). J. Cell Biol. 65, 237. Dipple, R. V. (1976). J. Cell Biol. 69, 622. Doenges, K . H. (1978). FEBSLert. 89, 157. Doenges, K. H., Nagle, B. W., Uhlmann, A,, and Bryan, J. (1977). Biochemisrry 16, 3455. Dustin, P. (1978). “Microtubules. ” Springer-Verlag, Berlin and New York. Dustin, P., Flament-Durand, J., and Hubert, J. (1976). J. Cell B i d . 70, 41a. Eigsti, 0. J., and Dustin, P., Jr. (1955). “Colchicine, in Agriculture, Medicine, Biology and Chemistry.” Iowa State Coll. Press, Ames. Eipper, B. A. (1972). Proc. Nurl. Acad. Sci. U.S.A. 69, 2283. Elzinga, M., Maron, B. J., and Adelstein, R. S . (1976). Science 191, 94. Euteneuer, U., Bereiter, J., and Schliwa, M. (1977). Cytobiologie 15, 169. Everhardt, L. P. (1971). J. Mol. Biol. 61, 745. Farell, K. W., and Wilson, L. (1978). J . Mol. Biol. 121, 93. Feit, H., Slusarek, L., and Shelanski, M. L. (1971). Proc. Narl. Acud. Sci. U.S.A. 68, 2028. Feit, H., Neudeck, U., and Baskin, F. (1977). J. Neurochem. 28, 697. Flament-Durant, J., Hubert, J . P., and Dustin, P. (1976). Exp. Cell Res. 99, 435. Flemming, W. (1880). Arch. Mikrosk. Anar. 18, 151. Forer, A., and Behnke, 0. (1972). Chromosoma 39, 145. Forer, A., and Goldman, R. D. (1969). Nature (London) 222, 689. Forer, A., and Goldman, R. D. (1972). J. Cell Sci. 10, 387. Forer, A., and Jackson, W. T. (1976). Cytobiologie 12, 199. Forer, A., and Zimmerman, A . M. (1974). J . Cell Sci. 16, 481. Forer, A., and Zimmerman, A. M. (1976a). J. Cell Sci. 20, 309. Forer, A., and Zimmerman, A. M. (1976b). J. Cell Sci. 20, 329. Forer, A., Kalnins, V. I., and Zimmerman, A. M. (1976). J. Cell Sci. 22, 115.

BIOCHEMISTRY OF THE MITOTIC SPINDLE

85

Franke, W. W., Seib, E., Osborn, M., Weber, K., Herth, W., and Falk, H. (1977). Cytobiologie 15, 24. Friedlander, M., and Wahrman, J. (1970). J. Cell Sci. 7, 65. Fronk, E., Gibbons, I. R., and Ogawa, K. (1975). J . Cell Eiol. 67, 125a. Fuge, H. (1978). Inr. Rev. Cytol. Suppl. 6, 1 . Fujiwara, K., and Pollard, T. (1976). J. Cell Eiol. 71, 848. Fujiwara, K., and Pollard, T. (1978). J. Cell Eiol. 77, 182. Fujiwara, K., Porter, M. E., and Pollard, T. (1977). J. Cell Eiol. 75, 267a. Fuller, G . M., Brinkley, B. R., and Boughter, J. M. (1975a). Science 187, 948. Fuller, G. M., Ellison, I., McGill, M., and Brinkley, B. R. (1975b). J. Cell Eiol. 67, 126a. Fulton, C., Kane, R. E., and Stephens, R. E. (1971). J . Cell Eiol. 50, 762. Fuseler, J. W. (1975). J . Cell Eiol. 67, 789. Fuseler, J. W., Jones, J. E., Fuller, G. M., and Brinkley, B. R. (1976). J. Cell Eiol. 70, 54a. Gallin, J. I., and Rosenthal, A. S. (1974). J . Cell Eiol. 62, 594. Garrels, J. I., and Gibson, W. (1976). Cell 9, 793. Gaskin, F., Kramer, S. B., and Cantor, C. R. (1974). FEES Lefr. 40, 281. Gawadi, N. (1971). Nature (London) 234, 410. Gawadi, N. (1974). Cytobios 10, 17. Gibbons, I. R. (1965). Arch. Eiol. 76, 317. Gibbons, I. R. (1975). In “Molecules and Cell Movement” (S. Inou6 and R. E. Stephens, eds.), p. 207. Raven, New York. Gibbons, I. R., and Fronk, E. (1972). J. Cell Eiol. 54, 365. Gibbons, I. R., and Rowe, A. J. (1965). Science 149, 424. Gibbons, I. R., Cosson, M. P., Evans, J. A,, Gibbons, B. H., Houck, B., Martinson, K. H.,Sale, W., and Tang, W. Y. (1978). Proc. Narl. Acad. Sci. U.S.A. 75, 2220. Goldman, R. D., and Rebhun, L. I. (1969). J . Cell Sci. 4, 179. Goldman, R. D., Lazarides, E., Pollack, R., and Weber, K. (1975). Exp. Cell Res 90, 333. Goldstein, L., KO, C., and Emck, J. (1977a). Cell Eiol. Int. Rep. 1, 5 1 1. Goldstein, L., Rubin, R., and KO, C. (1977b). Cell 12, 601. Goode, D., Roth, L. E. (1969). Exp. Cell Res. 58, 343. Gordon, D. J., Eisenberg, E., and Kom, E. D. (1976). J. Eiol. Chem. 251, 4778. Gould, R., and Borisy, G. G., (1976). J. Cell Eiol. 70, 43a. Govindan, V. M., and Wieland, T. (1975). FEES Left. 59, 117. Green, L. H., Brandis, J. W., Turner, F. R., and Raff, R. A. (1975). Eiochemisrry 14, 4487. Gross, J., and Marme, D. (1978). Proc. Nut/. Acad. Sci. U.S.A. 75, 1232. Gruenstein, E., and Rich, A. (1975). Eiochem. Eiophys. Res. Commun. 64, 472. Haber, J. E., Peloquin, J. G., Halvorson, H. O., and Borisy, G. G. (1972). J. CeNEiol. 55, 355. Harris, P. (1962). J. Cell Eiol. 14, 475. Harris, P. (1975). Exp. Cell Res. 94, 409. Harris, P. (1976). Exp. Cell Res. 97, 63. Harris, P., and Mazia, D. (1962). In “The Interpretation of Ultrastructure” (R. J. Harris, ed.), p. 297. Academic Press, New York. Harrison, C. M. H., Page, B. M., and Keir, H. M. (1976). Nature (London) 260, 138. Hart, J. W., and Sabnis, D. D. (1976a). Curr. Adv. PIanr Sci. 26, 1095. Hart, J. W., and Sabnis, D. D. (1976b). J . Exp. Eor. 27, 1353. Hartman, H., Puma, J. P., and Gurney, T. (1974). J. Cell Sci. 16, 241. Hartmann, J. F., and Zimmerman, A. M. (1968). Exp. Cell Res. 50, 403. Hartwell, L. H. (1978). J. Cell Eiol. 77, 627. Hartwig, J. H., and Stossel, T. P. (1975). J. Eiol. Chem. 250, 5695.

86

CHRISTIAN PETZELT

Hartwig, J. H., and Stossel, T. P. (1977). J . Cell Eiol. 75, 253a. Hasselbach, W. (1972). “The Sarcoplasmic Calcium Pump. Molecular Bioenergetics and Macromolecular Biochemistry. ’’ Springer-Verlag, Berlin and New York. Hatano, S., and Oosawa, F. (1966a). J. Cell. Physiol. 68, 197. Hatano, S., and Oosawa, F. (1966b). Eiochim. Eiophys. Acta 127, 488. Heath, I. B. (1975a). Protoplasma 85, 147. Heath, I. B. (1975b). Protoplasma 85, 177. Heidemann, S. R., and Kirschner, M. W. (1975). J . Cell Eiol. 67, 105. Heidemann, S. R., Sander, G., and Kirschner, M. W. (1977). Cell 10, 337. Heilbrunn, L. V. (1921). J . Exp. Zool. 30, 211. Hepler, P. K. (1976a). J. Cell Sci. 21, 361. Hepler, P. K. (1976b). I n “Mechanism and Control of Cell Division” (T. Rost and E. M. Gifford, Jr., eds.), p. 212. Dowden, Hutchinson & Ross, Inc., Strandsburg, Pennsylvania. Herman, I. M., and Pollard, T. D. (1978). Exp. Cell Res. 114, 15. Herzog, W., and Weber, K. (1977). Proc. Narl. Acad. Sci. U.S.A. 74, 1860. Himes, R. H., Burton, P. R., Karsey, R. N., and Pierson, G. (1976a). Proc. Natl. Acad. Sci. U.S.A. 73, 4397. Himes, R. H., Karsey, R. N., Ruscha, M., and Houston, L. L. (1976b). Eiochem. Eiophys. Res. Commun. 68, 1362. Himes, R. H., Burton, P. R., and Gaito, J. M. (1977). J . Eiol. Chem. 252, 6222. Hinek, A., Thyberg, J., and Friberg, U. (1977). J . Neurocytol. 6, 13. Hinkley, R., and Telser, A. (1974). Exp. Cell Res. 86, 161. Hoebeke, J., Van Nijen, G., and DeBrabander, M. (1976). Eiochem. Eiophys. Res. Commun. 69, 319. Holmes, R. P., and Stewart, P. R. (1977). Nature (London) 269, 592. Huxley, H. E. (1963). J . Mol. Eiol. 7, 281. Inoue, S., and Dan, K. (1951). J. Morphol. 89, 423. h o d , S., and Sato, H. (1967). J. Gen. Physiol. 50, 259. Inout, S., Borisy, G. G., and Kiehart, D. P. (1974). J . Cell Eiol. 62, 175. Inoue, S., Fuseler, J., Salmon, E. D., and Ellis, G. W. (1975). Eiophys. J. 15, 725. Ishikawa, H., Bischoff, R., and Holtzer, H. (1969). J. Cell Eiol. 43, 312. Ito, S. (1960). J . Eiophys. Eiochem. Cytol. 6, 433. Izant, J. G., and McIntosh, J. R. (1977). J . Cell Eiol. 75, 263a. Jackson, W. T., and Doyle, B. G. (1977). J. Cell Eiol. 75, 268a. Jacobs, M., and McVittie, A. (1970). Exp. Cell Res. 63, 53. Johnson, K. A., and Borisy, G. G. (1974). J. Cell Eiol. 63, 157a. Johnston, R. N., Paul, M. (1977). Dev. Eiol. 57, 366. Jorgensen, A. O., Kalnins, V. I., Zubrzycka, E., and MacLennan, D. H. (1976). J . Cell Eiol. 70, 279a. Kakiuchi, S., Yamazaki, R., and Nakajima, H. (1970). Proc. Jpn. Acad. 46, 587. Kakiuchi, S., Yamazaki, R., Teshima, Y., Venishi, K., and Miyamoto, E. (1975). Eiochem. J. 146, 109. Kane, R. E. (1962). J. Cell Eiol. 12, 47. Kane, R. E. (1965). J . Cell Eiol. 25, 137. Kane, R. E. (1967). J. Cell Eiol. 32, 243. Kane, R. E. (1975). J . Cell Eiol. 66, 305. Kane, R. E. (1976). J. Cell Eiol. 71, 704. Kane, R. E., and Forer, A. (1965). J. Cell Eiol. 25, 31. Kane, R. E., Petzelt, C., and Bryan, J. (1977). J. Cell Eiol. 75, 253a.

BIOCHEMISTRY OF THE MITOTIC SPINDLE

87

Kato, T., and Tonomura, Y. (1977). J . Biochem. (Tokyo) 81, 207. Kiehart, D. P., and h o d , S. (1975). Biol. Bull. Woods Hole, Mass.) 149, 433. Kiehart, D. P., and Inoue, S. (1976). J. Cell Biol. 70, 230a. Kiehart, D. P., Inoue, S., and Mabuchi, I. (1977). J. Cell Eiol. 75, 258a. Kinoshita, S., and Yazaki, I. (1967). Exp. Cell Res. 47, 449. Klein, I., Willingham, M., and Pastan, I. (1978). Exp. CeN Res. 114, 229. Kowitt, J . D., and Fulton, C. (1974). Proc. Narl. Acad. Sci. U.S.A. 71, 2877. Kubai, D. F. (1975). Int. Rev. Cyfol. 43, 167. Kuriyama, R. (1977). J. Biochem. (Tokyo) 81, 1 1 15. LaFountain, J. (1974). J. Cell Eiol. 60, 784. LaFountain, J . (1975). J. Cell Eiol. 67, 231a. LaFountain, J . , and Zobel, C. R. (1976). J. Cell Eiol. 70, 196a. Lambert, A.-M., and Bajer, A. S. (1977). Cyrobiologie 15, 1. Lawrence, J. H., and Wheatley, D. N. (1975). Cytohios 13, 167. Lazarides, E., and Lindberg, U. (1974). Proc. Natl. Acad. Sci. U . S . A . 71, 4742. Lazarides, E., and Weber, K. (1974). Proc. Narl. Acad. Sci. U.S.A. 71, 2268. Lee, J. C., and Timasheff, S. N. (1977). Biochemisrry 16, 1754. Lin, Y. M., Liu, Y. P., and Cheung, W. Y. (1974). J. B i d . Chem. 249, 4943. Lockwood, A. H. (1978). Cell 13, 613. Loew, A. G. (1952). J. Cell. Comp. Physiol. 40, 127. Longo, F. J . (1976a). Dev. Biol. 49, 347. Longo, F. J. (1976b). J. Cell Eiol. 69, 539. LudueAa, R. F., and Woodward, D. 0. (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 3594. LudueAa, R. F., Wilson, L., and Shooter, E. M. (1975). In “Microtubules and Microtubule Inhibitors” (M. Borgers and M. DeBrabander, eds.), p. 47. Elsevier. New York. Luykx, P. (1970). Int. Rev. Cyrol.. Suppl. 2, 1. Mabuchi, I. (1973). J. Cell Biol. 59, 542. Mabuchi, I. (1974). J. Biochem. (Tokyo) 76, 47. Mabuchi, I. (1976a). J . Mol. Eiol. 100, 569. Mabuchi, I. (1976b). J. Eiochem. (Tokyo) 80, 413. Mabuchi, I., and Okuno, M. (1977). J. Cell. Biol. 74, 251. McGill, M., and Brinkley, B. R. (1975). J. Cell Biol. 67, 189. McGuire, J . , Quinn, P., and Knutton, S. (1974). J. Cell Biol. 63, 217a. Mclntosh, J. R. (1974). J . Cell Eiol. 61, 166. McIntosh, J . R . (1977). In “Mitosis, Facts and Questions” (M. Little et al., eds.), p. 167. SpringerVerlag, Berlin and New York. Mclntosh, J. R., Hepler, P. K., and Van Wie, D. G. (1969). Narure (London) 224, 659. Mclntosh, J . R., Cande, W. Z., and Snyder, J. A. (1975). In “Molecules and Cell Movement” (S. Inoue and R. E. Stephens, eds.), p. 31. Raven, New York. Maekawa, S., and Sakai, H. (1978). J. Eiochem. (Tokyo) 83, 1065. Maller, J . , Poccia, D., Nishioka, D., Kidd, P., Gerhart, J., and Hartmann, H. (1976). Exp. Cell Res. 99, 285. Margolis, R. L., and Wilson, L. (1978). Cell 13, 1. Matthews, L. (1977). J . Cell Eiol. 72, 502. Maupin-Szamier, P., and Pollard, T. D. (1978). J. Cell Eiol. 77, 837. Mazia, D. (1937). J. Cell. Comp. Physiol. 10, 291. Mazia, D. (1961). In “The Cell” (J. Brachet and A. E. Mirsky, eds.), Vol. 3, p. 77. Academic Press, New York. Mazia, D. (1974). Proc. Narl. Acad. Sci. U.S.A. 71, 690.

88

CHRISTIAN PETZELT

Mazia, D., and Dan, K. (1952). Proc. Nurl. Acud. Sci. U.S.A. 38, 826. Mazia, D., Mitchison, J. M., Medina, H., and Harris, P. (1961a). J . Eiophys. Eiochem. Cyrol. 10, 467. Mazia, D., Chaffee, R. R., and Iverson, R. M. (1961b). Proc. Nurl. Acud. Sci. U.S.A. 47, 788. Mazia, D., Petzelt, C., Williams, R. O., and Meza, I. (1972). Exp. Cell Res. 70, 325. Mazia, D., Schatten, G . , and Sale, W . (1975). J. Cell Eiol. 66, 198. Meza, I., Huang, B., and Bryan, J. (1972). Exp. Cell Res. 74, 535. Milsted, A,, and Cohen, W. D. (1973). Exp. Cell Res. 78, 243. Milsted, A., Cohen, W. D., and Lampen, N. (1977). J. Cell Sci. 23, 43. Mohri, H. (1968). Nurure (London) 217, 1053. Mohri, H.(1976). Eiochim. Eiophys. Actu 456, 85. Mohri, H., Mohri, T., Mabuchi, I., Yazaki, I., Sakai, H., and Ogawa, K. (1976). Dev., Growth & Direr. 18, 391. Moore, L., and Pastan, I. (1977). J. Eiol. Chem. 252, 6304. Moore, P. B., and Carraway, K. L. (1978). Eiochem. Eiophys. Res. Commun. 80, 560. Mooseker, M. S., Fujiwara, K., and Pollard, T. D. (1977). J . Cell Eiol. 75, 258a. Morgan, J. L., and Seeds, N. W. (1975). J. Cell Eiol. 67, 136. Morgan, J. L., Holladay, C. R., and Spooner, B. S. (1978). Proc. Nutl. Acud. Sci. U.S.A. 75, 1414. Miiller, W. (1972). Chromosomu 38, 139. Murphy, D. B., and Borisy, G. G. (1974). J. Cell Eiol. 63, 236a. Murphy, D. B., and Borisy, G. G. (1975). Proc. Nurl. Acud. Sci. U.S.A. 72, 2696. Murphy, D. B., Vallee, R. B., and Borisy, G. G. (1977) Biochemistry 16, 2598. Nagle, B. W., Doenges, K. H., and Bryan, J. (1977). Cell 12, 573. Naruse, H., and Sakai, H. (1978). J . Eiochem. (Tokyo) 83, 1265. Nath, J., and Rebhun, L. I. (1976a). J. Cell Eiol. 68, 440. Nath, J., and Rebhun, L. I, (1976b). J. Cell Eiol. 70, 43a. Niederman, R., and Pollard, T. D. (1975). J. Cell Eiol. 67, 72. Nishida, E., Sakai, H. (1977). J. Eiochem. (Tokyo) 82, 303. Ogawa, K. (1973). Eiochim. Eiophys. Actu 293, 514. Ogawa, K., and Gibbons, I. R. (1976). J . Eiol. Chem. 251, 5793. Ogawa, K., and Mohri, H. (1975). J . Eiol. Chem. 250, 6476. Okuno, M., Ogawa, K., and Mohri, H. (1976). Eiochem. Eiophys. Res. Commun. 68, 901. Olmsted, J. B., and Borisy, G. G. (1973). Eiochemistry 12, 4284. Olmsted, J. B., and Borisy, G. G. (1975). Biochemistry 14, 2996. Olmsted, J. B., Marcurn, J. M., Johnson, K. A., Allen, C., and Borisy, G. G. (1974). J. Suprumol. Srruct. 2, 429. Osborn, M., and Weber, K. (1976). Proc. Nutl. Acud. Sci. U.S.A. 73, 867. Osbom, M., Webster, R. E., and Weber, K. (1978). J. Cell Eiol. 77, R28. Otto, J., Zigmond, S., and Bryan, J. (1977). J. Cell Eiol. 75, 262a. Parchman, L. G., and Stem, H. (1969). Chromosomu 26, 298. Paul, M., and Johnston, R. N. (1978). J. Exp. Zoo/. 203, 143. Paulson, J. C., and McClure, W. 0. (1975). J. CellEiol. 67, 461. Penningroth, S. M., and Kirschner, M. W. (1977). J . Mol. Eiol. 115, 643. Pepper, D. A., and Brinkley, B. R. (1977). Chromosomu 60, 223. Peterson, S. P., and Berns, M. W. (1977). J . Cell Eiol. 75, 282a. Petzelt, C. (1970). Chromosomu 29, 237. Petzelt, C. (1972a). Exp. Cell Res. 70, 333. Petzelt, C. (1972b). Exp. Cell Res. 74, 156. Petzelt, C. (1974a). Exp. Cell Res. 86, 404.

BIOCHEMISTRY OF THE MITOTIC SPINDLE

89

Petzelt, C. (1974b). J. Cell Biol. 63, 267a. Petzelt, C. (1976). Exp. CellRes. 102, 200. Petzelt, C., and Auel, D. (1977). Proc. Nurl. Acad. Sci. U.S.A. 74, 1610. Petzelt, C., and Auel, D. (1978). In “Cell Reproduction-Honoring Daniel Mazia” (D. Prescott and E. R. Dirksen, eds.), p. 487. Academic Press, New York. Petzelt, C., and von Ledebur-Villiger, M. (1973). Exp. Cell Res. 81, 87. Petzelt, C., Auel, D., and Sachsenmeier, W. (1979). In preparation. Pickett-Heaps, J. (1969). Cyrobios 3, 257. Pickett-Heaps, J . (1974). BioSystems 6, 37. Pollard, T. D. (1976). J. Cell Biol. 58, 579. Pollard, T. D., and Weihing, R. R. (1974). Crir. Rev. Biochem. 2, 1 . Pollard, T . D., Thomas, S. M., and Niedennan, R. (1974). Anal. Biochem. 60, 258. Pollard, T. D., Porter, M. E., and Stafford, W. (1977). J. Cell Biol. 75, 262a. Ponstingl, H., Hofer-Warbiiek, R., Kempf, T., Krauhs, E., and Little, M. (1978). Cytobiologie 18, 196. Porter, K. R. (1966). Princ. Biomol. Organ., Ciba Found. Symp., 1965 p. 308. Porter, K. R., and Machado, R. D. (1960). J. Biophys. Biochem. Cyrol. 7, 167. Prescott, D. M., and Bender, M. A. (1962). Exp. Cell Res. 26, 260. Raff, R. A., and Kaumeyer, J. S. (1973). Dev. Biol. 32, 309. Raff, R. A., Greenhouse, G., Gross, K. W., and Gross, P. R. (1971). J. Cell Biol. 50, 516. Raff, R. A., Brandis, J. W., Green, L. H., Kaumeyer, J. F., and Raff, E. (1975). Ann. N.Y. Acad. Sci. 253, 304. Rebhun, L. I. (1976). Am. Zool. 16, 469. Rebhun, L. I. (1977). fnt. Rev. Cyrol. 37, 1 . Rebhun, L. I . , and Sander, G. (1967). J . Cell Biol. 34, 859. Rebhun, L. I . , Rosenbaum, J., Lefebvre, D., and Smith, G . (1974). Nature (London) 249, 113. Rebhun, L. I . , Nath, J., and Remillard, S. (1975). Cell Moril., Coldspring Hurbor Conf. Vol. 4, p. 1343. Rebhun, L. I., Miller, M., Schnaitman, T. C., Nath, J., and Mellon, M. (1976). J. Suprumol. Srrucf. 5, 199. Remillard, S . , Rebhun, L. I., Howie, G. A,, and Kupchan, S. M. (1975). Science 189, 1002. Renaud, F. L., Rowe, A. J., and Gibbons, I. R. (1968). J. Cell Biol. 36, 79. Ridgway, E. B., and Durham, A. C. H. (1976). J. Cell Biol. 69, 223. Ridgway, E. B . , Gilkey, J. C., and Jaffe, L. F. (1976). J . Cell Biol. 70, 227a. Robbins, E., and Jentzsch, G. J. (1970). J . Cell Biol. 40, 678. Roberts, K. (1974). Prog. Biophys. M o l . Biol. 28, 373. Roobol, A,, Gull, K., and Pognos, C. I. (1976). FEBS Lerr. 67, 248. Rosenfeld, A. C., Zackroff, R. V., and Weisenberg, R. C. (1976). FEES Leu. 65, 144. Rubenstein, P. A,, and Spudich, J. A. (1977). Proc. Nurl. Acud. Sci. U.S.A. 74, 120. Rubin, R. W., and Cousins, E. H. (1976). Phyrochemistry 15, 1837. Rubin, R. W., Warren, R. H., Lukeman, D. S., and Clements, E. (1978). J. Cell Biol. 78, 28. Rustad, R. C. (1959). Experienriu 15, 323. Sakai, H. (1966). Biochim. Biophys. Acra 112, 132. Sakai, H., and Kuriyama, R. (1974). Dev., Growrh & DiHer. 16, 123. Sakai, H., Hiramoto, Y.,and Kuriyama, R. (1975). Dev.. Growrh & D @ r . 17, 265. Sakai, H., Mabuchi, I., Shimoda, S . , Kuriyama, R., Ogawa, K., and Mohri, H. (1976). Dev., Growth & Difler. 18, 21 1 . Salmon, E. D. (1975a). J. Cell Biol. 65, 603. Salmon, E. D. (1975b). Science 189, 884.

90

CHRISTIAN PETZELT

Salmon, E. D. (197%). Ann. N.Y. Acad. Sci. 253, 383. Salmon, E. D., and Jenkins, R. (1977). J. Cell Biol. 75, 295a. Salmon, E. D., Goode, D., Maugel, T. K.,and Bonar, D. B. (1976). J. Cell Biol. 69, 443. Sanger, J. W. (1975a). Proc. Natl. Acad. Sci. U.S.A. 72, 1913. Sanger, J. W. (1975b). Proc. Natl. Acud. Sci. U.S.A. 72, 2451. Sanger, J. W. (197%). Cell Tissue Res. 161, 431. Sanger, J. W. (1977). In “Mitosis, Facts and Questions” (M. Little et al., eds.), p. 98. SpringerVerlag, Berlin and New York. Sanger, J. W., and Sanger, J. M. (1975). J. Cell Biol. 67, 381a. Sanger, J. W., and Sanger, J . M. (1976). Cold Spring Harbor Conf. Cell Proliferation Vol. 3, p. 1295. Sato, H., Ellis, G. W., and h o d , S. (1975). J. Cell Biol. 67, 501. Sawada, N., and Rebhun, L. 1. (1969). Exp. Cell Res. 55, 33. Schatten, H. (1977). Dissertation, Universitat Heidelberg. Schliwa, M. (1976). J. Cell Biol. 70, 527. Schliwa, M. (1977). J . Submicrosc. Cytol. 9, 221. Schloss, J. A., Milsted, A,, and Goldman, R. D. (1977). J. Cell Biol. 74, 794. Schroeder, T. E. (1973). Proc. Narl. Acad. Sci. U.S.A. 70, 1688. Schroeder, T. E. (1976). Cold Spring Harborc Con$ Cell Proliferation Vol. 3, p. 265. Schulman, H., and Greengard, P. (1978). Nature (London) 271, 478. Shelanski, M. L., and Taylor, E. W. (1968). J . Cell Biol. 38, 304. Sherline, P., and Schiavone, K. (1978). J . Cell Biol. 77, R 9. Sisken, J. E., Wilkes, E., Dounelly, G. M., and Kakefuda, T. (1967). J. Cell Biol. 32, 212. Sluder, G. (1976). J. Cell Biol. 70, 75. Smith, D. S., Jiirlfors, U.,and Beranek, R. (1970). J . Cell Biol. 46, 199. Smith, G. W., and Rebhun, L. I. (1974). J . Cell Biol. 63, 321a. Smoake, J. A., Song, S. Y., and Cheung, W. Y. (1974). Biochim. Biophys. Acta 341, 402. Snell, W . J., Dentler, W. L., Haimo, L. T., Binder, L. I . , and Rosenbaum, J. L. (1974). Science 185, 357. Snyder, J. A., and McIntosh, J. R. (1975). J. Cell Biol. 67, 744. Snyder, J. A., and McIntosh, J. R. (1976a). Annu. Rev. Biochem. 45, 699. Snyder, J. A., and McIntosh, J. R. (1976b). J. Cell Biol. 70, 368a. Stadler, J., and Franke, W. W. (1972). Nature (London), New Biol. 237, 237. Stadler, J., and Franke, W. W. (1974). J . Cell Biol. 60, 297. Starger, J. M., Brown, W. E., Goldman, A. E., and Goldman, R. D. (1978). J. Cell Biol. 78, 93. Steams, M.E., Coundly, J. A., and Brown, D. L. (1976). Science 191, 188. Stephens, R. E. (1%7). J. Cell Biol. 32, 255. Stephens, R. E. (1968). J . Mol. Biol. 32, 277. Stephens, R. E. (1970). J . Mol. Biol. 47, 353. Stephens, R. E. (1972a). Biol. Bull. (Woods Hole, Mass.) 142, 145. Stephens, R. E. (1972b). Biol. Bull. (Woods Hole, Mass.) 142, 489. Stephens, R. E. (1973). J. Cell Biol. 57, 133. Stephens, R. E. (1976). J. Cell Biol. 70, 95a. Stephens, R. E., and Edds, K. T. (1976). Physiol. Rev. 56, 709. Storti, R. V., and Rich, A. (1976). Proc. Narl. Acad. Sci. U.S.A. 73, 2346. Storti, R. V., Coen, D. M., and Rich, A. (1976). Cell 8, 521. Stossel, T. P., and Hartwig, J. (1975). J . Biol. Chem. 250, 5706. Stossel, T. P., and Pollard, T. D. (1974). J. Biol: Chem. 248, 8288. Taylor, E. W. (1965). J . Cell Biol. 25, 145.

BIOCHEMISTRY OF THE MITOTIC SPINDLE

91

Telzer. B. R., Moses, M. J., and Rosenbaum, J. L. (1975). Proc. Natl. Acad. Sci. U.S.A. 72,4023. Teo, T. S . , and Wang, J. H. (1973). J. Biol. Chem. 248, 5950. Thyberg, J., and Hinek, A. (1977). J. Neurocytol. 6, 27. Tilney, L. G. (1975). J. Cell Biol. 69, 289. Tilney, L. G., and Detmers, P. (1975). J. Cell Biol. 66, 508. Tilney, L. G., and Gibbins, J . R. (1968). Protoplasma 65, 167. Tilney, L. G., Hatano, S., Ishikawa, H., and Mooseker, M. S . (1973). J . Cell Biol. 59, 109. Timourian, H., Jotz, M. M., and Clothier, G. E. (1974). Exp. Cell Res. 83, 380. Tucker, J. B. (1977). Nature (London) 226, 22. Vale, M. G. P., and Arselio, P. C. (1973). Biochim. Biophys. Acra 325, 29. Vallee, R. B., and Borisy, G. G. (1978). J . Biol. Chem. 253, 2834. Vanderkerckhove, J., and Weber, K. (1978). Proc. Narl. Acad. Sci. U.S.A. 75, 1106. Waisman, D., Stevens, F. C., and Wang, J. H. (1975). Biochem. Biophys. Res. Commun. 65,975. Wang, K., Ash, J. F., and Singer, S. J. (1975). Proc. N u t / . Acad. Sci. U.S.A. 72, 4483. Wang, R. W. (1974). Narure (London) 248,76. Wang, R. W., Rebhun, L. I., and Kupchan, S . M. (1977). Cancer Res. 37, 3071. Watson, E. L., Vincenzi, F. F., and Davis, P. W. (1971). Biochim. Biophys. Acra 249, 606. Watterson, D. M., Harrelson, J. R. W. G., Keller, P. M., Sharief, F., and VAnaman, T. C. (1976). J . Biol. Chem. 251, 4501. Weatherbee, J. A., Luftig, R. B., and Weihing, R. R. (1978). J. Cell Biol. 78, 47. Weber, K., and Groeschel-Stewart, U. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 4561. Weber, K., Pollack, R., and Bibring, T. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 459. Weber, K., Wehland, J., and Herzog, W. (1976a). J. Mol. Biol. 102, 817. Weber, K., Rathke, P. C., Osbom, M., and Franke, W. W. (1976b). Exp. Cell Res. 102, 285. Weber, K., Koch. R., Herzog, W., and Vandekerckhove, J. (1977). Eur. J. Biochem. 78, 27. Weber, K., Rathke, P. C., and Osbom, M. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 1820. Weihing, R. R. (1977). J . Cell Biol. 75, 251a. Weihing, R. R., and Korn, E. D. (1971). Biochemisrry 10, 590. Weihing, R. R., and Korn, E. D. (1972). Biochemistry 11, 1538. Weingarten, M. D., Lockwood, A. H., Hwo, S., and Kirschner, M. W. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 1858. Weisenberg, R. C. (1972). Science 177, 1104. Weisenberg, R. C. (1975). Ann. N . Y . Acad. Sci. 253, 573. Weisenberg, R. C., and Rosenfeld, A. C. (1975). J . Cell Biol. 64, 146. Weisenberg, R. C., and Taylor, E. W. (1968). Exp. Cell Res. 52, 372. Weisenberg, R. C., Borisy, G. G., and Taylor, E. W. (1968). Biochemisrry 7, 4466. Welsh, M. J., Dedman, J. R., Brinkley, B. R., and Means, A. R. (1977). J . Cell Biol. 75,262a. Welsh, M. J., Dedman, J. R., Brinkley, B. R., and Means, A. R. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 1867. M d e n , R. G . , Brulter-Browne, G. S . , and Gros, F. (1976). Proc. Narl. Acad. Sci. U S A . 73, 2018.

Wiche, G., and Cole, R. D. (1976). Proc. Narl. Acad. Sci. U.S.A. 73, 1227. Wiche, G., Corces, V. G., and Avila, J. (1978). Nature (London) 273, 403. Wick, S . M., and Hepler, P. K. (1976). J. Cell Biol. 70, A209. Wieland, T., and Govindan, V. M. (1974). FEBS Letr. 46, 351. Wilson, L., and Bryan, J. (1975). Adv. Cell Mol. Biol. 3, 21. Wilson, L., and Friedkin, M. (1967). Biochemistry 6 , 3126. Wilt, F. H., Sakai, H., and Mazia, D. (1967). J. Mol. Biol. 27, 1. Witman, G. B., Carlson, K., and Rosenhaum. J. L. (1972). J . Cell Biol. 54, 540.

92

CHRISTIAN PETZELT

Witman, G . B . , Cleveland, D. W . , Weingarten, M. D., and Kirschner, M. W . (1976). Proc. Nutl. Acad. Sci. U.S.A. 73, 4070. Wray, W. (1973). Methods Cell Biol. 6, 283. Wray, W . , and Stubblefield, E. (1970). Exp. Cell Res. 59, 469. Yamada, S., and Tonomura, Y. (1972). J . Biochem. (Tokyo) 71, 1101. Yerna, M . , Hartshorne, D. J . , and Goldman, R. D. (1977). J . Cell Biol. 75, 254a. Yu, B. P., Masoro, E. J . , and Bertrand, H. A. (1974). Biochemistry 13, 508. Zackroff, R., Rosenfeld, A., and Weisenberg, R. (1975). J . Cell Biol. 67, 469a.

INTERNATIONAL REVIEW OF CYTOLOGY, VOL 60

Alternatives to Classical Mitosis in Hemopoietic Tissues of Vertebrates VIBEKEE. ENGELBERT The Ramsay Wright Zoological Laboratories, University of Toronto, Toronto, Canada I. Introduction

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

11. Imprint, Smear, Fixation, and Staining Methods

111. IV. V. VI. VII. VIII.

IX.

X. XI. XII.

. . . . . . .

Methods Using Tritiated Thymidine . . . . . . . . . . . Fluorescence Method for DNA and RNA . . . . . . . . . Tissue Culture in Vitro of Hemopoietic Tissues . . . . . . . Behavior and Morphological Variations in Blast Cells and Their Nuclei . . . . . . . . . . . . . . . . . . . . . . The Occurrence and Significance of Two Spatially Separate Nuclear Masses in Blast Cells and in Differentiating Cells . . . . . . Fate of the Peripheral or Shell-like Nuclear Mass, and the Inner Nuclear Mass, in Differentiating Cells of Leukemic Mice (AKR Strain) . . . . . . . . . . . . . . . . . . . . . . Results following Injection of Tritiated Thymidine . . . . . . A . Reassociation of Vesicles and Nuclear Granules in Rabbit Spleens . . . . . . . . . . . . . . . . . . . . B. Grain Counts in Vesicular Nuclei of Rabbit Spleen and Bone Marrow . . . . . . . . . . . . . . . . . . . . Erythropoiesis in Blood of Vertebrates with Nucleated Erythocytes. Formation of Clone Cells from Nuclei of Young Mature Erythrocytes Summary . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . .

93 96 98 98 99 100 105

107 109 109 111

114

117 117 I I8

I. Introduction Cell reproduction in hemopoietic tissues of vertebrates, and especially of mammals, has for many years occupied research workers in the biological sciences. When one is concerned with cell proliferation, one naturally looks for the occurrence of mitosis. This process, which is so easily seen in teaching material such as blastulas of fish or onion root tips and in human blood when it is cultured in vitro with mitosis-stimulating substances, is difficult to find in slide preparations of hemopoietic tissues from healthy humans or other animals. 93

Copyright @ 1979 by Academic Ress, Inc. All rights of reproduction in any form reserved. ISBN 0-12-364360-0

94

VIBEKE E. ENGELBERT

Many early hematologists, as well as later ones, attempted to find a “blood mother cell,” a hemocytoblast. Downey (1932, 1938) defined the myeloblast of Naegli as “the undifferentiated non-granular, lymphoid stem cell” of the red bone marrow which functions as the indifferent, polyvalent parent cell of all the “myeloid” elements. Downey stated further that “the hemocytoblast of Maximow and Danchakoff and the lymphoid hemoblast of Jordan and Latta are polyvalent large lymphocytes. A cell of myeloblastic structure is not recognized. The hemocytoblast of Ferrata is identical with Naegli’s myeloblast, but it may produce lymphocytes when in lymphoblastic function and it occurs in normal lymphatic tissue as well as the marrow. The important point here is the recognition of a stem cell that is lymphoid in character, that is polyvalent, and that occurs in normal lymphatic tissue as well as in bone marrow. Jordan (1938), after describing the comparative hematology of lower vertebrates, writes in “Terminology of Lymphoid Cells, “In the foregoing pages the terms hemocytoblast, hemoblast, lymphoid hemoblast and lymphocyte were used synonymously as indicating functional identity of morphologically variable multipotential blood stem cell. Morphologically the hemocytoblast appears very different from the smaller lymphocytes. However, both have identical capacities to develop into erythrocytes, thrombocytes, monocytes and granulocytes, the specific route of differentiation being presumably determined by the impingement of specific differential stimuli. ” In the same section Jordan writes later, “The typical hemocytoblast is a relatively large cell, with large vesicular nucleus and a moderately basophilic cytoplasm. The most distinctive features concern the arrangements of the chromatin in the form of minute granules, delicate nuclear membrane and the presence of one or several nucleoli usually achromatic. But it must be emphasized that the cytologic features of a hemocytoblast are not static; the cytology is subject to considerable variation, allowing for such variability, a concomitant of simple metabolism the large lymphocyte has an identical morphology.” He claims that mitosis occurs only in the large- and medium-sized cells in lymphoid nodules of chicken bone marrow. In numerous illustrations of blood cells he shows only three mitotic figures of hagfish “normoblast,” one toad lymphocyte, and one toad erythroblast. He labels three lymphocytes exhibiting slight size differences as “hemocytoblasts. ” It is just a little over 200 years since lymphocytes from lymph nodes were first described and illustrated by Hewson (1777) who called them “cells” some 60 years before the cell theory of Schwann (1839) was presented. Lymphocytes were first defined as a distinct cell type in blood and lymph by Jones (1846) at Charing Cross Hospital, London. Their motility was described by Ranvier (1873, Arnold (1887), and Askanazy (1905). Maximow (1909) considered the lymphocyte as the common mother cell of the different elements of the blood both in the embryo and in the postfetal life of mammals. Yoffey (1932-1933) and Jordan (1935) believed that the circulating lymphocytes lodge in the bone marrow, where they transform into erythrocytes. ”

”

95

ALTERNATIVES TO CLASSICAL MITOSIS

Kindred (1938, 1940, 1942) published quantitative studies of hemopoietic organs of young albino rats. Although he quotes percentages of cells in mitosis, his plates do not illustrate this process. Andreasen (1959) reported “mitotic figures counted in suspensions of cell nuclei prepared from the whole organ. The counting included all phases of mitosis and differential counting of the mitotic phases is possible,” and “that under certain conditions the lymphoid tissue was characterized by high mitotic rates, whereas under other conditions the same tissue was marked by just the opposite process, namely degeneration of lymphocytes combined with phagocytic activity carried out by the reticulum. Hamilton (1954) and Hill (1959) both believed that lymphocytic nucleoprotein was reutilized. Trowell (1957) claimed that a process of phagocytosis of pycnotic lymphocytes was an important link in the “re-utilization. ” Yoffey et al. (1958), after labeling lymphocytes of guinea pigs with tritiated thymidine, concluded that lymphoid tissue shows active synthesis of DNA. However, their results did not support the concept “either of massive re-utilization or large scale recirculation. ” Counts of blood cells at different time intervals were done by many workers. Yoffey er al. (1958) reported counts of the number of lymphocytes “in the cellular migration stream,” especially in the thoracic duct. In this way they dealt with “the high level of lymphocyte production. This group used tritium-labeled thymidine, as “this is believed to be rapidly and specifically incorporated into newly formed DNA, and in view of its precise localization in radioautographs seems to be suitable for study of cell production.” They reported 1.83 to 6.57% labeled cells depending on the time after injection of the tritiated thymidine that the counts were made. Trowell (1958) writes, “there is little doubt that the small lymphocyte originates by mitosis, followed by a shrinkage type maturation from the medium and large ones. ” The process was followed in cultures of thoracic duct lymph by Hall and Furth (1938). In similar experiments Gowans (1957) found that the daughter cells produced by mitosis were, initially at any rate, rather larger than small lymphocytes. “The medium and large lymphocytes are actively mitotic but the small lymphocytes never or rarely divide.” Trowell states further, “my own experience in a variety of species, has been that the small lymphocytes in the intestinal crypt epithelium are the only ones which can ever be found in mitosis. Very rarely we have seen mitosis of a small lymphocyte in rat lymph-node cultures. Dustin (1959) discussed mitotic growth in bone marrow of the rat by the stathmokinetic (colchicine) method, in which metaphases were believed to be arrested and thus could be counted to give an estimate of mitotic activity. We have prepared spleens of newborn mice with the Feulgen (Schiff) method and gentle squash technique (Engelbert, 1960, 1961). The very brilliantly colored “chromosomal” bodies which are found in very large numbers in configurations similar to metaphases do not appear to us as typical metaphases. Configurations similar to anaphases are never seen. The chromosomal ‘‘meta”

”

”

96

VIBEKE E. ENGELBERT

phase” separates out into three, four, or more chromosomal groups, which remain connected; ring- or band-shaped nuclei are also found. These nuclear stages, we believe, are young stages of polymorphonuclear leukocytes. J. H. Morrison and G. B. Wilson (private communication, 1958) reported that they had treated spleens of 3-day-old rats with the Feulgen squash method. They too found no anaphases and believed that metaphase chromosomes passed directly into the telophasic state shortly to form band-shaped nuclei of young neutrophilic leukocytes. Studies in our laboratory using in vitro preparations without mitotic stimulants and a medium of calf serum and synthetic medium 1066 (Connaught Laboratories), or the animals’ own blood, were made on lymphoid tissue of normal mice with phase-contrast and cinematographic recording. Metaphase figures entering anaphase were not seen as most of the chromosomes fused forming a ringshaped nucleus with cytoplasm. In contrast, when we cultured lymphoid tissue from leukemic mice (AKR) in vifro the metaphases entered anaphase in most cases (Engelbert, 1968). Wintrobe (1967) quotes studies of bone marrow from nine healthy males in which 8.86 mitoses per 1000 cells were found, i.e., a mitotic index of 0.9%. Diggs et al. (1957) claimed that if hemopoietic tissues showed mitotic figures greater than 1% it is indicative of abnormal cell production. Classical mitosis thus presents a considerable puzzle to workers interested in the reproduction of blood cells. Experiments carried out in our laboratory with injections of tritiated thymidine showed that out of 10,000 labeled cells counted in spleens and bone marrow, only one well-labeled anaphase figure was present (Engelbert, 1967). Westermann (1974) identified and followed the developmental series of thrombocytes in four species of turtle. She states “while other (cell) types were sometimes found in mitosis no division stages were ever observed in thrombocytes. ” Wedlock (1974) reported that spleen imprints of chicks at about 16 days of incubation showed all stages of mitosis in erythrocytes. The same author injected chicks with bacteriophage OX174 at the time of hatching and at 4, 8, and 16 days after hatching. After 4 days of exposure to the phage the spleens showed young plasma cells in all stages of mitosis. In bone marrow of newly hatched chicks exposed to the phage, young plasma cells were found in groups and mitotic figures were sometimes seen in these cells. Wedlock writes that, apart from the examples mentioned above, mitosis was exceedingly rare.

11. Imprint, Smear, Fixation, and Staining Methods The imprint method, used by Downey (1938) and since by many others, was used by us as a gentle touch method (Engelbert, 1961). The cut surfaces of spleen, lymph nodes, thymus, and bone marrow were touched gently to the

ALTERNATIVES TO CLASSICAL MITOSIS

97

surfaces of thoroughly cleaned slides. If the bone marrow was too soft, a sterile camel-hair brush (Mills, 1964; Mills et al., 1969) was used to make smears of the marrow. Imprints of bone marrow were also made by touching a small sterile gelfoam sponge to the marrow surface, then touching the sponge to a slide, thus making an imprint (Engelbert, 1956). The dry imprint method (Shelley, 1961; (Mills, 1964; Mills et a l . , 1969) has been used extensively; afterward the dry slides are stained with the May-Griinwald-Giemsa technique, developed by Pappenheim. This stain has been our routine stain for many years. Blood smears were made as camel-hair brush smears (Mills, 1964). The brush smears avoided smudges, which occurred easily with the slide smear method, as young soft cells were damaged. In several cases we took biopsy specimens from spleens of rabbits, 900-2000 gm in weight. The rabbits were anesthetized with 1.5 to 2 cc of 1% Ibatal (sodium pentobarbital U.S.P. XIV, Ingram and Bell, Toronto) injected intravenously, using a marginal ear vein. The biopsy method produced excellent material both for dry imprints and for tissue culture specimens. The splenic incisions were packed with sterile gelfoam sponges. All but one of the rabbits recovered. Fixation in methyl alcohol was followed by MGG stain. We used this stain according to the technique of Jacobson and Webb (1952). It should be emphasized that fast fixation is imperative. Slow fixation such as one finds with formalin causes the nuclei and cells to contract thus causing artifacts. Fast fixation with acetic alcohol 1:3 for 5 to 8 seconds catches the cells and nuclei in activities that other methods may miss. Acetic alcohol is followed by two rinses in absolute ethyl alcohol and air drying. The slides are stained either with the Schiff or Feulgen method for DNA (Feulgen and Rossenbeck, 1924) or with toluidine blue. The latter method was carried out according to Momson (1958), who was the first to adopt Bonhag’s (1955) techniques for use on imprints of hemopoietic tissues. Imprints were stained in toluidine blue 0 (N.S.) at 37°C for 30-40 minutes in a 0.05% staining solution in McIlvaine ’s citric acid-disodium phosphate buffer, 1/10 strength, pH 4.0 (see Pearse, 1960). The various behavioral stages of blast cell nuclei to be described later were well stained (Engelbert, 1961, Figs. 1 and 4). We also used the Feulgen (Schiff) nuclear stain on Millipore imprints fixed in Zenker’s acetic solution. After the rinse in distilled water following fixation, the filters are put into NHCL at 60°C for 10 minutes, then rinsed in distilled water and put into Schiff’s reagent for 1 hour. This is followed by three 3-minute rinses in fresh S02-water; a 5-minute wash in running tap water; a quick rinse in 95% ethyl alcohol; 2 minutes in absolute ethyl alcohol and xylol (1 :1); finally three changes in pure xylol of 5 minutes each, followed by mounting with malinol. The Millipore filters remain transparent for a long time and allow observation with high-power objectives as well as photomicrography (Engelbert, 1961, Fig. 4). Imprints on Millipore filters have the advantage of consisting of more than

98

VIBEKE E. ENGELBERT

one cell layer, forming, as it were, a thin section, without the disadvantages of having cells cut by the microtome knife.

111. Methods Using Tritiated Thymidine Mammalian tissue was prepared as dry imprints. Tissues and blood with nucleated erythrocytes had to be fixed briefly in methanol and air-dried before radioautography, as it had been found (Smith, 1969) that the large erythrocytes were damaged by the warm photographic emulsion, in which the slides had to be dipped. Three-week-old rabbits weighing 500-600 gm were injected intravenously with tritiated thymidine (Engelbert, 1967), at a dosage of 0.5 pCilgm body weight (specific activity 6.7 Cilmole). The dry imprint method was used to prepare the hemopoietic tissues for radioautography. The radioautographic technique of Car0 (1964) and Kopriwa and Leblond (1962) were used with Ilford K5 emulsion. The slides were developed after 10 days of storage at 4°C in the dark. Avian tissue was stored for 14 days. The slides were developed in D-19 Developer at 20°C for 2 minutes, then transferred to a stop bath of 1% acetic acid for 10 seconds, fixed in Kodak Rapid Fixer with hardener for 2 minutes, and washed in running water for 5 to 10 minutes. When the slides were dry they were stained with MGG stain using increased staining times.

IV. Fluorescence Method for DNA and RNA Dry imprints were used also for the acridine orange stain (Edward Gurr, Ltd., London, England) prepared as a 0.1% solution by measuring 10.0 ml of a 1% aqueous stock acridine orange solution and adding phosphate buffer to 100 ml. The slides, one at a time, were immersed in rapid succession in a series of solutions: dipped in 1% acetic acid for 30 seconds, stained in 0.1% acridine orange stain (10 seconds for mammalian tissue, 20 seconds for avian tissue), rinsed in phosphate buffer for 3 seconds, transferred to 1 M calcium chloride (to allow for differentiation of the nucleic acids) for 3 to 10 seconds, and finally rinsed in phosphate buffer for 3 seconds. The slides were then mounted with a few drops of the buffer and covered with a zero-thickness cover glass sealed on with hot paraffin. A Leitz Ortholux microscope equipped for fluorescence microscopy, with a mercury vapor lamp and the necessary activating filters as well as protective orange shielding and protective filters in the oculars, were used. It was found that an interval of 30 minutes between staining and observation of the cells produced

ALTERNATIVES TO CLASSICAL MITOSIS

99

the best results with avian tissue (Smith, 1969). This agrees with Bushong et al. (1968). With ultrablue light the stained tissues displayed the colors of the nucleic acids. DNA fluoresced green to yellow while RNA fluoresced orange to brick red. Hemoglobin appeared black as this substance blocks the fluorescent light, because it absorbs monochromatic light (Nairn, 1962).

V. Tissue Culture in Vitro of Hemopoietic Tissues These techniques were carried out in a special sterile room and sterile techniques were maintained throughout. In the early work we were interested in the effect of foreign proteins. Rabbits of farm stock were injected with 10 ml of sterile normal neutral horse serum per 1800 gm body weight (McMillan, 1958; McMillan and Engelbert, 1963). The injection was given by way of marginal ear veins with a sterile 26-gauge hypodermic needle. Cultures in v i m were made from a small explant of spleen or lymph nodes or thymus, approximately 1 mm3, placed on a sterile glass slide with a drop of medium. The medium was either sterile normal neutral horse serum and Earl's modification of Tyrodes solution 1:1, or serum from blood of the donor animal. Later, when it became available, we used horse serum ultrafiltrate and also Connaught Laboratories synthetic medium 1066 (courtesy of Dr. R. C. Parker and Mr. Healey). The 1066 was used sometimes with calf serum 1:l. A 22 X 40-mm sterile cover glass of zero thickness was placed carefully over a drop of medium containing the cells. Sometimes the cover glass was lifted slightly by placing sterile pieces of zero coverslips under it. The cover glass was always sealed on with hot paraffin. The culture preparations were placed in an incubator at 37.5"C, or examined immediately on a warm stage registering 33" or 35°C mounted on a Reichert Zetopan microscope equipped with positive' and negative phase contrast. An Arriflex 16-mm motion picture camera loaded with 100 ft of Eastman Tri-X-Reversal safety film, Type F278, was mounted above. A 1OX Leitz Periplan ocular was mounted in a Micro Ibso attachment especially fitted for the Arriflex camera. The normal incubator temperature of 37°C would have made it necessary to use high speeds and high light intensities in order to follow the cell movements with cinematography. As we wished to avoid cell damage that might occur with high light intensities we lowered the temperature as explained. Another culture method (McMillan, 1958; McMillan and Engelbert, 1963) consisted of taking small whole fragments of mesenteric lymph nodes and culturing them in a Maximow slide, the depression filled with horse serum and Earl's modification of Tyrodes solution 1:1. A large sterile cover glass was placed over the culture chamber; the slide was then placed in a sterile petri dish with strips of

100

VIBEKE E. ENGELBERT

blotting paper soaked in sterile distilled water to minimize evaporation of the culture medium. The culture chambers were kept in an incubator at 37°C. Imprints were made five to one slide on the day the experiment was started. Six cultures were set up for each of five rabbits. Each day for 11 days imprints were made on sterile slides with the cultured tissue under sterile conditions as usual. In this way 20 or 30 sets of imprints were made from the same piece of tissue over a period of 11 days.

VI. Behavior and Morphological Variations in Blast Cells and Their Nuclei In 1938 Jordan wrote, “It must be emphasized that the cytologic features of a hemocytoblast are not static; the cytology is subject to considerable variation. ” Unfortunately, however, Jordan did not describe or illustrate the considerable variations. We believe that we have seen “the considerable variation” in practically every slide prepared, regardless of the method (we used all methods) or the type of mammal employed [we used rabbits, rats (Wistar strain), mice from Connaught Laboratories or inbred C57 mice from Jackson Laboratories, field mice, hamsters, guinea pigs, artic lemmings, a Canadian racoon, an American opossum and also leukemic mice (AKR) before the disease appeared]. One of the extreme variations consisted of a lengthening or stretching of individual nuclei; sometimes the stretching made the nucleus appear thin and threadlike. Besides the above variation we found the usual rounded nuclei. These often appeared with very little cytoplasm. The stretched nuclei always appeared with very little cytoplasm. Cultures in vitro and viewing of single cells or nuclei with high-power phase contrast optics plus cinematography allowed us to follow and photograph the changes and see the actual stretching of the nucleus. The movements exhibited by the nucleus were often in a spiraling fashion. Lewis (1931) showed a spiraling movement of myeloblasts which he cultured in v i m (see also Engelbert, 1956, 1958, 1960, 1961, 1967, 1971; McMillan, 1958; McMillan and Engelbert, 1960, 1963; Shelley, 1961; Engelbert and McMillan, 1962; Shelley et al., 1969; Westermann et al., 1970). While the nucleus in our cultures was extended or stretched, intranuclear divisions of small nuclear bodies constantly took place. Eventually nuclear granules, the products of the divisions, were released to the medium (Engelbert and McMillan, 1962; Engelbert, 1967). After being extended for a considerable time, the nucleus rounded up and a rim of cytoplasm appeared around it. The nucleus remained in this state for 10 minutes or more displaying as it were the textbook morphology. The rounding up stage appeared as a resting stage with no special activity visible until the nucleus again assumed the stretched out stage. These alternative changes went on for

ALTERNATIVES TO CLASSICAL MITOSIS

101

many hours and could go on for 1 day and 1 night or longer. In imprints the stretched nuclei could be seen radiating from the associated tissue, as pins from a pincushion. Figure 1 (McMillan and Engelbert, 1960) shows an elongated nucleus extended to a length of 120 pm. The twist one sees on this extended nucleus often looks as if the nucleus made a half-turn during the stretching period, part of the spiraling motion. In their study “The Development History of the Plasma Cells in the Lymph Node of the Rabbit,” McMillan and Engelbert (1963) present tables in which numbers of elongated and contracted blast cells are included in all the cell types found in normal rabbit lymph nodes as well as after injections of horse serum. They also present several graphical text figures in which the relative frequencies of elongated and contracted blast cells are shown in relation to the relative frequencies of other cells in rabbit lymph nodes. In their Fig. 333 Lucas and Jamroz (1961) show, in the thymus of a 35-day-old chick, elongated threadlike nuclei in two areas of the illustration. Both are labeled “smudged nuclei,” although there is no evidence of damage. In their Fig. 332 of a chick embryo thymus, a long threadlike nucleus passes over a red cell and over several lymphocytes from the upper middle of the figure toward the

FIG. 1 . Imprints from germinal centers in the white pulp of spleen of a rabbit injected 85 minutes previously with [3H]thymidine. Arrows indicate fusing vesicles. A very elongated well-labeled nucleus as well as a round well-labeled nucleus can be seen. (From Engelbert, 1967.)

102

VIBEKE E. ENGELBERT

left corner. None of these cells are described. Another stretched nucleus, not threadlike but fairly broad, is partly hidden by an unlabeled cell with a magenta nucleus and deep blue cytoplasm. Figure 331 of the spleen from a 35-day-old chick shows a threadlike nucleus reaching from the right lower part of the figure to the upper middle. Figure 329 of an embryo spleen shows an elongated nucleus in the upper left-hand portion .of the figure. A second elongated nucleus, partially damaged, is found in the upper middle of this figure. In experiments carried out in our laboratory by Dr. Jean E. Mills (Westermann), 12 young rabbits, weighing 500-600 gm, were injected intravenously with tritiated thymidine; 12 exposure times were maintained (see Table I). From each of the 12 exposure times, 25 labeled cells from randomly chosen areas in each of six imprints of spleen were counted and classified on the basis of morphological shape into three variants: round, oblong or irregular, and elongated. When these figures were analyzed statistically for correlation (rank difference method) (Davenport and Ekas, 1936), the results suggested that an inverse relation existed between the numbers of labeled rounded or elongated cells ( r = -0.3), i.e., where numbers of rounded cells or nuclei are high, those of the elongated nuclei are low, and vice versa. From these results we must assume that the round and elongated or extended nuclei are members of the same cell population, but each shape expresses a different activity stage, such as described earlier. From the evidence presented so far we believe we must acknowledge that nuclei of lymphoid cells normally change their shape during their life history and these changes, when they are “caught” with fixatives or with cinematography of TABLE I OCCURRENCE I N RABBIT SPLEENOF CELLSWITH VARYING MORPHOLOGY CARRYING LABEL AFTER DIFFERENT TIMESOF EXPOSURE TO TRITIATED THYMIDINE

Rabbit no.

Time exposed to tritiated thymidine

Cell variant

Round

1 2 3 4 5 6 7 8 9 10 11 12

5 minutes 40 minutes 85 minutes 2 hours 4 hours 8 hours 12 hours 24 hours 2 days 4 days 8 days 12 days

80 86 53 37 61 59 41 61 45 64 72 37

Oblong or irregular 36 43 40 78 48 52 52

40 31 55 45 63

Elongate

Total no. of cells

34 21 57 35 41 39 57 49 74 31 33 50

150 150 150 150 150 150 150 150 150 150 150 150

ALTERNATIVES TO CLASSICAL MITOSIS

103

cultures in vitro, cannot be “shrugged off” as degenerative stages or artifacts (Engelbert, 1967). During the stretching and twisting period the nucleus also releases vesicles that appear colorless with the MGG stain. Shelley (1961) demonstrated intranuclear vesicles in blast cells in rabbit lymph nodes with different fixatives and staining techniques. The study was limited to “seemingly naked nuclei” of round to ovoid form, with the long axis no greater than three times the short axis, since they make up about 77% of blast cell nuclei in a mesentric rabbit lymph node. Shelley defined an intranuclear vesicle as a pale-staining or nonstaining area, 1 p m or greater in diameter, which is enclosed by the nucleus for more than half its circumference. She found that with the May-GriinwaldGiemsa staining method, iodine vapor and methanol were equally valuable as fixatives for demonstrating intranuclear vesicles. Neutral formalin fixation and MGG staining of imprints showed statistically fewer vesicles in the nuclei. Neutral formalin kills the nucleus very slowly; the nucleus contracts squeezing out the vesicles. It is therefore necessary to fix the nuclei very quickly to maintain their morphology and contents (Engelbert, 1967). Pearse (1960) considers that formalin is the best protein fixative, but we find it is a very poor fixative for nuclei. When staining with hematoxylin following fixation with either iodine or formalin vapor, Shelley found no differences in the proportions of nuclei containing and not containing vesicles. Hematoxylin will stain vesicles containing protein alone. From the statistical analysis of intranuclear vesicles in lymph nodes of seven normal rabbits. Shelley et al. (1969) concluded that: (i) The rabbits used were homogeneous with regard to the number of blast cell nuclei displaying intranuclear vesicles. (ii) The size of the intranuclear vesicle was independent of the nuclear size. (iii) Nuclei with a larger nuclear index (width X length) tended to display a greater number of vesicles. When vesicles are viewed with high-power and negative phase-contrast optics they display a grayish color (Engelbert, 1960, Fig. 9). With the mercury bromphenol blue method (Mazia et al., 1953) the contents of vesicles stain blue indicating that the contents are protein (Pearse, 1968). During the process of elongation and stretching the nuclei release much of their nucleoplasm (Engelbert, 1958). The nucleoplasm when released has the typical appearance of ‘‘vacuolated cytoplasm or ‘‘degenerative cytoplasm (Dacie and White, 1949, Pl.IV, Figs. 2 and 4). The “vacuoles” these authors studied in erythropoiesis in human bone marrow we believe are “our vesicles,” which we shall eventually show play an important role in the formation of new cells and new nuclei in hemopoietic tissues. Downey and Weidenreich (1912) showed that lymphocytes released pieces of their cytoplasm and they believed it was a normal property of these cells. Weill (1913) and Williamson (1950) demonstrated similar phenomena. Nuclear fragments of lymphocytes were shown in culture preparations by Popoff (1927), Tschassownikow (1927), and Emmart (1936). ”

”

104

VIBEKE E. ENGELBERT

In our laboratory Blair (1955) observed 5590 normal lymphocytes from thymi of mammals and chick embryos. She counted 4591 cells that showed one or more tubes radiating from each cell; 551 nuclear bodies were found inside tubes; Engelbert (1953) found tubes extending from the thymic cells cultured in vitro. White (1947-1948) and Frank and Dougherty (1953) reported “cytoplasmic budding” from lymphocytes. In a paper published in 1882 by Watney on the thymus, one illustration clearly shows two cells with fairly long tubes. Watney did neither label nor describe the two cells, but the details of his drawing were carefully camed out. Westermann et al. (1970) described protoplasmic fragments in hematopoietic tissues and an analysis of intranuclear vesicles in lymph node blast cells of the rabbit. They demonstrated that vesicles and cytoplasmic fragments are found in greatest number in the lymph node and spleen and are least common in the thymus and bone marrow in the rabbit. Vesicles appear to originate by the extrusion of intranuclear and intracytoplasmic vesicles mostly from cells of the lymphoid series. Vesicles and cytoplasmic fragments are absent from blood smears and extremely difficult to recognize in sections or in areas of imprints where the cells are closely applied one to another. These authors also suggest that chromatin from stretched nuclei and smaller free chromatin masses may become transferred to free vesicles and this process may function in new cell formation. Wedlock (1974) carried out an extensive investigation of differentiation of hemopoietic cells in the thymus, bursa of Fabricius, spleen, and bone marrow of chick embryos and hatched chicks (Callus domesticus) from 14 days of incubation to 16 days after hatching. Further, the cytology of normally developing organs was compared to that of organs stimulated by the bacteriophage 0x174, and numbers of all the different cell types were counted in imprints. She also used fluorescence microscopy as well as radioautography after injections with tritiated thymidine. Five chicks were used as controls for each stage. Five were used for experiments with phage for each age. Each treatment or staining method was studied in chick embryos at 14, 16, and 18 days of incubation and immediately before hatching and in young chicks at 4, 8, and 16 days after hatching. She described nuclei of three morphological types: long extended nuclei containing vesicles, spherical nuclei with webbed chromatin, and spherical nuclei with homogeneous chromatin and two nucleoli. Her illustrations show the typical long extended nuclei we have described. She reports that in the thymus, 76%of the cells counted at 14 days of incubation have very little cytoplasm. This decrease to 20% at hatching and rises to 36% at 16 days after hatching. During the entire period the lymphocytes increase steadily from 8 to 55%.In the bursa of Fabricius, 57% of the cells counted at 14 days of incubation have very little cytoplasm. This decreases to 36% at hatching but increases to 53% at 16 days. Lympnocytes increase from 8% at 14 days of incubation to 38% at 16 days after hatching. In the spleen 54% of the cells counted at 14 days of incubation were

ALTERNATIVES TO CLASSICAL MITOSIS

105

nuclei with little cytoplasm. Their number decreases to 13% at 18 days of incubation, but at hatching they constitute 22% and at 16 days after hatching, 34%. In bone marrow at 14 days of incubation nuclei with little cytoplasm form 31%, at hatching 15%, and at 16 days after hatching 6%. Lymphocytes in the same period increase from 2.6 to 24% at 16 days after hatching. Wedlock tested the cell counts in normal and in phage-stimulated chicks with analysis of variance. Calculations were performed by a C.G.E. time-sharing computer service. The two-factor analysis used the counts of embryonic cells and the three-factor analysis of variance used the counts of cells of hatched chicks. Probability value of F ratio: (i) indicates a significant difference in cell populations at the 5 % probability level; (ii) indicates a significant difference in cell population at the 1% probability level; and (iii) indicates a significant difference in cell populations at the 0.1% probability level.

VII. The Occurrence and Significance of Two Spatially Separate Nuclear Masses in Blast Cells and in Differentiating Cells During early research work (Engelbert, 1956) it was often observed in in vitro cultures of normal rabbit spleens that some of the elongated nuclei released nuclear and cytoplasmic masses, through a tubelike opening at one end. In fixed imprints the same behavior was also observed. Sometimes the elongated nuclei were slightly broken and part of the nuclear wall lifted away. In such cases one could see small nuclear bodies lying inside the elongated nuclei. Rounded nuclei could at certain stages also be sufficiently “nonplastic” to crack open, even with a gentle imprint technique. In these cases one saw clearly that the damaged part of the nucleus constituted part of a shell-like or peripheral layer, inside which two well-developed nuclear masses were hidden (Engelbert, 1956, 1970). The shell-like or peripheral layer had small nuclear granules on both its inner and outer surfaces. These granules appeared to be produced by the “shell.” Various stages of such nuclei seen often over the years clearly demonstrated that all of the shell-like or peripheral nuclear portion finally became individual free granules and that the inner nuclear masses formed the nuclei of granulocytes. When the granules began to appear on the peripheral nuclear mass, its future existence as a cohesive mass seemed only of short duration. Often one sees only small pieces of the peripheral or shell-like nuclear mass. The granules lying in the cytoplasm are typical of those in granulocytes. In textbook illustrations of neutrophilic and eosinophilic “myelocytes” (Bloom and Fawcett,

106

VIBEKE E. ENGELBERT

1968), one sees mainly the granular mass; only in later stages can the nuclei be seen, as the granules become spread farther apart. In view of this evidence we decided to run experiments with tritiated thymidine in order to find out if the two separated nuclear masses would label and thus indicate DNA synthesis. We had observed that the blood of birds (Galfus domesticus) often carried a good number of immature granulocytes, which displayed the shell-like or peripheral nuclear layer. We therefore used birds of Leghorn stock for the experiments. Chick embryos, newly hatched chicks, and pullets weighing 1500-1800 gm were used. Three pullets were injected intravenously using a brachial vein. Tritiated thymidine was injected into the coelom of 16 newly hatched chicks and into the vascular bed behind the eye of 13 chick embryos. Ten embroys, eleven newly hatched chicks, and one pullet served as controls. Imprints were made of spleens and brush smears were made of the blood. After the injections of tritiated thymidine label was found on both elongated and rounded nuclei similar to that reported for mammals (Section VI) (Engelbert, 1967). Eighteen-day chick embryos carried heavy label on the shell-like or peripheral nuclear layer of differentiating cells. If large or small pieces of this mass remained the cells were called “immature. ” If granules were fully formed and no shell-like nuclear mass remained the cells were called “mature.” For each animal used 500 granulocytes or all of the granulocytes found in 6 to 10 samples were counted. After 30 minutes of exposure to tritiated thymidine imprints from 18-day embryo spleen had 100 granulocytes; 16 were mature with label on the granules, and 37 were immature cells with label on the granules and on the remains of the shell-like nuclear layer. From one pullet blood samples were taken 3, 6, and 24 hours after injection of the isotope. At 3 hours, 275 granulocytes were counted: 173 mature cells with label on the granules, 9 mature cells with no label, 75 immature cells with label, and 18 immature cells with no label. At 6 hours, 500 granulocytes were counted: 493 mature cells with label on the granules, and 7 mature cells without label. No immature stages were found. At 24 hours, 10 samples were scanned; some cells had label but most had none. A second pullet killed 1 hour after injection had so few granulocytes in 10 samples that no counts were made. Intense erythropoiesis was present however (Smith and Engelbert, 1969) (see later). A pullet killed 2 hours after the injection had 500 granulocytes: 88 mature with label, 74 mature without label, 166 immature with label, 172 immature without label. A pullet killed as a control without an injection of tritiated thymidine had 41 eosinophils, 29 of which were mature and 12 immature, and 160 heterophils, 132 of which were mature and 28 immature. There were so few basophils in the samples that they were not reported. The eosinophils have rounded granules; the heterophils have rice-grain-shaped granules. In control animals one could easily see the difference between heterophilic and eosinophilic granulocytes, but in slides prepared with radioautography it was

ALTERNATIVES TO CLASSICAL MITOSIS

107

often difficult to determine the shape of the granules. However, in a good number of cases we could see the difference and both rounded and rice-grainshaped granules carried label. The label on the granules consisted of a single small “cap” on each. The label on the shell-like nuclear mass appeared as paired grains or as several joined grains which indicates DNA synthesis while granules were being formed (Engelbert, 1970, Fig. 4). On the nuclei of the granulocytes, the inner nuclear mass did not carry any label. With the Feulgen or Schiff method of staining a small area on each granule is “Feulgen or Schiff positive”; this presumably is the DNA which labels with tritiated thymidine. The mature granulocytes eventually releases granules either individually or in a small mass. This can be seen clearly in in vitro cultures and also in imprints. Finally, the grain is detached from the granule and the small mass of labeled DNA becomes free in the blood. We have seen many granulocytes in mammals with label on the granules, but never with label on the nuclei. We believe that the peripheral or shell-like nuclear mass, which produces granules, is a morphological example of what Roels (1966), in his extensive review of “the variability of DNA,” calls “metabolic DNA.” Roels wrote, “one may explain these variations by accepting two types of DNA: a staple one with genetic function and a labile one with metabolic function. ” In a paper entitled, “Turnover of DNA and Function,” Pelc (1968) wrote, “the metabolic DNA of a given type of differentiated cell consists of extra copies of the genes which are active in the cell; the metabolic DNA is the working DNA, which regulates and performs the transcription of RNA and possibly other functions of DNA, while active molecules of metabolic DNA are subject to wear and tear and are periodically renewed. DNA can thus be labeled during three periods: premitotic synthesis, formation of metabolic DNA and renewal or repair. The inner nuclear mass which becomes the nuclei of granulocytes we regard as genetic DNA and in normal tissues we have not seen these nuclei with label.

”

VIII. Fate of the Peripheral or Shell-like Nuclear Mass, and the Inner Nuclear Mass, in Differentiating Cells of Leukemic Mice (AKR Strain) Imprints of spleen, thymus, and lymph nodes from AKR mice, with detectable enlargements of the inguinal or axiliary lymph nodes, had many immature plasma cells with a broad border of basophilc cytoplasm and the pale central area characteristic of the developmental stages of this cell. It is known to appear in response to antigens. The Gooch virus is considered the causative agent in murine leukemia (Metcalf, 1966). Fully developed plasma cells were rare or missing entirely. The immature plasma cells shed their cytoplasm to a large extent. The blast cell nuclei of

108

VIBEKE E. ENGELBERT

elongated shape were split lengthwise into individual nuclear strands in many cases. Intact blast cell nuclei extruded basophilic granular substance and also nucleoli. The vesicles produced were not the normal type which do not stain with MGG and thus appear colorless. Instead the blast cell nuclei produced a great many hemoglobin-containing vesicles. In cats with leukemia the blast cells also produced great numbers of hemoglobin-carrying vesicles (V. E. Engelbert, unpublished). In leukemic mice the peripheral nuclear mass labels with tritiated thymidine, but the label appears to consist of single grains, and not of several joined grains, as in normal animals. The normal relationship between the shelllike or peripheral nuclear mass and the inner nuclear mass is not maintained, even in these relatively early stages of murine leukemia. The peripheral nuclear mass “expels” as it were the inner nuclear mass. The latter has a ring-shaped or lobed nucleus surrounded by pale nongranular cytoplasm. The expelled mass is “the leukocyte” seen commonly in leukemias. The shell-like nuclear mass, now an empty shell, gradually breaks up and small labeled pieces or fragments can be found scattered over the imprint. In the later stages of murine leukemia extreme enlargements of thymus, spleen, and lymph nodes take place. At this time most of the cells in these organs undergo classical mitosis. The cells undergoing mitosis appear to be enlarged plasma cell nuclei; these cells have no cytoplasm. Elongated naked blast cell nuclei are still often found and may transform into plasma cells (McMillan and Engelbert, 1960, 1963). Other cell types are not found at this stage. The drug cytosine arabinoside, produced by the Upjohn Co. and tested by them on animals with leukemia, was used by us in order to see if any cytological effect could be found. Dr. E. L. Masson of the Upjohn Co. of Canada gave us a sample as well as a copy of the company’s unpublished records of its use and effect. After four daily intraperitoneal injections of 0.5 ml tripledistilled water in which 20 mg of cytosine arabinoside was dissolved, nine mice were examined as to weight of spleen, thymus, and lymph nodes. Imprints were made as usual from the three organs. Weights of the three organs were mostly normal or nearly normal. Imprints of the treated mice showed a much more scanty cell population; they were less dense than imprints of nontreated mice, as if a large part of the cell population had been expelled from the three organs. Plasma cells were reduced in number, but some were still found even in mice treated early (as soon as enlargement of lymph nodules could be detected). The plasma cell persisted especially in the thymus. The shedding of their cytoplasm also persisted. The splitting of the elongated blast nuclei was much decreased. The hemoglobin vesicles were often entirely absent and the normal colorless (with MGG) vesicles were back in large numbers. Classical mitosis was not seen in the treated “apparently healthy looking” mice. However, the most significant change was that the shell-like nuclear mass and the inner nuclear mass appeared to maintain their close relationship. Differentiation was thus not totally interrupted as in the nontreated mice, in which the two nuclear masses, peripheral and

ALTERNATIVES TO CLASSICAL MITOSIS

109

inner, became separated and the first one was destroyed. The label on cells in the spleen was in paired or small clumps of grains, which was a sign that differentiating cells “were working” on DNA synthesis. Normal granulocytes were found in bone marrow. The work with leukemic mice is quoted here because it sheds additional light on the importance of the two associated nuclear masses, the peripheral and the inner mass, and their importance in differentiation (Engelbert, 1971).

IX. Results following Injection of Tritiated Thymidine A. REASSOCIATION OF VESICLES A N D NUCLEAR GRANULES I N RABBITSPLEENS In Section VI, production of vesicles by blast cell nuclei was reported fully and the work of Shelley, Westermann, McMillan, and Engelbert described. The elongated nuclei release their vesicular contents into a central mass or “nest” of vesicles. The nuclei radiate out from the edge of this central mass. Soon however the nuclei move into the mass of vesicles and come to lie close to vesicular membranes. Figure 1 shows an elongated well-labeled nucleus “caught” as it moves between vesicles; one end of this nucleus adheres closely to a large partially stained vesicle on the left. In the same figure small vesicles can be seen fusing (arrows) and forming larger vesicles. In Fig. 2 the black masses surrounding the colorless large vesicles (Ve) are the labeled densely packed elongated nuclei. In a vesicle at the upper center (Ve) labeled nuclear granules can be seen entering the colorless vesicle. The individual grains cannot be seen on the surrounding nuclei. When labeled nuclei are condensed or contracted, they appear completely black as in Fig. 2. In Plate 3 of Engelbert (1967) one can see both condensed black areas and stretched areas of the same nucleus. In the stretched part one can count the individual grains. Figure 2 in the present paper presents a large vesicle at the right side with a good label (Lb). One can see the grains are two to four times the size of individual grains, indicating that further DNA synthesis is taking place. A vesicle in the lower right-hand comer appears to be entered by labeled nuclear granules around its periphery. It may be reasonable to assume that the nuclear granules enter the vesicles in vesicular blebs formed from the vesicular membrane. In Fig. 3 in the upper center a vesicle with a good label and stainable content (MGG) has along its lower periphery the remains (Nb) of one of the nuclei or “the” nucleus which contributed its contents of labeled DNA. We now call such vesicles “vesicular nuclei. Other almost colorless vesicles are surrounded by black rims of labeled nuclei. In Fig. 4 a well-labeled vesicular nucleus in DNA synthesis shows the remains (Nb) of a contributing nucleus on its upper left periphery. Stained but not labeled vesicles are also present. ”

110

VIBEKE E. ENGELBERT

FIG.2. Vesicles (Ve) with a tight rim of labeled nuclei (mows) around them. Gradual entrance of nuclear material into vesicles can be detected by grains above lightly labeled vesicles and staining of the vesicular substance-LLheavy label on a former vesicle that now has acquired labeled DNA. (From Engelbert, 1967.)

It should be mentioned at this time that we have seen the morphology and behavior of nuclear granules in our work with live preparations of in vitro cultures. Individual granules divide first forming clumps because they lie close together. Shortly after they part. We believe that the label on the vesicular nuclei means that these cells will differentiate, forming either granulocytes, erythrocytes, or lymphocytes. Thomas (1959) and Yoffey (1960) suggested that this label meant division of the whole cell. Andreasen (1959) reported ‘‘degeneration of lymphocytes” which we believe was due to his observation of the large nonstaining vesicles shown above. He also thought that the degeneration was “combined with phagocytic activity of the reticulum.” We have not seen cells of the reticulum often in imprints, but we have sometimes seen their nuclei in imprints on Millipore filters. The author wrote in 1960 that “reticular cells and other mesenchymal elements, which form connective tissue in the animal body, constitute a group of cells, where normal mitosis is

ALTERNATIVES TO CLASSICAL MITOSIS

111

FIG. 3. Nuclear rim (arrows) around vesicles (Ve). Nb, Remains of nuclear border or nuclear rim around a well-labeled vesicular cell. (From Engelbert, 1967.)

relatively easy to demonstrate. However, we still need to know how blood cells are formed. ” Yoffey said “the term reticular has been fraught with such difficulties that I avoid it. But you start off with a primitive cell, which goes through a number of divisions” (Yoffey et al., 1959, p. 58). In the same discussion Yoffey denies that “extensive pycnosis” is found in “germinal centers.” He says further, “Frankly we don’t believe that you find very many of these cells dying in healthy animals. ” We do not believe in the death of lymphocytes either. The idea of “pycnosis” may have originated from the fact that, in contracted form, the blast cell nuclei stain very dark and appear totally black when labeled. We believe that the stages of vesicles and “vesicular nuclei” are the cells Yoffey (1973) calls “transitional cells” and lists as “pale transitionals, “basophilic transitionals, ” and finally “blast cells. ”

”

B. GRAINCOUNTS IN VESICULAR NUCLEI OF RABBITSPLEEN AND BONE MARROW Figure 5 presents grain counts in individual cells from spleens of four rabbits killed 5 minutes, 85 minutes, 2 hours, and 1 day after injections with tritiated

112

VIBEKE E. ENGELBERT

FIG.4. Nb, Remains of nucleus that have contributed DNA to the well-labeled vesicular cell. (From Engelbert, 1967.)

thymidine (Engelbert, 1967). Total labeled cells counted are shown at the upper left in each square. The lower grain counts shown as short columns represent vesicles which have begun to accumulate nuclear granules from the blast cells surrounding them (see Fig. 2). The tall columns with grain counts above 10 grains present vesicular nuclei that have not only accumulated nuclear granules from blast cells surrounding them, but in which the nuclear granules have begun “mini-mitoses,’’ and thus have started the second growth phase of vesicular nuclei. The grain counts 1 day after injection of tritiated thymidine show that 60% of the vesicular nuclei examined have entered the second growth phase. If one examines Figs. 1-4, it is clear that vesicles with a high grain count are all in the second growth phase, increasing DNA content through mini-mitoses. Clumps of two, four, or more granules originate through such divisions. From the clumps of granules individual granules will move away and possibly divide again later (see basic nuclear units, Engelbert, 1956). The decrease of grain counts in Fig. 5 must therefore not be regarded as a “dilution effect,” where the

2or 113

ALTERNATIVES TO CLASSICAL MITOSIS

0

-

100

6RAIN COUNT: INTERVALS OF 20

FIG.5 . Grain counts in individual cells from spleens of four rabbits killed 5 minutes, 85 minutes, 2 hours, and I day after injections of tritiated thymidine. (From Engelbert, 1967.)

cells with lower grains counts are division products of cells with higher grain counts. The differences in grain numbers counted in the splenic cells from the four rabbits must be regarded and evaluated as cells in growth, in which the increase and accumulation of DNA is the main function taking place. This increase of DNA content eventually leads to various stages of differentiation whereupon different blood cells will be produced by the spleen. In tissue cultures I have seen vesicles gradually accumulate small granular bodies (V. E. Engelbert, unpublished). The most frequent grain counts in spleen and bone marrow cells of the four rabbits were compared in Fig. 15 of Engelbert (1967). At 1.5 hours after injection of tritiated thymidine, the cells of the spleen had the highest grain count, 110, the bone marrow cells, 25. At 3 hours and continuing to the second day the most frequent splenic grain count was 15 grains per cell. The bone marrow counts varied: 30 grains at 3 hours, 10 grains at 7 . 5 hours, 25 grains at 12 hours, and 15 grains at 1 day after injection of the isotope. The grain counts in both organs then declined steadily giving 5 grains or less on the 9th day. To obtain the best measure of DNA accumulation and consequential growth in hemopoietic cells, grain counts after one injection of tritiated thymidine should be taken early during the first day and not later than 24 hours after the injection.

114

VIBEKE E. ENGELBERT

Thomas (1959) and Yoffey (1960) interpreted cells with lower grain counts as products of direct mitosis of the cells with higher grain counts. Such an interpretation does not agree with the work presented here.

X. Erythropoiesis in Blood of Vertebrates with Nucleated Erythrocytes. Formation of Clone Cells from Nuclei of Young Mature Erythrocytes In 1960, in studies of blood semars of nucleated erythrocytes of birds, amphibians, and fishes, I suggested that the nucleus of the red blood cell in its younger stages behaved as a mother cell. Pouchlike extensions of the nuclear membrane with nuclear contents pinched off outside the mother cell. Recently, a paper by Komocki (1929) was brought to my attention. This author wrote “Uber die Abstammung der Erythrocyten der niederen Wirbeltiere von den sogenannt nackten Kernen. He had published several earlier papers on his work with turtles and, later, worked with salamanders. Morashita (1 957) believed that leukocytic cells were produced by extrusion of cytoplasm from the nucleated erythrocyte of the toad. He did not mention any involvement of the nucleus in this process. We have observed intranuclear divisions in living erythrocytes. The small nuclear granules divide one at the time; then one part moves away toward the nuclear membrane leaving a “whitish track” in the nucleoplasm. In imprint preparations it is very easy to see when these intranuclear divisions are taking place because the whitish tracks show up very well. Lucas and Jamroz (1961, Fig. 228) show in detail nuclei of erythrocytes from blood of a chick embryo heart on the 10th day of incubation. The small nuclear granules and also the whitish tracks which we regard as a sign that “mini-divisions” are taking place within the nuclear membrane are very clear in this illustration. These authors, however, use terms from mammalian hematology to describe the various erythrocytes. Smith (1969) described and illustrated the formation of small nuclear buds or blebs from erythrocytes in peripheral blood of chick embryos and hatched chicks. She used camel-hair brush smears and four staining methods, May-Griinwald-Giemsa, fluroescence with acridine orange stain, mercury bromphenol blue stain, and the Schiff (or Feulgen) method, as well as labeling with tritiated thymidine. She showed that more than one bleb could originate on a nucleus at the same time although one at a time seemed usual. The bleb finally became a pouch at its outer end. While the May-Griinwald-Giemsa method showed development and finally the free new cell very well, the fluorescence method with acridine orange stain presented the best detail. In the latter preparations the mother cell nucleus fluoresced yellow and the hemoglobin black as it absorbed monochromatic light (Nairn, 1962). The blebs, while still attached to the mother nucleus or finally free, also stained yellow. The free new cells, the ”

ALTERNATIVES TO CLASSICAL MITOSIS

115

“clone cells,” soon showed that the yellow material had concentrated and stained brighter yellow in the center in which the new nucleus was forming. The surrounding “cytoplasm” while still yellowish began to show portions of brick red color indicating the formation of RNA. Soon all the cytoplasm was brick red and the nucleus, dense and yellow, the same color and shape as in other cells. The cytoplasm gradually became black as hemoglobin was formed. The mercury bromphenol blue stained the mother cell, but left the blebs completely colorless. The Feulgen method stained the nuclei of erythrocytes well, but the early stages of clone cells although magenta in color were pale. However, the fact that they stained as DNA instead of protein (bromphenol blue) we believe is significant. Smith counted over 86,000 cells and only one red cell showed any resemblance to pro- or metaphases, but no ana- or telophases were found. The frequency of clone cells ranged from 3.7 to 6 and 8%. Lucas and Jamroz (1961) reported 0.2% cells in mitosis in chicken bone marrow. The same authors also show two clone cells (Fig. 226, Nos. 15, 16) still attached to the mother erythrocyte; both are called “smudged primary erythrocytes.” In Fig. 2 a clone cell between two erythrocytes (No. 6) is called a squashed erythrocyte nucleus. In our photographs the early stages of clone cells in Fig. 12 of Smith and Engelbert (1969) form as basket-like structures. This, however, is a temporary stage and soon the strands form a bleb. During differentiation of the clone cells into erythrocytes in chicks Smith found that approximately 50% of the clone cells remained adherent to the mother cell. Both erythrocytes and clone cells were labeled in peripheral blood of chicken exposed to tritiated thymidine for 24 hours. Deutsch (1970) investigated clone formation in the peripheral blood of the white sucker, Cutostomus commersoni, after labeling with tritiated thymidine for various exposure intervals. He found that 10% of the cells of peripheral blood were clone cells. Both clone cells and erythocytes labeled after exposure to the isotope and 13 to 18% of the saclike extensions of clone cells (called Stage 11) had mean grain counts of 235 to 301 grains per cell. The greatest uptake of tritiated thymidine was observed about the forty-eighth hour and the fifth day of exposure. Counts of clone cells in peripheral blood of seven species of New Zealand birds and one species of Canadian birds were made by Engelbert and Young (1970b). In blood from one kiwi, 18% of the 1500 cells counted were clone cells. In kiwi No. 2 (8 years old), of 2000 cells 41% were clone cells; in weka No. 1, of 1500 cells 23% were clone cells; in weka No. 2, of 2400 cells, 10% were clone cells; in weka No. 3, of 1500 cells, 14% were clone cells. In yellow-eyed penguin, of 2000 cells, 5% were clone cells. Blue or fairy penguin (on nest with two eggs) had, out of 2000 cells, 12% clone cells. Takahe (Notorismantelli) had, out of 2000 cells, 15%clone cells. h k e h o No. 1 (broken leg) had, out of 1500 cells, 5% clone cells. h k e h o No. 2 (healthy) had, out of

116

VIBEKE E. ENGELBERT

2000 cells, 12% clone cells. Kea had, out of 2000 cells, 7% clone cells. The young white-throated Canadian sparrow had 18% clones in 2000 cells; the mature white-throated Canadian sparrow had 12% clones in 2000 cells. In studies of erythropoiesis in peripheral blood of tuatara (Sphenodon punctatus) and turtle (Muluclemys terrapin), Engelbert and Young (1970a) found the following: Sphenodon punctatus

Clone cells Immature clone cells with hemoglobin forming Mature clone cells Lymphocytes Granulocytes

No. 1

No. 2

No. 3

No. 4

29 1

165

67

160

90 1653 13 9

I789 23 7

I880 37 6

1813 8 11

Malaclemys terrapin

Clone cells Immature clone cells with hemoglobin forming Mature clone cells Lymphocytes Granulocytes Thrombocytes were observed in both species but not counted

No. I

No. 2

No. 3

149

150

I83

235 I101

10

126 1213 Very few, not counted 10

100 1241 5

Recently I examined slides of blood from a New Zealand tree frog; the clone cells were so numerous that over 60% of erythrocytes on the slides had clone cells attached. Examination of numbers of clone cells in fish and birds might be important in estimating the condition of health of these animals; the same can be applied to reptiles and amphibians where ecology and the balance of nature are of interest. One can say with conviction that the clone cell has been present for a very long time as Sphenodon has a history of over 200 million years. It is an important feature of blood cell formation and should not be ignored as it has to date by biologists and hematologists; it could be of great help to them and to the species involved.

ALTERNATIVES TO CLASSICAL MITOSIS

117

XI. Summary 1 . The search for a blood mother cell, a hemocytoblast, as well as for evidence of mitosis by many workers. 2. Methods used in tracing of behavior and morphological variations in blood cells and their nuclei including in vitro methods and cinematography. 3. The production of vesicles with protein content by nuclei in hemopoietic organs. 4. Two spatially separate nuclear masses, one peripheral the other centrally located. Only the peripheral mass labels with tritiated thymidine in normal animals. 5 . The state of the two nuclear masses in mice with leukemia (AKR strain). 6 . Accumulation by free nuclear vesicles of small nuclear granules in spleen of rabbits. The granules when inside the vesicles undergo mini-mitoses and thus synthesis of DNA takes place. 7. Grain counts in cells of spleen of rabbits after injection of tritiated thymidine show lower and higher grain counts. The lower grain counts indicate the accumulation of nuclear granules by vesicles, thus the first growth stage. The second growth stage is the increase in DNA by mini-mitoses of the nuclear granules. Thus two phases of growth create large nuclear contents. Soon these cells differentiate to form various new blood cells. 8. Erythropoiesis in peripheral blood of vertebrates with nucleated erythrocytes takes place by formation of nuclear buds from young mature erythrocytes. The new cells are clone cells. They can be found in fish, amphibians, reptiles, chick embryos, and adult birds. Sphenodon punctatus and seven species of New Zealand birds were included in this study.

XII. Conclusion Alternatives to mitosis are: (a) metaphases changing directly to nuclei in mammals; (b) development of vesicular nuclei through two growth phases, accumulation of nuclear granules and rapid increase in DNA by mini-mitoses of these granules; and (c) development of clone cells from nuclear buds of nucleated erythrocytes.

ACKNOWLEDGMENTS

I am indebted to Dr. Jean E. M. Westermann, McMaster University, Canada, for valuable criticism, to Professor Donald B . McMillan, University of Western Ontario, Canada, Mrs. Jessica

118

VIBEKE E. ENGELBERT

Shelley, Mrs. Natasha Bikadoroff Smith, Mr. M. Deutsch, and Dr. Diana Wedlock for the use of their material. Table I and Figs. 1-5 are published with the kind permission of the editor of Haematologica Larim, Milano. I am sincerely grateful to principal secretary, Miss A. M. Sorensen, Museum of Natural History, Aarhus, for the secretarial work of Miss Karen Berg.

REFERENCES Andreasen, E. (1959). In ”The Kinetics of Cellular Proliferation” (F. Stohlman, Jr., ed.), p. 19. Grune & Stratton, New York. Andreasen, E., and Christensen, S. (1949). Anar. Rec. 103, 401. Arnold, J. (1887). Arch. Mikrosk. Anat. 30, 205. Askanazy, M. (1905). Zentralol. Allg. Parhol. Parhol. Anar. 16, 897. Blair, M. H. (1955). M.A. Thesis, University of Toronto, Toronto. Bloom, W., and Fawcett, D. W. (1968). “A Textbook of Histology,” 8th ed. Saunders, Philadelphia, Pennsylvania. Bonhag, P. F. (1955). J. Morphol. 96, 381. Bushong, S. C., Watson, J. A., and Atchison, R. W. (1968). Slain Technol. 43, 273. Caro, L. (1964). In “High Resolution Autoradiography” (D. M. Prescott, ed.), Vol. 1, p. 327. Academic Press, New York. Christensen, S. (1950). Acra Anar. 10, 233. Dacie, J. V., and White, J. C. (1949). J. Clin. Pathol. 2, 1. Davenport, C. B., and Ekas, M. P. (1936). “Statistical Methods in Biology, Medicine and Psychology.” Wiley, New York. Deutsch, M. (1970). M.Sc. Thesis, University of Toronto, Toronto. Deutsch, M., and Engelbert, V. E. (1970). Can. J . Zool. 48, 1241. Diggs, L. W., Sturm, D., and Bell, A. (1957). “The Morphology of Human Blood Cells.” Saunders, Philadelphia, Pennsylvania. Downey, H. (1932). In “Special Cytology. The Form and Functions of the Cell in Health and Disease” (E. V. Cowdry, ed.), Vol. 2, p. 653. Harper (Hoeber), New York. Downey, H. (1938). In “Handbook of Hematology” (H. Downey, ed.), Vol. 111, p. 1963. Hafner, New York. Downey, H., and Weidenreich, F. (1912). Arch. Mikrosk. Anat. 80, 360. Dustin, P., Jr. (1959). In “The Kinetics of Cellular Proliferation” (F. Stohlman, Jr., ed.), p. 50. Grune & Stratton, New York. Emmart, E. W. (1936). Anar. Rec. 66, 59. Engelbert, V. E. (1953). Can. J. 2001.31, 106. Engelbert, V. E. (1956). Can. J. Zool. 34, 707. Engelbert, V. E. (1958). Can. J. 2001.36, 131. Engelbert, V. E. (1960). Can. J . Zool. 38, 189. Engelbert, V. E. (1961). Can. J. 2001.39, 367. Engelbert, V. E. (1967). Haemarol. Lar. 10, 65. Engelbert, V. E. (1968). Haemarol. Lar. 11, 349. Engelbert, V. E. (1970). Haemarol. Lar. 13, 1. Engelbert, V. E. (1971). Haernarol. Lar. 14, 1. Engelbert, V. E., and McMillan, D. B. (1962). Can. J . 2001.40, 83. Engelbert, V. E., and Young, A. D. (1970a). Can. J. Zool. 48, 209.

ALTERNATIVES TO CLASSICAL MITOSIS

119

Engelbert, V. E., and Young, A. D. (1970b). Can. J . Zool. 48, 227. Feulgen, R., and Rossenbeck, H. (1924). Z. Phys. Chem. 135, 203. Frank, J. A., and Dougherty, T. F. (1953). Proc. Soc. Exp. Biol. Med. 82, 17. Gowans, J. L. (1957). Br. J . Exp. Parhol. 38, 67. Hall, J. W., and Furth, J. (1938). Arch. Parhol. 25, 46. Hamilton, L. D. (1954). J . Clin. Invest. 33, 939. Hewson, A. (1777). I n “Collected Works.” Sydenham SOC.,London (1846). Hill, M. (1959). Narure (London) 183, 1060. Jacobson, W., and Webb, M. (1952). Exp. Cell Res. 3, 163. Jones, T. W. (1846). Philos. Trans. R. Soc. London 136, 63. Jordan, H. E. (1935). Am. J . Anar. 57, 1. Jordan, H. E. (1938). In “Handbook of Hematology” (H. Downey, ed.), Vol. 2, p. 699. Hafner, New York. Kindred, J. E. (1938). Am. J. Anat. 62, 453. Kindred, J. E. (1940). Am. J. Anar. 67, 99. Kindred, J. E. (1942). Am. J. Anat. 71, 207. Komocki, W. (1929). Arch. Anat. Microsc. 22, 514. Kopriwa. B. M., and Leblond, C. P. (1962). J. Hisrochem. Cytochem. 10, 269. Lewis, W. H. (1931). Bull. Johns Hopkins Hosp. 49, 29. Lucas, A. M., and Jamroz, C. (1961). “Atlas of Avian Hematology.” US Govt. Printing Office, Washington, D.C. McMillan, D. B. (1958). Ph.D. Thesis, University of Toronto, Toronto. McMillan, D. B., and Engelbert, V. E. (1960). Can. J . Zool. 38, 613. McMillan, D. B., and Engelbert, V. E. (1963). Am. 1. Pathol. 42, 315. Maximow, A. (1909). Folia Haematol. (Leipzig) 8, 125. Mazia, D., Brewer, P. A,, and Alfert, M. (1953). Biol. Bull. (Woods Hole, Mass.) 104, 57. Metcalf, D. (1966). In “Recent Results in Cancer Research” (P. Rentschnick, ed.). SpringerVerlag, New York. Mills, J. E. (1964). Ph.D. Thesis, University of Toronto, Toronto. Mills, J. E., Westermann, J. E. M., and Engelbert, V. E. (1969). Can. J. Zool. 47, 1381. Morashita, K. (1957). Shika Gakuho 57 (11). Morrison, J. H. (1958). Ph.D. Thesis, University of Toronto, Toronto. Nairn, R. C. (1962). “Fluorescent Protein Tracing. ” Livingstone, Edinburgh. Pearse, A. G. E. (1960). “Histochemistry, Theoretical and Applied,” 2nd ed. Churchill, London. Pearse, A. G. E. (1968). “Histochemistry, Theoretical and Applied, 3rd ed., Vol. 1. ChurchillLivingstone, London. Pelc, S. R. (1968). Narure (London) 219, 162. Popoff, N. W . )1927). Arch. Exp. Zellforsch. Besonders Geweheznecht. 4, 395. Ranvier, L. (1875). “Traite technique d’histologie. ” Savy, Paris. Roels, H. (1966). Int. Rev. Cyrol. 19, 1. Sainte-Marie, G., and Leblond, C. P. (1958a). Proc. Soc. Exp. Biol. Med. 97, 263. Sainte-Marie, G., and Leblond, C. P. (1958b). Proc. Soc. Exp. Biol. Med. 98, 909. Schwann, T. (1839). “Mikroscopische Untersuchungen iiber die Uebereinstimmung in der Struktur und dem Wachsthum der Thiere und Pflanzen. ” Berlin. Shelley, J. L. G. (1961). M.Sc. Thesis, University of Toronto, Toronto. Shelley, J. L. G., Westermann, J. E. M., and Engelbert, V. E. (1969). Can. J . Zool. 47, 1414. Smith, N. (1969). M.Sc. Thesis, University of Toronto, Toronto. Smith, N., and Engelbert, V. E. (1969). Can. J . Zool. 47, 1269. Thomas, E. D. (1959). In “The Kinetics of Cellular Proliferation” (F. Stohlman, Jr., ed.), p. 118. Grune & Stratton, New York.

120

VIBEKE E. ENGELBERT

Trowell, 0. A. (1957). J . Biophys. Biochem. Cylol. 3, 317. Trowell, 0. A. (1958). Inr. Rev. Cytol. 7, 235. Tschassownikow, N. (1927). Arch. Exp. Zellforsch. Besondes Gewebeznecht. 3, 250. Watney, A. (1882). Philos. Trans. R. SOC. London) 173, 1063. Wedlock, D. (1974). Ph.D. Thesis, University of Toronto, Toronto. Weill, P. (1913). Arch. Mikrosk. Anar. 83, 305. Westermann, J. E. M. (1974). Rev. Can. Biol. 33, 255. Westermann, J. E. M., and Engelbert, V. E. ( (1969). Can. 1. Zool. 47, 1381. Westermann, J. E. M., Shelley, J. L. G., and Engelbert, V. E. (1970). Can. J . Zool. 48, 709. White, A. (1947-1948). Harvey Lect. 43, 43. Williamson, R. (1950). J. Parhol. Bacreriol. 62, 47. Wintrobe, M. M. (1967). “Clinical Hematology,” 6th ed. Lea & Febiger, Philadelphia, Pennsylvania. Yoffey, J . M. (1932-1933). J. Anat. 67, 250. Yoffey, J. M. (1960). “Quantitative Cellular Haematology. ” Thomas, Springfield, Illinois. Yoffey, J. M. (1973). HaemopoieticStem Cells, Ciba Found. Symp., 1972 No. 13 (new ser.), p. 7. Yoffey, J. M., Hanks, G . A,, and Kelly, L. (1958). Ann. N.Y. Acad. Sci. 73, 47. Yoffey, J. M., Everett, N. B., and Reinhardt, W. 0. (1959). In “The Kinetics of Cellular Proliferation” (F. Stohlman, Jr., ed.), p. 69. Grune & Stratton, New York.

I"An0NAL

REVIEW OF CrrOLooY,VOL.60

Fluidity of Cell Membranes-Current Trends

Concepts and

M. SHINITZKY The Department of Membrane Research, The W e i m n n Institute of Science, Rehovot. Israel

P. HENKART Immunology Branch, The National Cancer Institute, Bethesda, Maryland I. The Lipid Fluidity

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

II. The Protein Mobility

Spreading of Antigens in Heterokaryons . . . . . . . . Spreading of Locally Applied Antibody . . . . . . . . Rhodopsin . . . . . . . . . . . . . . . . . . . Bacteriorhodopsin . . . . . . . . . . . . . . . . Rotational Diffusion of Membrane Proteins . . . . . . . Fluorescence Photobleaching Recovery . . . . . . . . . Movement of Membrane Glycoproteins in an Electric Field . H. Interpretation of Diffusion Constants . . . . . . . . . . 111. Future Perspectives . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . A. 8. C. D. E. F. G.

121 130 132 133 133 135 135 137 142 143 144 145

I. The Lipid Fluidity A great deal of our knowledge on the dynamics and organization of lipid assemblies derives from studies of model lipid membranes of well-defined chemical composition. However, because of the heterogeneity in both the lipid composition and in the acyl chains of single phospholipids, which is inherent in biological membranes, extrapolations from the model system can only be qualitative in nature. The current notion of lipid organization in biological membranes is of a complex and heterogeneous structure. In areas where the protein to lipid ratio is small the membrane presumably behaves in accord with the fluid mosaic model of Singer and Nicolson (1972) where the lipids form a fluid pool in which proteins are embedded to different extents. The modes of diffusion of the proteins as asserted by this model are mostly passive and are therefore determined to a large extent by the fluidity of the lipid matrix. In other membrane areas the protein to lipid ratio may be greater and the structure may approach a loosely bound network of proteins among which lakes of lipids are spread. Since the proteins can change their position either by passive diffusion or by some 121

ISBN 0-12-364360-0

122

M. SHINITZKY AND P. I-ENKART

mechanism requiring metabolic energy, these types of structures may be constantly undergoing reorganization as dictated by the physiological state of the membrane. The thermal rearrangements of the lipid components are much more prominent than the redistribution of proteins. Thermal lateral motion of lipids is fast and provides the main mechanism of lipid mixing. Rotational movements of lipids, on the other hand, are practically restricted to axes perpendicular to the plane of the membrane. Thus an alternate exposure of a single phospholipid molecule to the two sides of the lipid bilayer by thermal motion [“flip-flop” transversion (Kornberg and McConnell, 1971)] is believed to be extremely slow (Rousselet et al., 1976). However, redistribution of lipids between the two layers can occur more rapidly with the assistance of special proteins or pores (Rothman and Kennedy, 1977). In addition to lateral and rotational processes lipid mixing may occur by random insertion and deletion of various components either by exogenous interaction with the serum or by intracellular mechanisms. Despite the processes of lipid mixing, the lipid distribution in biological membranes between the outer and the inner layers is quite asymmetric (Rothman and Lenard, 1977). This asymmetry of lipid bilayers poses a fluidity variable which relates to the degree of coupling between the two monolayers. Vesicles made of synthetic phosphatidylcholine of identical chain length (palmitoyl or myristoyl) were shown to be of minimum coupling. Perturbations in the head groups of either the inner or the outer layers were not transmitted to the other layer (Bystrov et al., 1971). Moreover, in systems where the inner and the outer layers were of different composition each layer was found to maintain its characteristic fludity and phase transition (Sillerud and Barnett, 1979). Of great interest is the finding that, when sphingomyelin is introduced into the lipid bilayer, coupling between the two layers is achieved, probably due to sphingomeyelin’s long a chain (24:O or 24:l) which can penetrate into the opposite layer (Schmidt et al., 1978). It is reasonable to assume that in biological membranes the coupling between the lipid layers is strong since the phospholipid acyl chains cover a wide range of lengths. For most biological membranes the term lipid fluidity will therefore maintain its original meaning which in principle relates to a strongly coupled system. However, the possibility of local decoupling between the inner and outer lipid layers, where the acyl chains are of similar length, should be borne in mind. In such regions the fluidity properties of the two lipid layers could be different (Schmidt et al., 1978). In addition to the asymmetry in the lipid bilayer some heterogeneity of lipid distribution in each monolayer can also prevail. The apparent surface distribution of lipids is the result of a steady-state equilibrium between lipid mixing and lipid segregation. Transient lipid clusters which are different in composition from the bulk lipid population can be formed by a series of processes. In the presence of an appreciable amount of a highly saturated phospholipid (e.g., sphingomyelin,

FLUIDITY OF CELL MEMBRANES

123

dipalmitoyl lecithin) and below the transition temperature, separation of crystalline isles of this phospholipid can occur (Shimshick and McConnell, 1973). Self-association of phospholipids can also be induced by Ca2+ bridging of two negatively charged phospholipid molecules (e.g ., phosphatidyl serine) (Papahadjopoulos et al., 1974; It0 et al., 1975) or by hydrogen bondings (e.g., phosphatidyl ethanolamine). Another type of segregation can occur by preferential interactions with cholesterol (Demel et al., 1977) or with specific proteins (Birre11 and Griffith, 1976). All these processes which promote lipid segregation will be in constant competition with the opposing lipid mixing processes. The constant thermal motion and shuffling of the lipid constituents is by definition a characteristic of a fluid system. However, the qualitative term which is often used to describe it-"lipid fluidity "-is to some extent ambiguous and oversimplified. By analogy to macroscopic liquids, fluidity, the reciprocal of viscosity, is proportional to the molecular free volume and is inversely proportional to the packing density of the fluid molecules. In lipid domains these factors can change considerably across the bilayer and one can discern the following principal regions: the hydrocarbon core, the hydrocarbon-water interface, and the hydrophilic boundary. The hydrocarbon core of lipid domains resembles a hydrocarbon fluid. In this region the energy of interaction between the hydrocarbon chains is relatively small and falls in the range of kT (about 1 kcal/mole). Hence, the fluidity of this region correlates with the partial specific volume of the hydrocarbon chain. This is the main reason why fully saturated acyl chains (e.g., palmitoyl, stearoyl) which occupy a relatively small volume confer rigidity, whereas acyl chains with cis double bonds (e.g., oleoyl) confer fluidity on lipid assemblies. In addition, the rigid planar structure of cholesterol is also of a low specific volume and will also contribute to reduction of fluidity. In the two other regions of the lipid layer, the hydrocarbon-water interface and the hydrophilic boundary, strong intermolecular associations prevail. These interactions are mainly ionic or hydrogen bondings and have a free energy much greater than kT. Thus, in these regions the fluidity is determined mostly by the intermolecular forces rather than by the specific volume of the constituents. These arguments imply that there is a gradient of fluidity when proceeding from the outer layer toward the core of the membrane. Nuclear magnetic resonance and electron spin resonance measurements of movements of methylene groups along the acyl chains of phospholipids have indeed demonstrated that the motional freedom becomes progressively hindered as the ester bond is approached (Seelig and Seelig, 1974). More detailed examinations have shown that at the upper hydrocarbon region, below C lo-C12, the rigidity is kept about constant, whereas at the lower layer, above Clz, it is progressively diminished (Stockton et al., 1977). Furthermore, the presence of cholesterol or integral proteins will impart rigidity mostly to the upper part of the hydrocarbon and the hydrocarbon-water interface. The cumulative intermolecular forces in the upper

124

M. SHINITZKY AND P. HENKART

regions of the membrane can be partially transmitted down the hydrophobic core of the lipids and thus can exert a significant reduction of fluidity in the hydrocarbon region. The above complications which are inherent in the term lipid fluidity are practically manageable and one can still apply to it quantitative parameters as long as they relate to a well-defined region of the membrane. In most reported cases this term was related to the hydrocarbon region in the bulk sense, and unless otherwise stated the average lipid fluidity in this region may be referred to as the “membrane fluidity.” A useful operational term for quantitative assessment of membrane fluidity is the term “microviscosity” ( f j ) . This term is derived from the fluorescence polarization measurements of the rotational diffusion of a fluorescent probe embedded in the analyzed region and is expressed in absolute macroscopic units (poise) for both self-consistency and comparison with related liquids. By inference from macroscopic fluids one can further use the empirical temperature dependence: fj =

AEIRT

(1)

Thus, by determining the change of f j with temperature one can derive the flow activation energy, A E , which can serve both as an energetic parameter and as a criterion for the degree of order in the system (Shinitzky and Inbar, 1976). For a comprehensive treatise on the method and its application the reader is referred to the recent review by Shinitzky and Barenholz (1978). Extensive studies, mostly with liposomal model membranes, have established a series of gross factors which determine the lipid microviscosity. These factors are: (a) the ratio of cholesterol to phospholipid, CIPL (Vanderkooi et af., 1974; Shinitzky and Inbar, 1976); (b) the degree of unsaturation and length of the phospholipid acyl chains (Cogan et af., 1973; Lentz et af., 1976); (c) the ratio of lecithin to spingomyelin (LIS) (Shinitzky and Barenholz, 1974; Schmidt et al., 1977); and (d) the ratio of lipid to protein (Shinitzky and Inbar, 1976). Cholesterol is the most abundant sterol in animal tissues. It consists of a rigid plane with specific residues [3(P)-OH, A5 double bond, and aliphatic chain at C( 17)] which establishes an alignment with the phospholipid chains (Huang, 1976, 1977). In most mammalian membranes under physiological conditions the presence of cholesterol will increase f j and decrease A E . In liposomes of the same phospholipid composition as human red blood cells, a system that can serve as a standard for fluidity measurements (see below), an increase of CIPL from 0 to 1.4 will increase f j (37°C) from 0.5 to 6 poise and decrease A E from 15 to 5 kcal/mole (Shinitzky and Inbar, 1976). This example elucidates the physiological role of cholesterol-increasing both the viscosity and the order of the lipid layer-and how it may be implicated in modulation of membrane function. In the

FLUIDITY OF CELL MEMBRANES

125

“solid” state of lipid domains, as studied with liposomes of synthetic phospholipids, cholesterol acts as a fluidizer by perturbing the structure regularity of the system (decrease in 3 and increase in A E ) . However, in most biological membranes under physiological conditions such solid lipid domains are not formed. Under normal conditions the CIPL of any biological membrane remains constant presumably to maintain optimal function. However, the constancy of CIPL is a steady-state result of a series of dynamic processes which may momentarily enrich or deplete the membrane cholesterol. At least half of the cholesterol molecules in cell plasma membranes can exchange with the cholesterol pool of the serum. When the CIPL in the membrane is sufficiently different from that of the serum this exchange will lead to a net translocation of cholesterol directed toward equalizing the CIPL in the two reservoirs. This process is especially pertinent to erythrocytes (Vanderkooi et al., 1974; Cooper et al., 1975) and platelets (Shattil et al., 1975), which lack metabolic processing of cholesterol, and leukocytes in the resting state (Shinitzky and Inbar, 1974) in which these processes are at a basal level. The direct contact with the serum in the case of blood cells may provide the main determinant of the CIPL in their outer membrane. In metabolically active cells, ingestion of the cholesterol-rich lipoproteins via a specific receptor (Brown and Goldstein, 1976) and intracellular synthesis can increase the cholesterol content and can compete with the loss of cholesterol either by passive translocation or by intracellular esterification (Arbogast et a l ., 1976). All these processes are internally regulated and provide an efficient maintenance of the membrane CIPL. The second gross determinant of membrane microviscosity is the degree of unsaturation of the phospholipid acyl chains (Chapman and Wallach, 1968). Double bonds of natural fatty acids are virtually all of the cis configuration and their presence in phosopholipds increases considerably its partial specific volume. The presence of double bonds therefore increases the fluidity and decreases the degree of order in the system. The net fluidizing effect of double bonds is far from being proportional to their number per molecule. Thus, replacement of stearic acid (18:O) with oleic acid (18:l) markedly increases the fluidity, but further replacement of oleic acid with linoleic acid (18:2) has only a small effect. Fully saturated phospholipids (e.g., distearoyl phosphatidylcholine) above the transition temperature still possess 3 values which are considerably greater than in analogous systems with unsaturated acyl chains (e.g . , dioleoyl phosphatidylcholine, egg lecithin) (Cogan et al., 1973; Lentz et al., 1976). This difference is diminished as the acyl chains become longer (Lentz et al., 1976). In the extreme case of a high proportion of saturated chains, phase transition and phase separation may occur. The degree of unsaturation of phospholipid acyl chains can be efficiently modulated by intracellular metabolism which is now believed to be the

126

M. SHINITZKY AND P. HENKART

main regulatory mechanism of membrane microviscosity in adaptation to temperature (Hazel and Prosser, 1974; Sinensky, 1974; Cossins, 1977) and metabolic or nutritional disorders (Cooper, 1977). The third parameter, the ratio of lecithin to sphingomyelin (LIS), deserves special attention. These two phosphorylcholine phospholipids constitute more than 50% of the phospholipids in mammalian membranes (Rouser et al., 1968) and about 90% of mammalian serum phospholipids (Nelson, 1967). Because of inherent structural differences, their fluidity properties are markedly different. Natural lecithin bears highly unsaturated acyl chains, especially at the /3 position, and therefore imparts high fluidity to lipid domains. Natural sphingomyelin, on the other hand, is highly saturated and forms lipid domains of low fluidity which also display a broad phase transition between 25" and 35°C (Shinitzky and Barenholz, 1974; Schmidt et al., 1977). Furthermore, at 37"C, where sphingomyelin is mostly in a fluid phase it still possesses fj values about 6-fold higher than natural lecithin (Shinitzky and Barenholz, 1974). The rigidifying effect of sphingomyelin is only partially due to its highly saturated hydrocarbon chains. Inter- and intramolecular hydrogen bonds of its amide linkage and the free hydroxyl group, in addition to the sphingosine trans double bond, condense the hydrocarbon-water interface region which presumably confers rigidity on the hydrocarbon region as well. Passive translocation of lecithin or sphingomyelin between the serum and cell membranes is extremely slow (days or weeks) and in most cases can be neglected. In patients with abetalipoproteinemia,however, the LIS is about half the normal level, in both the serum and the red blood cells, which suggests an exchange mechanism for sphingomyelin (Cooper ef al., 1977). Another instance of such an equilibration is when sheep erythrocytes, which have a very low level of lecithin, LIS -0.01, are incubated in human serum in the presence of EGTA (Borochov et al., 1977). Within 24 hours of such incubation the sheep erythrocyte membrane increases its lecithin level by more than 5-fold. The incorporation of lecithin in this particular case is presumably facilitated by a membrane-bound lecithinase (Kramer et al., 1974). Two important biological processes have been associated with changes in LIS. Both aging and arteriosclerosis are characterized by substantial decreases in LIS (Eisenberg et al., 1969), whereas maturation of the fetal lung is associated with an increase in LIS of the pulmonary surfactant (Gluck et al., 1971). All the above cases, in which the membrane LIS is modulated, are associated with the expected change in the membrane microviscosity. The dynamic characteristics of the membrane lipid layer are indirectly affected by the presence of proteins-the fourth determinant of microvisocity. The effect of proteins is qualitatively similar to that of cholesterol; it increases the microviscosity and decreases the flow activation energy. These effects are more prominent at low cholesterol levels and in a way compete with the effects of choles-

FLUIDITY OF CELL MEMBRANES

127

terol. This competition probably accounts for the fact that the AE value of most mammalian membranes is approximately invariant and falls around 7 kcal/mole regardless of the membrane lipid composition (Shinitzky and Inbar, 1976). The interrelation between membrane proteins and lipid microviscosity can be illustrated by the concept of vertical displacement of membrane proteins (Borochov and Shinitzky, 1976; Shinitzky and Rivnay, 1977). Accordingly, as the lipid microviscosity increases, the bulk membrane proteins become more exposed to the aqueous medium. Conversely, as the membrane fluidity increases (e.g., CIPL decreases), the bulk proteins will sink into the membrane interior. This physical process may be of physiological significance, since it provides a subtle mechanism for modulation of membrane receptors, antigens, and enzymes, and it may also extend to functional lipids as well (Shinitzky, 1979; Yasuda et al., 1977; Brulet and McConnell, 1977). In addition to the general and gross factors which were described above each individual lipid head group may exert a specific, though subtle, effect on the bulk lipid fluidity of the membrane. For assessment of the contribution of each membrane lipid head group to the overall microviscosity it can be assumed that each lipid component has more or less a typical saturation-unsaturation profile in its hydrocarbon region as in lecithin, sphingomyelin, and glycolipids. One can then employ the phospholipid mixture of human erythrocytes (HEP) (Shinitzky and Inbar, 1976) as a basic fluidity unit and measure the relative change in microviscosity in the presence of known amounts of the various lipid components. Assuming that the change is additive in nature, this approach can be pursued by defining a microviscosity index, y , as the relative contribution of a lipid component (L) to the overall microviscosity, of HEP. This defintion is formally expressed in the equation,

where fL and fHEp are the mole fractions of L and HEP in the membrane cfL + = l), +j(HEP L) and +j(HEP) are the apparent microviscosities of HEP with and without L, and yLis the microviscosity index of the lipid L. Accordingly, by definition y H E p = 1 which provides the relative scale for the various lipid components. Lipids with y < I can thus be considered as “fluidizers” whereas lipids with y >1 can be considered as “rigidifiers.” Table I presents the y values, related to the membrane hydrocarbon core, obtained for various membrane lipids in mixtures with HEP at 25” and 37°C. The table distinctively displays a group of phospholipids which act as fluidizers (PC, PS, PG, and Car) and a group of lipids which act as rigidifiers (PE, PI, G, Sph, and Chol). Except for gangliosides, sphingomyelin, and cholesterol, the index y for all other phospholipids remains approximately constant at the two measured temperatures. This indicates that the effect of these phospholipids on AE of HEP

fHEP

+

128

M. SHINITZKY AND P. HENKART TABLE I MlCROVlXOSlTY INDEX, y , OF MEMBRANE LIPIDS PRESENTED ON A SCALE RELATIVETO HUMAN ERYTHROCYTE PHOSPHOLIPID MIXTURE Y (25°C)

Lipid HEP" PC PS PG Car PE PI

G SPh Chol

y (37°C)

1 .o 0.5 0.4

0.6 0.5

0.4 0.4 1.5 2.2 2.5 6.6 4.0

0.5 0.6 1.7 2.2 1.4 3.4 12.1

1 .o

"HEP, human erythrocyte phospholipid mixture.; PC, phosphatidyl choline (lecithin) from egg yolk; PS, phosphatidyl serine from bovine brain; PG, phosphatidyl glycerol from egg yolk; Car, cardiolipin from bovine heart; PE, phosphatidyl ethanolamine from egg yolk; PI, phosphatidyl inositol from soybean; G, mixed gangliosides from bovine brain; Sph, sphingomyelin from bovine brain; Chol, cholesterol.

is relatively small, namely, they do not affect much the degree of order in the system. Cholesterol, on the other hand, has about a 4-fold greater effect at 37" than at 25°C. This is undoubtedly due to the marked reduction of A E and the increase in the degree of order which cholesterol exerts on lipid layers (Shinitzky and Inbar, 1976). The two other components which do not maintain a constant y value are gangliosides and sphingomyelin which undergo a phase transition between 25" and 37°C (Shinitzky and Barenholz, 1974; Schmidt et al., 1977) and therefore lose much of their rigidifying effect at 37°C. It is interesting to note that in the hypothetical extreme case of fL = 1 the extrapolated value of q(L) = yL ?(HEP) is similar in magnitude to the experimental q(L) value which is obtained with liposomes made of pure L. It is therefore reasonable to assume that the index yLobtained with HEP L mixtures will also hold for membranes (M) of different lipid composition. Equation (2) can be thus generalized to the form,

+

which may be applied for estimation of the change in membrane microviscosity upon enrichment with a known amount of L. It should be stressed, however, that some variability in yLof a phospholipid obtained from different sources is expected due to fatty acid variability.

129

FLUIDITY OF CELL MEMBRANES

The apparent microviscosity in biological membranes, as derived from the fluorescence depolarization of 1,6-diphenyl 1,3,5-hexatiene (DPH), ranges for most cases from 1 to 10 poise at 0-40°C. Since DPH, as well as other hydrocarbon fluidity probes, partitions indiscriminately into various lipid phases (for a review, see Shinitzky and Barenholz, 1978), the apparent microviscosity represents a weight average of all the membrane lipid domains. The most viscous membranes recorded to date by this method are the intestinal microvillus membranes from rat which display ij (25°C) of 13k1 poise (Schachter and Shinitzky, 1977). Lateral diffusion of phospholipids, as determined by electron spin resonance (Hubbell and McConnell, 1971) or fluorescence bleaching recovery (Schlessinger et al., 1977a), display a diffusion rate constant in the range of D -lo-* cm2/secondwhich for molecules with an effective volume of 1000 A3 means an ambient microviscosity in the range of several poise. Furthermore, rates of lateral and rotational movements of proteins (see next chapter) also suggest a semifluid lipid environment with a viscosity in the poise range. In many instances biological membranes have been asserted to display a phase transition at temperatures in the range of 10-40°C. In most of these cases the "phase transition" was inferred from kinetic data which are only indirectly related to the lipid fluidity. In studies using the fluorescence polarization of hydrocarbon probes, a method in which measurements are specifically confined to the hydrocarbon region of the lipid layer, only two actual lipid phase transitions in biological membranes have been reported. These cases are the sheep erythrocyte membrane (Borochov et al., 1978) and the microvillus membrane of the rat intestine (Schachter and Shinitzky, 1977). Both membranes display a broad phase transition around 25°C presumably because of a high content of sphingomyelin. In addition to the above factors, which are all integral constituents of the membrane, it is of relevance to discuss the ambient factors, temperature and pH, which may also act as modulators of membrane microviscosity. The effect of temperature on +j can be directly assessed with the aid of Eq. (l), provided that the membrane lipids do not undergo a phase transition. Taking A E = 7 kcall mole, a value close to that determined for most mammalian membranes, a decrease of temperature from 37" to 25" or 4°C will increase +j by a factor of 1.6 and 3.8, respectively. Unlike mammalian membranes, which can function normally only around 37"C, plant membranes are designed to maintain their functions over a relatively wide range of temperatures. In the plasmalemma of rose petals, a typical plant membrane, AE is 3.5 kcal/mole (Borochov et al., 1976) which reduces considerably the effect of temperature on lipid microviscosity . Accordingly, decreasing the temperature of this membrane from 37" to 25" or 4°C increases q by only a factor of 1.2 and 1.9. The intestinal microvillus membrane (Schachter and Shinitzky, 1977) and the plasma membrane of neuroblastoma cells (deLaat et al., 1978) were shown to

-

130

M . SHINITZKY AND P. HENKART

display a marked increase in 4 when the pH was increased from 6 to 8. In both cases the effect of pH seems to be cooperative and the microviscosity reaches a maximum around pH 7.5. Liposomes made of lipid extracts of these membranes show almost no change in 4 with pH, which suggests that the effect originates from the membrane proteins. Reduction of local pH can occur in situ upon an excess production of acids like lactic acid. Thus, changes in glucose metabolism can modulate the intracellular pH as well as the acidity in the immediate extracellular surrounding. The change in pH can then modulate the membrane microviscosity which can markedly alter the membrane function.

11. The Protein Mobility

The analysis of protein mobility in biological membranes faces a series of fundamental difficulties which relate to heterogeneity problems. One can first distinguish between two types of mobility-the passive diffusion and the metabolically driven mobility. The latter type of movement is associated with the assemblage of a complex network of microtubules and microfilaments and is only indirectly affected by the lipid fluidity. Hence, the metabolically driven movements cannot be classified as a fluidity related mobility and therefore will not be discussed in this article. Passive diffusion of membrane proteins can, in principle, be either rotational or translational. In the heterogeneous lipid matrix these two types of diffusion can be of different rates, but at small rotational displacements the movement approaches a translational movement and the difference between the two mobilities is small. Since lipid fluidity in biological membrane is about two orders of magnitude smaller than in the surrounding water, only the fraction of the protein mass embedded in the lipid layer should be considered as the effective moving volume. Thus, the basic question as to how changes in lipid fluidity may affect the passive diffusion of membrane proteins is not simple since a decrease in lipid fluidity will, on one hand, oppose the mobility but, on the other hand, will decrease the effective volume of the moving proteins (Borochov and Shinitzky, 1976). Except for these complications the physical nature of lateral diffusion in the two-dimensional array of the membrane only partially obeys the classical expressions of three-dimensional diffusion. Two-dimensional diffusion is of a formidable complexity which is not yet fully resolved (Richter and Eigen, 1974; Soffman and Delbriick, 1975; Hardt, 1979). The above difficulties can be partially overcome by discussing the problem of passive diffusion of each membrane protein separately. But even this approach is hampered by the finding that specific membrane proteins have a substantial fraction which is practically immobile (see Table 11).

TABLE I1 PASSIVE DIFFUSION CONSTANT, D, OF MEMBRANE COMPONENTS AT 25°C

Component Transplantation antigens

Rhodopsin

Cell or membrane Human-mouse fused cells Cultured myotubes Cultured fibroblasts Rod segment

D (cm2/second)

Immobile fraction

Reference

5x

Frye and Edidin (1970)

1-3x 10-9

Edidin and Fambrough (1973)

2-3 x

0.55

Schlessinger et al. (1977b)

Po0 and Cone (1974); Liebman and

3-4 x 10-9

Entine (1974) -1

RaziNaquietal. (1973);Cherryetal. (1977a)

0.64

Jacobson et al. (1976)

3-4x

0.35

Schlessinger et al. (1976a)

Mast cells

2 x 10-10

0.2-0.5

Schlessinger et al. (1976b)

Acetyl choline receptor

Rat muscle cells

4-5x 10-l'

0.25

Axelrod et al. (1976b)

Lipid analogs

L6 Lipid bilayer

8x 3-4x 10-9

-0 -0

Schlessinger er al. (1977a) Wolf et al. (1977)

Bacteriorhodopsin

Halobacterium halobium

Concanavalin A receptor

Mouse fibroblasts Rat myoblasts

Fc receptor

10-I I

132

M. SHINITZKY AND P.HENKART

This section on protein mobility is presented by discussing the mobility of specific proteinic sites as determined by the various spectral techniques which were recently introduced into membrane research. The techniques are briefly outlined throughout the text and mobilities obtained with each technique are discussed separately. An up-to-date summary of the various diffusion parameters accumulated for membrane components is given in Table 11. In line with the diffusion constant given in the table one can divide membrane components into three classes: freely diffusing components (D = cm2/second);components with restricted mobility (D = 10-9-10-11 cm2/second); immobile components (D < lo-” cm2/second). A. SPREADING OF ANTIGENS IN HETEROKARYONS The spreading of surface antigens in newly fused heterokaryons was first measured by Frye and Edidin (1970). This series of experiments provided a dramatic demonstration of the ease with which membrane proteins could diffuse in the plane of the membrane and was a pivotal piece of evidence leading to the fluid mosaic model of membrane structure. The experimental design is conceptually simple: Induce two types of cells to fuse together (using Sendai virus) and then observe the distribution of specific antigens of each type using immunofluorescence for a period of time after the fusion event. Frye and Edidin used cultured cell lines of mouse and human origin. The H-2 major transplantation antigens, known to be glycoproteins, were used as the specific antigen marker for the former, while less well-defined but clearly specific xenogeneic human cell antigens were used as the other marker. The experiments showed that in the first few minutes after fusion over 90% of the double-staining cells still showed antigen segregation: Roughly half of the cell membrane was “human” while the other half was “mouse.” However, upon further incubation at 37”C, antigenic mixing occurred; at 25 minutes roughly half the heterokaryons had antigens which were completely mixed; at 40 minutes roughly 90% of the cells showed such patterns. This rapid mixing was not affected by inhibitors of ATP production or protein synthesis but was strongly blocked at temperatures between 0” and 15°C. Although this technique has given a graphic and straightforward analysis of antigen lateral diffusion in the membrane, a number of disadvantages have discouraged its widespread use: (a) There is a need to use Sendai virus or other ‘‘fusogen” which may itself alter membrane properties; (b) there is difficulty in quantitating the results in terms of a diffusion constant (see, however, Huang, 1973); (c) measurements require somewhat tedious examination of large numbers of cells under two illuminations; (d) the results relate to populations of cells rather than to single cells. In a recent paper, applying this technique to lateral diffusion of the band 3 membrane proteins of human erythrocytes, Fowler and Branton (1977) showed

FLUIDITY OF CELL MEMBRANES

133

that these proteins, when labeled directly by fluorescein isothiocyanate (FITC), show lateral diffusion after heterokaryon formation. These proteins had previously been shown by fluorescence photobleaching recovery (see below) to be incapable of rapid movement (Peters et al., 1974). These authors studied fusion induced by Sendai virus and by PEG using mixtures of FITC-labeled and unlabeled cells. Even mixing of these proteins required about 2 hours at 37"C, rather than 40 minutes as found by Frye and Edidin for tissue culture cells. In the case of erythrocyte ghosts (hemolysis accompanies fusion) the rate of formation of uniformly labeled cells is slightly but significantly enhanced by ATP. The temperature dependence of label redistribution was generally similar to that observed by Frye and Edidin, with lateral diffusion decreasing sharply below 23°C.

B. SPREADING OF LOCALLY APPLIEDANTIBODY Edidin and Fambrough (1973) measured the rate of diffusion of fluorescently labeled Fab (monovalent) antibody directed against muscle cell surface proteins after local application of the antibody with a micropipet. The initial area of staining was limited to a well-defined spot on the elongated cultured muscle cell. The spot size was determined by measuring the length of the stained area on photographs made immediately and at various times afterward. Although the data showed considerable variation from cell to cell, it was clear that marked spreading of the spot occurred over a period of minutes, with a calculated D of 1-3x lop9cm2/second (at 37°C). Cold temperatures inhibited this diffusion by 10-fold or greater. The observed spreading did not appear to be due to dissociation of the labeled Fab fragments, since they could not be displaced from the membrane by subsequent incubation with unlabeled antibody, and since no diffusion was observed when prefixed cells were stained. The antigens recognized by these antibodies were shown to be heterogeneous and mostly of molecular weight over 200,000. One interesting observation was that anti-immunoglobulin crosslinking of the first (bivalent) membrane-bound antibody gave markedly slower diffusion times. Furthermore, if labeled Fab antibody was used to stain the membrane and later an excess of intact unlabeled antibody of the same type was added, diffusion was retarded. Thus, for whatever reasons, bivalent antibody appeared to retard difussion of membrane antigens. C. RHODOPSIN A number of unique features of rhodopsin have made possible the estimation of the rotational and translational diffusion rates of this membrane protein by spectroscopic means. The molecular chain of events leading to vision begins with the absorption of a photon by the rhodopsin chromophore 11-cis-retinal. The

134

M. SHINITZKY AND P. HENKART

retinal chromophore is covalently bound to the protein, opsin, via a Schiff base linkage and appears to be bound into a hydrophobic cleft in the protein. The rhodopsin molecules are located in the membranes of the “discs” of the outer segments of the rod cells of the retina. The discs are stacked in a parallel array within the rod outer segments which are themselves parallel to each other, with their long axes along the path of incoming light. It has been known since 1933 that the rods are dichroic when viewed from the side with visible light, and it was therefore clear that the retinal chromophores were not free to rotate freely in three dimensions. The most obvious explanation of this was that retinal was not free to rotate with respect to opsin, which in turn could not rotate around an axis parallel to the plane of the membrane. Another unique feature of rhodopsin is the bleaching process which follows the absorption of a photon by retinal. The initial photochemical event is the photoisomerization of 11-cis-retinal to the all-trans species, which is rapidly followed by a series of spectroscopically defined rhodopsin intermediates, after which rhodopsin finally dissociates. Regeneration of rhodopsin is a time- and energy-consuming process. When viewed from above (and on), rods fail to show dichroism. Hagins and Jennings (1959) found that they could not induce dichroism in this orientation by bleaching with polarized light and suggested rapid rotation of rhodopsin about an axis perpendicular to the membrane as the most likely explanation. This interpretation was later confirmed by Brown (1972) who showed that such dichroism could be induced by bleaching with polarized light after glutaraldehyde fixation to prevent rhodopsin rotation. It remained for Cone (1972) to measure the rapid transient photodichroism induced in unfixed rods after a polarized bleach and thus estimate the rate of rotational diffusion. This was accomplished by constructing a specialized flash photometer apparatus capable of a brief ( 5 nsec) pulsed laser bleaching flash and a photometry circuit capable of measuring the ensuing small changes in absorbance with a time resolution of less than 100 nsec. A few nanoseconds after the bleach, the dichroic ratio (ratio of absorbance of light polarized parallel to the plane of the bleaching flash to light polarized perpendicular to it) was found to be about 2, but this ratio decayed to 1.0 with a half-life of 3.0 nsec at 20°C. The decay process had a Qlo of 3.0. This rapid rotational diffusion was calculated to be consistent with an effective membrane viscosity of about 2 poise (see Section I). The lateral diffusion of rhodopsin in the plane of the membrane was measured by Po0 and Cone (1974) and later by Liebman and Entine (1974). These experiments also used the ordered arrangement of the retinal discs within the rod. In the Po0 and Cone study, a rectangular region, almost half of the cross-sectional area on one side of the rod, was bleached by a bright flash of light of appropriate wavelength. The absorption decrease from this area was measured and then the bleaching flash stopped and a much weaker (1OOOX less) intensity used to measure rhodopsin absorption in the bleached area. The rectangular beam was

FLUIDITY OF CELL MEMBRANES

135

then used to scan the width of the rod so that the unbleached rhodopsin levels could be measured quantitatively across the discs with time. The initial scan after the bleach showed distinct bleached and unbleached halves of the rod, but equilibration of rhodopsin was evident within 15 seconds. The half-time for equilibration of the rhodopsin absorbance across the disc at 70°C was 35 seconds for frog rods (8 p m diameter) and 23 seconds for mudpuppy rods (12 pm diameter). The Q,,, for these recoveries were 3.7 and 3.1, respectively. The lateral diffusion constants calculated from these measurements were 3.5 X and 3.9 x cm2/second, respectively. Several controls established that the measured recovery was actually due to protein lateral diffusion, as was the case with rotational diffusion. Prior glutaraldehyde fixation prevented the recovery after bleaching. Furthermore, in the presence of hydroxylamine, a scavenger for free retinal, the recovery proceeded as usual, indicating that diffusion of the free chromophore was not responsible for the recovery. Lastly, no lateral diffusion was measurable along the length of the rod, i.e., from disc to disc. These control experiments were consistent with a process of diffusion in the plane of the membrane.

D. BACTERIORHODOPSIN The purple membrane of Halobacterium halobium contains virtually a single protein, bacteriorhodopsin, in which retinal is convalently bound to a lysine side chain (for a review, see Henderson, 1977). The physiological and photochemical functions of this protein resemble those of vertebrate rhodopsin with the important difference that in bacteriorhodopsin the prosthetic group does not dissociate upon photochemical isomerization. This protein is therefore a unique case of a membrane protein in which a reversible photochromic probe is built in (see below). In contrast to rhodopsin, bacteriorhodopsin exists in the membrane in a planar hexagonal array (Blaurock and Stoeckenius, 1971) of a crystalline form (Henderson, 1975; Henderson and Unwin, 1975). Obviously, such a structural arrangement will practically freeze the thermal movement of the protein as was indeed observed by spectral means (Razi Naqui et al., 1973; Sherman et al., 1976; Cherry et al., 1977a). Rotational diffusion of bacteriorhodopsin could be detected only after disintegration of the crystalline structure (Cherry et al., 1977b). E. ROTATIONAL DIFFUSION OF MEMBRANE PROTEINS The classical and most efficient technique for determination of rotational diffusion of proteins in solution is fluorescence depolarization (Weber, 1953; Steiner and Edelhoch, 1962; Stryer, 1968). This method is based on the Perrin equation [Eq. (4)] which describes the dependence of the measured fluores-

136

M. SHINITZKY AND P. HENKART

cence polarization of a fluorescence probe attached to the protein, on the hydrodynamic properties of the system: r o = 1 + -37 (4) r P In this expression, r and ro are the measured and limiting fluorescence anisotropy (a measure of fluorescence polarization), T is the excited-state lifetime, and p is the rotational relaxation time. The latter term is inversely proportional to the rotation diffusion constant and is approximately equal to the average time for a displacement of about 60". Most fluorescence probes employed in this technique possess r -20 nsec, which for rotations of small proteins in aqueous solution approaches the optimal experimental condition of p -37. In biological membranes, however, the rotational motion is more than 10-fold slower than in water and the fluorescence depolarization technique becomes impractical. Nevertheless, with sensitive instruments which can detect small changes in r one can still observe changes in rotational mobility of membrane proteins by fluorescence depolarization in a semiquantitative manner. Such studies were carried out on fluorescently labeled lectins attached to normal and malignant transformed cells (Shinitzky et al., 1973; Inbar et al., 1973) where a good correlation between the apparent rotational mobility of the lectin receptors and the tendency for cap formation was observed. In these studies the rotational displacements which were monitored were very small and practically matched the translational movement. For monitoring the rotational motion of most membrane proteins one is bound to employ a probe with a transient lifetime much longer than the fluorescence lifetime. For this purpose a process analogous to fluorescence polarizationpolarized photochromism-can be employed. Photochromism is a photochemical process in which a reversible intramolecular isomerization between a parent compound and a transient of a distinctively different absorption spectrum takes place. The photochromic process can be physical, like a singlet-triplet transition, or chemical, like cis-trans isomerization. Since in many cases the phototransient possesses a half-life in the range of 10-3-10-s seconds, photochromism with polarized beams can be used for monitoring slow rotations like those of membrane proteins. As in fluorescence polarization, polarized photochromism can be monitored by a steady-state technique (Pasternak et al., 1978), which employs very mild operational conditions but yields average-type information, and a decay technique (Cherry, 1975) which employs strong photolysis flash and, in principle, can furnish more detailed information. The methods employed for measurement of the rotational diffusion of rhodopsin and bacteriorhodopsin (see above) belong to polarized photochromism. This method could, in principle, be used for other membrane proteins which bear a photolyzable or photochromic group which cannot itself rotate with respect to the protein. For example Junge (1972) attempted to measure the rotation of cyto-

FLUIDITY OF CELL MEMBRANES

137

chrome A3 in intact mitochondria in suspension. The cytochrome A3-CO complex was formed, and then a short pulse from a dye laser photolyzed the complex, giving rise to a positive absorption change at 445 nm due to the appearance of reduced free cytochrome As. Using instruments with a time resolution of 10 nanoseconds, no dichroism could be observed unless the mitochondria were fixed or the experiment was done in a viscous solution. However, no conditions were used which gave rise to a decaying dichroism. A rapid protein rotation in the membrane was interpreted as being he most likely explanation for the results. The study of rotational diffusion of most proteins by similar spectroscopic means is not feasible because of the lack of such chromophores. One possible approach to this problem is to introduce suitable chromophores covalently onto various proteins. Cherry et al. (1976) have taken this approach with the erythrocyte band 3 proteins, which were selectively modified by eosin-isothiocyanate. This dye molecule can be excited by a photon with the resulting formation of a relatively long-lived triplet state, which has absorption properties distinct from the gound-state molecules. By polarizing the exciting light, the slower rotational diffusion of proteins can be measured with this dye. The eosin-labeled band 3 showed a dichroism which had two components: one which decayed with a half-life of approximately 0.5 msec; and a second component which was time independent over the course of measurement (several milliseconds). The authors attribute the first component to rotation of band 3 molecules within the plane of the membrane and the second component to either aggregated band 3 molecules which do not rotate or to the nonrotation of band 3 molecules around axes in the plane of the membrane. Since band 3 molecules had been thought to be linked in some way to spectrin-depleted ghosts, it was of interest to study the rotational motion of band 3 in spectrin-depleted ghosts. Although spectrin depletion itself had no effect on rotational diffusion of band 3, subsequent treatment at pH 5.4, which aggregated the intramembranous particles containing band 3, resulted in a loss of the time-decaying component of the observed dichroism. This experiment serves as a control to eliminate a number of possible objections to the authors’ interpretation of these results. Taken as a whole, the results indicate that band 3 rotates in the plane of the membrane and make it clear that spectrin cannot be attached to all band 3 molecules in a way which eliminates rotation of the latter. The authors suggest that band 3 may be trapped in a spectrin and actin lattice on the cytoplasmic surface of the membrane. F. FLUORESCENCE PHOTOBLEACHING RECOVERY 1. Method

The principle utilized in the fluorescence photobleaching recovery (FPR) technique is that of the Po0 and Cone experiment (1974) on rhodopsin. It takes

138

M. SHINITZKY AND P. IENKART

advantage of the rapid photobleaching reactions of common fluorescent dyes such as fluorescein and rhodamine, on which such photobleaching reactions are irreversible. Measurements are then made of the rate of recovery of fluorescence in the area of the bleached spot. Given the sensitivity of fluorescence as a technique and the prior experience of using fluorescein- and rhodamine-labeled antibodies, this technique has aroused the interest of several laboratories and is at present the principal source of quantitative estimates of the diffusion of membrane proteins. The instrumentation has evolved somewhat from that initially used by Peters et al. (1974); currently lasers are being used to provide an extremely intense round spot of light for bleaching, with a diameter of less than 2 pm. The same light beam is used for quantitatively following the intensity of the fluorescent dye molecules, but it is attenuated by a factor of 103-104 using neutral density filters. The methodology and theoretical treatments of these techniques are given by Peters et al. (1974), Edidin et al. (1976), Jacobson et al. (1976), and in somewhat more detail by Axelrod et al. (1976a). 2. Results with Labeling Surface Amino Groups on Membrane Proteins The first use of fluorescencebleaching recovery to examine the lateral diffusion of membrane proteins was that of Peters et al. (1974). Intact human red cells were first labeled with fluorescein isothiocyanate, and ghost membranes were then prepared. SDS-gel electrophoresis of the membrane proteins showed that the fluorescein had principally labeled band 3, namely, spectrin and probably glycophorin. Fluorescein conjugates of lipid (phosphatidyl ethanolamine) or free fluorescein compounds were also observed on the gels. For bleaching experiments, the ghosts were flattened and immobilized by sandwiching them between a coverglass and microscope slide. In these experiments, Peters et al. inserted a diaphragm into the microscope so that a sharply defined line cut the field in half, allowing bleaching of exactly half of a cell in the center of the field. Before and after bleaching, the fluorescence signal from each of the two halves of the cell was measured by moving the slide so that first one half of the cell and then the other was in the 9x4.5-pm measuring field. Unlike the rhodopsin experiments, no change in the fluorescence of either side of the cell was detectable over 20 minutes; according to their theoretical treatment of diffusion under these conditions, the authors estimated 3 x cm2/secondas the upper limit of the diffusion coefficient. These results should be compared to the experiments of Fowler and Branton (1977) described previously and also to the study of rotational motion of band 3 (Cherry et al., 1976). Fluorescein-modified surface proteins of cultured mouse fibroblasts (L cells) were studied by Edidin et al. (1976). In these experiments, the cells were modified with FITC at pH 9.5, a procedure which clearly damaged sparsely grown cells and probably caused temporary damage to others. Fluorescein was

FLUIDITY OF CELL MEMBRANES

139

shown to be on surface proteins by demonstration that less than 10% of the fluorescence was extractable with chloroform-methanol and by removal of the fluorescent label with papain. Furthermore, most of these proteins were integral proteins, since 70% of the fluorescein was not extractable by salt or chelating agents. Bleaching was carried out on spherical cells which remained attached to the dish. For the first 2 hours after labeling, the cells contained cytoplasmic fluorescence, as well as the prominent “ring” staining associated with surface labeling of spherical cells; during this period no recovery after bleaching was observed. After several hours of culture, fluorescence in the cells appeared only on the surface. When spots on these cells were bleached, recovery was seen within a few seconds, with recovery curves close to those predicted from diffusion, and a mean of D =2.6x 10-lo cm2/secondwith a range of D=O.5-5 X 10-lo was calculated. After several minutes a high percentage (90%) of the original fluorescence intensity of the spot was recovered. Schlessinger et al. (1977a) measured the motion of fluorescein-labeled membrane protein on cultured rat myoblasts. In this study, labeling was done at physiological pH and no injurious effects were noted. It was assumed that FITC labeled proteins by analogy with the results of Edidin et al. (1976) noted above. The diffusion constant for these proteins was 2 . 2 10-lo ~ cm2/second, in excellent agreement with the fibroblast results of Edidin et al. One significant difference was that on myoblasts only about 40% of the initial fluorescence intensity was recovered after diffusion, which was interpreted to mean that 60% of the labeled proteins were immobile. In this study, a second labeling method was also used. Surface proteins of myoblasts were TNP modified by the reagent TNBS, and FITC anti-TNP was used to monitor the motion of these molecules in a manner analogous to that generally used for specific receptor studies. The results were very similar to those obtained by direct FITC labeling-D was found to be 1.9X 10-lo cm2/second-and greater than 50% of the fluorescence signal was restored with time. These results provide some reassurance that labeling with bivalent antibody can give meaningful estimates of receptor diffusion. With this label, as in the case with Con A receptors (Schlessinger et al., 1976a), the percentage recovery increased after the first bleach at the same spot. 3. Diffusion of Lectin “Receptors” Zagyansky and Edidin (1976) studied mouse fibroblasts labeled with LlTCsuccinyl Con A (divalent) and with FITC-Con A (tetravalent). These results showed the potential complexities introduced by crosslinking of surface molecules with the fluorescent labeled detection reagents. When long-term cultured fibroblast lines were studied no recovery of fluorescence was observed with either Con A or succinyl Con A labels, but some patching and capping was seen. In the case of primary fibroblast cultures recovery could be observed, but was

140

M. SHINITZKY AND P. HENKART

dependent on the Con A valency and the time after application of lectin. With tetravalent Con A a slow recovery was observable in the first hour after labeling, but after that time no recovery was measured. With succinyl Con A, recovery was observed, but its rate appeared to be slower at longer times after application of the succinyl Con A. No diffusion constants were calculated, nor were percentage recoveries given. Jacobson er al. (1976) studied the lateral diffusion of FITC-succinyl Con A in glycerol-water mixtures (of various viscosities and hence varying rates of diffusion) and on the surface of cultured fibroblasts. In the former case, a thin layer of solution in a microcuvet was used, bleaching recovery curves were close to those predicted from theory, and the rates obtained were within an order of magnitude of those predicted from the medium viscosity and the Stokes-Einstein equation. In these experiments, the extent of recovery was always greater than 80%. In the experiments with Con A bound to fibroblasts, the recovery rate was slower than that found for 98% glycerol and was estimated to be between 10-lo and lo-" cm2/second. The fractional recovery averaged 36%, indicating that most of the Con A markers were so immobile that they did not diffuse during the time of the measurements (10-20 minutes). An FPR study of Con A bound to cultured rat myoblasts was reported by Schlessinger et al. (1976a). The effect of time of incubation with Con A and its valency were apparently similar to those observed by Zagyansky and Edidin (1976). Furthermore, the rate of recovery was slower as the Con A concentrations increased. The most rapid recoveries for either form of Con A corresponded to D = 3 . 8 ~ 1 0 - l cm2/second. ~ Recoveries after bleaching were roughly 65% initially, increasing after the second bleach at the same spot. The temperature dependence of the recovery rate showed that the rate of diffusion of Con A slowed substantially below 2WC, but was somewhat less temperature dependent above 20°C. For succinyl Con A the data correspond to a Qlo of 15 below 20°C and of 6 above 20°C. Neither the mitochondria1oxidative phosphorylation inhibitor sodium azide nor colchicine treatment affected the rate of recovery. However, treatment with cytochalasin B slowed the recovery by roughly 3-fold. 4. Other Surface Markers

The IgE-specific Fc receptor of mast cells is the membrane trigger for degranulation and release of histamine and other mediators and hence plays a key role in immediate hypersensitivity phenomena. This receptor, which can be followed by using fluorescently labeled IgE, is monovalent in the sense that a single Fc receptor appears to bind only a single IgE molecule. The lateral motion of Fc receptors on rat peritoneal mast cells labeled in this way was measured by Schlessinger et al. (1976b). They found a lateral diffusion constant of 2 x 10-lo cm2/second and a recovery of 50-80%. As was the case with bound Con A, Fc

FLUIDITY OF CELL MEMBRANES

141

receptor diffusion was not influenced by colchicine or azide, but was retarded about 6-fold by cytochalasin B. Axelrod ef al. (1976b) examined the behavior of rhodamine-labeled a-bungarotoxin bound to the nicotinic acetyl choline receptors on cultured rat muscle cells. These receptors are normally localized at the neuromuscular junction, where they detect the neurotransmitter released by nerve impulses. During embryonic development and in tissue cultures, with no nerve cells present, muscle cells possess acetyl choline receptors diffusely distributed over the cell surface as well as localized in discrete concentrated patches. The protein a-bungarotoxin from snake venom is known to bind tightly to the isolated receptor, with a 1:l stoichiometry, and hence is not expected to crosslink the receptor on the membrane. When the bleaching recovery technique was used in this sytem, it was found that no recovery was observed in the patches of greater receptor density. The diffusely distributed receptors did show lateral motion, withD = 4 . 5 lo-” ~ cm2/second at 22°C and 7-16X lo-” cm2/second at 35°C. The average percentage recovery was 75% at both temperatures. The difference between diffuse and concentrated receptors could originate from phosphorylated subpopulations (Teichberg et al., 1977). Schlessinger et al. (1977b) studied the lateral diffusion of chemically undefined surface antigens on cultured fibroblasts with the intention of testing the “anchorage modulation” hypothesis proposed earlier by Edelman (1976). In these experiments, labeled Fab fragments of rabbit antibodies against a crossreacting tumor were found to have a lateral diffusion constant of 2 . 6 10-lo ~ cm2/second, or close to the generalized protein labels used earlier. The fractional recovery averaged 45%. When Con A-coated platelets were bound to the surface of these fibroblasts at some distance from the spot of the bleaching measurements, the lateral motion of the mobile antigens decreased 6-fold, while the percentage recovery remained unchanged. Colchicine and vinblastine, best known as microtubule-disrupting drugs, partially reversed the diffusion-retarding effect of Con A-platelets. These remarkable results had at least qualitatively been predicted by the anchorage modulation hypothesis, and various ideas as to its mechanism were discussed in the paper. 5 . Macromolecular Probe A study of membrane-bound synthetic lipopolysaccharides (designed as models for membrane proteins) on planar lipid bilayers was reported by Wolf et al. (1977). Dextran (MW 80,000) which was modified by addition of fatty acid, rhodamine, and hapten, was shown to bind to lipid bilayers and to undergo lateral diffusion as measured by FPR. A diffusion constant of 3 X lo-’ cm2/second was shown for membrane densities of 100 molecules/pm*, while increasing the twodimensional concentration to lo4 molecules/pm2 slowed the diffusion constant to 5x cm2/second. Crosslinking the dextran molecules with antibody to the

142

M. SHINITZKY AND P. HENKART

attached hapten caused patching at low dextran concentrations, but at higher membrane dextran concentrations the molecules were immobilized into a lattice without visible patching. 6. Conclusionsfrom FPR Studies The FPR results discussed above have demonstrated that membrane proteins are capable of varying degrees of movement. The technique has the advantage that it is relatively simple and is applicable to a wide range of membrane molecules via markers such as antibodies. A number of potential problems with the technique are apparent: (a) Damage to the cells due to the laser bleaching pulse might be expected to be a problem. However, since most cells do not appreciably absorb light at wavelengths used for bleaching, Axelrod (1977) has argued on theoretical grounds that the temperature increase is less than 10°C. Furthermore, the lipid probe in myoblast cell membranes gives the same diffusion constant as measured by FPR and the technique of fluorescence polarization (which does not involve bleaching), and the protein diffusion constants obtained by FPR are close to those estimated from data using heterokaryon fusion techniques. Nevertheless, with pigmented cells laser-induced damage may still be a possibility. (b) If antibodies are used, it should be established that the bleaching recovery is not due to dissociation of the antigen-antibody complex and diffusion in the medium. If fixed cells are then stained with antibodies and the bleaching recovery does not occur, such a possibility seems unlikely. (c) Crosslinking of antigens by bivalent antibody may retard diffusion as shown by Wolf et al. (1977) and Edidin and Fambrough (1973). In considering the quantitative aspects of the FPR results, the most surprising finding is that all of the membrane proteins observed have shown incomplete recovery of the fluorescent signal after bleaching. This has generally been interpreted as due to an immobile fraction of the labeled molecules, and this appears to be the only plausible explanation of results such as those obtained by Schlessinger et al. (1976a,b) in which repeated bleaches at the same spot give successively higher percentage recovery. An immobile fraction could be due to a linkage of this fraction to an immobile component(s), or could also be due to the endocytosis of this fraction during the time between membrane labeling and the bleach. This latter possibility does not appear to have been carefully considered in spite of rapid endocytosis in fibroblasts, macrophages, and presumably other cells.

G . MOVEMENT OF MEMBRANE GLYCOPROTEINS IN AN ELECTRIC FIELD Due to the fluidity of the membrane, it might be predicted that not only can membrane molecules undergo Brownian movement, but that charged molecules

FLUIDITY OF CELL MEMBRANES

143

would also be free to move in response to an electrical field. A rearrangement of Con A receptors on cultured muscle cells in response to an external electric field has been demonstrated by Po0 and Robinson (1977). This rearrangement was shown to occur in the presence of metabolic inhibitors. A surprising result of this study is that the Con A receptors behaved as if they were positively charged, while most proteins, especially sialic acid-containing glycoproteins, would be expected to bear a net negative charge. H. INTERPRETATION OF DIFFUSION CONSTANTS A realistic theoretical calculation of the diffusion constants expected of membrane proteins would be useful in estimating the extent to which even the recoverable fraction of membrane proteins is retarded in its diffusion. The FPRderived diffusion constant of 2 x 10-lo cm2/second obtained for FITC-labeled membrane proteins and the mast cell Fc receptor is still more than an order of magnitude slower than the diffusion constant of 4 X lo+’ cm2/secondobtained by Po0 and Cone for rhodopsin. It seems unlikely that this difference is due to experimental error on the calculation of D. If it were possible to calculate theoretical diffusion constants from given models of the membrane structure, it would be useful to compare the expected D with that obtained experimentally. Soffman and Delbriick (1975) have presented calculations for the lateral diffusion constant of membrane proteins assuming that they are cylinders which span the membrane but do not protrude into either aqueous phase. Although these calculations need to be refined to include more realistic assumptions, one interesting prediction is that the lateral diffusion constant is very insensitive to shape and molecular weight. The sensitive parameters in diffusional processes therefore remain the distance and the geometry of the track. A simple and elegant approach to this difficult problem was recently proposed by Hardt and Cone (Hardt, 1979; Hardt and Cone, 1979). Their method employs classical expressions of passively diffusing particles for calculating the average diffusion time, T ,in a complex structure like a biological membrane. A general expression was derived, 7 = b2 *f/2D, where T is the measured average diffusion time (as the one measured in the FPR method), b is the maximal diffusion distance, and f is a characteristic geometry factor of the membrane. This simple and intuitive approach may facilitate the interpretation of experimental data on protein mobility. Finally, it seems likely that even the “freely mobile” fraction of plasma membrane proteins is diffusing more slowly than rhodopsin or theoretical expectations. What is then retarding the diffusion? One obvious possibility is that since the membrane is not a dilute solution of proteins floating in a lipid bilayer, the proteins interact with each other in “cis” membrane interactions in addition to the better known “trans” membrane interactions.

144

M. SHINITZKY AND P. HENKART

111. Future Perspectives

The tremendous development which we have witnessed in the last 5 years in the understanding of the mobility of membrane components has led to important refinements in the previous concepts of membrane structure and dynamics. Currently, the biological membrane is viewed as a loosely bound protein network among which fluid lipid domains are spread. Most of the lipid components are at a constant thermal motion and shuffling while the proteins are roughly divided into two equal subgroups: the freely diffusing and the practically immobile. Most of the proteins in the latter group are associated, chemically or physically, with an underlying texture of actin-myosin filaments. In this heterogeneous but highly integrated structure the old concept that proteins carry the specific functions and lipids merely provide the proper fluidity does not hold any longer. Both the lipids and the proteins influence each other’s dynamics and each of them can bear specific functions. In such a complex network local perturbations can be transmitted over a long range which may be implemented in physiological processes. It is almost self-evident that the overt cellular functions and morphology are determined to a large extent by lipid fluidity and protein mobility. It has indeed been observed that these parameters change along the cell cycle, during differentiation and upon malignant transformation. On the other hand, under normal stationary conditions, lipid fluidity and protein mobility are maintained at specific levels which are designed to carry out properly the various physiological functions. However, under environmental stresses a battery of regulatory enzymes provides the cells with an efficient tool which can adjust the lipid fluidity to the proper level. When the change in environment is maintained for a long period of time the cells will eventually adapt to it with the aid of these regulatory enzymes. But when the stress exceeds a certain threshold, like in hyperlipidemic diseases, the lipid composition of the cellular membranes will adopt an improper fluidity which may exert marked deviations in cellular functions. The effect of lipid fluidity on membrane function is now one of the best approaches available for studying the mechanisms of membranal processes. As described above, lipid fluidity determines the lateral and rotational movements of mobile membrane proteins and in addition can modulate their degree of exposure. A combination of these effects will obviously be reflected in the active expression of specific antigens, receptors, or enzymes. Changes in lipid composition of cell membranes can now be achieved by various means which provide the key tools for reversible modulation of membrane function. We, therefore, feel that the correlation between lipid fluidity and activity of specific membrane sites will soon become one of the most investigated topics in modem cytology.

FLUIDITY OF CELL MEMBRANES

145

REFERENCES Arbogast, L. Y., Rothblat, G. H., Leslie, M. H., and Cooper, R. A. (1976). Proc. Narl. Acad. Sci. U.S.A. 73, 3680. Axelrod, D. (1977). Eiophys. J. 18, 129. Axelrod, D., Koppel, D. E., Schlessinger, J. E., Elson, E., and Webb, W. W. (1976a). Eiophys. J . 16, 1055. Axelrod, D., Ravdin, P., Koppel, D. E., Schlessinger, J., Webb, W. W., Elson, E. L., and Podleski, T. R. (1976b). Proc. Natl. Acad. Sci. U . S . A . 73, 4594. Birrell, G. B., and Griffith, 0. H. (1976). Biochemistry 15, 2925. Blaurock, A., and Stoeckenius, W. (1971). Nature (London), New Eiol. 233, 152. Borochov, A., Halevy, A. H.,and Shinitzky, M. (1976). Nature (London) 263, 158. Borochov, H., and Shinitzky, M. (1976). Proc. Narl. Acad. Sci. U.S.A. 73, 4526. Borochov, H., Zahler, P., Wilbrandt, W., and Shinitzky, M. (1977). Biochim. Biophys. Acra 470, 382. Borochov, H., Barenholz, Y., and Shinitzky, M. (1979). Cell Eiophys., in press. Brown, M. S., and Goldstein, J. L. (1976). Science 191, 150. Brown, P. K. (1972). Nature (London) 236, 35. Brulet, P., and McConnell, H. M. (1977). Biochemistry 16, 1206. Bystrov, V. F., Dubrovina, N. I., Barsukov, L. I., and Bergelson, L. D. (1971). Chem. Phys. Lipids 6, 343. Chapman, D., and Wallach, D. F. H. (1968). In “Biological Membranes” (D. Chapman, ed.), p. 125. Academic Press, New York. Cherry, R. J. (1975). FEES Lett. 55, 1. Cherry, R. J., Burkli, A., Busslinger, M., Schneider, G., and Parish, G. R. (1976). Nature (London) 263, 389. Cherry, R. I., Heyn, M. P., and Oesterhelt, D. (1977a). FEES Lert. 78, 25. Cherry, R. J., Muller, V., and Schneider, G. (1977b). FEES Lett. 80, 465. Cogan, U., Shinitzky, M., Weber, G., and Nishida, T. (1973). Eiochemisrry 12, 521. Cone, R. A. (1972). Nature (London) 236, 39. Cooper, R . A. (1977). N . Engl. J . Med. 297, 371. Cooper, R. A,, Amer, E. C., Wiley, J. S., and Shattil, S. J. (1975). J . Clin. Invest. 55, 115. Cooper, R. A,, Durocher, J. R., and Leslie, M. H. (1977). J. Clin. Invest. 60, 115. Cossins, A. R. (1977). Eiochim. Eiophys. Acta 470, 395. deLaat, S. W. et al. (1979). Submitted for publication. Demel, R. A,, Jansen, J. W., van Dijk, P. W. M., and van Deenen, L. L. M. (1977). Biochim. Biophys. Acra 465, 1. Edelman, G. M. (1976). Science 192, 218. Edidin, M., and Fambrough, D. (1973). J. Cell Eiol. 57, 27. Edidin, M., Zagyansky, Y., and Lardner, T. J. (1976). Science 191, 466. Eisenberg, S., Stein, Y., and Stein, 0. (1969). J. Clin.Invest. 48, 2320. Fowler, V . , and Branton, D. (1977). Nature (London) 268, 23. Frye, L. D., and Edidin, M. (1970). J. Cell Sci. 1, 319. Gluck, L., Kulovich, M. V., and Borer, R. C. (1971). Am. J . Obstet. Gynecol. 109, 440. Hagins, W. A,, and Jennings, W. H. (1959). Trans. Faraday SOC. 27, 180. Hardt, S. L. (1979). J. Membrane Eiol., in press. Hardt, S. L., and Cone, R. A. (1979). Eiophys. J., in press. Hazel, J. R., and F’rosser, C. L. (1974). Physiol. Rev. 54,620. Henderson, R. (1975). J. Mol. Biol. 93. 123.

146

M. SHINITZKY AND P. HENKART

Henderson, R. (1977). Annu. Rev. Biophys. Bioeng. 6, 87. Henderson, R., and Unwin, P. N. T. (1975). Nature (London) 257, 28. Huang, C. (1976). Nature 259,242. Huang, C. (1977). Chem.Phys. Lipids 9, 150. Huang, H. W. (1973). J . Theor. Biol. 40, 11. Hubbell, W. L., and McConnell, H. M. (1971). J . Am. Chem. SOC. 93, 314. Inbar, M., Shinitzky, M., and Sachs, L. (1973). J . Mol. Biol. 81, 245. Ito, T., Ohnishi, S., Ishinaga, M., and Kito, M. (1975). Biochemistry 14, 3064. Jacobson, K. E., Wu, E., and Poste, G. (1976). Biochim.Biophys. Acta 433, 215. Junge, W. (1972). FEES Lett. 25, 109. Kornberg, R. D., and McConnell, H. M. (1971). Biochemistry 10, 1111. Kramer, R., Jungi, B., and Zahler, P. (1974). Biochim. Biophys. Acra 373, 404. Lentz, R. B., Barenholz, Y., and Thompson, T. E. (1976). Biochemistry 15, 4521. Liebman, P. A., and Entine, B. (1974). Science 185, 457. Nelson, G. J. (1967). Lipids 2, 323. Papahadjopoulos, D., Poste, G., Schaeffer, B. E., andVail, W. J. (1974). Biochim. Biophys. Acta 352, 10. Pastern&, C., Slifkin, M., and Shinitzky, M. (1978). J. Chem.Phys. 68, 2669. Peters, R., Peters, J., Tews, K. H., and Bohn, W. (1974). Biochim. Biophys. Acta 367, 282. Poo, M., and Cone, R. A. (1974). Nature (London) 247, 438. Poo, M., and Robinson, K. R. (1977). Nature (London) 265 602. Razi Naqui, K., Gonzalez-Rodriguez, J., Cherry, R. J., and Chapman, D. (1973). Nature (London) 245, 249. Richter, P. H., and Eigen, M. (1974). Biophys. Chem. 2, 255. Rothman, J. E., and Kennedy, E. P. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 1821. Rothman, J. E., and Lenard, J. (1977). Science 195, 743. Rouser, G., Nelson, G. J., Fleischer, S., and Simon, G. (1968). I n “Biological Membranes” (D. Chapman, ed.), p. 5 . Academic Press, New York. Rousselet, A., Guthmann, C., Matricon, J., Bienvenue, A., and Devaux, P. F. (1976). Biochim. Biophys. Acta 426, 357. Schachter, D., and Shinitzky, M. (1977). J . Clin. Invest. 59, 536. Schlessinger, J., Koppel, D. E., Axelrod, D., Jacobson, K., Webb, W. W., and Elson, E. L. (1976a). Proc. Natl. Acad. Sci. U.S.A. 73, 2409. Schlessinger, J . , Webb, W. W., Elson, E. L., and Metzger, H. (1976b). Nature (London) 264,550. Schlessinger, J., Axelrod, D., Koppel, D. E., Webb, W. W., and Elson, E. L. (1977a). Science 195, 307. Schlessinger, J., Elson, E. L., Webb, W. W., Yahara, J., Rutishaman, U., and Edelman, G. M. (1977b). Proc. Natl. Acad. Sci. U.S.A. 74, 2909. Schmidt, C. F., Barenholz, Y., and Thompason, T. E. (1977). Biochemistry 16, 2649. Schmidt, C. F., Barenholz, Y., Huang, C., and Thompason, T. E. (1978). Nature (London) 271, 775. Seelig, A., and Seelig, J. (1974). Biochemistry 13, 4839. Shattil, S. J., Anaya-Galindo, R., Bennett, J., Coleman, R. W., and Cooper, R. A. (1975). J . Clin. Invest. 55, 636. Sherman, W. V., Sliflrin, M. A., and Caplan, S. R. (1976). Biochim. Biophys. Acta 423, 238. Shimshick, E. J., and McConnell, H. M. (1973). Biochemistry 12, 2351. Shinitzky, M. (1976). Bull. Schweiz. Akad. Med. Wiss. 32, 203. Shinitzky, M. (1979). In “Physical Chemical Aspects of Cell Surface Events in Cellular Regulation” (C. DeLisi and R. Blumenthal, eds.), p. 173. Elsevier North-Holland, New York. Shinitzky, M., and Barenholz, Y. (1974). J . Biol. Chem. 249,2652.

FLUIDITY OF CELL MEMBRANES

147

Shinitzky, M., and Barenholz, Y. (1978). Biochim. Biophys. Acru Rev. Biomembr. (in press). Shinitzky, M., and Inbar, M. (1974). J . Mol. Biol. 85, 603. Shinitzky, M., and Inbar, M. (1976). Biochim. Biophys. Actu 433, 133. Shinitzky, M.,and Rivnay, 8 . (1977). Biochemistry 16, 982. Shinitzky, M., Inbar, M., and Sachs, L. (1973). FEBS Leu. 30, 125. sillemd, L. O., and Barnett, R. E. (1979). Biochemistry (in press). Sinensky, M. (1974). Proc. Natl. Acud. Sci. U.S.A. 71, 522. Singer, S. J., and Nicolson, G. L. (1972). Science 175, 720. Soffman, P. G., and Delbriick, M. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 3111. Steiner, R. F., and Edelhoch, H.(1962). Chem. Rev. 62, 457. Stockton, G. W., Johnson, K. G., Butler, K. W., Tulloch, A. P.,Boulanger, Y.,Smith, I. C. P., Davis, J. H.,and Bloom. M.(1977). Nature (London) 269, 267. Stryer, L. (1968). Science 162, 526. Teichberg, V. I., Sobel, A., and Changeux, J. P. (1977). Nature (London) 267, 540. Vanderkooi, J., Fischkoft, S., Chance, B., and Cooper,A. (1974). Biochemistry 13, 438. Weber, G. (1953). Adv. Protein Chem. 8, 415. Wolf, D. E., Schlessinger, J., Elson, E. L., Webb, W. W., Blumenthal, R., and Henkart, P.(1977). Biochemistry 16, 3476. Yasuda, T., Dancey, G. F., and Kinsky, S. C. (1977). Proc. Natl. Acud. Sci. U.S.A. 74, 1234. Zagyansky, J., and Edidin, M. (1976). Biochim. Biophys. Actu 433, 209.

This Page Intentionally Left Blank

INTERNATIONAL REVIEW OF CWOLOGY, VOL. 60

Macrophage -Lymphocyte Interactions in Immune Induction MARCFELDMANN ICRF Turnour Immunology Unit, Department of Zoology, University College, London, England

ALANROSENTHAL Department of Immunology, Merck Institute for Therapeutic Research, Rahway, New Jersey

PETERERB Institute of Microbiology, University of Basel, Basel, Switzerland I. Introduction . . . . . . . . . . . . . . . . . . . . 11. Macrophage Function in Antigen-Specific T-cell Proliferation . . A. Antigen Uptake by Macrophages . . . . . . . . . . . B. How Do Macrophages Interact with Lymphocytes? . . . . C. Immune Response Gene Function in Macrophage-T-Lymphocyte Interaction . . . . . . . . . . . . . . . . . . . D. Genetic Analysis of Determinants Mediating Macrophage-T-Lymphocyte Interaction. . . . . . . . . 111. Macrophage Function in Helper T-cell Induction . . . . . . A. Nature of Macrophage-T-cell Interactions in Helper Cell . . . . . . . . . . . . . . . . . . . Induction B. Genetic Restrictions on Macrophage-T-cell Cooperation. . . IV. Macrophage-B-Lymphocyte Interactions in Antibody Production . V. Concluding Discussion . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . Note Added in Proof . . . . . . . . . . . . . . . . .

149 151 152

153 155

158 163 163

167 171 172 176 178

1. Introduction

The immune system functions in vivo to differentiate acceptable from nonacceptable antigenic substances. There are two major types of lymphocytes bearing antigen-specific receptors, thymus-derived lymphocytes (T cells) and bone-marrow derived lymphocytes (B cells), and (2) macrophage-like accessory cells (Fig. 1) which interact with antigen and present approximately configured immunogenic signals to the lymphocytes. Such cell interactions provide an unique opportunity 149

Copyright @ 1979 by Academic Press, Inc. reproduction in any form reserved.

All rights of

ISBN 0-12-36436@0

150

MARC FELDMANN ET AL.

FIG.1. Macrophage from a guinea pig peritoneal washout.

for regulation. Our remarks focus on the extensive literature on macrophage-Tcell interactions. Other functions of macrophages are discussed in The Immunobiology of the Macrophage, edited by David Nelson (1976). In most species studied T-cell responses depend on macrophage function. This is most clearly demonstrated in vitro, but there are also examples in vivo, such as the delayed hypersensitivity response. Antigen-induced proliferation (Waldron et al., 1973), helper cell induction (Erb and Feldman, 1975a), proliferation in

MACROPHAGE-LYMPHOCYTE INTERACTIONS

151

response to mitogens such as concanavalin A (Con A) and phytohemagglutinin (PHA) (Wahl et al., 1975; Habu and Raff, 1977), the mixed lymphocyte reaction (MLR) e.g. (Greineder and Rosenthal, 1975a; Rode and Gordon, 1974; Schumacher et al., 1975), the induction of killer cells (Wagner et al., 1972), and the release of lymphokines all require the participation of macrophages. This participation may vary in both quantitative and qualitative aspects, and the mechanism of macrophage-T-cell interaction with soluble antigens is possibly different than with cell-bound antigens, as in the MLR or in killer cell induction. It is unclear whether T-cell responses which have not yet been shown to require macrophages, such as suppressor cell induction (Feldmann and Kontiainen, 1976), merely require far fewer, or perhaps qualitatively different (less adherent?) macrophage-like cells. This article is chiefly concerned with antigen-specific aspects of T-cell recognition with the exception of the rather provocative surface carbohydrate oxidation models of T-cell activation described in the mouse by Novogrodsky and Katchalski (1976) and in the guinea pig by Greineder and Rosenthal ( 1975b), and some comparisons with macrophage-B-cell interactions.

11. Macrophage Function in Antigen-Specific T-cell Proliferation Analysis of the precise cellular and molecular events which underlie macrophage function has begun in earnest. It is clear, for example, that antigenspecific activation of mediator production and clonal expansion by T lymphocytes requires the presence of macrophages and that the effects of macrophages extend far beyond simple maintenance of in vitro culture conditions. Indeed, in both the mouse and guinea pig, it has been shown that the recognition of soluble protein antigens by T lymphocytes is preceded by an initial uptake of antigen by the macrophages (Waldron et al., 1974). The biological complexities inherent in multicellular immunogenic recognition units are obvious. Nonetheless the definition of both the cellular and immunological specificity of the requirement for macrophages is not fully understood. In the simplest experiments the proliferative capacity of populations of lymphoid cells depleted of macrophages is restored by macrophages but not by lymphocytes, granulocytes, or fibroblasts (Wahl et al., 1975). This requirement for macrophages in T-cell antigen recognition is not determined by the immune state of the animal from which the macrophages were obtained, although macrophages and lymphocytes must share genetic identity at some portion of the major histocompatibility complex (MHC) for successful detection of the immunogenic signal bound by the macrophages. The response of T lymphocytes to nonspecific plant mitogens such as PHA and Con A also requires accessory cells (Wahl et al., 1975; Habu and Raff, 1977) but, unlike that in soluble protein antigen recognition, their function can be performed by fibroblasts or increased cell density. However, these experiments

152

MARC FELDMANN ET AL.

are not easy to interpret; for example, fibroblasts may be contaminated by macrophages, unless multiple passages are performed, and increased cell density facilitates interaction with the residual macrophages. A. ANTIGENUPTAKEBY MACROPHAGES

Since the antigen selects the relevant lymphocytes for activation, it is important to determine the nature of the antigen uptake process and the physical state of macrophage-associated antigen. The functional uptake of protein antigens is assessed by their capacity to activate primed T cells. Uptake occurs at 4°C via an initial binding step not requiring metabolic energy and is followed by a second temperature-dependent, metabolically active event. When the latter step is carried out in a glucose-free medium, immunological activity is reduced to the level seen at 37°C in the presence of sodium azide or 2-deoxyglucose, agents which inhibit aerobic and anaerobic metabolism, respectively, and is approximately equal to that observed when macrophages are pulsed at 4°C (Rosenthal, unpublished data). Cytochalasin B, an agent active against a variety of membrane functions which also blocks classic pinocytosis, does not influence the uptake of immunologically relevant protein (Lipsky and Rosenthal, 1976). Trypsinization of macrophages pulsed with antigen at 4°C removes all immunological activity without altering the ability of such macrophages to take up new antigens. If, however, trypsinization is progressively delayed by prior culture at 37"C, macrophages pulsed with antigents at 4°C become progressively resistant to its effects such that by 120 minutes macrophage immunological potential is unaltered (Waldron et al., 1974). It is surprising that antibody against antigens (as high as 0.5 mg/ml) does not affect the ability of antigen-pulsed macrophages to initiate T-cell proliferation (Ellner and Rosenthal, 1975). This may be due to the redistribution of antigen and internalization, as evidenced by the recent studies of Thomas and Shevach, using trinitrophenol (TNP)-conjugated macrophages. Anti-TNP antibody inhibited the antigeninduced lymphocyte proliferation by such macrophages immediately after conjugation, but antiserum inhibition was lost when the macrophages were cultured 24 hours prior to mixture with T cells. Recent evidence (E. M. Shevach, personal communication) suggests that antibody inhibition fails to occur because there is too little antigen on the surface for divalent binding. The proliferation-inducing potential of macrophage-bound antigen decayed in a linear fashion (half-life of 1.5 days) when the macrophages were cultured prior to mixing with specifically immune lymphocytes. However, no concordant decrease in cell-associated 1251-labeledprotein antigen was noted (by quantitation of the amount of 1251-labeledantigen remaining) after the initial 24 hours of preculture (Ellner and Rosenthal, 1975, 1977). Thus these data are inconsistent with a free and random distribution of antigens on the macrophage membrane, a model of antigen handling originally proposed for the mouse (Unanue and Cerottini,

MACROPHAGE-LYMPHOCYTJ2 INTERACTIONS

153

1970). Obstacles to further analysis of the molecular character of antigen associated with macrophages include (1) ignorance of the structure of the immunological determinants of the antigens, (2) the need for viable macrophages in antigen-specific proliferation, (3) uncertainties concerning the location and function of the immune response (Ir) gene products, and (4) uncertainties concerning the nature of the genetic restrictions observed in macrophage-lymphocyte interactions in in vivo and in vitro primed lymphoid cell populations. These issues are among the major topics of this article and are some of the key unresolved issues in immunology today. Their resolution should provide an insight into whether T cells possess single or dual receptors for the antigens or complexes which stimulate them.

B. How Do MACROPHAGES INTERACT WITH LYMPHOCYTES? Macrophage-lymphocyte interaction may occur either indirectly, such as via the secretion of antigen complexed to soluble macrophage products cytophilic for lymphocytes [e.g., the genetically restricted factor (GRF),a complex of I-region-associated (Ia) antigen and immunogen of molecular weight -55,0001 or, alternatively, by direct cell interaction. Evidence for the first exists for the generation of specific helper T-cell activity in the mouse (Erb and Feldmann, 1975) and is described in Section III. The presence of factors active in vitro does not necessarily imply that the same mechanism is operative in vivo. It is possible that the same factor molecule works just as effectively (or even more so) if still attached to the membrane of its cell of origin. Soluble macrophage products have not been found which elicit antigen-specific lymphocyte proliferation in the guinea pig (Ellner and Rosenthal, 1977). Instead, at least two types of physical interactions are seen between guinea pig macrophages and lymphocytes. The first is not dependent on the presence of antigen. This step requires active macrophage (but not lymphocyte) metabolism, divalent cations (either Ca2+ or Mg2+), and a trypsin-sensitive macrophage site. This binding does not distinguish T or B lymphocytes and is reversible with an equilibrium between cellular association and dissociation (Lipsky and Rosenthal, 1973). When this nonspecific interaction has brought immune lymphocytes into apposition with the appropriate antigen-bearing macrophages, a second type of binding results, which is dependent upon the presence of antigen and a sharing of histocompatibility-linked gene products on the lymphocytes and macrophages. This association is not easily reversed and eventually results in proliferation of the bound lymphocyte (Lipsky and Rosenthal, 1975) (Fig. 2). A linkage between the physical and functional macrophage-lymphocyte interactions in the antigen recognition process was found by using cytochalasin B, a reagent which inhibits the recognition of antigen-specific proliferation, since it does not interfere with uptake of antigen by macrophages or inhibit PHA-induced

154

MARC FELDMANN ET AL.

FIG.2. Macrophage-lymphocyte interaction of CFA-primed guinea pig cells from a peritoneal washout after 48 hours culture with PPD.

proliferation; this indicated that antigen recognition itself was perturbed, not the machinery of DNA synthesis. The effects of cytochalasin B thus appear to result from its inhibition of the antigen-independent phase of macrophage-lymphocyte interaction (Rosenthal et al., 1975). Inhibition of this step would prevent the development of subsequent antigen-dependent interactions assayed both physically and functionally. Similar antigen-dependent clusters are seen in humans (Clive and Swett, 1968) as well as mice (Mosier, 1969).

MACROPHAGE-LYMPHOCYTE INTERACTIONS

155

c. IMMUNE RESFQNSEGENEFUNCTION IN MACROPHAGE-T-LY MPHOCYTE INTERACTION The response to many antigens, such as synthetic polypeptides, or to low doses of protein antigens is under the control of Zr genes linked to the MHC in many species including guinea pigs and mice. These genes control T-cell-dependent responses, such as the antibody response, delayed hypersensitivity, and antigeninduced proliferation. A diagram of the mouse, guinea pig, and human MHC is shown in Fig. 3. Because of the Iack of recombination the guinea pig MHC is less well understood. The use of antigens under the control of H-linked Ir genes and specific alloantisera which recognize cell-associated antigenic determinants which are Mouse: 17th chromosome Regions

0.5 cm , K

I

I

H - i/'~

A-ZG

D I

H-2D

\

, I-E

I-J

I-B

G I

\ '

/ ,/-A,

S

I

I

I - i

MLR

Human: -1.8 cm

6th chromosome

- - -

b

HLA-D

Bf

-

Ilr

HLA-A

HLA-C

HLA-B

MLR

Guinea pig:

Ir Genes Ir Genes

Ir Genes

Ir Genes

I

PLL,GA

Z

DNP,BA 3,5

7 I

'

4,5,(6)

(6)

I

strain-13 Guinea Pig

I

I

I

4

GT

7 I

strain-2 Guinea Pig

7

I

I

1,6

FIG. 3. MHC for mouse [based on Schreffler et al. (1977)],human [based on Van Rood et al. (1977)], and guinea pig [based on Shevach et al. (1977)J.GPA, Guinea pig albumin.

156

MARC FELDMA" ET AL.

also products of H-linked genes provides useful biological probes of the function and cellular location of molecules which play critical but as yet incompletely defined roles in T-lymphocyte activation. Alloantisera blocked the proliferative responses controlled by Ir genes linked to the gene coding for the alloantigen but had little effect on the response controlled by an Ir gene which was a product of the other parental haplotype in an F1 animal (Shevach et al., 1972; Braendstrup et al., 1978). One interpretation of such findings is that Ir genes control the production of a cell surface product which plays a role in the mechanism of antigen recognition by T lymphocytes. Because of the specificity of Ir gene effects, the inhibitory activity of the alloantisera was initially presumed to be directed solely against the proliferating T lymphocytes. However, our current understanding of the central role of macrophages in the activation of T-cell proliferation suggests that alloantisera act on the macrophages, perhaps by blocking their interactions with the T cells. Unfortunately experiments which attempt to define the cellular site of action of the alloantisera using combinations of antigen-pulsed macrophages and T cells are difficult to perform, because interactions of macrophage-associated antigens and T cells usually occur only when the macrophages and T lymphocytes share identity of gene products linked to the guinea pig MHC (Shevach and Rosenthal, 1973). Notable exceptions are the recent studies of Thomas and Shevach (1976, 1977), using an in vitro primary proliferative assay in the guinea pig in which alloantisera were used as antigens, or TNP-conjugated macrophages. It is unlikely that the histocompatibility restrictions observed for macrophage-T-cell interactions were the result of an active process of rejection that occurred whenever histoincompatible cells were mixed together, as significant T-cell proliferation was observed when F1 macrophages were added to parental T cells, and vice versa. There are, however, several special conditions under which macrophages can paradoxically induce allogeneic lymphocyte proliferation. These circumstances show that the failure to recognize soluble protein antigens bound to allogeneic macrophages is not a primary inability of such cells to present signals but may instead reflect a basic feature of antigen handling by macrophages. Both human (Schumacher et al., 1975) and guinea pig (Ellner and Rosenthal, 1975, 1977) macrophages are very effective stimulators of allogeneic lymphocyte proliferation. In the guinea pig MLR, little stimulatory activity can be attributed to T lymphocytes, although in the mouse both B cells and macrophages are effective stimulators whereas T cells are less effective. Thus, while soluble proteins are displayed in an immunologically relevant fashion only on syngeneic or semisyngeneic macrophages, it is the histoincompatible macrophages that induce T-cell proliferation in the MLR. This finding suggests that the critical difference lies in the display of alloantigen as opposed to that of foreign proteins. In another model system, chemical alteration of either macrophage-depleted lymphocytes or macrophage plasma membrane glycoproteins by treatment with

MACROPHAGE-LYMPHOCYTE INTERACTIONS

157

either sodium metaperiodate or neuraminidase-galactose oxidase generates free aldehyde groups on different terminal sugar groups (Novogrodsky and Katchalsky , 1972). Subsequent and rather massive lymphocyte proliferation will ensue if and only if modified lymphocytes are combined with unmodified macrophages or modified macrophages are combined with unmodified lymphocytes. Moreover, proliferation occurs irrespective of the histocompatibility difference between the macrophages and lymphocytes and, upon mixing modified and unmodified cells, a striking clustering of lymphocytes about macrophages is also observed. The latter observations have led to the suggestion that immune recognition may in fact consist of at least two separate, but linked, events. The first phase of this interaction involves the selective binding of antigen-specific T cells by macrophages bearing that antigen. The second or activation phase of macrophage-lymphocyte interaction is a relatively nonspecific event which follows as a consequence of stabilization of a physical macrophage-lymphocyte interaction initiated by the antigen-specific selection process. We can assess, at a functional level, the site of action of alloantisera in situations in which allogeneic macrophages are capable of activating T-cell proliferation. Thus the MLR between two inbred strains is markedly inhibited by an alloantiserum directed against the stimulatory macrophages, while the same serum has no effect on the responding lymphocytes. In a similar fashion, when T lymphocytes are treated with sodium periodate and mixed with untreated macrophages, the resultant stimulation is markedly inhibited by an alloantiserum directed against the macrophages but not against the proliferating T cells. Although these studies do not define the site of action of the alloantisera for soluble protein antigen-induced proliferation, they suggest that certain T-cell proliferative responses can be markedly inhibited by alloantisera acting solely on the stimulatory macrophages. Thomas and Shevach (1976, 1977) also found that the alloantisera inhibited the macrophages. Another approach to localizing the site of alloantisera blockage is to examine directly the effect of alloantisera on antigen-induced macrophage-lymphocyte interaction. Binding of lymph node cells by antigen-pulsed macrophages was determined after 20 hours of culture at 37°C. Anti-strain-2 serum completely inhibited the binding of strain-2 lymphocytes to strain-2 macrophages pulsed with a dinitrophenylated copolymer of glutamic acid and lysine (DNP-GL) to a greater extent than to macrophages pulsed with purified protein derivative (PPD) (Braendstrup et al., 1978). This result suggested that the inhibition of T-cell proliferation observed by measurement of th~midine-~H incorporation was secondary to a blockage of the physical interaction of macrophages and T cells, rather than to an interference with T-cell proliferation subsequent to antigenspecific binding. Again, as with th~midine-~H incorporation, the site of action of this inhibition of antigen-specific binding cannot be determined. One can, however, examine the binding of strain- 13 lymphocytes to strain-2 neuraminidase-

158

MARC FELDMANN ET AL.

galactose oxidase-treated macrophages. In such experiments, aldehydedependent binding is completely inhibited by an alloantiserum directed solely against the macrophages (Greineder et al., 1977). D. GENETICANALYSIS OF DETERMINANTS MEDIATING MACROPHAGE-T-LYMPHOCYTE INTERACTION An additional restriction on the interaction of antigen-pulsed macrophages with immune T cells was observed when macrophages, obtained from a parent that lacked a given Ir gene, were pulsed with an antigen controlled by that gene and then mixed with nonresponder x responder F1 T cells. Thus macrophages from nonimmunized strain-2 and strain-2 x strain-13 F, guinea pigs pulsed with DNP-GL (an antigen the response to which is controlled by a strain 2-linked Zr gene) activated immune F, T-lymphocyte proliferation equally, while the magnitude of stimulation observed when strain-13 DNP-GL-pulsed macrophages were used was approximately 1/10 that seen when strain-2 or strain-2 X strain-13 F, macrophages were used. One possible explanation for the failure of the nonresponder macrophages to activate the nonresponder X responder F, T cells is an intrinsic defect in Ir gene product function in the nonresponder macrophages. The most direct approach to the analysis of the antigenic determinant controlling macrophages and/or T cells is to use cells derived from an animal which bears a recombinant chromosome in which the genes which code for the alloantigens and Ir gene products have been separated. Unfortunately, no animals possessing recombinant chromosomes have been found. These observations have recently been extended to the mouse by Schwartz and co-workers (1978). Another explanation is suggested by studies on the mouse, which have shown that T cells sensitized to haptens (Shearer et al., 1975) or viruses (Zinkernagel and Doherty, 1977) are primarily cytotoxic for targets which are H - 2 D or H-2K compatible (H-2 is the mouse MHC). The explanation offered for those results was that T cells were sensitized to a new antigenic determinant or NAD, in the form of complexes of hapten or virus-altered “self” antigens. These altered self antigens are viewed as the products of the H-2K or H - 2 D , or of closely linked genes. Although the products of the I region appear to play no role in the expression of specificity of T-cell cytotoxicity , the explanation offered for these results might have relevance for an understanding of the requirements for macrophage-T-lymphocyte interaction in the induction of T-cell proliferation. One might thus postulate that the T cell is incapable of recognizing native antigen and can be sensitized only to antigen-altered self MHC products; the failure of interaction between allogeneic macrophages and T cells could be regarded as a failure of recognition rather than interaction. Strain- 13 macrophages pulsed with PPD fail to activate immune strain-2 T cells not because the strain-2 animals have never previously recognized PPD, but rather because during the primary im-

MACROPHAGE-LYMPHOCYTE INTERACTIONS

159

munization they have been sensitized to “PPD-altered self”; in our experiments the self antigens appeared to be products of the I region of the guinea pig MHC (Paul et al., 1977). Unfortunately, the antigen-altered self hypothesis does not adequately explain the restriction observed in studies in which antigen-pulsed nonresponder macrophages failed to activate response x nonresponder F1 T cells. T cells derived from an immunized F, animal should be equally stimulated by the antigen-altered Ia complex of both parental haplotypes and should respond equally well to antigen-pulsed macrophages of either parent. Indeed, such a response is observed when F1 immune T cells are mixed with parental macrophages pulsed with PPD, an antigen not under Ir gene control. In order to explain the defect in nonresponder macrophages one must postulate that they lack the Ia antigen which can be suitably altered by the antigen under study. Additional limitations of the NAD model for cellular selection in the recognition of soluble protein antigens are: (1) It is difficult from a chemical point of view to postulate a family of proteins or glycoproteins whose plasticity is such that a sufficient range of conformational changes can be induced in them by binding antigens to explain the exquisite sensitivity of the T-cell recognition process. (2) A NAD thesis implies that some of the antigenic determinants recognized by the T cell are exclusive, since antibody maintains the ability to interact directly with antigen in free solution, that is, out of context of self, while the T-cell receptor does not recognize antigen except in the context of self. Recently, another model of dual recognition has been proposed in which the T cell recognizes a compound determinant of self (Ia) plus antigen. Such models present difficulties as well, since antibody or antigen should block T-cell functions. While one can clearly demonstrate blockage of T-cell proliferation with anti-Ia sera, no reproducible inhibition of soluble protein antigen-induced T-cell phenomena has been observed with antibodies directed against the antigenic contributor of the compound determinant. Rosenthal and co-workers (1978) have favored a simpler and perhaps more unifying hypothesis which states that a given macrophage’s repertoire of Ir gene products functions to specify determinant selection. This hypothesis can be tested most directly by assessing the contribution of antigen structure, that is, amino acid sequence and conformation, on macrophage function. Most complex natural antigenic proteins such as PPD and synthetic polymers such as GL, have neither defined secondary nor tertiary structure and thus do not permit precise intramolecular mapping of the areas responsible for immunogenicity. In contrast, polypeptides of known structure such as insulin (Blundell, 1972) offer a unique opportunity to characterize the regions of the molecule recognized by either the T-cell receptor or antibody. The insulin monomer, for example, consists of two chains connected by interchain disulfides (Fig. 4). The A chain has two helical portions separated by a loop region (a loop) formed by an

160

MARC FELDMANN ET AL.

FIG.4. Beef insulin A and B represent chains, numbers refer to amino acids numbered from N terminal. A8, A9, and A10 comprise the amino acids of the (I loop (Blundell, 1972).

intrachain disulfide between cysteines 7 and 11, while the B chain consists of a j3 pleated-sheet segment at the amino terminus separated from the carboxy terminal j3 chain by a helical region. The immune response to insulin, in both the mouse (Keck, 1975; Rosenthal et af., 1978) and guinea pig (Barcinski and Rosenthal, 1977; Rosenthal et af., 1978) is under the control of MHC-linked Ir genes and is summarized in Table I. When immunized with either pig or beef insulin in complete Freund’s adjuvant (CFA), both strain-2 and - 13 guinea pigs respond by antigen-specific lymphocyte proliferation and synthesis of specific antibody. The specificity of the elicited antibodies is indistinguishable in these inbred strains; however, the precise determinants are not yet known. In contrast, strain-2 T cells recognize a distinct region of the A chain (a-loop responder), and H-2d (B-chain responder) mice similarly discriminate which area of the molecule is recognized by their T lymphocytes. The function of the Ir gene in,both the guinea pig and mouse appears to determine an intramolecular selection of discrete regions within the antigen for

MACROPHAGE-LYMPHOCYTE INTERACTIONS

161

TABLE I GENETICS OF THE IMMUNE RESPONSE TO INSULIN AS MEASURED BY T-CELLPROLIFERATION" ~~

~~

Insulin

Guinea pig Strain 2 Strain 13 F, (strain 2

Mouse H-2b H-2d H-2k H-2"

X

strain 13)

Determinant selected

Pig

Beef

B chain

+ +

+ + +

-

a loop B chain a loop and B chain

-

+ +

-

a loop B chain none none

+ +

-

-

+ + + -

"Data from Keck (1975) and A. S. Rosenthal ef al. (unpublished).

recognition by the T cell and that this function operates at the level of the macrophage and not at the level of the T cell. The data which support such an interpretation are as follows. If Ir gene function operates exclusively at the level of the responding lymphocyte, both determinants should be simultaneously recognized by F, T cells independently of which parental macrophage presents the antigen. Conversely, if the definition of which determinant is actually recognized depends on the genetic profile of the antigen-presenting macrophage, Ir gene function must also operate at this cell level. Macrophages from both inbred strains of guinea pig can present pig insulin to pig insulin-immune strain-2 x strain-13 F1 lymphocytes. As with PPD, a strong proliferative response is observed when pig insulin-pulsed macrophages from either strain of guinea pig are added to F, peritoneal exudate lymphocytes (PELs). However, when different species variants of intact insulin are presented to the same pig insulin immune F, PELs by either strain-2 or -13 macrophages, two different patterns of response are seen. An A-chain loop pattern of cross-reactivity is observed when the antigens are presented on strain-2 macrophages. This pattern of cross-reactivity is characterized by partial crossreactivity between pig and beef insulin (partial identity in the loop) and little cross-reactivity between pig and sheep insulin (no identity in the loop). In order to corroborate the above findings and to show that a specific B-chain determinant is being selected when the whole insulin molecule is presented by strain-13 macrophages, we studied the response of oxidized insulin B-chain-immune strain-2 x strain-13 F, PELs to parental macrophages pulsed with native insulin

162

MARC FELDMA" ET AL.

or isolated B chain. Only strain-13 macrophages can present native insulin to B-chain-immune strain-2 x strain-13 F, PELS. Despite the ability of strain-2 macrophages to present the same insulin to insulin-immune F, T cells, they cannot initiate DNA synthesis on B-chain-immune F1 T cells. Such studies raise the possibility that, in strain-2 X strain-13 F, animals immunized to insulin, two distinct sets of clones of T cells are generated: one that recognizes the a-loop antigenic determinant processed by or presented by the strain-2-specificcell structures in the F, macrophage, and one that recognizes the B-chain determinant generated by the strain-13 counterpart of the system. This hypothesis has been examined using bromodeoxyuridine (BdU) and light elimination of experiments in F1 guinea pigs. Initial culture of insulin-bearing parental macrophages with insulin-immune F1 T cells in the presence of BdU and light selectivity decreases responsiveness on subsequent exposure to insulin-bearing macrophages identical to those originally used to elicit cell activation. Thus preculture of F1 T cells with strain-2 macrophages bearing insulin significantly depresses (65%) their response to antigen on strain-2 macrophages on subsequent culture without comparably effecting responsiveness to insulin presented by strain-13 macrophages. Conversely, the exposure of F, T cells to strain-13 macrophages bearing pig insulin eliminates 87% of their responsiveness to pig insulin on strain-13 macrophages on repeat culture but does not effect the response of F, T cells to strain-2 macrophages bearing pig insulin. Such data show that at least two distinct clones of insulinreactive T cells exist in F, animals. Although immune specificity is traditionally considered a lymphocyte function (Burnet, 1959), the following observations support our interpretation that macrophages have a discriminatory function in the antigen recognition process. 1. The response of strain-2 x strain-13 F, guinea pigs primed to the synthetic random copolymers GL and L-glutamic acid-L-tryosine (GT) is under the control of MHC-linked Zr genes. T lymphocytes from such animals efficiently recognize antigen bound only to parental macrophages of responder origin (strain-2 for GL and strain-13 for GT) while macrophages from both parental strains can present, to F1T cells, antigen not under unigenic control (Rosenthal and Shevach, 1973). 2. When both the Zr gene and its linked alloantigen were deficient in the macrophage donors, macrophages pulsed with an antigen (the response to which was controlled by the Zr gene) failed to activate proliferation of T cells in animals that possessed both the alloantigen and the appropriate Zr gene. Thus T-cell proliferation did not occur when macrophage and T cell shared only other Ia antigenic specificities normally found in the same haplotype but not closely linked to the Zr gene under study (Shevach, 1976). 3. Alloantisera recognizing MHC-linked cell surface determinants, Ia antigens, can block T-cell proliferation in response to antigen-bearing chemically

MACROPHAGE-LYMPHOCYTE INTERACTIONS

163

modified and allogeneic macrophages (Greineder et al., 1977). In the latter two situations, DNA synthesis could be initiated across an MHC barrier and thus allowed definition of the site of alloantisera blockage. Such experiments showed that the inhibitory activity of the alloantisera was directed against the Ia antigenic specificities of the participating macrophages and not against those of the responding lymphocytes. 4. Studies in mice (Pierce et al., 1976) and guinea pigs (Thomas and Shevach, 1977) have shown that secondary T-cell responses are elicited not against the intact priming antigen alone but rather against antigen in association with MHC-linked cell surface structures present on the original antigenpresenting cell. While not exluding the possibility that Ir genes also function at the T-cell level, our data suggest that macrophages play a fundamental role in selecting, in a complex antigen, the moiety(ies) recognized by immune T cells. Generation or display of such antigenic determinants is thus a function of Ir genes operating at the level of the antigen-presenting cell. Two general mechanisms by which macrophages might serve such a function can be proposed. One possibility is that Ir genes define a class of receptors of broad specificity which recognize molecular shape and thus have the unique ability to focus or orient distinct regions of the antigen for presentation to the T cell. A second possibility is that Ir gene products are, or regulate, the activity of families of enzymes which modify or metabolize polypeptide antigens. In the latter situation, the repertoire of Ir genes associated with a given haplotype would define restricted areas of the molecule available for display to the T-cell receptor.

111. Macrophage Function in Helper T-cell Induction

A. NATURE OF MACROPHAGE-T-CELL INTERACTIONS IN HELPER CELL

INDUCTION

In vivo systems for helper cell induction have not been analyzed for their macrophage requirement, although this could be attempted using techniques for macrophage depletion in vivo such as those employing silica or carrageenan. In vitro, helper cells can be induced by low doses of antigen in Marbrook-type flasks (Kontiainen and Feldmann, 1973). The process is notably dependent on the presence of adherent cells; the passage of spleen, lymph node, or cortisoneresistant T cells through glass or polystyrene beads, nylon wool, or carbonyl iron abrogates their capacity to yield helper cells (Erb and Feldmann, 1975b). Restoration of the helper cell response with adherent, radioresistant, phagocytic

164

MARC FELDMANN ET AL.

peritoneal exudate cells indicates that the adherent cells involved are indeed macrophages, although it is by no means clear whether all macrophages can participate in this way. The mechanism of macrophage function in helper cell induction in vitro was investigated. The first question investigated was whether macrophage-T-cell contact was essential. Two approaches were employed. First, cultures in a double chamber, separated by a cell-impermeable membrane, were used; T cells were cultured in one compartment and macrophages in the other. Second, supernatants were generated from macrophages cultured in the presence of antigen and tested for their capacity to replace macrophages (Erb and Feldmann, 1975b,c). Both approaches indicated that in vitro contact was not essential. However, these results do not have much bearing on the question whether macrophages-T-cell contact is an alternative which is possibly more efficient. It should be pointed out that, in the mouse, in vivo and in vitro, there is much morphological evidence that macrophage-T-cell contact occurs, as indicated above in the guinea pig studies (Section 11,A). Lohmann Matthes et al. found that T lymphocytes associated with antigen-bearing macrophages in the omentum (Lohmann Matthes et al., 1971). Mosier, Pierce, and others (e.g., Mosier, 1969) reported the presence of, and requirement for, clusters of cells for the induction of antibody response in vitro. The existence of macrophage coooperative activity in a cell supernatant permits identification of the component(s) involved. At least two categories have been identified so far in our studies: (1) those which are antigen-specific by virtue of their antigen and are called GRF (Erb and Feldmann, 1975) and (2) those which are not antigen-specific and are not genetically restricted. These induced only helper cells in the presence of particulate antigens and were termed nonspecific macrophage factor (NMF) (Erb and Feldmann, 1975b,c). These factors are summarized in Table 11. Many other macrophage factors have been identified by other workers, including lymphocyte activity factor (LAF), involved in augmenting the PHA response (Gery and Waksman, 1972), and others which stimulate proliferation and differentiation of lymphocytes (e.g., Calderon et al., 1975) (Table 111). It is conceivable that LAF, GRF, BAF, and MCF are all closely related or identical molecules. Attempts to characterize NMF have so far been difficult, suggesting a high degree of heterogeneity, and its relationship to other mediators, such as LAF, is not clear. GRF has been easier to purify and was found to consist of two components; a fragment of the immunogen linked noncovalently to an Ia antigen (Erb et al., 1976a). The total molecular weight of the complex, as determined by Sephadex chromatography, was -55,000. As yet there is no information on the detailed biochemical nature of the Ia component-its molecular weight, number of chains, and so on. Serological analysis, however,

165

MACROPHAGE-LYMPHOCYTE INTERACTIONS TABLE I1 CHARACTERISTICS OF THE MACROPHAGE FACTORS GRF AND NMF PropeflY Source Production requirement Effect

Nature I subregion Molecular weight Heat stability Enzyme sensitivity Target cells Receptor on target Genetic restriction

GRF Peritoneal macrophages Antigen Induces helper cells Ia-Ag complex I-A 50,000-60,000

Labile Trypsin, papain Ly 1+2+3+T cells Yes, I-A control Yes

NMF Peritoneal macrophages Medium Induces helper cells with particulate antigens Unknown None <30,000 Labile Trypsin, papain Ly l+2-3-T cells Unknown No

suggested that the factor [CBA mouse strain reactive to keyhole limpet hemocyanin (KLH)] was a product of the Z-A subregion (Erb et al., 1976). The mechanism of action of the factors has been investigated. GRF acts on T cells, and it only binds to certain T cells. On functional analysis it was shown to bind, in Cantor and Raff’s nomenclature, the short-lived pool of T cells (Erb et al., 1977). These cells are not helper cell precursors but are T cells which interact with helper cell precursors (see Fig. 5 ) . They are termed amplifier cells and have the Ly l+2+3+membrane phenotype (Feldmann et al., 1977a). More recently, evidence was found that GRF acted also on long-lived helper cell precursors (Ly l+2-3-), in experiments with supernatants of Ly 123 cells which had been activated by GRF. These stimulated Ly 1 cells effectively only if the latter were also incubated with GRF (Erb and Feldmann, 1979). Events subsequent to the binding of GRF have yet to be clearly delineated. While it is clear that these T cells are activated and release mediators involved in the activation of helper cell precursors (Feldmann et al., 1977), the mechanism of activation of these precursors is not yet understood. It is puzzling that alterations in the physical form of the antigen, by coupling antigen, say KLH, onto Sepharose beads causes drastic changes in the cellular pathway of helper cell induction. For example, the genetic restriction on macrophage-T-cell interactions was abolished, as was the need for T,-cell-T,-cell interaction. This clearly implies a heterogeneity of T-cell induction pathways, as also exists for B-cell induction (e.g., Feldmann and Basten, 1971), and indicates that the NMF needed with particulate antigens acts on helper cell precursors. It is not known whether NMF also acts on T, cells, but the capacity to respond to antiantigen or anti-Ia immune absorbent purified GRF, presumably free of NMF, suggests that TI cells do not need NMF. There is a clear analogy between the simpler mechanism of

TABLE III CHARACTERISTICS OF THE MACROPHAGE FACTORS LAF, BAF, Property Source

Production requirements Effect

Target cells Nature Molecular weight Heat stability Enzyme sensitivity 'Gery and Waksman (1972).

bWood et al. (1974). %dderon and Unanue (1975).

AND

MCF

LAF"

BAFb

Peritoneal macrophages, spleen adherent cells, human monocytes Endotoxin, PHA Potentiates mitogen response of T cells

Peritoneal macrophages, human monocytes

Peritoneal macrophages

Endotoxin, PHA Potentiates humoral response

Phagoc ytosis Potentiates cell proliferation, mitogen response of T cells, humoral response Thymocytes, T cells, B cells Unknown 15,000-2 1,000 Labile Chymotrypsin, pepsin

Thymocytes, T cells Unknown -15,000 Labile Not tested

B cells Unknown -18,000 Not tested Trypsin

MCFr

MACROPHAGE-LYMPHOCYTE INTERACTIONS

Macrophage

MW

-

55,ooO

167

Factor

FIG. 5. Helper cell induction.

helper cell induction with polymeric antigens and the thymus- and macrophageindependent pathway of B-cell activation with such antigens.

B. GENETIC RESTRICTIONS ON MACROPHAGE-T-CELL COOPERATION Rosenthal and Shevach (1973) first demonstrated that there were restrictions on the genotype of the macrophages which interacted with T cells in a proliferation assay using antigen-pulsed macrophages and primed peritoneal T cells. Thus it was of interest to determine whether another example of macrophage-T-cell interaction, in the induction of helper cells, also had genetic restrictions mapping in the MHC. Thus allogeneic macrophages and T cells were cultured. No helper cells were induced with macrophages differing at the H-2 complex, regardless of the ratio of T cells or macrophages, antigen concentration, or time of assay in our studies or in analogous ones by Gorczynski and McDougall (Erb and Feldmann, 1975b,c; Gorczynski, 1977; McDougall and Gordon, 1977). The genetic region involved in this interaction was defined by using congenic resistant lines of mice and mice bearing recombinations within the H-2 complex. For two antigens, this region was the I-A subregion of the H-2 complex. It is not yet clear whether this is true for all antigens, for example, those under control of Ir genes mapping in the I-C region. This possibility is being tested. The products of the I-A gene(s) involved in macrophage-T-cell cooperation were investigated. Macrophages produce GRF, which is one product of the I-A

168

MARC FELDMANN ET AL.

genes involved. The receptor for GRF, or part of it, is also controlled by the I-A subregion; only T cells absorb GRF and only if they are I-A-compatible with the source of GRF (Erb et al., 1976). Attempts have been made to identify this receptor for GRF by performing blocking experiments with anti-Ia antisera shown to react with GRF. It was found that no inhibition of GRF binding (and subsequent T-cell activation) occurred after pretreatment with anti-Ia antisera (Erb et al., 1977). These results suggest that the receptor for GRF does not have the same antigenic determinants expressed as GRF, suggesting that it may be the product of another gene. This conclusion is supported by the fact that macrophages which make GRF do not absorb it (i.e., do not have receptors), and T cells which absorb GRF do not make it. The significance of these genetic restrictions is not at all clear. While other immunological assay systems have provided evidence for similar MHC genetic restrictions, there are possible exceptions. Genetic restrictions have been found in secondary antibody responses to macrophages pulsed with GAT, a synthetic polypeptide (Pierce et al., 1976), in proliferative responses of primed mouse PELS to antigen-pulsed macrophages (Yano et al., 1977), and in the transfer of delayed-type hypersensitivity (Miller et al., 1976). In contrast, no genetic restrictions were found in the primary antibody response to antigen (GAT)-pulsed macrophages (Pierce et al., 1976). Unfortunately, these various reports of genetic restrictions or of their lack are difficult to interpret in an unequivocal manner, as the assay systems were not unambiguous. Three aspects of the question of genetic restriction can be considered separately: (1) whether genetic restrictions can be induced by priming with antigen associated with certain I-region products; (2) whether these restrictions must involve shared I regions; and (3) whether restrictions occur prior to priming. 1. Induction of Generic Restrictions

The induction of genetic restrictions by priming has been shown in various systems (Pierce et al., 1976), in the antibody response in vitro (Miller et al., 1977), in the delayed-type hypersensitivity response in vivo, and in earlier studies using F1 T cells and parental macrophages in the guinea pig (Shevach and Rosenthal, 1973). No dissenting reports have come to our attention. We have also examined this in the generation of helper cells, using F1 T cells and parental macrophages. Clones of helper cells were induced which reacted preferentially with the same macrophages (Erb et al., 1978), as judged by their interaction with spleen cells of the appropriate parental type. This concept of induction of genetic restrictions by priming is not really new. It merely restates the concept of immunological memory, with the relevant antigen not being just the immunogen but a complex of immunogen and Ia antigen.

MACROPHAGE-LYMPHOCYTEINTERACTIONS

169

2. Genetic Restriction and I-Region Compatibility Different results have been obtained in different systems concerning the question of whether genetic restriction involves I-region compatibility. Pierce reported that there was no need for I-region compatibility. C57BL/6 mice primed with GAT-pulsed DBN2 mice macrophages could be restimulated only by the latter. This result at first appears to be a simple example of immunological memory, but it is not. Why cannot self (C57BL/6) macrophages induce a primary response as they can prior to the injection of allogeneic macrophages? There appeared to be some degree of suppression, and preliminary results reported by Pierce (1978) suggested that this was indeed the case. Other examples require I-region compatibility; the transfer of delayed-type hypersensitivity in vivo does not occur in allogeneic situations (Hiller et al., 1976), nor does the induction of helper cells (Erb and Feldmann, 1975c) or antigen-induced proliferation in the mouse (Schwartz and Paul, 1974; Yano et al., 1977) and in the guinea pig. The transfer of delayed-type hypersensitivity could occur in allogeneic situations, provided cells from tetraparental bone marrow chimeras were used, which were tolerant to the alloantigens of the host (Miller et al., 1977). This is again compatible with the concept that genetic restrictions may be due to suppression. Analogous experiments have been performed in a helper system with such chimeras to test whether helper cells can be induced with allogeneic macrophages to which the T cells are tolerant in a system where suppressor cells cannot be generated. The results of these experiments were unequivocal-the restrictions were not broken (Erb et al., 1978). Nor have experiments attempting to break restrictions by treating helper cells with antisera which kill suppressor cells (anti-Ly 2, anti-Ly 3, anti-Ia antigen) yielded different results. Thus there is no consensus as to whether I-region compatibility is needed. It is conceivable that situations not requiring compatibility may stimulate by a different pathway, using the potent stimulatory action of the allogeneic reaction. Alternatively, the T cells involved may be sufficiently different to require a different mechanism of macrophage-T-cell interaction. It was reported by Pierce et al. (1976) that syngeneic and allogeneic antigenpulsed macrophages induced equivalent primary responses. This was interpreted as evidence that genetic restrictions do not occur in primary responses. Other interpretations can be proposed to explain these findings-it is possible that allogeneic macrophages stimulate an allogeneic reaction, with the liberation of stimulatory factors. This would be a mechanism of induction different from that with syngeneic macrophages and invalidates the concept that allogeneic macrophages can function just as effectively. In the induction of helper cells using unprimed T cells, allogeneic macrophages do not function even if used in both the helper cell induction and expression phase (i.e., with the potential for re-

170

MARC FELDMA" ET AL.

stimulation). However, in preliminary experiments using allogeneic macrophages in the helper cell induction and expression phase, genetic restrictions were broken provided T cells capable of recognizing the macrophages were present. This experiment suggests that this breaking of genetic restrictions was due to allogeneic stimulation (P,Erb, unpublished data). The problem of allogeneic stimulation in testing whether allogeneic macrophages can function suggested the use of other techniques to investigate genetic restriction. Since the binding of GRF to T cells correlates well with subsequent stimulation, it is possible to use a binding assay (not susceptible to allogeneic stimulus) to study the receptors on T cells. It was found that unprimed spleen cells bound only I-A-compatible GRF. This result may not be conclusive, since it is possible that the spleen cells were primed by other antigens coupled to self Ia antigen and thus the clones of cells recognizing self I-A were augmented. Thus thymus has also been used, yielding analogous results (P. Erb, unpublished). These results indicate that the frequency of receptors recognizing self Ia antigen (I-A-compatible)is greater than for foreign Ia. It does not exclude some recognition of foreign Ia antigen, and perhaps the simplest reconciliation of conflicting data is to propose that, while there are fewer clones of cells (capable of generating helper cells) recognizing allo-Ia antigen, they may be expanded more effectively under the stimulus of allogeneic factors, eventually yielding the same number of helper cells. Recently, Gorczynski (1976, 1978) attempted to analyze some of the conflicting results concerning genetic restrictions. He found, using size separation procedures, that different macrophages were involved in responses with syngeneic and allogeneic lymphocytes. Small (4-7 mm per hour sedimentation velocity) macrophages (anti-0-treated and irradiated normal peritoneal cells) cooperated only in the antibody response to sheep red cells (SRC) with syngeneic cells. However, large macrophages (10-14 mm per hour) were effective in reconstituting the response to SRC of allogeneic lymphocytes, which were tolerant to the alloantigens of the macrophage donor. It was of interest that, when nontolerant lymphocytes were used with the consequent allogeneic effect, small macrophages were also effective in restoring the SRC response of allogeneic lymphocytes. Similar results were obtained, using supernatants of macrophages, with small macrophages releasing supernatant factors capable only of stimulating a SRC response to syngeneic lymphocytes. Large macrophages released factors which stimulated both syngeneic and allogeneic lymphocytes. The factors released by these macrophages were analyzed. All released molecules with a molecular weight of 30,000-50,000 and 50,000-70,000 only stimulated syngeneic cells; but only large macrophages released a product of MW 10,000-15,000 which reconstituted both syngeneic and allogeneic cells, the latter product being responsible for the different functional effects. Both the factors described were inactivated by proteolytic enzymes such as pepsin.

MACROPHAGE-LYMPHOCYTE INTERACTIONS

171

It is noteworthy that in Gorczynski’s studies the responses assayed were to SRC, a particulate antigen for which macrophage restrictions have not been found when unfractionated macrophages are used (Gorzynski, 1977; Shortman and Palmer, 1971). These studies have not only provided good evidence for functional macrophage heterogeneity, but have also demonstrated that the different results obtained are due to a heterogeneity of stimulatory mechanisms. It will be of interest to compare the macrophage factors which act on SRC (or cytotoxic) responses with those which act on protein responses, such as GRF and NMF.

IV. Macrophage-B-Lymphocyte Interactions in Antibody Production The antibody response to thymus-dependent antigens depends on macrophages. The previous discussion on the requirement for macrophages in helper cell induction indicated one site of macrophage function, but there is also evidence for a second locus involving macrophage-B-cell interaction (Feldmann, 1972). The basic experimental model for these studies involves the use of antigen-specific T-cell helper factor (HF) (Feldmann, 1972; Howie and Feldmann, 1977). HF stimulates T-cell-depleted spleen cell populations in vitro, but not macrophage-depleted populations (Howie and Feldmann, 1978). This result applies to responses which are or are not under Ir gene control. Further analysis revealed that macrophages, incubated with HF and antigen and subsequently washed, stimulated B cells as effectively as HF admixed with B-cell antigen and macrophages (Howie and Feldmann, 1978). Unlike the macrophage-T-cell interaction in the same culture system, the macrophage-Bcell interaction appears to require cell contact, as attempts to demonstrate macrophage-B-cell interaction in a double-chamber culture flask with the two cell pools separated by a cell-impermeable membrane have not succeeded (unpublished data). Recently analysis of the genetics of the HF-macrophage-B-cell interaction began. Experiments using responder x nonresponder B cells, and macrophages derived from either responder x nonresponder F,, and parental, responder or nonresponder mouse strains have revealed a close parallel to previous results with macrophage-T-cell interaction. The fact that nonresponder macrophages cannot stimulate F, B cells indicates that Ir genes are expressed at the macrophage level in this response (Howie and Feldmann, 1978). This raises interesting questions: Is the macrophage-B-cell interaction I region restricted? If so, which subregion? Attempts are being made to answer these questions. Published studies on this point have been difficult to interpret, as allogeneic effects often are not excluded. Kapp et al. (1973) found no genetic restriction in response to GAT, nor did Palmer and Shortman (1971) in the primary response to SRC. In

172

MARC F E L D M A " ET AL.

/r

TABLE IV GENEEXPRESSION I N MACROPHAGE-9-CELL INTERACTION^.^ IgM AFC/culture

Cells cultured"

Treatment

Anti-(T,G)-A--L

Anti-DNP

None Plus HF Minus macrophage plus HF Minus macrophage plus F, MI#J plus HF Minus macrophage plus B10 MQ plus HF Minus macrophage plus 9IO.Br MQ plus HF

13 210 10

31 200 30

203

253

300

263

23

217

~~

910 x 9 I O . B r spleen

910 x 91O.Br spleen 910 x 91O.Br spleen 910 x 9IO.Br spleen

B10 x 91O.Br spleen 910 x 9IO.Br spleen

"Data from Howie and Feldmann (1978). bResponse of responder x nonresponder F, spleen depleted of macrophages to HFTCALin the presence of responder or nonresponder F, macrophages is shown. Note that the response to the DNP group of DNP-(T,G)-A--L occurs with any macrophage, but to (T,G)-A--L only in the presence of responder macrophages.

the secondary response, the results of Katz and Unanue (1973), who stimulated DNP-KLH-primed mice with antigen-pulsed syngeneic or allogeneic macrophages, also suggest a lack of genetic restriction. However, the initial macrophages were not removed, and the results are thus not interpretable. Another complexity in the macrophage-B-cell interaction is the question of determinant selection in the HF-macrophage-B-cell interaction. This is illustrated in Table IV, using DNP-(T,G)-A-L as antigen and HFTCAL to stimulate spleen cells from responder or nonresponder mice. It was found that in responder mouse spleen cells HFTCAL induced a response to both DNP and (T,G)-A-L determinants, but in nonresponder mice there was only a response to DNP (Howie and Feldmann, 1978). The reasons for this are not clearly understood, but it may be that the model of HF action of B cells proposed by Munro and Taussig (1975) is too simple, as it cannot account for this observation. It is tempting to speculate that the function of the macrophage Ir genes in HFmacrophage-B-cell cooperation may be to orient the antigen-HF complex in an appropriate way, as proposed by Rosenthal and colleagues (1978) for insulin and the macrophage-T-cell interaction.

V. Concluding Discussion Much attention has been paid to the fact that T cells recognize antigen preferentially associated with alloantigens of the MHC. For killer T cells, against

MACROPHAGE-LYMPHOCYTE INTERACTIONS

173

viruses, these are the H-2K and H-2D antigens present on essentially all cells of the body. For helper cells and cells proliferating in response to antigen the relevant alloantigens are those of the I region, which contain the Zr genes. These studies have aroused much excitement because they suggest some insight into the function of the highly polymorphic H-2 and I-region antigens and yield clues about the most fundamental aspect of the immune mechanism: How do immunologically specific receptors recognize antigen in an immunogenic, that is, effective, way and initiate the complex events of the immune response? Many theories have been proposed to account for the mechanism of recognition by T cells of antigen linked to MHC products (McDevitt, 1978). It should be stressed that so far there is no conclusive evidence for B-cell recognition of antigen linked to MHC products, implying that B-cell activation pathways must fundamentally differ from T-cell activation pathways. One theory, which for simplicity has been termed the altered self theory, proposes that the target for T-cell recognition (leading to killing, proliferation, and so on) is modified self antigen, modified by a virus, chemical coupling, or antigen binding. This is a vague concept, as obviously the nature of the alteration or the binding is unspecified. One prediction initially was that recognition of altered self antigen involved a single receptor which recognized both aspects (see Section 11,D). Another concept has been termed the cell interaction molecule hypothesis. This is a dual-interaction or -recognition hypothesis which proposes that T cells have a receptor which recognizes the antigen, and another molecule which recognizes the alloantigen. It is possible in this scheme for the alloantigen to be recognized by itself on another cell-a form of "like-like" recognition which would adequately explain the requirement for genetic restrictions and MHC homology. However, this form of like-like recognition has been virtually excluded by the work with F, and parental cells-for various T-cell functions, such as cytotoxicity (Zinkernagel and Doherty, 1977), proliferation (Paul et al., 1977), and helper function (Erb et al., 1978). In these experiments F, T cells are restimulated only by parental macrophages which were used for the initial priming; the other parental macrophage, while sharing alloantigens, since F, cells express both MHC alloantigens, does not stimulate. This hypothesis can be partly rescued if it is proposed (ad hoc) that only one of the two MHC region alloantigens is expressed in the F, T cell-allelic exclusion. A pictorial representation of these concepts is shown in Fig. 6. Yet another possibility is a hybrid between these theories. It proposes that a complex is formed between an alloantigen and an antigen-the compound antigenic determinant concept. This complex is recognized by two receptors-ne for the alloantigen and another for the antigen. This theory does not have the disadvantages of the former and fits in much better with the available evidence on receptors, based on shared idiotype (combining site antigens) on receptors. Basically the theoretical advantages of a two-receptor model is that it fits in with shared idiotypes on T and B cells (Janeway et al., 1976), and also it permits a much smaller receptor repertoire and

174

MARC E L D M A ” ET AL. MACROPHACt

SINGLE RECEPTOR

DUAL RECEPTOR

CELL INTERACTION MOLECULE SHARING

TCELL IFORCYTOXICTCELLS. IAREPLACtDBY H 2 1

FIG.6. T-cell recognition.

gene pool to fulfill receptor functions. These considerations are discussed in more detail elsewhere (Feldmann, 1980). One of the problems of the dual-receptor hypothesis is the nature of the variable portion of both T-cell receptors. Are both of these immunoglobulin V genes? The nature of the constant part of T-cell receptors is even more difficult to understand, as despite predictions based on the function of Ir genes, Ia antigens were not detected in the idiotypic receptors isolated by Binz and Wigzell (1977) and Krawinkel et al. (1977). What have the macrophage studies contributed to our general understanding of T-cell receptor function? First, they have permitted the generalization that T-cell recognition of antigen involves MHC alloantigens, described also on the basis of genetic restrictions of T-cell-B-cell cooperation and of T-cell cytotoxicity. Furthermore, they have made clear that the type of alloantigen involved ( I region or H-2 region) depends on function and not on species or the exact antigen in question. They have also contributed to the virtual exclusion of one model-the cell interaction molecule hypothesis-unless fanciful ad hoc postulates are introduced. But what of more detailed aspects? For many purposes, GRF is very useful for analysis of T-cell receptors. Because in GRF two components are linked (the Ia antigen and the immunogen fragment) which are both recognized, certain constraints are placed on these recognition structures, which must be a single receptor, two linked receptors, or two receptors which must be very close together in order to function-as they must recognize a complex of molecular weight 55,000. Furthermore, competition studies may be performed to explore the nature of these receptors further. It has been found that antigen alone can inhibit the action of GRF (Feldmann, et a!., 1977b). This implies that there is a receptor site with a high affinity for the protein, arguing against the notion that a new antigenic determinant is created. Clearly more work is needed to clarify the nature of these T-cell receptors. Experiments in progress with idiotype markers may clarify this issue. In the discussion thus far, we have not discriminated between data obtained with the proliferation assay and with helper cell induction. The reason for this is

MACROPHAGE-LYMPHOCYTE INTERACTIONS

175

obvious-there are close analogies between the two systems. However, there may be some differences. In the mouse, the cell which proliferates in response to macrophage-bound antigens is not the cell which responds to GRF. The latter was phenotyped as an Ly l+2+3+ cell, whereas the former has been difficult to characterize but may be an Ly 1 + cell (R. Schwartz, personal communication). The macrophage-T-cell interaction has similar aspects in regard to the delayedtype hypersensitivity response described by Miller and his colleagues (1976), which is genetically restricted except where cells are used from chimeric mice which are tolerant to another haplotype. Under these circumstances, genetic restrictions are lost, but they are not if the analogous chimeric experiment is done for the induction of helper cells. The reasons for these differences are not known, but it is worth considering that there may be a whole continuum of similar interactions, each within the same basic framework but with idiosyncracies, which may explain some of the conflicting data concerning genetic restrictions and other issues. One of the most significant outcomes of the renewed interest in the macrophage-T-cell interaction has been the insight gained into possible mechanisms of Ir gene action. The first experiments of Rosenthal and Shevach have been amply confirmed with other antigens and extended to mouse Ir genes. Despite the fact that it is apparent that Ir genes are important in the macrophage-T-cell interaction, there is still much ignorance as to the nature of Ir gene products and the mechanism of Ir gene effects. Current data (vide supra) are compatible with a macrophage Ia antigen being the Ir gene product, which since it must express some degree of antigen specificity binds or otherwise interacts with antigen. Thus GRF is a possible Ir gene product (Erb et al., 1979) and GRF is not made by nonresponder animals of the type whose macrophages do not stimulate T-cell proliferation or produce helper cells. There is a correlation between GRF and Ir gene action; the first step is to determine whether GRF was an Ir gene product. The next point would be more critical and difficult to establish. If Ir genes act in an antigen-specific manner and GRF is an Ir gene product, then GRF must react with antigen in an antigen-specific manner; that is, it must have some type of antigen-combining site. Attempts to demonstrate antigen specificity of binding of the Ia component of GRF have not revealed any. However, these involved only a single pair of antigens, KLH and (T,G)-A-L, and were not quantitative. Antigens with a slight variation in structure detected by Ir genes would be needed to perform these experiments adequately, such as the variant insulins described earlier. If GRF, that is, macrophage Ia antigen, which binds immunogen is not a macrophage Ir gene product, what are the possible candidates? Perhaps they are enzymes which metabolize polypeptide antigens in a selective manner. Several other issues are raised by the studies discussed here. One of these is the nature of the macrophage-like antigen-presenting cell which interacts with T

176

MARC FELDMA” ET AL.

cells. It is known that the heterogeneity of lymphocytes is extensive. There is a whole series of lymphocyte alloantigens specific for lymphocytes, which define various functional subsets of T cells and B cells. Thus helper T cells with the Ly 1+ antigen are different from suppressor T cells with Ly 2+3+ antigens. Evidence for the heterogeneity of macrophages is still premature or preliminary. Even though there are multiple macrophage functions, it is not known whether they are all mediated by cells of one lineage or whether the one label “macrophage” encompasses, as with lymphocytes, a whole series of related cells. If in fact some macrophage functions like those mediating Ir gene function are of broad antigen specificity, are there clones of macrophages capable of presenting some antigens but not others? Another possibility is that the receptors used by macrophage antigen-presenting cells to mediate Ir gene function are the descendants of receptors used by coelomocytes in worms and other invertebrate species, which antedate the development of the classic immune system in vertebrates. If this were the case, the macrophage-T-cell interaction would represent an evolutionary crossroads where primitive receptors and more sophisticated immunoglobulin receptors interact.

REFERENCES Barcinski, M. A,, and Rosenthal, A. S. (1977). J. Exp. M e d . 145, 726. Binz, H., and Wigzell, H. (1977). Contemp. Top. Imrnunobiol. 7, 113. Blundell, T. (1972). Adv. Protein Chem. 26, 279. and Rosenthal, A. S. (1978). Submitted for publicaBraendstrup, O., Shevach, E. M., Werdlin, 0.. tion. Burnet, F. M. (1959). “The Cloral Selection Theory of Acquired Immunity.” Cambridge Univ. Press, London and New York. Calderon, J., Kiely, J. M., Lefko, J., and Unanue, E. R. (1975). J . Exp. Med. 142, 151. Clive, M. J., and Swett, U. C. (1968). J . Exp. M e d . 128, 1309. Ellner, J. J., and Rosenthal, A. S. (1975). J. Immunol. 114, 1563. Ellner, J. J., and Rosenthal, A. S. (1977). J. Immunol. 118, 2053. Erb, P., and Feldmann, M. (1975a). J. Exp. Med. 142, 460. Erb, P.,and Feldmann, M. (1975b). Nature (London) 254, 352. Erb, P.,and Feldmann, M. (197%). Eur. J . Immunol. 5, 759. Erb, P.,and Feldmann, M. (1979). “Proceedings of the Second Lymphokine Conference” (A. de Weck and M. Landy, eds.), Academic Press, New York (in press). Erb, P., Feldmann, M., and Hogg, N. M. (1976a). Eur. J . Immunol.6, 365. Erb, P., Meier, B., and Feldmann, M. (1976b). Nature (London) 263, 601. Erb, P., Vogt, P., Meier, B., and Feldmann, M. (1977). J. Immunol. 119, 206. Erb, P., Meier, B., Kraus, D., van Bohmer, H., and Feldmann, M. (1978). Eur. J. Immunol. 8, 786. Erb, P., Meier, B., Matsinaga, T., and Feldmann, M. (1979). J. Exp. Med. 149, 686. Erb, P., Vogt, P., Masinaga, T., Rosenthal, A. S., Rees, A., and Feldmann, M. (1980). In “Regulating Role of Macrophages in Immunity” (A. S. Rosenthal and E. R. Unanue, eds.), Academic Press, New York (in press).

MACROPHAGE-LYMPHOCYTE INTERACTIONS

177

Feldmann, M. (1972). J. Exp. Med. 136, 737. Feldmann, M. (1980). “Cell Interactions in Antibody Production. ” Academic Press, New York (in press). Feldmann, M., and Basten. A. (1971). J. Exp. Med. 134, 103. Feldmann, M., and Kontiainen, S. (1976). Eur. J. Immuml. 6, 302. Feldmann, M., Beverley, P., Woody, J., and McKenzie, I. F. C. (1977a). J . Exp. Med. 145,793. Feldmann, M., Beverley, P. C. L., Erb, P., Howie, S., Kontiainen, S., Maoz, A., Mathies, M., McKenzie, I. F. C., and Woody, J. (1977b). Cold Spring Harbor Symp. Quant. Biol. 41, 113. Gery, I., and Waksman, B. (1972). J. Exp. Med. 136, 143. Gorczynski, R. M. (1977). Scand. J. Immunol. 5. Gorczynski, R. M. (1978). Submitted for publication. Greineder, D. K., and Rosenthal, A. S. (1975a). J. Immunol. 114, 1541. Greineder, D. K.,and Rosenthal, A. S. (1975b). J . Immunol. 115, 932. Greineder, D. K., Shevach, E. M., and Rosenthal, A. S. (1977). J. Immunol.117, 1261. Habu, S., and Raff, M. C. (1977). Eur. J . Immunol. 7, 451. Howie, S., and Feldmann, M. (1977). Eur. J. Immunol. 7, 417. Howie, S., and Feldmann, M. (1978). Nature (London) (in press). Janeway, C., Wigzell, H., and Binz, H. (1976). Scand. J. Immunol. 5, 993. Kapp, J., Pierce, C. W. C., and Benacerraf, B. (1973). J. Exp. Med. 138, 1121. Katz, D. H., and Unanue, E. (1973). J. Exp. Med. 137, 967. Keck, K. (1975). Narure (London) 254, 78. Kontiainen, S., and Feldmann, M. (1973). Nature (London), New Biol. 246, 285. Krawinkel, U., Cramer, M., and Imanishi-Kari, T.(1977). Eur. J . Immunol. 7, 566. Lipsky, P. E.,and Rosenthal, A. S. (1973). J. Exp. Med. 138, 900. Lipsky, P. E.,and Rosenthal, A. S. (1975). J. Exp. Med. 141, 138. Lipsky, P. E., and Rosenthal, A. S. (1976). J . Immunol. 115, 868. Lohman Matthews, M., Ax, W., and Fischer, H. (1971). In “Cell Interactions and Receptor Antibodies in Immune Responses” (0. Makela, A. Cross, and T. U. Kosunen, eds.), p. 15. Academic Press, New York. McDevitt, H. 0. (1978). “Ir Genes and Ia Antigens.” Academic Press, New York. McDougall, J. S., and Gordon, D. S. (1977). J. Exp. Med. 145, 676. Miller, J. F. A. P., Vadas, M. A,, Whitelaw, A., and Gamble, J. (1976). Proc. Natl. Acad. Sci. U.S.A. 73, 2486. Mosier, D. E. (1969). J . Exp. Med. 129, 351. Munro, A. J., and Taussig, M. J. (1975). Nature (London) 256, 103. Nelson, D. (ed.) (1976). “The Immunobiology of the Macrophage.” Academic Press, New York. Norogrodsky, A,, and Katchalski, E. (1972). Proc. Nail. Acad. Sci. U.S.A. 69, 3207. Norogrodsky, A., and Katchalski, E. (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 1824. Paul, W. E.,Shevach, E.M., Thomas, D. W., Pickeral, S. F., and Rosenthal, A. S. (1977). J. Exp. Med. 145, 618. Pierce, C. W. (1978). In “Ir Genes and Ia Antigens” (H. 0. McDevitt, ed.), p. 357. Academic Press, New York. Pierce, C. W., Kapp, J . A,, and Berracerrof, B. (1976). J. Exp. Med. 144, 371. Rode, J., and Gordon, D. S. (1974). Cell. Immunol. 13, 87. Rosenthal, A. S., and Shevach, E. M. (1973). J . Exp. Med. 138, 1194. Rosenthal, A. S., Blake, J. T., and Lipsky, P. E. (1975). J . Immunol. 115, 1135. Rosenthal, A. S., Rosenwasser, L. J., Baskin, B. Z., Schroer, J., Thomas, J. W., and Blake, J . T. (1978). In “Immunobiology of Proteins and Peptides” (A. Z. Atassi, ed.) (in press). Schreffler, D. et al. (1977). Cold Spring Harbor Symp. Quant. Biol. 41,477. Schumacher. V., Pena-Martinez, J., and Festenstein, H. (1975). Narure (London) 255, 155.

178

MARC FELDMANN ET AL.

Schwartz, R. H. (1978). In “Proceedings of the Ir Gene Workshop” (H. 0. McDevitt, ed.). Academic Press, New York (in press). Schwartz, R. H., and Paul, W. E. (1974). J. Exp. Med. 143, 529. Schwartz, R. H., Yano, A., and Paul, W. E. (1978). Immunol. Rev. 40, 153. Shearer, G.,Rehn, T. G., and Garbarino, C. A. (1975). J . Exp. Med. 141, 1348. Shevach, E. M. (1976). J . Immunol. 116, 1482. Shevach, E. M . , and Rosenthal, A. S . (1973). J. Exp. M e d . 138, 1213. Shevach, E. M., Paul, W. E., and Green, I. (1972). J . Exp. Med. 136, 1207. Shevach, E. M., e t a l . (1977). J. Exp. Med. 146, 561. Shortman, K . , and Palmer, J. (1971). Cell. Immunol. 2, 399. Thomas, D., and Shevach, E. M. (1976). J. Exp. M e d . 145, 907. Thomas, D., and Shevach, E. M. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 2104. Unanue, E. R., and Cerottini, J. C. (1970). J . Exp. Med. 131, 71 I . Van Rood, J. J. er al. (1977). Cold Spring Harbor Symp. Quant. Biol. 1, 41. Wagner, H., Feldmann, M., Boyle, W., and Schrader, J. W. (1972). J. Exp. Med. 136, 331. Wahl, S. M., Wilton, J. M., Rosenstreich, D. L., and Oppenheim, J. J. (1975). J . Immunol. 114, 12%. Waldron, J. A,, Horn, R. G., and Rosenthal, A. S. (1973). J. Immunol. 111, 58. Waldron, J . A,, Horn, R. G.,and Rosenthal, A . S. (1974). J . Immunol. 112, 746. Yano, A. Schwartz, R., and Paul, W. (1977). J . Exp. M e d . 146, 828. Zinkernagel, R. M., and Doherty, P. C. (1977). Cold Spring Harbor Symp. Quant. Biol. 41, 505.

NOTEADDEDIN PROOF.Several aspects of macrophage-T cell interaction have recently become clarified by experiments using chimeric mice. Thus two stages of T helper cell differentiation were defined, one depending on the MHC type of the host tissues, presumably the thymus, and the other depending on the macrophage pool, and thus the stem cell donor MHC type (Erb et a l . , 1979). These studies have also demonstrated that the primary genetic restrictions of the macrophage-T interaction are not strictly for self (i.e., antigens on T cells), but for the MHC antigens present at both of the two stages of T cell maturation. Thus Fl stem cells injected into one parental strain (PI) irradiated recipients differentiated into T cells capable of recognizing PI presenting cells but not Rr, despite the fact that (PI x R2) Fl T cells bear both sets of MHC alloantigens, and, if matured in an F, environment, can recognize both. Furthermore, PI stem cells differentiating in an Fl (PI X Rr)-irradiated host generates T cells capable, in certain circumstances, of recognizing both P, and Rr macrophages, indicating that MHC antigens not represented in the T cell genotype or surface can be recognized effectively. Evidence for the second, macrophage-dependent step came from PI + Fl chimeras, T cells of which only responded to PI macrophages unless the animals were also injected with P2 macrophages (Erb et al., 1979). A further analysis of Ir gene involvement in this interaction was performed using such chimeras made with combinations of responder and nonresponder cells/irradiated host environment. Using three different antigens under Ir gene control, [(T,G)-A-L, beef and pork insulin] and various types of chimeras-allophenic or irradiation Fl (R X NR) NR, P(NR) + Fl (R X NR), PI (R) R2 (NR) + Fl (N X NR); in each instance NR genotype stem cells differentiated, if the host was of responder phenotype, into responder T cells. In contrast no such differentiation of NR genotype stem cells into responder antigen-presenting cells ever occurred, suggesting that Ir gene expression is more intimately linked to macrophage presentation than to T cell function.

+

INTERNATIONAL REVIEW OF CYTOLDGY, VOL. 60

Immunohist ochemistry of Luteinizing Hormone-Releasing Hormone-Producing Neurons of the Vertebrates JULIENBARRY U.156 INSERM and Laboratory of Histology and Embryology, Faculty of Medicine, Lille, France I. Introduction . . . . . . . . . . . . . . 11. Techniques of Study . . . . . . . . . . A. Preparative Histological Techniques . . . B. LH-RH Immunocytochemistry . . . . . 111. Morphology of LH-RH-Reactive Perikarya . . IV. Topography of LH-RH-Reactive Perikarya . . A. Mammals . . . . . . . . . . . . B . Birds . . . . . . . . . . . . . . . C. Amphibians and Fishes . . . . . . . V. Hypothalamohypophyseal LH-RH Tracts . . A. Mammals . . . . . . . . . . . . B. Birds . . . . . . . . . . . . . . . C. Amphibians anf Fishes . . . . . . . . VI. Preopticoterminal LH-RH Tract . . . . . . VII. Extrahypothalamic LH-RH Tracts . . . . . VIII. General Discussion and Conclusions . . . . References . . . . . . . . . . . . . .

. . . . . .

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

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

179 180 180 182 188 1%

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

194 199 200

. . . . . . .

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

201

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

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . .

202 206 206 207 209 213 214

I. Introduction Numerous studies have established the existence of a hypothalamic control of prehypophyseal gonadotrophic secretion of a neurovascular link between the hypothalamus and the prehypophysis (for references, see Harris, 1955; Benoit and Assenmacher, 1955; Szentagothai et al., 1962; Green, 1969), and of the stimulating action of hypothalamic extracts upon prehypophysial secretion of follicle-stimulating hormone (FSH) and of luteinizing hormone (LH) (‘ ‘LHreleasing factor,” LH-RF, or LRF, McCann et al., 1960; Hams, 1960; Campbell et al., 1961; Courrier et al., 1961; “FSH-releasing factor,” FSH-RF, or FRF, McCann et al., 1964; Gellert et al., 1964; Mittler and Meites, 1964). On the other hand, various experimental studies (notably localized hypothalamic lesions, hypothalamic disconnection, and intrahypothalamic pituitary grafts: Szentagothai et al., 1962; Flerko, 1970; Mess et al., 1970) have shown that a relatively well-developed hypothalamic region, including the anterior and the mediobasal hypothalamus (the so-called “hypophysiotrophic area, Halasz et ”

179

Copyright 0 1979 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-3643600

180

JULIEN BARRY

al., 1962), allows the maintenance of “PAS-positive cells” (notably gonadotrophs) in pituitary grafts and, for this reason, probably contains all, or a part of, the “gonadotrophic structures” of the hypothalamus. The nature of the “luteinizing hormone-releasing hormone” (LH-RH or LHRF)-producing cells has been long discussed: small neurons of the hypophyseotrophic area (“parvocellular neurosecretory system,” Szentagothai et al., 1962; Szentagothai, 1964); neurosecretory cells of the laterodorsal hypothalamic interstitial nucleus (Barry, 1960, 1967); primitive neurons of the periventricular brain (Stumpf, 1970); or specialized ependymal cells of the medial prechiasmatic gland (Leveque, 1972). The isolation of LH-RH and the determination of its chemical formula (Amoss et al., 1971; Burgus et al., 1971; Matsuo et al., 1971a; Schally et al., 1971; Sievertsson et al., 1971), followed by its synthesis (Monahan et al., 1971; Matsuo et al., 1971b), allowed the preparation of specific anti-LH-RH antisera. Beginning at the end of 1972, use of these antisera made possible the direct immunocytochemical study of the hypothalamic structures producing LH-RH. The characterization of hypothalamic LH-RH axons was easily carried out by all groups engaged in the study, notably at the level of the infundibulum (see Section V,A). On the other hand, the demonstration of the perikarya has been much more difficult and was obtained for the first time only in the guinea pig (Bany et al., 1973a,b). As the majority of these cell bodies are located before (and not in) the hypophyseotrophic area, these studies were criticized by groups which had obtained negative results with their own animals (generally the rat: Kordon et al., 1974; Setalo et al., 1975) until Sttalo et al. (1976a) and Silverman (1976) verified them in the guinea pig. At the present time, the demonstration of LH-RH-reactive perikarya is not a subject of debate and has been extended to various vertebrate species (see Section 111), particularly mammals. Finally, the divergencies observed between the results of various laboratories have led to a deeper analysis of the antiserum specificities, as a function of the various molecular conformations of somatic LH-RH (see Section IV).

11. Techniques of Study

A. PREPARATIVE HISTOLOGICAL TECHNIQUES Theoretically, the goal of histological preparative techniques is to allow the best immunocytochemical detection possible of tissular LH-RH, on slides or ultrathin sections. They should thus prevent LH-RH from being lost (or reduced)

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

181

or from being changed into the nonreactive state; allow the unmasking of its nonreactive forms; or even link LH-RH to nondiffusible cellular components. Nevertheless, the greater part of tissular LH-RH is probably lost, or destroyed, during the preparative histological techniques. Since the immunocytochemical reactions are always carried out in aqueous media, they can only reveal “nondiffusible” reactive LH-RH which represents only 5 to 10% of the LH-RH present in vivo (Goldsmith and Ganong, 1975). The sensitivity of the immunocytochemical techniques therefore becomes a decisive factor. 1. Fixation Fixation can be carried out by immersion (I) or by perfusion (P), generally through the left ventricle, with or without postfixation. The most common fixatives used are Bouin [King et al., 1974 (P); Kordon et al., 1974 (P); Zimmerman et al., 1974 (P); Naik, 1975a,b (P); Baker and Dermody, 1976 (I); Goos et al., 1976 (I); King and Gerall, 1976 (I); McNeill et al., 1976 (I); Silverman, 1976 (I); Zimmerman and Antunes, 1976 (P)], Bouin-Hollande, with or without acetic acid or sublimate [Barry et al., 1973a (I); Leonardelli et al., 1973a (I); Doerr-Schott and Dubois, 1975a, 1976 (I); Dubois, 1973 (I); Dubois et al., 1974 (I); de Reviers and Dubois, 1974 (I); Kordon et al., 1974 (1,P); Bugnon et a/., 1976a,b, 1977a (I)], formaldehyde [Alpert et al., 1975; Barry, 1976a, 1977a (I); Zimmerman and Antunes, 19761, PAF [picric acid formaldehyde according to Zamboni and de Martino, 1967; Mazzuca and Dubois, 1974 (1,P); Setalo et al., 1975 (I); Bugnon e tal. , 1976a, 1977b (I)], paraformaldehyde [Calas et al., 1973 (P); Kordon et al., 1974 (P); Pelletier et al., 1974 (P)], or glutaraldehyde [Pelletier et al., 1974 (P); Goldsmith and Ganong, 1975 (P); Naik, 1975b (I); Silverman and Desnoyers, 1976 (P); Bugnon et al., 1977b (I); Mazzuca, 1977 (PI]. It must be noted that the protein-LH-RH complexes present in hypothalamic extracts are dissociated by lowering the pH and are reconstituted at neutral pH (Shin and Howitt, 1977, in the rat). For light microscopy, the best fixatives appear to be Bouin-Hollande (without acetic acid and sublimate, but with formol added 1:10, Barry and Carette, 1975b) or Bouin pH 3.8, preferentially by perfusion (Hoffman et al., 1978a, after comparative tests with Bouin-Duboscq, Carnoy , Bouin-Hollande, with or without sublimate, and Stieve with or without acetic acid). For electron microscopy, fixation is generally carried out with glutaraldehyde (see above), or a mixture of glutaraldehyde and formaldehyde, followed by postfixation with buffered solutions of OsO,.

+

2 . Freeze-Substitution The freeze-substitution technique has been utilized by Silverman and Desnoyers (1976) and Silverman and Zimmerman (1978), after infiltration of tissues with 2%

182

JULIEN BARRY

glycerol + 0.5% dimethylsulfoxide to prevent the formation of ice crystals during cooling. 3. Embedding For light microscopy, samples are usually embedded in paraffin or Paraplast (the majority of authors cited above) and occasionally in Polywax 1000 (Setalo et a l . , 1976a). For ultrastructural immunocytochemical studies, embedding is usually carried out in Araldite (Pelletier et al., 1974; Naik, 1975b; Silverman et al., 1977; Bugnon et al., 1976c, 1977b; Mazzuca, 1977; Silverman and Zimmerman, 1978) and occasionally in Spurr medium (Goldsmith and Ganong, 1975). 4. Microtomy Sections are usually performed with the classical techniques (microtome or ultramicrotome) and occasionally with the cryostat (Calas et al., 1973; Mazzuca and Dubois, 1974) or with the Sorvall TC 2 (Calas et al., 1974). Access of the antibodies to LH-RH molecules present in the tissue is facilitated by the thinness of the slices and by a 5-minute incubation in 0.4% Triton X-100 (Hoffman et al., 1978b) and, after Araldite embedding, by a 5-minute exposure to 5% hydrogen peroxide. B. LH-RH IMMUNOCYTOCHEMISTRY

1. Primary Antisera LH-RH decapeptide [Pyro-Glu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-Gly (CONH,)] of low molecular weight (1182.5) is very poorly antigenic. It is thus generally coupled with carrier proteins, with which it can form effective immunogenic complexes (see Table I). Of the great number of possible coupling agents (Likhite, 1967; Kennedy et al., 1976; Schuurs and Van Weenen, 1977), only a few have been utilized for the coupling of LH-RH to proteins, notably the following. a. Glutaraldehyde, 0 =C(H)-CH2-CH2-CH2-(H)C=0, which preferentially reacts with free amine groups (80 to 90%), with the imidazole ring of histidine (30 to 40%), and with the phenol ring of tyrosine (20%) when it is linked to glycine (Habeeb and Hiramoto, 1968). b. Toluene diisocyanate, where the S=S=N end groups react with OH- or NH2 groups (Schick and Singer, 1961), for example, the primary alcohol of serine and the phenol group of tyrosine. c. Carbodiimide derivatives (R-N = C = N - R) which react with the amino, or with the carboxyl in Gly 1, Gly 10, or Tyr 5 , or with the azo-tyrosyl or azo-histidyl derivatives obtained by the attachment of p-diazonium phenylacetic acid (Koch et al., 1973).

TABLE I MAINIMMUNOGENS USEDFOR PRIMARY ANTISERUM PREPARATIONS

Immunogens LH-WHSA LH-Wguinea pig Ig LH-WBSA LH-RWHSA LH-RWBSA LH-RWHSA LH-WBSA LH-WBSA LH-WBSA LH-RWHRP LH-WTGB LH-Wlimpet hemocyanin LH-RH LH-RH

Presumed site of conjugation

Coupling agent

Authors

Reference

Tyr 5 ? Tyr (His) Tyr 5 ?

Glutaraldehyde Glutaraldehyde Glutaraldehyde Toluene diisocyanate Carbodiimide derivative Carbodiimide derivative Carbodiimide derivative bis-Diazotized benzidine bis-Diazotized benzidine bis-Diazotized benzidine

No No

Diazotized PABA No (adsorbed on PVP) No (adsorbed on PVP)

Dubois Kerdelhue Sorrentino (E,F) Dubois Arimura (710) Arimura (743, 744) Goos et al. Nisswender (38) Nett, Niswender (42) Stemberger Alpert Baker and Dermody (154) Arimura (422) Barry (21, 22)

Barry er al. (1973a); Dubois (1973) Calas et al. (1973); Kordon er al. (1974) Hoffman et al. (1978a) Barry et al. (1973a); Dubois (1973) Setalo et al. (1975) King et al. (1974); Setalo er al. (1976a) Goos et al. (1976) Zimmerman et al. (1974) McNeill er al. (1976) Alpert er al. (1975); Knigge er al. (1978) Alpert er al. (1976) Baker and Dermody (1976) Barry and Carette (1975a) Barry (1976b)

His 2 ? His 2 ? His 2 Gly 10 Glu 1

184

JULIEN BARRY

d. Bis-diazotized benzidine, with N=N(Cl) end groups, or diazo-paraaminobenzoic acid (PABA), whose carboxyl groups allow it to be coupled to amino groups at basic pH (Gross et al., 1968). Appropriate precautions should be taken when handling the benzidine because it is a carcinogen. Primary antisera have, however, been obtained without covalent coupling, for example, after absorption on polyvinylpyrrolidone(PVP), according to Worobec et al. (1972) (see Table I). Or course, preparations of synthetic LH-RH should contain only the decapaptide, thus being free of any impurities or LH-RH cleavage products. But even under these conditions, the obtained antigens may form a heterogeneous molecular population after coupling at different sites on the LH-RH molecule. The antigens thus obtained are emulsified with Freund’s adjuvant (complete, incomplete, or modified) and injected into rabbits (or guinea pigs) by various techniques. The best techniques are probably multiple immunization site techniques (Worobec et al., 1972) and intrasplenic injection (Dubois and Renoux, 1971) along with intradermic injections with iv boosters. as necessary in randomly bred, unrelated animals. Primary antisera obtained starting with conjugated LH-RH must be absorbed out for antibodies directed against the carrier protein (precipitation with centrifugation, or saturation with this protein, followed by affinity chromatography). Of course, the antisera may also have numerous “nonspecific” antibodies reacting against various components of Freund’s adjuvant (or other adjuvants), against impurities in the carrier-protein preparations, against antigenic determinents adjacent to the area of coupling, or even against PVP, when the antigen is nonconjugated LH-RH. In any given primary antiserum, the “specific” antibodies (that is, directed against LH-RH) probably form a heterogeneouspopulation where the “antibody” sites may react against various antigenic portions (specificities) of the LH-RH molecule. Theoretically they might well be purified by isoelectric focusing or the immunoadsorbent techniques. Finally, specific antibodies directed against these same antigens may vary from one animal to another and within the same animal, from bleeding to bleeding. 2. Immunocytochemical Techniques The immunocytochemical detection of tissular LH-RH molecules is obtained by the fixation of specific antibodies from the primary antiserum on these molecules (Figs. l a and b). These primary antibodies are then visualized by specific anti-Ig antibodies, covalently linked to a marker molecule [fluorescein in the “fluorescein-labeled antibody technique” of Weller and Coons (1954); peroxidase in the ‘‘peroxidase-labeled antibody technique of Avrameas and Lespinats (1967) or Nakane and Pierce (1967)l; or by binding of the detector ”

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

185

n o r e o c t i o n w i t h PAP

Specific Reaction

A

N

A

Ic\ E n z y m e - L a b e l e d A n t i b o d y Specific Reaction

Nonspecific Reaction

A

N

A

Id) Enzyme-Lobeled Antibody Nonspecific Reaction

FIG.1. Unlabeled-antibody enzyme method: (a) Tissue antigen (A) is localized by specific antibody in primary antiserum (P). This localization is followed by reaction with antiimmunoglobulin in a secondary antiserum ( S ) , peroxidase-antiperoxidase complex (PAP), and cytochemical reaction for peroxidase. (b) Nonspecific antigens in tissue (N) react with antibodies that contaminate the secondary antiserum. This nonspecific reaction remains undetected in the unlabeled-antibody enzyme method, as contaminant antibodies cannot react with PAP. Indirect enzyme-labeled antibody method: (c) Specific antibody in the primary antiserum (P) is followed by a secondary antiimmunoglobulin chemically conjugated with peroxidase (SPOC). (d) Antibodies that contaminate the secondary antiimmunoglobulin react with nonspecific antigens (N) in the tissue. Since their antibodies are labeled with peroxidase, a nonspecific background reaction ensues. (From Sternberger, 1977.)

molecule by immunological linkage ( “unlabeled-antibody enzyme technique, Sternberger et al., 1970). a. Fluorescein-Labeled Antibody Technique. This technique requires the use of a fluorescence microscope and gives excellent results, notably for examination by reflected light (for details of the manipulations, see Barry et al., 1973a; Leonardelli et al., 1973a; Dubois, 1973, 1976a,b; de Reviers and Dubois, 1974; Naik, 1975a; Doerr-Schott and Dubois, 1976; Goos et al., 1976; Bugnon et a l . , 1977a,b,c; Bons et al., 1978a,b). It is however, possible that marked nonspecific antibodies present in the secondary antiserum may react nonspecifically (Fig. Id), diminishing the signal-to-noise ratio; this problem does not occur with the technique of Sternberger et al. (1970) (Fig. lb). It is also possible that the covalent labeling of the secondary antiserum antibodies modifies their specificity, but in practice, this risk is minimal. ”

186

JULIEN BARRY

b. Peroxiduse-Labeled Antibody Technique. This technique allows the study of slides with the light microscope (after detection of peroxidase activity of the secondary antibodies by derivatives of diaminobenzidine or chloronaphtol: for details of manipulations, see Kordon et a l . , 1974; Mazzuca and Dubois, 1974; Naik, 1975b) and with the electron microscope (after treatment with OsOl to obtain an electron-dense deposit). This technique has the advantage of giving very durable preparations, which may be photographed at low enlargements (which is not the case with fluorescent microscopy); it is also more sensitive, since the quantity of the visible product may be increased by prolonging the reaction time. Nevertheless, nonspecific reactions are possible, as is the case with the fluorescein-labeled antibodies (Fig. Id). In addition, there is the danger of modification in specificity due to covalent labeling. Finally, the diaminobenzidine derivatives have carcinogenic properties, and appropriate precautions should be taken when handling them. c. Unlabeled-Antibody Enzyme Technique. This technique utilizes immunologic linkage to antibodies not submitted to covalent binding (Fig. la). The primary antibodies bound to tissular LH-RH (first step) are detected by specific secondary antibodies (generally anti-rabbit Ig) (second step). These antibodies then react with an excess of peroxidase-antiperoxidase antibody complexes (PAP, according to Sternberger et al., 1970; or Kawaoi and Nakane, 1973). This third step is followed by the revealing of peroxidase activity (for details of manipulations, see King et a l . , 1974; Pelletier et a l . , 1974, 1976; Zimmerman et al., 1974; Goldsmith and Ganong, 1975; Baker and Dermody, 1976; Hoffman et al., 1976, 1978a; Picard and Silverman, 1976; Silverman, 1976; Silverman and Desnoyers, 1976; Bugnon et a l . , 1976a,b, 1977a-e, 1978; McNeill et a l . , 1976; Mazzuca, 1977; Silverman and Zimmerman, 1978). This technique is equally valuable for light microscopy or electron microscopy (generally with postembedding coloration); it is 100 to 1000 times as sensitive as the immunofluorescence technique (Sternberger, 197X; Pelletier et a l . , 1976), with a very high signal-to-noise ratio. Finally, it does not give a reaction when secondary antiserum nonspecific antibodies react with tissular nonspecific antigens (Fig. lb). 3, Specificity of Immunocytochemical Reactions This problem has been the subject of numerous discussions (for references, see Sternberger, 1973, 1974, 1977; Petrusz et a l . , 1976, 1977; Swaab et a l . , 1977; Knigge et al., 1978; Sternberger and Hoffman, 1978). It depends principally upon the properties of the immunologic linkage. a. Properties of the Immunologic Linkage. The antigen-antibody complexes result from weak interactions (electrostatic interactions, hydrogen bonds, van der Waals forces) between complementary surfaces, which allow a strong linkage with a high rate constant (approximately lo-* M-' second-'), probably

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

187

excluding large-scale conformational rearrangements. The region with the antigen-binding capacity specifically recognizes certain sequential or conformational groups of the antigen, or of a given region (not greater than lo3 of that antigen. For LH-RH, specificity increases with the length of the recognized sequence (or better still, the whole decapeptide), but it can be prevented by conformational rearrangements of LH-RH (see Section IV). Finally, if the specificity of the immunologic link is high, it is not absolute and does not exclude crossed reactions. For these various reasons, it is important to carry out a detailed serological study of the specific antisera, notably by immunoprecipitation, immunoelectrophoresis, complement fixation, passive imrnunohemalysis, and radioimmunoassays (RIA) (for technical details, see Weir, 1978). It should be noted, however, that the specific antibody titers (evaluated by RIA) are less important from the immunocytochemical point of view than are their affinities for tissular LH-RH, which may have a conformation different from that of LH-RH in vitro. b. Criteria of Specificity. We cannot here enter into a detailed discussion of the various possible eventualities (true- or false-positive reactions; true- or false-negative reactions) and of their interpretations. “True”-positive reactions result from a specific linkage by the anti-LH-RH antibody to tissular components having antigenic sites identical to those of the hormonal decapeptide (or some of its conformations, possibly incorporated into precursor molecules, or linked to cellular substrates). “True” reactions should be considered as only those positive reactions obtained with primary antisera which are sufficiently diluted (1:lOO to 1:lOOO in light microscopy; 1:400 to 1 :10,000 in electron microscopy). These reactions should disappear in the absence of incubation with these antisera; after saturation of primary antisera with synthetic LH-RH (principal criterion for specificity); following substitution of primary antisera with nonimmune sera (or preimmune sera) of the species used for their preparation; after treatment with unmarked secondary antibodies before incubation with the marked anti-Ig antiserum (modification of the Goldman, 1975, inhibition method); or (for the PAP technique) following a second treatment with the primary antiserum, after incubation with the secondary antiserum. On the other hand, addition to primary antisera of neuropeptides other than LH-RH should not suppress the true-positive reactions (for details of manipulations, see publications cited above). When all the criteria for specificity have been fulfilled, there is always the possibility that an unknown molecule, other than LH-RH (or its precursor), may give a true-positive reaction with the antibody used. Under these conditions (and in the absence of decisive biochemical proofs which would require complex and time-consuming techniques), I believe that structures producing LH-RH should not only contain sufficient quantities of a

Az)

188

JULIEN BARRY

substance(s) which reacts specifically with the anti-LH-RH antibodies, but should also behave as we would expect LH-RH to act in various physiological or experimental circumstances (Barry et al., 1973a).

111. Morphology of LH-RH-Reactive Perikarya The identification of LH-RH neuronal perikarya has been difficult and controversial, particularly in mammals (see Section IV). Nevertheless, in all species in which these perikarya have been demonstrated (see Table 11), they present the same morphologic characteristics, although the quantity of reactive material which they contain varies considerably according to the species and to the physiological or experimental conditions (Figs. 2-5). The form of the perikarya (rounded, oval, fusiform, very rarely stellar) depends on the angle of the section, as well as on the cellular dimensions: 10 to 30 pm in the guinea pig, the cat, the dog, and the rabbit; up to 40 p m in the monkey and in man, with a nucleus as large as 10 to 14 pm (Barry et al., 1973a; Barry and Dubois, 1974b, 1975; Barry, 1976a-c, 1972a,b; Barry and Carette, 1975b). The unreactive nucleus is usually well distinguished from the cytoplasm. The cytoplasm is frequently finely granular; it may occasionally contain a few lipofuscins; in monkeys it at times has reactive, rounded inclusions. The dendrites are slightly branched, if at all; they are rarely visible in the guinea pig under physiological conditions. They are, however, easily observed in the monkey and in man, where they can be followed for 200 pm or more (Barry, 1976a, 1977a). In man, reactive perikarya can be seen in subjects where the hypothalamus is almost completely devoid of reactive axons; this is very infrequent in other species. Under certain very favorable conditions one can observe the emergence of the axon from the cell body, or from the base of a dendrite, but these images are infrequent, the reactive material only rarely accumulating in the segment proximal to an axon (Barry and Dubois, 1974a). Analogous observations have been made in other species of mammals: by Naik (1975a), Setalo et al. (1976a), King et al. (1977), Kozlowski and Hostetter (1978), and Flerko et al. 1978) in the rat, where the reactive perikarya measured 15 to 25 pm; by Zimmerman et al. (1974) in the mouse (rounded neurons, with reactive material frequently located in the periphery, but at times filling all the cytoplasm); by Zimmerman and Antunes (1976), in the rhesus monkey, where the perikarya have a diameter half that of magnocellular neurons; by Hoffman (1976) and Hoffman et al. (1976, 1978a) in the rat, the sheep, the cat, and the dog; by Leonardelli et al. (1974), Silverman and Desnoyers (1976); Weindl and Sofroniew, 1978) in the guinea pig; and finally by Knigge et al. (1977) and Toran-Allerand (1978) in primary cultures of dispersed basal hypothalamic cells. In birds, LH-RH-reactive perikarya may have a rounded form (McNeill et al.,

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

189

TABLE I1 IDENTIFICATION OF LH-RH-REACTIVE PERIKARYA I N VERTEBRATES Class Mammals

Species Guinea pig

Mouse Rat

Cat Dog Rabbit Sheep Garden dormouse Cercopithecus Cebus apella Saimiri sciureus Rhesus Man

Reference Barry et al. (1973a); Leonardelli et al. (1973b); Setalo et al. (1976 ) Silverman (1976); Hoffman (1976); Weindl and Sofroniew (1978) Zimmerman er al. (1974); Silverman (1976); Hoffman ( 1976) Naik(1974);Setaloetal.(l976a); Hoffman (1976);King and Gerall (1976); Kozlowski and Hostetter (1978); Kami et al. (1977); Krisch (1978) Barry and Dubois (1974~) Barry and Dubois (1974~);Hoffman et al. (1978a) Barry (1976); Flerko et al. (1978); Hoffman et al. (1978a) Hoffman er al. (1978a) Richoux and Dubois (1976) Barry et al. (1975) Barry and Carette (1975a) Barry and Carette (1975a); Mazzuca (1977) Barry et al. (1975); Zimmerman and Antunes (1976); Silverman et al. (1977) Bugnon et al. (1976a); Paulin et al. (1977)-fetuses Barry (1977a)-neonates, infants, adults, elderly

Birds

Mallard duck Chicken Pheasant

McNeill et al. (1976); Bons et al. (1977) Hoffman et al. (1978a) Hoffman et al. (1978a)

Amphibians

Bufo vulgaris Xenopus laevis Rana pipiens Ranu remporaria Rana esculenta

Doerr-Schott and Dubois (1975) Doerr-Schott and Dubois (1976) Alpert et al. (1976) Alpert er ul. (1977) Goos et al. (1976)

Fishes

Salmo gairdneri

Goos and Murathanoglu (1977)

1976, in the mallard duck), or an oval, bipolar shape, with a diameter of 15 to 25 p m (Bons et al., 1977, 1978a,b, in the same species; Hoffman et al., 1978b, in the chicken and the pheasant). Similar perikarya have been observed in the Amphibia, notably in the toad (Doerr-Schott and Dubois, 1975a); Xenopus (Doerr-Schott and Dubois, 1976), the frog (Goos et al., 1976), and in fishes (Goos and Murathanoglu, 1977). In this last class of vertebrates the granular material of LH-RH-reactive perikarya is very low (Fig. 3e). At the present time, there has been very little electron microscopic documentation of the perikarya of LH-RH neurons. After immunocytochemical identifica-

190

JULIEN BARRY

FIG.2. LH-RH-Reactive neurons. (a) Retrochiasmatic area of the squirrel monkey (x380); (b) lamina terminalis of the squirrel monkey ( X 380); (c) infundibular nucleus of Cebus upella ( x 380); (d) paraolfactory area of the rabbit (X380). Bar = 10 pm. an, Arcuate nucleus; ax, axons; de, dendrites, pt, pars tuberalis; sg, subglial rostra1 layer. Perikarya are indicated by white arrows. (Original documents, Barry.)

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

191

FIG.3. LH-RH-Reactive neurons. (a) Infundibular nucleus of man ( ~ 6 8 0 ) .(b) Lamina ter. Dorsolateral portion of the arcuate nucleus of the mallard minalis of the squirrel monkey ( ~ 2 5 5 )(c) duck, dark-field micrograph. Unlabeled-antibody enzyme technique and antibody 42 of Niswender and Nett (X475). With kind permission of Pr. McNeill. (d) Preoptic area of Xenopus laevis (X255). With kind permission of Mrs. Doerr-Schott. (e) Area dorsalis pars medialis of the telencephalon f X300). With kind permission of Pr. Goos. Bar = 10 pm. ax, axon (thin arrows); de, dendrites; el, ependymal lining; sg, subglial rostra1 layer. Reactive perikarya are indicated by thick arrows. [(a) and (b), original documents, Barry.]

FIG.4. LH-RH neurons of the vascular organ of lamina terminalis, in the squirrel monkey. (a) Reactive perikaryon: labeled granules (PAP technique) localized at the periphery of the cell along the plasma membrane, ( x 85OO). (b) Immunoreactive cell process: labeled granules always at the periphery ( X 12,750). (c) Nervous process in classical electron microscopy; compare with (b) 192

FIG. 5 . Ultrastructural study of LH-RH neurons, 24-week-old fetus. Glutaraldehyde-fixed hypothalamus. (a, b) The same LH-RH body identified by immunocytology (peroxidaseantiperoxidase complex) on a semithin section (a), and stained with lead citrate and uranyl acetate on the adjacent thin section. Note the corresponding agranular areas (black arrows) in (a) and (b). Bar = 1 pm. (From Bugnon et al., 1978.)

( x 13,600). (d) General aspect of a neurosecretory LH-RH cell in classical electron microscopy; . Neurosecretory cell compare with (b). Neurosecretory granules are clearly visible ( ~ 8 5 0 0 ) (e) process, contacted by a presynaptic ending(s) ( X 13,600). (f) Details of a neurosecretory LH-RH cell: neurosecretory granules are also present in the Golgi region ( X 18,700). Bar = 0.5 pm. (Original documents; with kind permission of Pr. Mazzuca.) 193

194

JULIEN BARRY

tion, in the rat, Naik (1975b) described 90- to 130-nm granules, which are frequently associated with a Golgi apparatus and occasionally dispersed in the cytoplasm. In the squirrel monkey, at the level of the lamina terminalis, Mazzuca (1977) observed rounded cell bodies 10 to 15 p m in diameter, with an oval nucleus, a large nucleolus, and a nuclear membrane with deep invaginations (Fig. 4). The reactive granules, 90 to 130 nm, are generally peripheral and occasionally associated with dictyosomes of the Golgi apparatus. The ergastoplasm forms flattened cistemae, with ribosomes in groups of six to nine. The cytoplasm contains numerous microtubules, which are also visible in the neuronal prolongations. In the human fetus, Bugnon et al. (1976a-c, 1977a-e, 1978) have observed similar features (Fig. 5 ) .

IV. Topography of LH-RH-Reactive Perikarya A. MAMMALS

The topography of LH-RH neurons was first determined in mammals, which have been the subject of most studies. It was obviously necessary to identify a sufficient number of reactive perikarya, and from the beginning, numerous difficulties as well as prolonged discussions were encountered. As I noted at the time, “The perikarya of these neurons remain almost invisible under usual experimental conditions. Three reasons may be advanced for this. It may be due to proximal concealment of the immunoreactive substance in question (through bonding with a substance, or a synthesis in the form of nonreactive prohormone), insufficient intracellular accumulation (because of axon evacuation without preliminary intracellular accumulation), or synthesis taking place distal of the perikarya (a very remote possibility)” (Barry et al., 1973a). As various attempts to “unmask” perikaryal LH-RH have not been successful, and the hypothesis of a distal (axonal) synthesis of LH-RH does not seem very probable, I have carried out various experiments in order to increase the intracellular concentrations of LH-RH in the reactive perikarya. Castration plus injections of colchicine into the lateral ventricle have been shown to be particularly effective in the guinea pig. Under these conditions (Barry et al., 1973a,b), the number of reactive cell bodies may reach several hundred. They are principally located in the suprachiasmatic, the anterior hypothalamic, and the rostral pericommissural, septal, and paraolfactory areas, generally in the paramedial region. Some perikarya can also be observed in the retrochiasmatic and arcuate regions and in the rostral mesencephalon (Fig. 6). In the normal adult guinea pig, on the other hand (as in the majority of other species), the number of reactive perikarya is very small in the majority of experimental or physiological circumstances; it is more or less augmented near the end of fetal life, in the neonatal

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

195

iset

FIG.6 . Schematic sagittal drawing of LH-RH-reactive systems in guinea pig brain. (1) Mediobasal hypothalamic reactive perikarya (black dots); (2) supraoptic reactive perikarya; (3) rostral pericommissural reactive perikarya; (4) parolfactory reactive perikarya. Reactive tracts are indicated by broken lines. Bar = 1 mm. A, Adenohypophysis; ab, axon branchig; brt, basal rhinencelphalic tract; hip, habenulointerpeduncular tract; hit, hypothalamoinfundibular tract with radiating collateral endings (arrowheads); hnt, hypothalamoneurohypophyseal tract; M, mammillary body; MS, mesencephalon; N, neural lobe; 0, optic chiasm; pst, preopticosupraoptic tract; R, rostral commissure; S,septum; sc, short axon collaterals; set, septoepithalamic tract; T, thalamus; V, third ventricle. (Original drawing, Barry.)

period (Barry and Dubois, 1974a), and, in pregnant females, toward the end of gestation (Barry and Dubois, 1973a, 1974b). Experimentally, the injection of melatonin (Barry et al., 1974, in the castrated animal), serotonin (Leonardelli et al., 1974, in the castrated animal), sulpiride (Leonardelli er al., 1973b), or metaraminol and sodium dimethylthiocarbonate (Leonardelli and Dubois, 1974, for the mediobasal region), prolactin (Leonardelli et al., 1977; Leonardelli, 1977), and various other drugs (Leonardelli and Dubois, 1974, 1976; Leonardelli, 1976) tends to increase the number of reactive perikarya in the same species. The topography which I had proposed for LH-RH neurons in the guinea pig has been confirmed by numerous authors (Silverman, 1976; Setalo et al., 1976a; Silverman and Desnoyers, 1976; Silverman and Zimmerman, 1977; Hoffman et al., 1978a; Weindl et al., 1976; Weindl and Sofroniew, 1978). In the cat (Barry and Dubois, 1974b, 1975), the topography of the LH-RH neurons is very similar to that observed in the guinea pig. In the dog (Barry and Dubois, 1974c, 1975) the LH-RH-reactive neurons are abundant in the prepubertal period and are concentrated in the mediobasal

196

JULIEN BARRY

hypothalamus and in the postinfundibular, premamillary , preoptic, suprachiasmatic, septal, and anterior pericommissural regions. A few disseminated perikarya can be observed in the periventricular, retromammillary , and rostral mesencephalic regions (Fig. 7). In the rhesus monkey and Cercopithecus (Barry et al., 1975), reactive perikarya are seen in the mediobasal hypothalamus and in the rostral pericommissural region. In the squirrel monkey and Cebus apella (Barry and Carette, 1975a,b; Barry et al., 1976a,b) LH-RH-reactive perikarya are principally concentrated in the mediobasal hypothalamus (retrochiasmatic and infundibular regions and postinfundibular and premammillary areas), in the anterior hypothalamus, and in the lamina terminalis [observations confirmed by Hoffman et al. (1978a) in the same species; results similar to those of Zimmerman and Antunes (1976) in Rhesus]. Rostra1 pericommissural neurons have been also identified in Cebus apella and the squirrel monkey (Barry and Carette, 1975b; Barry et al., 1976a,b), as well as in the septal, epithalamohabenular, posterior subcommissural, and periaqueductal areas (Barry, 1976d, 1978a,b; Fig. 8).

8’.

MS

FIG.7. Schematic sagittal drawing of LH-RH-reactive systems in dog brain. (1) Mediobasal hypothalamic reactive perikarya (black dots); (2) preoptic reactive perikarya; (3) pericommissural reactive perikarya; (4) rostral mesencephalic perikarya. L, Lamina terminalis. Other abbreviations as in Fig. 6. Bar = 1 mm. (Original drawing, Barry.)

IMMUNOHISTOCHEMISTRY O F LH-RH-PRODUCING NEURONS 9,-

.*

197

,r

6 ,?," 0,)'

FIG.8. Schematic sagittal drawing of LH-RH-reactive systems in squirrel monkey (Saimiri sciureus) brain. (1) Mediobasal hypothalamic reactive perikarya (black dots); (2) preopticoterminal reactive perikarya; (3) rostral pericommissural reactive perikarya; (4) paraolfactory reactive perikarya; ( 5 ) rostral mesencephalic perikarya; (6) rostral limbic reactive perikarya; (7) septoepithelamic reactive perikarya; (8) habenular reactive perikarya; (9) mediothalamic reactive perikarya. E, Epiphysis; F, fornix; H, mediohabenular ganglion; P, posterior commissure; SO, subfornical organ. Other abbreviations as in Fig. 6. Bar = 1 mm. (Original drawing, Barry.)

Injection into the cerebral ventricle of serotonin, noradrenaline, glutamate (Barry et al., 1976b), and prolactin (Barry and Poulain, 1976; Barry et al., 1978) causes a more or less notable increase in the number of LH-RH-reactive perikarya in the squirrel monkey. In the female, cyclic modifications in the number of reactive perikarya are observed during the course of the estrus cycle (Barry and Croix, 1978). In the rabbit (Barry, 1976b,d) the LH-RH-reactive perikarya are principally concentrated in the mediobasal hypothalamus (retrochiasmatic, infundibular, and premamillary), the suprachiasmatic region, the subcommissural organ, the anterior hypothalamic area, the rostral pericommissural region, the rostral mesencephalic and mamillary regions, the paraolfactory region, and the limbic

198

JULIEN BARRY

telencephalon. These observations have been partially confirmed by Setalo et al. (1976b) and Flerk6 et al. (1978), according to whom the majority of the reactive cell bodies are concentrated in the preopticosuprachiasmatic area, but are also present in the mediobasal area of the hypothalamus. In the human fetus (Bugnon et al., 1976a-c, 1977a-e, 1978; Bloch, 1977; Paulin et al., 1977) and in man (Barry, 1976c, 1977a,b), LH-RH-reactive perikarya are topographically very close to the maximal concentrations in the mediobasal region (retrochiasmatic, infundibular, postinfundibular, and premamillary areas), in the preoptic region and lamina terminalis, in the septum, and in the rostral pericommissural area. A few reactive perikarya can also be observed in the suprainfundibular region and in the rostral and ventral mesencephalon. The demonstration of LH-RH-reactive perikarya in the rat was particularly controverted, due to the failure of various groups (Kordon et al., 1974; Baker et al., 1974, 1975; King et al., 1974; Pelletier et al., 1974; Goldsmith and Ganong, 1975) and the criticisms of Naik’s work (1975a,b). Nevertheless, after administration of Nembutal to proestrus rats, or to animals with the mediobasal hypothalamus frontally isolated, Setalo et a!. (1976a) were able to identify reactive perikarya in the preoptic area. Similar results were obtained in normal rats by Flerk6 et al. (1978, after injection of sulpiride or of reserpine), by King and Gerall (1976), Kozlowski and Hostetter (1977; immunoserum No. 4305 A from Stemberger), and Kami et al. (1977 after injection of reserpine). In newborn rats (King et al., 1978), the number of reactive perikarya is higher in males than in females; it is diminished in males treated with estrogens. It is augmented in the preoptic area (but not the arcuate nucleus) after dehydration (Krisch, 1978). In the mouse, the first observations by Zimmerman et al. (1974) showed the presence of reactive perikarya in the region of the arcuate nucleus. These results were criticized due to the positive reactions which were also found in the infundibular tanicytes. Several years passed before systematic research was carried out by the Rochester group, using several antisera of various specificities, from different laboratories. These investigations have shown that reactive perikarya in the rat and mouse are made up of two distinct populations with different immunoreactivities. Neurons of type I (retrochiasmatic and infundibular) react only with antisera E and F from Somentino and Sundberg (1975) (see Table I) prepared against histidyl-conjugated LH-RH. Type I1 neurons (mediopreoptic and septa1 regions) do not react with antiserum F, but react with Arimura antisera 743 and 710 (see Table I), prepared against LH-RH conjugated at the terminal glutamyl and glycyl residues, respectively (Hoffman, 1976; Hoffman et al., 1976, 1977, 1978a; Bennett-Clarke, 1977; Knigge et al., 1978). These same authors have shown that

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

199

the perikarya from the guinea pig, the squirrel monkey, and the dog (Type 111 neurons) react (like axons in all species) with antisera F, 743, and 710, as well as with a few other antisera (thus confirming my earlier observations in the guinea pig with Dubois antisera, with antisera 743 and 422 from Arimura, and with my antisera 21 and 22) (see Table I). Finally, in the sheep (Hoffman et af., 1978a), reactive bipolar neurons of type I1 are dispersed from the septa1 region to the arcuate nucleus and the premammillary area. In summary, in mammals, LH-RH-reactive perikarya are very widespread, outside of the hypothalamus, and do not respect the limits of the classical nuclei. The number of reactive perikarya with a given antiserum varies considerably (from several cells to several hundred cells per hypothalamus) according to the species, and, in the same species, according to experimental and physiological conditions. The variable reactivity of perikarya to LH-RH according to species, and within the same species according to topography, suggests the existence of various molecular conformations (or linkages) of somatic LH-RH (probably linked by its radical N terminal, or possessing a substituted amino acid in one of the first residues, for neurons of type I, and probably linked by its N- or -C extremities for neurons of type 11: Hoffman et al., 1978b). Antisera produced by antigenic challenge with LH-RH conjugated to bovine serum albumin at Tyr 5 do not show the presence of reactive perikarya (Baker et al., 1975; Goldsmith and Ganong, 1975; Baker and Dermody, 1976; Gross, 1976; Gross and Baker, 1977). The picture presented above has been modified very recently (G. E. Hoffman, personal communication) because it has been shown that both antisera E and F contained antibodies against ACTH, -24 (and therefore, that the majority of type I neurons are ACTHI-,,, not LH-RH-containing cells) and that true LH-RH cells (regardless of type I1 or 111) are concentrated either in the rostra1 hypothalamus or in the mediobasal area, or in both areas, according to the species.

B. BIRDS In the mallard duck reactive perikarya have been observed in the dorsolateral region of the arcuate nucleus (McNeill et al., 1976) as well as in the usually bipolar oval perikarya (diameter 15 to 25 pm) in the dorsal region of the preoptic periventricular nucleus (Bons et af., 1977, 1978a) and in the infundibular area (Bons et al., 1978b; Fig. 9). In the pheasant and the chicken (Hoffman et a / . , 1978a), type I1 reactive neurons are seen in the mediobasal hypothalamic region, notably, in the periventricular area. These neurons are generally bipolar (though occasionally multipo-

200

JULIEN BARRY

NPA

FIG.9. Schematic sagittal drawing of LH-RH neurosecretory systems in duck hypothalamus. Localization of LH-RH neurons (0) and fibers (0-0). AH, Adenohypophysis; CO, optic chiasm; CSO, supraoptic commissure; CT, tuberal complex; EM, median eminence; NH, neurohypophysis; NPA, anterior preoptic nucleus; 3V, third ventricle. (From Bons er al., 1978b.)

lar) and are widely dispersed in the medioseptal region, the preoptic area, the paraolfactory lobe, and the olfactory bulb. C. AMPHIBIANS A N D FISHES

In Xenopus laevis, Doerr-Schott and Dubois (1976) have identified numerous perikarya, which form three groups: a telencephalic dorsal group, whose elements are dispersed on either side of the area around the ventricle; a dorsal group, near the preoptic recessus; and a third group near the floor of the infundibulum (Fig. 10). In the toad Bufo vulgaris, after hypophysectomy (Doerr-Schott and Dubois, 1975b), reactive perikarya are observed in a paired nucleus, situated in the caudal region of the diencephalon. In the frogs Rana pipiens and Rana temporaria, Alpert et al. (1976), and in the adult green frog Rana esculenta, Goos et al. (1976) have identified reactive perikarya in the interior of an impaired nucleus, situated in front of the preoptic recessus. Finally, in the trout Salmo gairdneri (animals sacrificed in September) Goos and Murathanoglu (1977) were able to show small perikarya which were faintly reactive, on both sides of the ventriculis communis, in the area dorsalis pars medialis of the telencephalon.

IMMUNOHISTOCHEMISTRY O F LH-RH-PRODUCING NEURONS

20 1

Preoptic reces

lnfundibular recess

Median eminence

FIG. 10. Diagrammatic horizontal section through the forebrain of Xenopus luevis indicating the positions of the perikarya and the LH-dH-like neurosecretory fiber pathways. (A) Two groups of LH-RH-like cells in the sepal zone of the telencephalon. (These more dosal cells are superimposed.) (B)LH-RH-Like cells oh either side of the preoptic recess among the neurosecretory fibers. (C) LH-RH-Like cells in the infundibulum. (1) Paired neurosecretory bundles running dorsoventrally through the telencephalon and forming a single bundle (2) between the posterior edge of the telencephalic furrow and the anterior edge of the preoptic recess. The bundle divides into two (3) and the two resulting bundles fan out on either side of the preoptic recess running above the optic chiasma along the ventral edge of the infundibulum; (4) the fibers terminate in the external zone of the medial eminence. (From Doerr-Schott and Dubois, 1976.)

V. Hypothalamohypophyseal LH-RH Tracts The reactive LH-RH present in the axons reacts with all specific primary antisera used. Consequently, LH-RH-reactive axons have been identified by all authors who have attempted their demonstration, both those who have been able to characterize reactive perikarya (references cited above in Section IV and. in Table 11) and those who have not managed to do so (Dubois, 1973, in the ram, the boar, and the bull; Baker et al., 1974, 1975; King et al., 1974; Pelletier et

202

JULIEN BARRY

a l . , 1974; Goldsmith and Ganong, 1975, in the rat; Bugnon et al., 1976d, in the guinea pig, the cat, and the dog; de Reviers and Dubois, 1974, in the cock; Calas et a l . , 1974, in the duck; Sharp et a l . , 1975, in the green finch). In all species where LH-RH-reactive axons have been identified, they present the same nodular appearance and contain variable quantities of reactive material, generally most abundant in their extremities. It is in the infundibulum and the hypophyseal stalk that they are generally most numerous and most easily seen.

A. MAMMALS 1 . Hypothalamoinfundibular LH-RH Tract (HIT) This tract constitutes the major efferent axonal route for LH-RH neurons, and it ends in the area of the capillaries of the primary portal plexus of the pituitary (intercalar plexus and intrainfundibular loops), generally after having traversed the entire external infundibular zone. This tract is made up of various components (anterior or ventral, lateral, dorsal or posterior, periaxial) whose importance, trajectories, and complexity vary according to species (see Figs. 11 and 12). The axons appear to end, in order of decreasing importance, on the lateral, posterior, and anterior faces of the infundibulum. A few may end at the level of the capillaries of the postinfundibular eminence. The LH-RH released by the axon endings at the level of the fenestrated capillaries of the primary portal plexus reaches the prehypophyseal capillaries by the portal vessels and can thus act upon the FHS/LH gonadotrophic cells (for hypothalamoprehypophyseal relations and the corresponding references, see Barry, 1966; Duvernoy et a l . , 1971; Page and Bergland, 1977). With the electron microscope, 90- to 130-nm specific granules were observed in reactive axons (Pelletier, 1976; Pelletier et al., 1974, 1976, 1977; Goldsmith and Ganong, 1975, in the rat; Silverman and Desnoyers, 1976, in the guinea pig; Bugnon et al., 1977b-e, 1978; Bloch, 1977, in the human fetus) as well as in their pericapillary extremities, where their diameter is between 60 and 130 nm in the guinea pig (Silverman and Desnoyers, 1976). The question as to the origin of the axons of the hypothalamoinfundibular tract is not yet resolved, and results may vary according to species. In the guinea pig, our initial hypothesis of a preoptic perikaryal origin (that is to say, a “preopticoinfundibular tract,” Barry et al., 1973a, 1974; Barry and Dubois, 1974d) was based upon the fact that the majority of reactive perikarya were located rostrally in this species and that the frontal hypothalamic isolation caused, in our animals, a decrease in infundibular reactive material (however the number of ‘‘interrupted” axons observed remained relatively small). Furthermore, the existence of a preopticoinfundibular tract seems well established in the guinea pig (Poulain and Partouche, 1973, by antidromic stimulation of the external median

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

203

FIG. 11. Hypothalamoinfundibular LH-RH tract (HIT) of the guinea pig. (a) Sagittal section, 2-week-old animal ( X 185). (b) Coronal section of the dorsolateral part of the median eminence, young adult (X280). (c) Coronal section of the ventral part of the median eminence (X280). Bar = 50 p m . (1) Ventral component of the HIT; (2) dorsal component of the HIT; (3) lateral component of the HIT; al, anterior lobe. of the hypophysis; an, arcuate nucleus; il, intrainfundibular capillary loop; pt, pars tuberalis; t, tuber; V, third ventricle. (Original documents, Barry.)

FIG.12. Hypothalamohypophyseal LH-RH-reactive tracts in monkeys. (a) Posterior component (2) of the hypothalamoinfundibular LH-RH-reactive tract (HIT) on sagittal section, in Cebus apella (x280). (b) Anterior component (1) of the HIT, on sagittal section in Cebus apellu (X280). (c)

Hypothalamoneurohypophyseal LH-RH-reactive tract (hnt), on sagittal section, in the squirrel monkey (x280). Bar = 50 pm. ah, Adenohypophysis; ame, anterior median eminence; ip, intercalar plexus; pie, postinfundibular eminence; pme, posterior median eminence; pt, pars tuberalis. (Original documents, Barry.)

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

205

eminence; Poulain, 1977, after retrograde axonal transport of horseradish peroxidase), even if it has not been unequivocably proved to come from LH-RH neurons. On the other hand, experiments with hypothalamic disconnection (Silverman, 1976) and lesions of the preoptic area (Silverman and Krey, 1977, in the guinea pig) have not demonstrated a notable decrease in the quantity of reactive infundibular material, an observation which supports a mediobasal neuronal origin for HIT. However, in the rat and the mouse analogous experiments show a marked decrease in mediobasal hypothalamic LH-RH (Weiner et al., 1975; Kalra, 1976; Kalra et al., 1977; Brownstein et al., 1976) and a major decrease in the number of reactive fibers in the median eminence (Sttalo et al., 1976a). It is nevertheless possible that these modifications are linked to a suppression of certain afferent connections of the hypothalamic mediobasal reactive neurons, but in the dehydrated rat, Krisch (1978) observed an increase in the number of reactive perikarya in the preoptic area (but not in the arcuate nucleus), as well as evidence for axonic connections between the preoptic area and the median eminence. In the primates (notably the squirrel monkey and man), few reactive axon tracts are observed between the rostra1 hypothalamus and the basal hypothalamus. This could indicate that the great majority of reactive axons of the HIT originate from the mediobasal hypothalamic group of reactive perikarya (Fig. 8). The hypothalamoinfundibular LH-RH tract shows characteristic charge modifications during ontogenesis (Barry and Dubois, 1974a, in the guinea pig fetus; Bugnon et al., 1976a,b,c, 1977a-e, 1978; Bloch, 1977, in the human fetus), during the neonatal period (Barry and Dubois, 1974a, in the guinea pig; Paull, 1978, in the rat), during the estrus cycle (Barry and Dubois, 1974b,d, in the guinea pig; Naik, 1976; Ibata et al., 1978, in the rat; Barry and Croix, 1978, in the squirrel monkey), with a charge maximum at the end of diestrus and the beginning of proestrus and a charge minimum at ovulation. Modifications are also seen during gestation (Barry and Dubois, 1973a, 1974b, in the guinea pig), after castration (Barry et al., 1974; Leonardelli et al., 1975), and after dehydration (Krisch, 1978, in the rat). A satisfactory correlation exists between these observations and the variations in hypothalamic LH-RH content: in the rat, during the estrus cycle, assayed by biological methods (Chowers and McCann, 1965; Ramirez and Sawyer, 1965) or radioimmunoassays (Kalra et al., 1973; Araki et al., 1975), and during the postnatal period (Araki et al., 1976); in the sheep, during the course of the estrus cycle (Crighton et al., 1973); in the female rat with the variations of LH-RH plasma levels during the estrus cycle (Meyer et al., 1974); and in the guinea pig, with the plasma variations of FSH and LH during the neonatal and the postnatal periods (Croix, 1976), during the estrus cycle (Croix and Franchimont, 1975), and after castration (Croix , 1977).

206

JULIEN BARRY

2. Hypothalamoneurohypophyseal LH-RH Tract (HNT) Reactive axons may traverse the hypophyseal stalk, reaching the hilum and the neural lobe. Their number is generally small, but varies greatly according to species. These axons have been described in the dog (Barry and Dubois, 1974c, 1973, various monkeys (Barry and Carette, 1975a,b; Barry et al., 1976a,b), the human fetus (Bugnon et al., 1977a-e, 1978; Bloch, 1977), man (Barry, 1976c, 1977a,b), and guinea pig (Silverman and Zimmerman, 1978). See Fig. 12c. B. BIRDS In birds the axons of the hypothalamoinfundibular LH-RH tract are directed vertically toward the external zone of the median eminence and are distributed to the capillaries of the intercalary plexus at the level of the rostra1 and caudal areas of the median eminence (Calas et al., 1973, 1974, 1975a,b; de Reviers and Dubois, 1974; Sharp et al., 1975; McNeill et al., 1976; Bons et al., 1977, 1978a,b) (Fig. 9). This distribution suggests that LH-RH may be secreted into the portal circulations of both the cephalic lobe and the caudal lobe. C. AMPHIBIANS A N D FISHES

In the toad Bufo vulgaris, the axons coming from the reactive perikarya of the caudal telencephalon (Doerr-Schott and Dubois, 1975b) reach the median eminence, where the majority end; some of these axons may end in the vicinity of the paired pars tuberalis or the pars nervosa. In Xenopus laevis (Doerr-Schott and Dubois, 1976), the axons leaving the telencephalic cellular groups form paired tracts which traverse the telencephalon in a dorsoventral fashion and then form a single bundle. This bundle then divides into two bundles which are joined by axons of the preoptic and infundibular cellular groups, before terminating in the external zone of the median eminence (Fig. 10). At this level its fibers contain irregularly shaped neurosecretory granules with an average diameter of 890 A (Doerr-Shott et a l . , 1978). In Rana pipiens and Rana catesbeiana (Alpert et al., 1976) and in Rana esculenta (Goos et al., 1976), the axons coming from the reactive cell bodies are situated in front of the preoptic recessus, forming a single tract which passes above the optic chiasm. This tract then divides into two symmetrical bundles, which reunite just before penetrating the median eminence. These axons terminate near the capillaries of the external zone of the median eminence. In Salmo gairdneri (Goos and Murathanoglu, 1977), the axons coming from the reactive cell bodies do not form a compact bundle, but proceed in a dispersed

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

207

fashion in a caudoventral direction, toward the hypophyseal stalk. Their endpoints have not been determined in a precise manner.

VI. Preopticoterminal LH-RH Tract The presence of LH-RH-reactive axons in the vascular organ of the lamina terminalis (“organum vasculosum of the lamina terminalis, OVLT) has been observed since the first studies, notably in the guinea pig (Barry et al., 1973a), the mouse (Barry et al., 1973a; Zimmerman et al., 1974), and the rat, (Barry et al., 1973a; Kordon et al., 1974; King et al., 1974; Naik, 1975a,b). These observations were later confirmed in the rat (Pelletier, 1976; Pelletier et al., 1976, 1977; Barry and Dubois, 1976; King and Gerall, 1976; Kozlowski et al., 1977; Weindl and Sofroniev, 1978) and the guinea pig (Silverman, 1976; Silverman and Desnoyers, 1976) and extended to other species: monkeys (Barry and Carette, 1975a,b; Barry et al., 1976a,b,; Zimmerman and Antunes, 1976; Silverman et al., 1977), rabbit (Barry, 1976b; Flerko et al., 1978), human fetus (Bugnon et al., 1976a-c, 1977a-e, 1978; Bloch, 1977; Paulin et al., 1977), man (Barry, 1976c, 1977a), and Tupia (Barry, 1978b). The perikarya of these axons are principally situated in the lamina terminalis and neighboring regions, the preoptic region and the anterior hypothalamus (Barry and Carette, 1975a; Barry et al., 1976a,b, in the squirrel monkey; Zimmerman and Antunes, 1976, in Macacus rhesus). This was later also shown for man and Tupia (Barry, 1976c, 1977a, 1978b). In the rat, Palkovits et al. (1978) have shown an analogous origin, by injection of horseradish peroxidase at the level of the lamina terminalis. These results are in accordance with the hypothalamic disconnection experiments (Weiner et al., 1975, in the rat) which do not cause a notable depletion of LH-RH at the level of OVLT, this area containing high quantities of reactive LH-RH (Kizer el al., 1976). In some cases, supraependymal perikarya, which resemble the “liquor contacting neurons” (Vigh and Vigh-Teichmann, 1973), have been observed in the OVLT (J. Barry, unpublished results, 1977). See Fig. 13. It appears that the axons (or axon collaterals) which issue from the abovementioned perikarya form a true preopticoterminal LH-RH tract (Barry and Carette, 1975a) which constitutes, after the hypothalamohypophyseal LH-RH tract, the principal efferent axonal pathway for LH-RH cell bodies. This tract forms an important plexus in the hypendyma and gives rise to numerous collaterals which terminate mainly along the short external capillaries and the long, internal vascular loops of the OVLT. These axons have been studied by electron microscopic immunocytochemistry in the rat (Pelletier et al., 1976, 1977) and in the squirrel monkey (Mazzuca, 1977) and probably correspond to a part of the

FIG.13. Preopticoterminal LH-RH-reactive tract in squirrel monkey in: (a) rostral coronal section of the lamina terminalis ( x 180); (b) middle coronal section of the lamina terminalis ( X 180); (c) sagittal section of the upper third of the lamina terminalis (X300); (d) sagittal section of the lower third of the lamina terminalis (X255). Bar = 50 pm. e, Ependyma; fs, fibrous septum; hp, hypendymal plexus; rc, radiating collaterals; rp, rostral plexus; sg, subglial rostral layer. Reactive perikarya (c) are indicated by arrows. (Original documents, Barry.)

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

209

nervous fibers previously described at this level by Rohlich and Wenger (1969), Le Beux (1971, 1972), and Le Beux et al. (1971) in the rat. Certain axons of the preopticoterminal LH-RH tract reach and cross the subglial rostral layer. Other axons cross between the flattened ependymal cells covering the ventricular side of the lamina terminalis and give rise to thin intraventricular nerve endings (Barry, 1978b). These structures may correspond to the neurosecretory endings observed with the electron microscope in the golden hamste; (Weindl and Schinko, 1974) and in the rat (Le Beux, 1972). In electron microscopy, a complex network of perivascular spaces has been described in the OVLT (Wenger and Toro, 1971; Wenger, 1976, Mazzuca, 1977; Palkovits et a l . , 1978), as well as fenestrated capillaries (references in Bouchaud, 1975; Weindl and Sofroniew, 1978). The preopticoterminal LH-RH tract shows characteristic variations of granular material (Wenger, 1976; Naik, 1976) or immunoreactivity under various circumstances (Barry et a l . , 1976a,b; Naik, 1976; Setalo et al., 1976b; Wenger, 1976; Gross and Baker, 1977; Bugnon e t a l . , 1976a,b,c, 1977 a-e, 1978; Bloch, 19771, with marked depletion during the ovulatory period (Naik, 1976, in the rat; J. Barry, in the squirrel monkey, unpublished results, 1978). Taking these data into account, we believe that the preopticoterminal tract has an important functional role and that it does not represent a vestigial structure, as suggested by Kizer et al. (1976) and Palkovits et al. (1978). The morphological features of this system suggest that LH-RH of the preopticoterminal tract may act through the intermediary of the internal or external cerebrospinal fluid, but especially through the blood of the OVLT. Nevertheless, despite numerous studies (Hofer, 1958; Duvernoy and Koritke, 1961; Duvernoy et al., 1969; Weindl and Joynt, 1972; Weindl and Sofroniew, 1978; Palkovits et a l . , 1978), the destination of the venous blood from the OVLT remains debatable: general circulation (Duvernoy et a l . , 1969), capillaries of the mediopreoptic region (Palkovits et a l . , 1978), or capillaries of the rostral pericommissural area (Weindl and Sofroniew, 1978). The majority of authors have rejected the suggestion of direct venous connection with the hypophyseoportal system.

VII. Extrahypothalamic LH-RH Tracts Reactive LH-RH axonic pathways have been observed outside of the hypothalamus since the first immunochemical studies (Barry et al., 1973a). I proposed to designate them “extrahypophyseal, or “extrahypothalamic” LHRH pathways (Fig. 14). If we except a few dispersed axons, which reach the ventricular cavity at various levels (hypothalamoventricular LH-RH tracts) we can classify these extrahypothalamic pathways as follows: ”

210

JULIEN BARRY

FIG.14. Extrahypophyseal LH-RH-reactive tracts in the squirrel monkey. (a) Small bundle of reactive axons in the dorsal epithalamus (x255). (b) Reactive axon running along the ventral side of the supracommissural fornix ( ~ 5 8 5 ) (c, . d) Thickened reactive preterminal segments of LH-RH axons in the mediohabenular ganglion ( x 585). Note the numerous pericellular knobs (arrows).Bar = 10 hm. ax, Axons; cp, choroid plexus; e, ependyma; sf, supracommissural fornix. (Original documents, Bany.)

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

21 1

i. F’reopticosupraoptic LH-RH tract (Barry et al., 1973a, 1974; Barry and Dubois, 1974b, in the guinea pig; tracts I and I1 of Silverman, 1976). This tract is formed by axons issuing from rostra1 and pericommissural perikarya, which terminate principally in the region of the supraoptic crest, but also in the suprachiasmatic region. ii. Basal rhinencephalic LH-RH tracts (Barry and Dubois, 1976, in the guinea pig; minor tracts of Silverman, 1976, in the guinea pig; Barry, 1976b; Flerko et af., 1978, in the rabbit; Silverman and Zimmerman, 1978, in the mouse, where reactive axons were observed surrounding vessels up to the level of the olfactory bulb). iii. Rostra1 limbic LH-RH tract, formed by axons running from the white anterior commissure, toward the telencephalon (Barry et al., 1976a, in the guinea pig; Barry, 1976b; Flerko et al., 1978, in the rabbit; Barry, 1976d, 1978b, in the squirrel monkey). Very rarely, reactive fibers could be observed in the subfornical organ (Barry et al., 1976a, in the squirrel monkey; Weindl and Sofroniew, 1978, in the rat). iv. Septoepithalamohabenular and thalamic LH-RH tract. This tract is formed by reactive axons coming from septal, epithalamic, and habenular perikarya. They travel the length of the stria-medullaris and terminate principally in the mediohabenular ganglion and the mediodorsal thalamic nucleus (Barry, 1978a, in the squirrel monkey). Axons with a thalamic destination (paratenial thalamic nucleus and mediodorsal thalamic nucleus) have recently been described by Knigge et al. (1978) and Hoffman et al. (1978b) in the mouse. On the other hand, reactive axons observed by Silverman and Zimmerman (1978) in the guinea pig, in the epithalamic and habenular areas, proceeded the length of the fasciculus retroflexus and terminated in the region of the interpedoncular ganglion. v. Hypothalamoamygdaloid LH-RH tract (Barry et al., 1973a, in the guinea pig; Barry and Carette, 1975a,b; Barry et al., 1976a, in the squirrel monkey). This tract is formed by axons which terminate at the level of the basal and medial amygdaloid nuclei, after having followed a ventral (ventral amygdaloid tract, Leonardelli and Poulain, 1977, in the guinea pig) or dorsal route. In the squirrel monkey, some of these axons may come from periamygdaloid reactive perikarya (Barry, 1976d, 1978b). vi. Mesencephalic LH-RHtract, represented by retromammillary and mesencephalic axons, which are often difficult to follow (Barry et al., 1973a; Silverman, 1976, tract IV in the guinea pig; Barry and Dubois, 1974c,d; Hoffman et al., 1978a; in the dog; Barry, 1976b; Flerko et al., 1978, in the rabbit; Barry and Carette, 1975a,b; Barry et a l . , 1976a,b; Hoffman et al., 1978a, in the squirrel monkey; Bugnon et al., 1976a-c, 1977d-f, 1978; Bloch, 1977, in the human fetus; Barry, 1976c, 1977a, in man).

212

JULIEN BARRY

These various tracts are difficult to study and resemble the “extrahypophyseal neurosecretory pathways” which I described previously (Barry, 1954a,b), using the techniques of Gomori, as originating in paraventricular and supraoptic nuclei in the bat, Rhinolophus. The general topography of the “extrahypothalamic LH-RH pathways” (and those of extrahypothalamic reactive perikarya) is in keeping with the significant quantities of LH-RH present in the various circumventricular organs (Kizer et al., 1976; Okon and Koch, 1977) and in the central nervous system (Currie et al., 1977; Aksel and Tyrey, 1977; Jackson, 1978; Gilmore et al., 1978). The majority of these pathways (if not all of them) probably originate in extrahypothalamic perikarya. An origin from the mediobasal hypothalamic perikarya, however, remains possible, based on both axonic divisions observed in immunocytochemistry (Barry et af., 1974; Weindl and Sofroniew, 1978) and above all, neurophysiological evoked potential experiments (Renaud, 1977a,b, 1978). Images observed at the end (or along the length of) certain extrahypothalamic LH-RH pathways suggest that LH-RH could act through the intermediary of local vascular systems, or by the cerebrospinal fluid (references in Ben-Jonathan et al., 1974; Uemura et af., 1975; Knigge et al., 1977, 1978). With this in mind, it is of interest to note that numerous extrahypothalamic LH-RH-reactive perikarya and axons are located in the proximity of blood vessels. In addition, numerous images of “neurosecretory synapses” (which should be verified by electron microscopy) or of diffuse projection systems have been observed in certain target regions, principally the supraoptic crest, the mediohabenular ganglia, the dorsomedial thalamic nuclei, the amygdala, and the interpeduncular ganglion. These observations suggest that LH-RH could act at this level as either a neuromodulator or a neurotransmitter. This hypothesis fits well with the results obtained after intravenous or microiontophoretic infusion of LH-RH (Moss and McCann, 1973; Pfaff, 1973; Dyer and Dyball, 1974; Renaud et af., 1975, 1976; Kelly and Moss, 1976; Moss and Foreman, 1976; Poulain and Carette, 1977; Koranyl et af., 1977; Moss et al., 1978; Renaud, 1978) or anti-LH-RH antisera (Sakuma and Kawakami, 1976; Kozlowski and Hostetter, 1978). A specific role for certain extrahypothalamic pathways is particularly probable in terms of sexual behavior, notably for the preopticosupraoptic tract and the mesencephalic pathways. The preoptic area is implicated in male sexual behavior in the rat (Heimer and Larsson, 1967, references in Pfaff, 1978); the mediobasal hypothalamus and certain regions of the mesencephalon are implicated in the control of female sexual behavior (references in Pfaff, 1978); finally, LH-RH has various effects upon sexual behavior in animals and in man (references in Moss, 1978). It is, for instance, interesting to note that the majority of extrahypothalamic LH-RH-reactive axons became invisible after copulation in the female rabbit

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

213

(Setalo et al., 1978), and that the septoepithalamohabenular tracts are generally more visible in the male squirrel monkey during the summer period of low sexual activity (J. Barry, unpublished results, 1977).

VIII. General Discussion and Conclusions The hypothalamic topography of the LH-RH-reactive perikarya giving rise to the hypothalamoinfundibular tract, and to the preopticoterminal LH-RH tract, fits closely with the hypothalamic LH-RH topography determined by biological or biochemical assays (Crighton et al., 1970, 1973; Palkovits et al., 1974; Estes et al., 1977), by radioimmunoassays (King et al., 1974; Benveniste et al., 1975; Wheaton et al., 1975; Kizer et al., 1976; Styne et al., 1977, in the rat; Bird et al., 1976; Okon and Koch, 1976, 1977, in man; Siler-Khodr and Khodr, 1978, in the human fetus; Estes et al., 1977, in the cow), or after hypothalamic disconnection (Brownstein et al., 1975, 1976; Weiner et a l ., 1975; Kalra, 1976; Kalra et al.. 1977; Kobayashi et al., 1978). The modifications of immunoreactivity of the hypothalamoinfundibular and the preopticoterminal LH-RH tracts, as well as of the extrahypophyseal LH-RH pathways (see Sections V, VI, VII), suggest that the reactive material shown is in fact LH-RH, which is, at least in part, synthesized in the reactive perikarya, possibly in the form of a high-molecular-weight precursor, as suggested by Millar et al. (1977). For the time being LH-RH neurons do not appear to make any other neurohormone and seem to be distinct from cells which produce other neuropeptides (Dubois et al., 1974; Pelletier, 1976; Bugnon et al., 1976d; Hokfelt et al., 1978), or from monoaminergic neurons (Kizer et al., 1976). A substance crossreacting with anti-ACTH,,_,, has, however, recently been detected in guinea pig LH-RH perikarya and axons (Tramu et al., 1977). Observations of various LH-RH-reactive axonal tract endings suggest that, besides its major, direct, prehypophysotrophic action through the hypophyseoportal system, LH-RH may also be distributed by the blood of the vascular organ of the lamina terminalis (does it finally reach the hypophysis and other peripheral target organs by the systemic circulation? or by various rostral hypothalamic regions: medial preoptic area? rostral commissural area?). It may also be secreted into the internal or external cerebrospinal fluid, or it may act as a neurotransmitter or a neuromodulator (through “neurosecretory synapses” or diffuse peri- or interneuronal projections), in various processes of behavior or neuroendocrine integration. LH-RH neurons are thus seen not only as classical neurosecretory cells, with solely pericapillary axon endings, but also as neurons with multiple, portal, and

214

JULIEN BARRY

“extrahypophyseal” efferents (Barry et al., 1973a), that is to say, as “neuroendocrine integrator neurons” (Gorski, 1977). The characterization of the complex system formed by these neurons raises numerous problems, notably, the classification of the neuronal populations giving rise to the various LH-RH tracts (infundibular tract, preopticoterminal tract, extrahypophyseal pathways); the localization of the target sites and the functions of the extrahypophyseal LH-RH pathways [some of which may be immunocytochemically negative: ‘‘silent pathways” of Sternberger and Hoffman (1978) detectable by a systematic study of central LH-RH receptors]; the possible participation of short LH-RH axon collaterals (Barry and Dubois, 1973b) in the constitution of local retroactions [as in the intranuclear circuits postulated by Matsumoto and Arai (1978) for the arcuate nucleus]; and the interactions between the two rostra1 and basal LH-RH neuronal populations. In the coming years, the current “immunomorphologic ” description of LH-RH neurons should be extended to a greater number of species (particularly in reptilians and fishes) and complemented by “histophysiological” in vivo and in vitro studies; the results of immunocytochemical investigations should be ‘‘quantified”; the specific neurotransmitters, as well as any other neuropeptides eventually present in these neurons, should be characterized. Finally, we hope to see the identification of the specific membrane receptors which regulate the function of these neurons and of the mechanisms which control their differentiation. The theoretical and practical interest of such progress need not be underlined.

REFERENCES Aksel, S., and Tyrey, L. (1977). Ferril. Steril. 28, 1067. Alpert, L. C., Brawer, J. R., Jackson, M. D., and Patel, Y. (1975). Fed. Proc., Fed. Am. SOC.Exp. Biol. 34, 239 (abstr.). Alpert, L. C . , Brawer, J. R., Jackson, M. D . , and Reichlin, S. (1976). Endocrinology 98, 910. Amoss, M., Burgus, R., Blackwell, R., Vale, W . , Fellow, R . , and Guillemin, R. (1971). Biochem. Biophys. Res. Commun. 44, 205. Araki, S., Ferin, F., Zimmerman, E. A., and Van de Wiele, R . L. (1975). Endocrinology 96, 644. Araki, S., Toran-Allerand, C. D., Ferin, F., and Van de Wiele, R. L. (1976). Endocrinology 97, 693. Avrameas, S., and Lespinats, G. (1967). C.R. Hebd. Seances Acad. Sci. 265, 1149. Baker, B. L., and Dermody, W. C. (1976). Endocrinology 98, 1116. Baker, B. L., Dermody, W. C., and Reel, J . R. (1974). Am. J . Anar. 139, 129. Baker, B. L., Dermody, W. C., and Reel, J . R . (1975). Endocrinology 97, 125. Barry, J . (1954a). Bull. SOC.Sci. Nancy 13, 126. Barry, J. (1954b). Arch. Anar. Microsc. Morphol. Exp. 43, 310. Barry, J. (1960). C.R. Seances SOC. Biol. Ses Fil. 154, 1250. Barry, J. (1966). Pathol. Biol. 13, 866. Barry, J . (1967). Neurosecretion Int. Symp., 4th. 1967 pp. 56-59.

IMMUNOHISTOCHEMISTRY OF LH-RH-PRODUCING NEURONS

215

Barry, J. (1976a). In “Hypothalamus and Endocrine Functions” (F. Labrie, J. Meites, and G. Pelletier, eds.), p. 451. Plenum, New York. Barry, J. (1976b). Neurosci. Letr. 2, 201. Barry, J. (1976~).Neurosci. Lett. 3, 287. Barry, J. (1976d). C.R. Hebd. Seances Acad. Sci. 283, 1531. Barry, J. (1977a). Cell Tissue Res. 181, 1. Barry, J. (1977b). Ann. Endocrinol. 38, 11 1. Barry, J. (1978a). Brain Res. 151, 183. Barry, 1. (1978b). lnr. Symp. Neuroendocr. Regul. Mech. 1979 pp. 13-250. Barry, J., and Carette, B. (1975a). C.R. Hebd. Seances Acad. Sci. 281, 735. Barry, J., and Carette, B. (1975b). Cell Tissue Res. 164, 163. Barry, J., and Croix, D. (1978). Cell Tissue Res. 192, 215. Barry, J., and Dubois, M. P. (1973a). C . R . Seances SOC. Biol. Ses Fil. 167, 1812. Barry, J., and Dubois, M. P. (1973b). Ann. Endocrinol. 34, 735. Barry, J., and Dubois, M. P. (1974a). Brain Res. 67, 103. Barry, J., and Dubois, M. P. (1974b). In “Neurosecretion-The Final Neuroendocrine Pathway” (F. Knowles and L. Vollrath, eds.), pp. 148-153. Springer-Verlag, Berlin and New York. Barry, J., and Dubois, M. P. (1974~).Arch. Anat. Microsc. Morphol. Exp. 63, 363. Barry, J., and Dubois, M. P. (1974d). Neuroendocrinology 15, 200. Barry, J., and Dubois, M. P. (1975). Neuroendocrinology 18, 290. Barry, J., and Dubois, M. P. (1976). Acra Anat. 94, 497. Barry, J., and Poulain, P. (1976). J . Physiol. (Paris) 72, 14B (abstr.). Barry, J., Dubois, M. P., and Poulain, P. (1973a). Z. Zellforsch. Mikrosk. Anat. 146, 351. Barry, J., Dubois, M. P., Poulain, P., and Leonardelli, J . (1973b). C . R . Hebd. Seances Acad. Sci. 276, 3191. Barry, J., Dubois, M. P., and Carette, B. (1974). Endocrinology 95, 1416. Barry, J., G i r d , C., and Dubois, M. P. (1975). Bull. Assoc. Anat. 59, 103. Barry, J., Poulain, P., and Carette, B. (1976a). Ann. Endocrinol. 37, 227. Barry, J., Poulain, P., and Carette, B. (1976b). In “Neuroendocrinologie” (R. Mornex, ed.), p. 263. INSERM, Paris. Barry, J., Croix, D., and Poulain, P. (1978). In “Progress in Prolactin Physiology and Pathology” (C. Robyn and M. Harter, eds.), p. 253. Elsevier, Amsterdam. Ben-Jonathan, N., Mical, R. S.,and Porter, J. C. (1974). Endocrinology 95, 18. Bennett-Clarke, C. (1977). Anat. Rec. 187, 536. Benoit, J., and Assenmacher, I. (1955). J . Physiol. (Paris) 47, 427. Benveniste, R., Frohman, L. A., and McCann, S. M. (1975). Endocrinology 97, 30. Bird, E. D., Chiappa, S. A., and Fink, G. (1976). Nature (London) 260, 536. Bloch, B. (1977). Ann. Sci. Univ. Besancon 5, 213. Bons, N., Kerdelhuk, B., and Assenmacher, I. (1977). C.R. Hebd. Seances Acad. Sci. 285, 1327. Bons, N., Kerdelhuk, B., and Assenmacher, I. (1978a). Cell Tissues Res. 188, 99. Bons, N., Kerdelhue, B., and Assenmacher, I. (1978b). C . R . Hebd. Seances Acad. Sci. 287, 145. Bouchaud, C. (1975). J. Microsc. Biol. Cell. 24, 45. Brownstein, M. J., Palkovits, M., and Kizer, J. S . (1975). Life Sci. 17, 679. Brownstein, M. J., Arimura, A,, Schally, A. V., Palkovitz, M., and Kizer, J. S. (1976). Endocrinology 98, 662. Bugnon, C., Bloch, B., and Fellmann, D. (1976a). J. Physiol. (Paris) 72, 7 8 (abstr.). Bugnon, C., Bloch, B., and Fellmann, D. (1976b). Bull. Assoc. Anat. 169, 269. Bugnon, C., Bloch, B., and Fellmann, D. (1976~).C . R . Hebd. Seances Acad. Sci. 282, 1625. Bugnon, C., Bloch, B., Fellmann, D., and Gouget, A. (1976d). C.R. Seances SOC.Biol. Ses Fil. 170, 83.

216

JULIEN BARRY

Bugnon, C., Bloch, B., and Fellmann, D. (1977a). Actn Endocrinol. Suppl. 212, 58 (abstr.). Bugnon, C., Bloch, B., and Fellmann, D. (1977b). J. Steroid Eiochem. 8, 565. Bugnon, C., Bloch, B., and Fellmann, D. (1977~).Cell Tissue Res. 183, 319. Bugnon, C., Bloch, B., and Lenys, D. (1977d). J. Eiol. Cell 129, 34 (abstr.). Bugnon, C., Bloch, B., Lenys, D., and Fellmann, D. (1977e). Brain Res. 137, 175. Bugnon, C., Bloch, B., Lenys, D., and Fellmann, D. (1978). Brain-Endocr. Interact., Neural Horm. Reprod., Int. Symp. 3rd, I977 p. 183. Burgus, R., Butcher, M., Ling, N., Monahan, M., Rivier, J., Fellows, R., Amos, M., Blackwell, R., Vale, W., and Guillemin, R. (1971). C . R . Hebd. Seances Acad. Sci. 273, 161 1. Calas, A,, Kerdelhue, B., Assenmacher, I., and Jutisz, M. (1973). C . R . Hebd. Seances Acad. Sci. 277, 1765. Calas, A., Kerdelhue, B., Assenmacher, I . , and Jutisz, M. (1974). C.R. Hebd. Seances Acad. Sci. 278, 2557. Calas, A., Kerdelhue, B., Assenmacher, I., and Jutisz, M. (1975a). J. Physiol. (Paris) 70, 9B (abstr.). Calas, A,, Dubois, M. P., and Assenmacher, I. (1975b). J. Physiol. (Paris) 70, 10B (abstr.). Campbell, H. J., Fever, G.,Garcia, J., and Harris,G. W. (1961). J. Physiol. (London) 157, 30P (abstr .) . Chowers, I., and McCann, S. M. (1965). Endocrinology 76, 700. Courrier, R., Guillemin, R., Jutisz, M., Sakiz, E., and Aschheim, P. (1961). C.R. Hebd. Seances Acad. Sci. 253, 922. Crighton, D. B., Schneider, H. P. G., and McCann, S. M. (1970). Endocrinology 87, 323. Crighton, D. B., Hartley, B. M., and Lamming, G. E. (1973). J. Endocrinol. 58, 377. Croix, D. (1976). C.R. Hebd. Seances Acad. Sci. 282 949. Croix, D. (1977). J. Reprod. Fertil. 51, 149. Croix, D., and Franchimont, P. (1975). Neuroendocrinology 19, 1. Cunie, R. L., Warberg, J., Folkers, K., Bowers, C. Y., Paulsen, E., Mulder, J., and Paulsen, F. (1977). Eiochem. Eiophys. Res. Commun. 79, 1207. de Reviers, M., and Dubois, M. P. (1974). Horm. Metab. Res. 6, 94 (abstr.). Doerr-Schott, J., and Dubois, M. P. (1975a). C . R . Hebd. Seances Acad. Sci. 278, 1397. Doerr-Schott, J., and Dubois, M. P. (1975b). C . R . Hebd. Seances Acad. Sci. 280, 1285. Doerr-Schott, J., and Dubois, M. P. (1976). Cell Tissue Res. 172, 477. Doerr-Schott, J., Clauss, R. O., and Dubois, M. P. (1978). C . R . Hebd. Seances Acad. Sci. 286, 477. Dubois, M. P. (1973). J . Reprod. Fertil. 35, 595. Dubois, M. P. (1976a). Ann. Histochim. 21, 269. Dubois, M. P. (1976b). Ann. Eiol. Anim., Biochim., Biophys. 16, 177. Dubois, M. P., and Renoux, G. (1971. Ann. Inst. Pasteur, Paris 121, 711 (abstr.). Dubois, M. P., Barry, J., and Leonardelli, J. (1974). C . R . Hebd. Seances Acad. Sci. 279, 1899. Duvernoy, H.,and Koritke, J. G. (1961). C.R. Assoc. Anat. 112, 294. Duvernoy, H., Koritke, J. G., and Monnier, G. (1969). Z . Zellforsch. Mikrosk. Anat. 102, 49. Duvernoy, H., Koritke, J. G., and Monnier, G.(1971). J . Neuro-Visc. Relar. 32, 111. Dyer, G., and Dyball, R. E. J. (1974). Nuture (London) 252,486. Estes, K. S . , Padmanabhan, V., and Convey, E. M. (1977). Biol. Reprod. 17, 706. Flerk6, B. (1970). In “The Hypothalamus” (L. Martini, M. Motta, and F. Fraschini, eds.), p. 351. Academic Press, New York. Flerk6, B., Setalo, G., Vigh, A., Arimura, A., and Schally, A. V. (1978). Brain-Endocr. Interact. Neural Horm. Reprod. Int. Symp., 3rd, I977 p. 108. Gellert, R. J., Bass, E., Jacobs, C., Smith, R., and Ganong, W. F. (1964). Endocrinology 74,664.

IMMUNOHISTOCHEMISTRY O F LH-RH-PRODUCING NEURONS

217

Gilmore, D. P., Dobbie, H. C., McNeilly, A. S., and Mortimer, C. H. (1978). J. Reprod. Fertil. 52, 355. Goldman, M. (1975). J. Exp. Med. 105, 557. Goldsmith, P. C., and Ganong, W. F. (1975). Brain Res. 97, 181. Goos, H. J. T., and Murathanoglu, 0. (1977). Cell Tissue Res. 181, 163. Goos, H. 3. T., Ligtenberg, P. J. M., and Van Oordt, P. G. W. J. (1976). Cell Tissue Res. 168,325. Gorski, R. A. (1977). In “Hypothalamic Peptide Hormones and Pituitary Regulation” (J. C. Porter, ed.), p. 319. Plenum, New York. Green, J. D. (1969). In “The Hypothalamus” (W. Haymaker, E. Anderson, and W. J. H. Nauta, eds.), p. 276. Thomas, Springfield, Illinois. Gross, D. S. (1976). Endocrinology 98, 1048. Gross, D. S., and Baker, B. L. (1977). Am. J . Anar. 148, 195. Gross, S. J., Campbell, 0. H., and Weetall, H. H. (1968). Immunochemisfry 5, 55. Habeeb, A. F., and Hiramoto, S. A. (1968). Arch. Biochem. Biophys. 126, 16. Halasz, B., Pupp, L., and Uhlarik, S. (1962). J. Endocrinol. 25, 147. Harris, G. W. (1955). In “Neural Control of the Pituitary Gland.” Arnold, London. Harris, G. W. (1960). In “Control of Ovulation” (C. A. Villee, ed.), p. 56. Pergamon, Oxford. Heimer, L., and Larsson, K. (1967). Brain Res. 3, 248. Hofer, H. (1958). Verh. Dtsch. 2001.Ges., Frankfurt an Main p. 201. Hoffman, G. E. (1976). Anat. Rec. 184, 429. Hoffman, G. E., Moynihan, J. A,, and Knigge, K . M. (1976). Neurosci. Abstr. (SOC.Neurosci.) 2, 673. Hoffman, G. E., Knigge, K. M., and Joseph, S. A. (1977). Neurosci. Abstr. (SOC.Neurosci.) (in press). Hoffman, G. E., Melnyk, V., Hayes, T . , Bennett-Clarke, C., and Fowler, E. (1978a). BrainEndocr. Neural Horm. Reprod. Inr. Symp., 3rd, 1977 p. 67. Hoffman, G. E., Knigge, K. M., Moynihan, J. A,, Melnyk, K., and Arimura, A. (1978b). Neuroscience 3, 219. Hokfelt, T., Elde, R., Fuxe, K., Johansson, 0..Ljungdahl, A , , Goldstein, M., Luft, R., Efendic, S., Nilsson, G., Terenius, L., Ganten, D., Jeffcoate, S. L., Rehfeld, J., Said, S., Perez de la Mora, M., Possani, L., Tapia, R., Teran, L., and Palacios, R. (1978). In “The Hypothalamus” ( S . Reichlin, R. J. Baldessarini, and J. B. Martin, eds.), p. 69. Raven, New York. Ibata, Y., and Watanabe, K. (1977). Arch. Histol. Jpn. 40, 303. Ibata, Y . , Watanabe, K., Kimura, H., Sin, S., and Imagawa, K. (1978). Endocrinol. Jpn. 25, 141. Jackson, I. M. D. (1978). In “The Hypothalamus” ( S . Reichlin, R. J. Baldessarini, and J. B. Martin, eds.), pp. 217-231. Raven, New York. Kalra, S. P. (1976). Endocrinology 99, 101. Kalra, S. P., Krulich, L., and McCann, S. M. (1973). Neuroendocrinology 12, 321. Kalra, S. P., Kalra, P. S., and Mitchell, E. 0. (1977). Endocrinology 100, 201. Kami, K., Makino, T., Shiina, M., and Wada, M. (1977). Okajimas Folia Anar. Jpn. 54, 259. Kawaoi, H., and Nakane, P. K. (1973). Fed. Proc., Fed. Am. SOC. Exp. Biol. 32, 840 (abstr.). Kelly, M. J., and Moss, R. L. (1976). Neuropharmacology 15, 325. Kennedy, J . H., Kricka, L. S., and Wilding, P. (1976). Clin. Chim. Acta 70, 1. King, J. C., and Gerall, A. A. (1976). J . Hisrochem. Cyfochem. 24, 829. King, J. C., Parson, J. A,, Erlandsen, S. L., and Williams, T. H. (1974). Cell Tissue Res. 153,211. King, J. C., Elkind, K. E., Gerall, A. A,, and Millar, R. P. (1978). Brain-Endocr. Interact., Neural Horm. Reprod., Inf. Symp., 3rd, 1977 p. 97. Kizer, J. S., Palkovits, M., and Brownstein, M. J. (1976). Endocrinology 98, 311. Knigge, K. M., Hoffman, G., Scott, D. E., and Sladek, J. R. (1977). Brain Res. 120, 393.

218

JULlEN BARRY

Knigge, K. M., Joseph, S. A,, and Hoffman, G. E. (1978). In “The Hypothalamus” ( S . Reichlin, R. J. Baldessarini, and J. B. Martin, eds.), pp. 49-67. Raven, New York. Kobayashi, R. M., Lu, K. H., Moore, R. Y.,and Yen, S. S. C. (1978). Endocrinology 102, 98. Koch, Y., Wilchek, M., Chobsieng, P., Zor, U., and Lindner, H. R. (1973). Biochem. Biophys. Res. Commun. 55, 616. Kordon, C., Kerdelhue, B., Pattou, E., and Jutisz, M. (1974). Proc. SOC. Exp. Biol.Med. 147, 122. Kozlowski, G. P., and Hostetter, G. (1978). Brain-Endocr. Interact., Reprod.. Inr. Symp.. 3rd. 1 9 7 7 ~ 138. . Kozlowski, G . P., Brownfield, M. S., and Hostetter, G. (1978). In “Current Studies of Hypothalamic Function” (K. Lederis and W. Veale, eds.). Karger, Basel (in press). Krisch, B. (1978). Cell Tissue Res. 186, 135. Le Beux, Y. J . (1971). Z. Zellforsch. Mikrosk. Anar. 114, 404. Le Beux, Y. J. (1972). Z . Zellforsch. Mikrosk. Anar. 127, 439. Le Beux, Y. J., Langelier, P., and Poirier, L. J. (1971). Z . Zellforsch. Mikrosk. Anar. 118, 147. Leonardelli, J. (1976). C . R . Seances Soc. Biol. Ses Fil. 170, 602. Leonardelli, J. (1977). Ann. Endocrinol. 38, 235. Leonardelli, J., and Dubois, M. P. (1974). Ann. Endocrinol. 35, 639. Leonardelli, J., and Dubois, M. P. (1976). In “Cellular and Molecular Bases of Neuroendocrine Processes” (E. Endroczi, ed.), pp. 89-1 10. Akademiai Kiadd, Budapest. Leonardelli, J., and Poulain, P. (1977). Brain Res. 124, 538. Leonardelli, J., Barry, J., and Dubois, M. P. (1973a). C.R. Hebd. Seances Acad. Sci. 276,2043. Leonardelli, J., Hermand, E., and Tramu, G. (1973b). C . R . Seances Soc. Biol. Ses Fil. 167, 1815. Leonardelli, J . , Dubois, M. P., and Poulain, P. (1974). Neuroendocrinology 15, 69. Leonardelli, J., Tramu, G., and Hermand, E. (1975). LiNe Med. 20, 505. Leonardelli, J., Croix, D., Tramu, G., and Hermand, E. (1977). Ann. B i d . Anim. Biochim., Biophys. 17, 403. Leveque, T. F. (1972). Brain-Endocr. Inreract., Median Eminence: Srrucr. Funcr., Inr. Symp., 1971 p. 298. Likhite, V. (1967). Methods Immunol. Immunochem. 1, 150-151. McCann, S. M., Taleisnik, S., and Friedman, H. M. (1960). Proc. Soc. Exp. Biol.Med. 104,432. McCann, S . M., Ramirez, V. D., and Igarashi, M. (1964). Metab., Clin, Exp. 13, 1177. McNeill, T. H., Kozlowski, G. P., Abel, J. H., Jr., and Zimmerman, E. A. (1976). Endocrinology 99, 1323. Matsumoto, A., and Arai, Y. (1978). Exp. Neurol. 59, 404. Matsuo, H., Baba, Y., Nair, R. M. G., Arimura, A,, and Schally, A. V. (1971a). Biochem. Biophys. Res. Commun. 43, 1334. Matsuo, H., Arimura, A., Nair, R. M. G., and Schally, A. V. (1971b). Biochem. Biophys. Res. Commun. 45, 822. Mazzuca, M. (1977). Neurosci. Lett. 5, 123. Mazzuca, M., and Dubois, M. P. (1974). 1. Hisrochem. Cytochem. 22, 993. Mess, B., Zanisi, M., and Tima, L. (1970). In “The Hypothalamus” (L. Martini, M. Motta, and F. Fraschini, eds.), p. 259. Academic Press, New York. Meyer, M. H., Masken, J. F., Nett, T. M., and Niswender, G. D. (1974). Neuroendocrinology 15, 32. Millar, R. P., Aehnelt, C., and Rossier, G. (1977). Biochern. Biophys. Res. Commun. 74, 720. Mittler, J. C., and Meites, J. (1964). Proc. Soc. Exp. Biol. Med. 117, 309. Monahan, M., Rivier, J., Burgus, R., Amoss, M., Blackwell, R., Vale, W., and Guillemin, R. (1971). C . R . Hebd. Seances Acad. Sci. 273, 508. Moss, R. L. (1978). In “Psychopharmacology: A Generation of Progress” (M. A. Lipson, ed.), pp. 431-440. Raven, New York.

1MMUNOHlSTOCHEMlSTRY OF LH-RH-PRODUCING NEURONS

219

Moss, R. L., and Foreman, M. M. (1976). Neuroendocrinology 20, 176. Moss, R. L., and McCann, S. M. (1973). Science 181, 177. Moss, R. L., Dudley, C. A,, and Kelly, M. J. (1978). Neuropharmacology 17, 87. Naik, D. V. (1975a). Cell Tissue Res. 157, 423. Naik, D. V. (1975b). Cell Tissue Res. 157, 437. Naik, D. V. (1976). Cell Tissue Res. 173, 143. Nakane, P. K., and Pierce, G.B., Jr. (1967). J. Cell Biol. 33, 307. Okon, E., and Koch, Y. (1976). Nature (London) 263, 345. Okon, E., and Koch, Y. (1977). Nature (London) 268, 445. Page, R. D., and Bergland, R. M. (1977). Am. J. Anar. 148, 345. Palkovits, M., Arimura, A . , Brownstein, M., Schally, A. V., and Saavedra, J. (1974). Endocrinology 95, 554. Palkovits, M., Mezey, E., Ambach, G . , and Kivovics, P. (1978). Brain-Endocr. Interact., Neural Horm. Reprod.. Int. Symp., 3rd, 1977 p. 302. Paulin, C., Dubois, M. P., Barry, J., and Dubois, P. M. (1977). Cell Tissue Res. 182, 341. paull, W. K. (1978). Brain-Endocr. Interact.. Neural Horm. Reprod., Int. Symp., 3rd, 1977 p. 16. Pelletier, G.(1976). In “Hypothalamus and Endocrine Functions” (F.Labrie, J. Meites, and G. Pelletier, eds.), p. 433. Plenum, New York. Pelletier, G . , Labrie, F., Puviani, R., Arimura, A,, and Schally, A. V. (1974). Endocrinology 95, 314. Pelletier, G.,Leclerc, R., and Dube, D. (1976). J . Histochem. Cytochem. 247, 864. Pelletier, G.,Leclerc, R., Dube, D., Arimura, A,, and Schally, A. V. (1977). Neurosci. Left. 4,27. Petrusz, P., Sar, M., Ordronneau, P., and Dimeo, P. (1976). J . Histochem. Cytochem. 24, 1110. Petrusz, P., Sar, M., Ordronneau, P., and Dimeo, P. (1977). J . Hisrochem. Cytochem. 25, 390. Pfaff, D. W. (1973). Science 182, 1148. Pfaff, D. W. (1978). In “The Hypothalamus” (S. Reichlin, J. B. Baldessarini, and J. B. Martin, eds.), pp. 245-253. Raven, New York. Picard, G.E., and Silverman, A. J. (1976). Neurosci. Abstr. (Soc. Neurosci.) 2, 679. Poulain, P. (1977). Bruin Res. 137, 150. Poulain, P., and Carette, B. (1977). Brain Res. 137, 154. Poulain, P.,and Partouche, C . (1973). C . R . Hebd. Seances Acud. Sci. 277, 737. Ramirez, V. D., and Sawyer, C. H. (1965). Endocrinology 76, 282. Renaud, L. P. (1977a). J. Physiol. (London) 264, 541. Renaud, L. P. (1977b). BrainRes. 121, 201. Renaud, L. P. (1978). In “The Hypothalamus” (S. Reichlin, J. Baldessarini, and J. B. Martin, eds.), p. 268. Raven, New York. Renaud, L. P., Martin, J. B., and Brazeau, P. (1975). Nature (London) 255, 233. Renaud, L. P., Martin, J. B., and Brazeau, P. (1976). Pharmacol., Biochem. Behuv. 5, Suppl. 1, 171. Richoux, J. P., and Dubois, M. P. (1976). C.R. Seances SOC.Biol. Ses Fil. 170, 860. Rohlich, P., and Wenger, T. (1969). Z. Zellforsch. Mikrosk. Anar. 102, 483. Sakuma, Y., and Kawakami, M. (1976). In “Cellular and Molecular Bases of Neuroendocrine Processes” (E. Endroczi, ed.), p. 291. Akadkmiai Kaido, Budapest. Schally, A. V., Nair, R. M. G.,Redding, T. W., and Arimura, A. (1971). J. Biol. Chem. 246, 7230. Schick, A. F., and Singer, S. J. (1961). J. Biophys. Biochcm. Cytol. 9, 519. Schuurs, A. H. W. M., and Van Weenen, B. K. (1977). Clin. Chim. Acta 81, 1. SCtalo, G.,Vigh, S . , Schally, A. V., Arimura, A,, and Flerko, B. (1975). Endocrinology 96, 135. Setalo, G., Vigh, S., Schally, A. V., Arimura, A,, and Flerk6, B. (1976a). Bruin Res. 103, 597.

220

JULIEN BARRY

SCtalo, G . , Vigh, S., Schally, A. V., Arimura, A., and Flerko, B. (1976b). Acra Biol. Acad. Sci. Hung. 27, 75. Setalo, G., Flerko, B., Vigh, S., Arimura, A,, and Schally, A. V. (1978). In “Neurosecretion. Neurosecretion and Neuroendocrine Activity” ( W . Bargmann, A. Oksche, and A. Polenov, eds.), pp. 287-289. Springer-Verlag, Berlin and New York. Sharp, P. J., Hasse, E., and Fraser, H. M. (1975). Cell Tissue Res. 162, 83. Shin, S. H., and Howitt, C. (1977). Neuroendocrinology 24, 14. Sievertsson, H., Chang, J. K., Cunie, B. L., Bogentoft, C., Folkers, K., and Bowers, C. Y. (1971). Fed. Proc. Fed. Am. Soc. Exp. Biol. 30, 1193. Siler-Khodr, T., and Khodr, G. (1978). Am. J. Obsret. Gynecol. 130, 795. Silverman, A, J., (1976). Endocrinology 99, 30. Silverman, A. J., and Desnoyers, P. (1976). Cell Tissue Res. 169, 157. Silverman, A. J., and Krey, L. C. (1977). Endocrinology 100, Suppl. 261 (abstr.). Silverman, A. J., and Zimmerman, E. A. (1978). Brain-Endocr. Interact., Neural Horm. Reprod., Int. Symp., 3rd. 1977 p. 83. Silverman, A. J., Antunes, J. L., Ferin, M., and Zimmerman, E. A. (1977). Endocrinology 101, 134. Somentino, S . , Jr., and Sundberg, D. K. (1975). Neuroendocrinology 17, 274. Sternberger, L. A. (1973). In “Electron Microscopy of Enzymes” (M. A. Hayat, ed.), Vol. 1, p. 150. Van Nostrand-Reinhold, Princeton, New Jersey. Sternberger, L. A. (1974). “Immunochemistry. ” Prentice-Hall, Englewood Cliffs, New Jersey. Sternberger, L. A. (1977). In “Peptides in Neurobiology” (H. Gainer, ed.), pp. 61-96. Plenum, New York. Sternberger, L. A., and Hoffman, G. E. (1978). Neuroendocrinology 25, 111. Sternberger, L. A., Hardy, P. H., Cuculis, J. J., and Meyer, H. G. (1970). J. Hisrochem. Cytochem. 18, 315. Sternberger, L. A., Petrali, J. P., Joseph, S. A., Meyer, H. G., and Mills, K. R. (1978). Endocrinology 102, 63. Stumpf, W. E. (1970). Am. J . Anat. 129, 207. Styne, D. M., Goldsmith, P. C., Burstein, S. R., Kaplan, S. L., and Brumbach, M. M. (1977). Endocrinology 101, 1099. Swaab, D. F., Pool, C. W., and Van Leeuwen, F. W. (1977). J . Histochem. Cytochem. 25, 388. Szentigothai, J . (1964). Prog. Brain Res. 5 , 135. Szentigothai, J., Flerko, B., Mess, B., and Halasz, B. (1962). In “Hypothalamic Control of the Anterior Pituitary: An Experimental Morphological Study,” p. 76. Akidemiai Kiadb, Budapest. Toran-Allerand, C. D. (1978). Brain Res. 149, 257. Tramu, G., Leonardelli, J., and Dubois, M. P. (1977). Neurosci. Lett. 6 , 305. Uemura, H., Asai, T., Nozaki, M., and Kobayashi, H. (1975). Cell. Tissue Res. 160, 443. Vigh, B., and Vigh-Teichmann, I. (1973). Z. Zellforsch. Mikrosk. Anat. 144, 139. Weindl, A., and Joynt, R. J. (1972). Brain-Endocr. Interact., Median Eminence: Struct. Funct., Int. Symp., 1971 p. 281. Weindl, A., and Schinko, I. (1974). In “Neurosecretion. The Final Neuroendocrine Pathway” (F. Knowles and L. Vollrath, eds.), p. 327. Springer-Verlag, Berlin and New York. Weindl, A., and Sofroniew, M. V. (1978). Brain-Endocr. Interact., Neural Horm. Reprod., Int. Symp., 3rd. 1977 p. 117. Weindl, A., Sofroniew, M. V., and Schinko, I. (1978). In “Neurosecretion and Neuroendocrine Activity” (W. Bargmann, K. Oksche, and A. Polenov, eds.), p. 312. Springer-Verlag, Berlin and New York. Weiner, R. I., Pattou, E., Kerdelhue, B., and Kordon, C. (1975). Endocrinology 97, 1597. Weir, D. W., ed. (1978). “Handbook of Experimental Immunology,” 3rd ed.Blackwell, Oxford.

IMMUNOHISTOCHEMISTRY O F LH-RH-PRODUCING NEURONS

22 1

Weller, T. H., and Coons, A. H.(1954). Proc. SOC.Exp. Eiol. Med. 86, 789. Wenger, T. (1976). Neurosci. Lerr. 3, 29. Wenger, T., and Toro, I. (1971). Acra Eiol. Acad. Sci. Hung. 22, 331. Wheaton, J. E., Krulich, L., and McCann, S. M. (1975). Endocrinology 97, 30. Worobec, R. B., Wallace, J. H., and Huggins, C. G. (1972). Imrnunochernisrry 9, 229. Zamboni,L., and De Martino, C. (1967). J . Cell Eiol. 35, 148A (abstr.). Zimmerman, E. A,, and Antunes, J. L. (1976). J . Histochem. Cytochem. 24, 807. Zimmerman, E. A , , Hsu, K. C., Ferin, M., and Kozlowski, G. P. (1974). Endocrinology 5, 1.

This Page Intentionally Left Blank

INTERNATIONAL REVIEW OF CYTOLQGY. VOL. 60

Cell Reparation of Non-DNA Injury V. YA. ALEXANDROV Laboratory of Cytoecology and Cytophysiology, Komarov Botanical Institute, USSR Academy of Sciences, Leningrad, USSR I. Introduction . . . . . . . . . . . . . . . . . . . . 11. Recovery during the Action of Injurious Agents (Reparatory Adaptation) . . . . . . . . . . . . . . . . . . . . III. Repair after Elimination of Injurious Agent . . . . . . . IV. Repair of Thermal Injuries . . . . . . . . . . . . . A. Characteristic of Cell Thermal Injury . . . . . . . . B. Repair of Separate Functions and the Cell as a Whole . . C. RepairZone . . . . . . . . . . . . . . . . . D. Recovery Speed . . . . . . . . . . . . . . . . E. Conditions Affecting Repair of Heat Injury . . . . . . F. Stimulation of Cell Reparability . . . . . . . . . . V. Mechanism of Heat Injurious Action . . . . . . . . . . VI . Resynthesis or Reactivation? . . . . . . . . . . . . . VII . Some Evidence of Protein Renativation . . . . . . . . . VIII. What Happens in the Cell? . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . .

223

.

. .

. .

. .

.

. . . .

225 229 230 232 232 235 240 243 246 25 2 257 259 263 264

I. Introduction Structural and functional integrity of cells is in a considerable measure maintained by weak chemical bonds. Regulation of the activity of enzymes is commonly accomplished through conformational changes caused by rearrangements of weak bonds. The weak bonds are susceptible to fluctuations of various physical and chemical factors, which do not exceed the physiological boundary. Therefore, the fluctuations of the external and internal media, when altering the activity of enzymes, may cause considerable changes in the sphere of covalent bonds, the break and formation of which are catalyzed by enzymes. The enzymes act as mediators between the spheres of weak interactions and fast covalent bonds. This endows cells with the ability to respond even to the slightest of changes in environmental factors by physiological reactions, which comprise chemical processes with the participation of covalent bonds. The integrity of cells is maintained on condition of incessant reversible changes in the state of cell components. Furthermore, most of the normal physiological functions of the cell 223

Copyright 0 1979 by Academic Press, Inc. All rights of repduction in any form reserved. ISBN 0-12-364360-0

224

V. YA. ALEXANDROV

are cyclic. Thus the ability to restore the original state of structures and functions after any deviation is an inherent property of the living cell. The ability of the cell to restore its original state is most vividly expressed at the action of injurious agents putting the state of cell components outside the physiological norm. This ability for recovery is an essential element determining the resistance of the cell to many injurious agents. The investigation of the reparatory activity of injured cells enables us to detect (in an exaggerated form) some processes that may play the primary role during normal cell functioning. At the same time, the progress in this field is quite one-sided. Thousands of publications are devoted to the subject of repair of DNA damage induced by radiation and various genetically active chemical agents. Great discoveries have been made in this field. It is found that after various kinds of damage the recovery of DNA is brought about (1) via photoreactivation and (2) via different types of dark reactivation. Dozens of enzymes are involved in these processes, the investigation of which has widened significantly our knowledge of mechanisms protecting the genome and controlling genotypic variability during the cell’s everyday life. The results obtained are of great practical importance for medicine and agriculture. An extensive study of this side of cell reparatory activity so much relegated to the background other aspects of the recovery problem that a great number of authors do not even mention DNA when using such terms as “repair,” “reactivation,” or “recovery” in the titles of their books and papers. Not infrequently, however, many of the agents that inflict reversible injuries on cells choose as an application site not the DNA but other components of the cell, protoplasmic proteins among them. Next to nothing is known about cell recovery after the action of such agents. It is also unknown whether the cell possesses special mechanisms promoting reactivation of proteins following the example of DNA mending “repair shops.” The purpose of the present review is to provoke interest in this side of cell reparatory activity since the author believes that elaboration of the problem will contribute considerably to our knowledge of cell life. The appeal is timely for, despite the fact that investigators studying various aspects of the cell are often confronted by phenomena of repair not related to DNA breaks, publications devoted to this particular question are still unjustifiably scarce. In this review special attention is paid to the recovery of injuries induced by heating. The choice is dictated not only by the author’s experience in this field of research but also by the essential role the recovery of the altered state of protoplasmic proteins plays after such injury. Diverse forms of cell repair may be conditionally divided into two categories: (1) recovery of lost parts of the cell body-regenerations; and (2) elimination of injuries not associated with obvious destruction of some parts of the cell body. The second form of repair may be called reactivation. Many cells (Protozoa, neurons) are capable of restoring, through regeneration, large regions of the

CELL REPARATION OF NON-DNA INJURY

225

body, but consideration of this most inquisitive question exceeds the bounds of this paper. Repair of reproducing cell populations may be accomplished either by restoration of the original state of its cells or by compensation of its composition due to propagation of more stable cells, which were less affected or suffered no damage at all. Confusion of these two reasons may give rise to controversy (see Heinmets et al., 1954; Hurwitz et al., 1957; Iandolo and Ordal, 1966). Henceforth we shall dwell only on cases of cell recovery after injury.

11. Recovery during the Action of Injurious Agents (Reparatory Adaptation) At appropriate doses of an injurious agent reactivation occurs both after removal of the agent and during its application. In some instances, cell recovery in the presence of poison is accounted for by the fact that, by some way or the other, the cell starts to decrease the concentration of poison in the cytoplasm. This may be attained by retention of the influx of poison into the cell, by activation of its efflux, and by its chemical or physical detoxication. If this be the case, repair in the presence of poison in the medium is practically similar to that occurring after its elimination. In certain instances, however, repair is found to proceed in cells despite the presence in the protoplasm of an injurious agent in the active form. The easiest way to demonstrate this is to trace cell repair during prolonged sublethal heating. Below are given observations of epidermal cells in a piece of the leaf sheath of the reed Phragmites communis immersed in water maintained at a temperature of 41°C (Alexandrov, 1956). The protoplasmic streaming seen in all cells before heating serves as an index of the cell state. Time of heating (hours)

Observation

13 17 24 48

Streaming in many cells Streaming in single cells No Streaming Streaming in a number of cells

If the action of heating is more durable the protoplasmic streaming is finally inhibited. It is obvious that cells in a piece of tissue immersed in heated water are not able to neutralize heating in any of the above ways. At prolonged heating the streaming is found to resume also in the epidermal cells of Campanula persicifolia leaves (Alexandrov, 1964). Resumption of growth was observed on Rhizopus nigricans mycelium (Preobrazhenskaya and Shnoll, 1965). Favard (1963) recorded normalization of the state of mitochondria after injury at con-

226

V . Y A . ALEXANDROV

tinuous heating of pieces of Drosera leaves. Recovery of nerve conductivity in frog at a temperature that cuts it off was described by Thomer (1919). Recovery during continuous application of an injurious agent takes place only when the intensity of the latter is not very high. Initiation of repair can be easily detected by a curve describing the dependence between the time of onset of injury and the agent intensity. The semilogarithmic curve in Fig. 1 indicates the persistence of protoplasmic streaming in epidermal cells of C. persicifolia leaf as a function of the heating temperature. When pieces of the leaf are placed in water previously heated in the range of 46"-42"C the streaming is inhibited gradually until it stops completely without any sign of repair. The lower the temperature the longer the time required to stop the protoplasmic streaming. All three points min

640

t \

320

I60

00

I I \ I I I \

I

40

20

10

5

2.5

41

42

43

44

45

-c

46 OC

FIG. 1. Heat resistance of epidermal cells of rosette leaves of Campanula persicifolia. Abscissa: temperature of heating; ordinate: duration of protoplasmic streaming in minutes, logarithmic scale. (From Alexandrov, 1964.)

227

CELL REPARATION OF NON-DNA INJURY

lie on one straight line. With a further decrease in temperature, however, the curve shows a sharp break upward. The protoplasmic streaming lasts longer than would be expected from extrapolation of the right straight-line portion of the curve toward a lower temperature (broken line). In this very region, heating at 41°C permits observation of the first phase of inhibition, after which the streaming nearly regains its normal rate. The secondary final cessation of protoplasmic streaming occurs only after a long time. At even weaker heating the first phase of inhibition is either not noticeable or absent. The break in the intensity-effect curve demonstrates a range of temperatures triggering some mechanisms of cell resistance to the action of injurious agents. Curves with similar breaks caused by heating were obtained for cells of a great number of plants (Alexandrov, 1977) and also for some animal cells (Arronet, 1963; Skholl, 1965). Other agents yield analogous curves in cases when the cell is able to actively resist their injurious action in the range of weaker concentrations. Nasonov (1959) gives logarithmic curves with breaks indicating the dependence of time of onset of muscle nonexcitability on the agent concentration for 13 substances, among them salts, acids, alkalis, alcohols, esters, quinine, and chloralhydrate. Along such curves, we often find curves where breaks or bends in the range of lower concentration are directed in the opposite direction, i.e., downward (Fig. 2). These breaks show that in the range of lower intensities the cell is injured more rapidly than would be expected if the right straight-line portion of the curve

I

1000

I

1

29

,

31

I

33

35

37

I

I

I

I

39

41

43

45

47

-OC

FIG.2. Heat resistance of skeletal muscle of white rat. Abscissa: temperature of heating; ordinate: retention time of electric excitability in minutes, logarithmic scale. (From Skholl, 1963.)

228

V. Y A. ALEXANDROV

were extrapolated in the region of lower intensities. The breaks result from the prevalance of destructive aftereffects over reparatory efforts of the cell. In experiments in which tissue is detached from organ, of mainly warm-blooded animals, not only the primary injury inflicted by the basic agent but also the disturbances in the normal viability of cells due to detachment of a tissue from an organ can be a source of destructive aftereffects. The intensity-time curves with downward breaks may also be obtained upon the action of other agents. Initiation of repair during the action of an injurious agent by itself increases the resistance of the cell, i.e., leads to adaptation. This can be accompanied by stabilization of the protoplasmic components against the injurious agent. Therefore I suggest that the process of cell normalization during prolonged action of an agent should be called “reparatory adaptation” (Alexandrov, 1965). This term seems to be more adequate than the term “endogenous recovery” used by Wunderlich and Peyk (1969). Reparatory adaptation has been long described for many objects treated with various physical and chemical agents. For instance, the excitability of frog muscles placed into a Ringer solution with urea is quickly lost, but is restored in a few tens of minutes (Ilynskaya, 1960; Stabrovskaya, 1967). In a solution containing urethan or Verona1 the rate of protoplasmic streaming in the cells of Elodea leaves is first repressed but then increases again (Regnier and Bazin, 1940). In the ciliate Tetrahymena pyriformis, divisions stopped by cycloheximide, streptomidone (Roberts and Orias, 1974), colchicine, and colcymide (Wunderlich and Peyk, 1969) resume a few hours later in the presence of the same inhibitors. The transfer of Bacillus subtifis from 37” to 15°C destroys the ultrastructure of their mesosomes and the cell wall. Later, however, cells eliminate these defects and grow normally (Neale and Chapman, 1970). We could list many such examples, but even the above cases show that reparatory adaptation occurs in the tissue cells of plants and animals and in protozoa and bacteria in response to the action of the most diverse agents. When observing cells that recover their inhibited function in the presence of an injurious agent, one may gain the impression that cell reparatory ability becomes intensified while in progress. In this connection the evidence of Davies and Exworth (1973) is of interest. The authors poisoned a cell culture of rose with cycloheximide, an inhibitor of protein synthesis, and traced [ l 4 Clleucine incorporation into proteins. Immediately after the onset of the action of cycloheximide, at concentrations of lop3, M , the incorporation of the amino acid and was inhibited significantly. When concentrations of M and higher were used protein synthesis resumed 4 to 10 hours later and regained its normal rate by 20 hours, despite the presence of the poison. Paradoxical is the fact that at a lower concentration ( M) of cycloheximide the increase in the rate of synthesis on the first day was negligible. The authors believe that the toxic agent itself stimulates processes underlying cell function recovery. This requires a definite

CELL REPARATION OF NON-DNA INJURY

229

dose of poison beneath which the agent can injure the cell but is not able to spur its recovery. The nature of processes providing for the reparatory adaptation, as well as their stimulation mechanism, is unknown. Similarly, it is not clear whether in this instance we observed intensification of repair or an increase in the resistance of cell elements to poison. In T. pyriformis, cycloheximide at a concentration of 0.25 mg/ml cuts off division, but after a 3-hour lag the culture passes on to the logarithmic phase of reproduction. When the ciliates are first incubated in the presence of cycloheximide at a concentration of 0.25 ng/ml, which does not affect the rate of division, and then are transferred to a concentration of 0.25 mg/ml, the lag period decreases 3-fold (Roberts and Orias, 1974). In this case also it is hard to decide what is stimulated by cycloheximide at such a low concentration-adaptation or recovery processes? Stimulation of cell reparatory ability will be discussed in detail in Section IV,F. In some cases cell function resumes in the presence of an injurious agent due to initiation or intensification of a roundabout or vicarious metabolic path without recovery of the injured metabolic link [e.g., replacement of blocked oxidative phosphorylation for glycolysis (Errera, 196l)] . Consequently, reactivation of the cell does not necessarily imply reactivation of its injured components.

111. Repair after Elimination of Injurious Agent To study reparation not complicated by a simultaneous transition of cells to a higher resistance level, one must consider cell recovery after termination of the action of an injurious agent. The investigation of the deleterious effect of dissolved substances shows that their elimination from the medium does not indispensably involve cessation of their action inside the cell. The use of such physical factors as temperature, high hydrostatic pressure, centrifugation, etc. , provides great advantages in this respect. Furthermore, it enables us to avoid the complicated question of the extent of penetration of the agent across cell membranes . For analyzing processes that form the basis of repair it is essential to know what the agent is and how it injures the cell. The discovery of reparatory mechanisms eliminating DNA damage became possible owing to our knowledge of the nature of these disturbances. Unfortunately, with the exception of the specific inhibitors of metabolism, our information of primary injuries induced in the cell by other agents not aimed at DNA is rather scarce. In this context, it would be of interest to investigate the process of recovery after heating since we are a little better informed about the mechanism of cell thermal injury.

230

V . YA. ALEXANDROV

1V. Repair of Thermal Injuries Among the works concerned with the subject of repair of heat injuries the investigations performed on bacteria are predominant. They are in great measure dictated by practical uses of sterilization. Many aspects of the recovery problem, however, are easier to understand when nondividing tissue cells are used for the experiment. The first descriptions of recovery after heat injury were made by Sachs (1864) and Kuhne (1864). Sachs traced the reactivation of protoplasmic streaming stopped by heating in the cells of some plants. Kuhne gave a detailed description of a reversible thermal coagulation of the protoplasm for Amoeba dzfluens, Actinophrys lichhornii, and Plasmodium myxomycetes and for the cells of Tradescantia virginica. In all these, protoplasmic movement was stopped by heat coagulation. After cessation of heating the protoplasm resumed in its normal state and started moving. Gradually, new data on the recovery of various cell functions and properties that deviate from the norm as a result of heat treatment have been accumulated. Recovery of reproductive ability, restoration of the lost tolerance to some substances, resumption of the ability to grow on minimum medium, and reconstruction of injured ribosomes were demonstrated repeatedly on bacteria (Sogin and Ordal, 1967; Beuchat and Lechowich, 1968; Clark and Ordal, 1969; Baldy et al., 1970; Mukherjee and Bhattacharjee, 1970; Pierson et al., 1971, 1974; Tomlins and Ordal, 1971a,b; Gray et a l . , 1973; Iandolo, 1974). The recovery of the lost enzyme inducation was described by David (1976). Reversible cessation of protoplasmic streaming was most frequently studied in plants. Additionally, there are data on the resumption of chloroplast phototaxis (Lomagin et al., 1966, 1967) and photosynthesis, which were completely or partially inhibited by heating (Montfort et al., 1955, 1957; Lutova, 1962; Semikhatova et al., 1962; Semikhatova, 1964; Lange, 1965; Lomagin and Antropova, 1966; Bauer, 1970, 1972; Bauer et al., 1975; Benzioni and Itai, 1972, 1973, 1975; Egorova, 1975, 1976; Semikhatova and Egorova, 1976; Falyzova and Sagatov, 1977); the reversibility of chloroplast swelling induced by heat (Daniel1 et al., 1969) and of inhibition of the Hill reaction in chloroplasts (Rakhimov, 1977); elimination of injuries in mitochondria and nucleoli (Favard, 1963; Lomagin, 1978); reversibility of breaks in the ability of leaves to yield photoinduced lasting afterlurinescence (Alexandrov and Dzhanumov, 1972); and the recovery of disturbed permeability (Olejnikova, 1964; Benzioni and Itai, 1973). Some evidence of the reactivation of thermal injury has been obtained for animal cells. It was found that the excitability of certain muscles (Vinogradova and Dzhamusova, 1963) and nerves (Thorner, 1919) lost as a result of heating is regained and that ciliary movement is restored in ciliary epithelial cells (Arronet, 1963) and in Paramecium caudatum (Alexandrov, 1948). Moreover, disturbances in the ultrastructure of nucleoli and of cytoplasmic organelles were found to be eliminated and RNA synthesis was found to resume in the cells of tissue

23 1

CELL REPARATION OF NON-DNA INJURY

cultures (Simard and Bernard, 1967; Simard et al., 1969; Heine et al., 1971). In animal and plant cells heating, along with other injurious agents, leads to a decrease of dispersion in the protoplasm and to an enhancement of its viscosity. In a number of cases these changes are reversible (Lepeschkin, 1923; VanHerwerden, 1927; Nasonov and Alexandrov, 1940; Alexandrov, 1948). After cessation of heating, at a reversibly injurious dose, cells more or less rapidly resume their normal state. In some instances, however, the recovery is preceded by a period of increasing aggravation of their state (Semikhatova et al., 1962). In the leaves of the Sakhalin buckwheat, Polygonum sachalinense, immediately after cessation of a 15-minute heating at 41"C, photosynthesis is reduced by 35%. During the following 2 hours, however, photosynthesis decreases progressively, but only after this period of destructive aftereffect does the function begin to resume (Fig. 3, curve 1). In a similar experiment with maize leaves ( Zea mays) no destructive aftereffect was discovered (Fig. 3, curve 2), and photosynthesis started to reactivate immediately after removal of the leaves from the thermostat (Egorova, 1976). In Paramecium caudatum, movement slows down after 5 minutes of heating at about 38°C. After 1 hour at a normal temperature it stops to resume afterward in some individuals (Irlina's experiments). Belikov and Melechov (1975) demonstrated in bean leaves that, depending on the temperature and the length of heating, the recovery of disturbed photosynthesis may vary with time. Specifically, after an injury of moderate intensity, the time curve of photosynthetic recovery assumes a wavelike shape as it approaches the initial level. After heavier doses, an overshoot occurs prior to the

90 -

100

80

-

70 -

60

-

50-

40

-

30 -/ I

J

I

*

I

I

1

1

1

'

*

'

c--L--c

232

V. YA. ALEXANDROV

attainment of a new level. After severe injuries photosynthesis smoothly reaches a new but very low level. OF CELLTHERMAL INJURY A. CHARACTERISTIC

Compared to inhibitors of metabolism which switch off only some definite metabolic links, heat belongs to injurious agents of general action. Such division is to a certain extent conventional, since with increases in concentration many of the specific inhibitors hit additional targets. Furthermore, due to integrity of the cell system, the focus of primary injury involves in the process of destruction some cell components and metabolic chains which have not been directly affected by the specific inhibitor. This is well illustrated by cells poisoned with inhibitors of ATP synthesis. On the other hand, after weak heat treatment, only single metabolic reactions may switch off (see Alexandrov, 1977, and references therein). As the heating temperature increases, various cell functions are commonly injured in a definite order owing to variation in the thermostability of different protoplasmic components. Figure 4 illustrates the sequence of injury of various functions in Trudescantiu leaf cells as the temperature of a 5-minute heating is elevated. The left boundary of the band corresponds to the temperature at which the f i s t manifestations of damage of a given function begin to appear. The right boundary of the band indicates the temperature of heating efficient for complete suppression of the function concerned. As seen from the figure, protoplasmic streaming, phototaxis of chloroplasts, and photosynthesis are more susceptible to heating than respiration or selective permeability of protoplast identified via plasmolysis. Differences in the thermoresistance of these functions reach 15°C. Such correlations during short-term intensive heating are detected in cells of many higher plants. Relationships between the thermostabilities of different functions in animal cells and microorganisms have been less studied. Some observations show, however, that in animal cells various functions are inhibited more synchronously as the heating temperature is increased.

B. REPAIROF SEPARATE FUNCTIONS AND

THE

CELLAS

A

WHOLE

It is natural to expect that thermolabile functions may be more easily repressed reversibly than the more resistant ones since for inhibition the latter functions require a heating does that affects profoundly the cell system as a whole. In fact, short-term intensive heating makes it possible to arrest reversibly the protoplasmic streaming and to inhibit photosynthesis, phototaxis of chloroplasts (Fig. 4), and photoinduced lasting afterluminescence. This failed, however, to restore heat-inhibited respiration in plant cells. Only in rare cases is the cell ability for plasmolysis inhibited reversibly.

Manifestation of injury

All tissues

Uncoupling of oxidative phosphorylation Leakage of electrolytes Suppression of respiration

m

Dzhanumov. 1 Alexandrov and Alekseeva, 1912 Experiments of Derteva Alexandrov, 1955

FIG.4. Sequence of suppression of cellular function in Trndescantiafluminensis leaves with increasing temperature of 5-minute heating. Explanation in the text.

234

V. YA. ALEXANDROV

FIG.5 . Recovery of the ability for phototaxis in chloroplasts of Tradescantiupurninensis leaves after I minute heating at 51°C. Phototaxis was estimated by changes in light transmission when weak illumination was replaced for strong illumination (in percentage of the initial level). Curve 1: unheated leaves; curve 3: heated leaves. (From Lomagin et al., 1967.)

Recovery of a single function does not mean that the whole cell regains its normal original state. The cell is able to restore some of the functions while the others are inhibited irreversibly. After a 5-minute heating at 49°C photosynthesis in C. persicifolia leaves is inhibited irreversibly, but protoplasmic streaming resumes in these cells even after exposure to 51°C (Lutova, 1962). Pieces of the Tradescantia jluminensis leaf survive in water changed daily for 70 days. The temperature of the minimum 5-minute heating, which stops protoplasmic streaming in epidermal cells, is about 44"C, and complete recovery of streaming is observed after exposure to 48.5"C. As seen in Fig. 6, after heating in the temperature range 42" to 48.5"C, the survival time of cells decreases progressively, despite complete repair of the motor functions. Consequently, resumption of protoplasmic streaming proceeds in the background of other noneliminated disturbances. For quantitative evaluation of the cell's ability to restore broken functions two indices are used: (a) speed of recovery of a given function after definite degree of inhibition; (b) interval of intensity of an injurious agent within which inhibition of the function becomes reversible after a definite time of exposure. This characteristic

235

60 -

70

50-

40

-

30 -

20

-

10

42

44

46

48

50

52

c OC

FIG.6. Effect of temperature of 5-minute heating of Tradescanfiafluminensis leaves on cell survival time. Abscissa: temperature of heating; ordinate: duration of protoplasmic streaming in days (curve 1) and persistence of anthocyanin (curve 2). Horizontal broken line: duration of cell life in unheated leaves. (A) Zone of complete reactivation of protoplasmic streaming in all of the cells. (From Alexandrov, 1955 .)

is called repair zone. Using these criteria one may estimate various influences on the reparatory ability of cells and compare this ability in different cells.

C. REPAIRZONE The repair zone has been most thoroughly studied using heat-induced reversible stoppage of the protoplasmic streaming by way of illustration. The repair zone identified through recovery of protoplasmic streaming after 5 minutes of heating varies greatly in cells of different plants (Table I). It may differ considerably in cells of one species depending on the developmental stage, season, growth temperature, etc. In cells of both multicellular and unicellular plants this zone is commonly larger than in tissue cells of animals. Thus after a 5-minute heating of pieces from the palate of the frog Rana temporaria reversible stoppage of ciliary movement in cells of the ciliary epithelium occurs at 42.0"C.However, heating at 42.7"C stops it irreversibly. For ciliary cells of veligers of the mollusk Lacuna divaricata the repair zone is less than 3°C (Arronet, 1963). Only in rare cases is protoplasmic streaming arrested reversibly in Paramecium after a 5-minute heating. Therefore the repair zone is lacking for the heat-injured motor

236

V. YA. ALEXANDROV

TABLE I REPAIR ZONESFOR REACTIVATION OF DIFFERENT TYPES OF MOTION AFTER 5 MINUTES OF HEATING

Speices

Elodea densa Caiabrosa aquatica Allium cepa Zebrina pendula Tradescaniia fluminensis Campanula persicifolia Dactylis glomeraia Ficus radicans Chlorophytum elaium Phormidium auiumnale Oscillatoria ienuis Chlamydomonas eugametos

Type of motion Protoplasmic streaming in the cells of Leaf epidermis Leaf epidermis Bulb scale epidermis Leaf epidermis Leaf epidermis Leaf epidermis Leaf epidermis Leaf epidermis Leaf epidermis Oscillatory movement Oscillatory movement Cell movement

Repair zone ("C)

0.0 3.8 4.7 6.3 7.1 7.4 7.5 8.8 9.0 4.2 4.8 8.0

function (Irlina's experiments). At the same time, in the unicellular alga Chlamydomonas eugametos, a 6-minute heating in the range of 34" to 42°C reversibly stops the beating of flagella (Fig. 13D). It is not unlikely that a lesser repair zone in animal cells is accounted for by the fact that heat injury of different protoplasmic components proceeds more synchronously. Among plants, however, there occur cells that are incapable of restoring the protoplasmic streaming arrested by short-term heating. In Elodea densa even incomplete stoppage of the streaming is not reversible. Cells in which protoplasmic streaming is inhibited considerably but not completely as a result of a 5-minute heating are doomed to a rapid death. Instead of repair their develop destructive aftereffects leading to total coagulation of the protoplasm (Zavadskaya and Antropova, 1978). In accordance with this evidence, the thermostability curve for protoplasmic streaming in E. densa cells in the region of weaker heating bends downward and not upward as it does in plants with a well-expressed reparability (Fig. 7). Compared to E. densa cells, E . canadensis cells are less heat resistant, but unlike E . densa, they show some (although insignificant) reparability. In the case of a 5-minute heating the repair zone for protoplasmic streaming is 1°C. Accordingly, the heat resistance curve for protoplasmic streaming in E . canadensis differs notably from that in E . densa (Fig. 7). The extent of the repair zone depends on the length of heating time. Figures 8A-C present curves (abc) indicating time of cessation of protoplasmic streaming in Dactylis, Tradescantia, and Campanula cells after heating of different intensity. Horizontal lines demonstrate the extent of repair zones after heating of dif-

237

CELL REPARATION OF NON-DNA INJURY

I

L

37

38

39

40

41

42

43

44

45

46

OC

FIG. 7. Heat resistance of Elodea cells. Abscissa: temperature of 5-minute heating; ordinate: duration of protoplasmic streamining in minutes, logarithmic scale. (From Zavadskaya and Antropova, 1979.)

ferent duration. We see that, after some definite time of exposure, the longer the heating, and, accordingly, the lower the temperature at which protoplasmic streaming is arrested, the lesser the thermal range within which its cessation is reversible. Thus after 5 minutes of heating Trudescantia leaves (Fig. 8B) the repair zone is 7.1"C. After 1000 minutes of heating, protoplasmic streaming is stopped reversibly only within 1.4"C. In Moms alba leaves the repair zone is 5.O"C (46.O"-51.O0C) and 1.3"C (42.7"-44.OoC) after 5 and 640 minutes of heating, respectively (Shukhtina's experiments). This regularity was also observed in animal cells (Fig. 8D). What induces the decrease in the repair zone upon lengthening the time of heating? The semilogarithmic curves abc indicate the dependence of the onset of a repair zone on the duration of heating (Fig. 8). In all three plants the curves have bends we have already mentioned above. Owing to introduction of cell resistance mechanisms to regions of more prolonged and weaker heating, the beginning of the repair zones is shifted toward the range of higher temperatures (sections bc). Curves indicating the dependence between the duration of heating and the end of the repair zone (de) show no significant deviation, and, as a result, they draw together. This implies that during prolonged moderate heating thermoresistance of the function of protoplasmic streaming increases notably due to cell reparatory activity. In this case, protoplasmic streaming does not stop before all reparatory resources of the cell are essentially exhausted. During short-term

rnin I000 64C

min

640

32C

320

160

I60

BC

80

4c

4c

20

2c

10

IC

5

W N

OD

A . . . . . . . . . . . . . 41 42 43 44 45 46 47 40 49 50 51 52 53 54 O C

m

rnin

640 320 160

80 40 W N

W

20

10

5. 2

38 39 40 41 42 43 44 45 46 47 48 4 9 50 51 52'C

D. 28

30

32

L

34

36

38

40

"C

FIG. 8. Thermostability curves of protoplasmic streaming (abc and a,b,c,) and of reparatory function (de) of leaf epidermal cells (A-C) and of mollusk ciliary cells (D). Abscissa: temperature of 5-minute heating; ordinate: duration of protoplasmic streaming (A-C) or of cilia beating (D) for the curves abc, and minimum time of heat exposure inhibiting the ability of cells to reactivate heat-arrested protoplasmic streaming (A-C) or cilia beating (D) for the curves de. Time in minutes, logarithmic scale. Horizontal lines: repair zones of inhibition of motion. (A) Leaves of Dactylis glomerata; (B) leaves of Tradescanriajuminensis; (C) leaves of Cumpunulu persicifolio; abc and de: greenhouse plants; repair zones: dotted horizontal lines: alblcl and d,e,: out-of-door plants taken from under the snow; repair zones: broken horizontal lines. (D) Ciliary cells of veliger of mollusk Lacuna divuricutu. [(A-C) From Alexandrov et af.. 1963; (D) from Arronet, 1963.1

240

V. YA. ALEXANDROV

intensive heating thermolabile protoplasmic streaming ceases while a number of other thermostable structures and functions are still retained. As demonstrated by the experiment on Tradescanria, after long-term weak heating, the moment of cessation of protoplasmic streaming nearly coincides with the time of appearance of symptoms of injury of such thermostable functions as loss of plasmolytic capacity and anthocyanin efflux. During application of an injurious agent and after an appropriate decrease in its intensity destruction of various components of the plant cell proceeds more synchronously (Alexandrov er al., 1963). The fact that, after intensive short-term or weak prolonged exposure to heat, protoplasmic streaming occurs under quite different intracellular conditions is vividly illustrated by the findings of Gorban (1964a). She demonstrated in Zebrina pendula leaves that heating at 41" to 47°C for 5- to 80-minute periods arrests protoplasmic streaming without inducing accumulation of ammonium in tissue. At a temperature of 40°C protoplasmic streaming is stopped in lo00 minutes and is accompanied by considerable enhancement of the ammonium content in cells. This permits the essential conclusion that, if in case of different ratios between the time length and intensity of action of an injurious agent a given vital function is inhibited to an equal degree, the inhibition may occur in the background of varying injury of the cell as a whole. Isodoses by a given index are not necessarily isodoses for the cell as a whole. This assumption was supported by the experiments of Shukhtina performed in our laboratory on T. Juminensis. The investigator traced the time of recovery of protoplasmic streaming in leaf epidermal cells after minimum heating, which arrests protoplasmic streaming after 5 and 1000 minutes of exposure. At a 5-minute exposure to heat protoplasmic streaming stops at 45°C and resumes 40 minutes after the termination of heating. After a 1000-minute heating, streaming ceases at 38.6"C. In this instance, however, protoplasmic streaming resumes only 1500 minutes later. The repair zones for different indices may vary considerably. Experiments on parenchymatous cells of C. persicifolia leaves show that after a 5-minute exposure the repair zone of photosynthesis is narrower than that of protoplasmic streaming-2' and 7°C respectively (Lutova, 1962). The cells of E. densa are not able to resume the heat-inhibited protoplasmic streaming identified by replacement of spherosomes. However they are capable of restoring a function much more sensitive to heating, an ability to respond by chloroplasts movement to intense light (Zavadskaya and Antropova, 1979). D. RECOVERYSPEED Speed of recovery of protoplasmic streaming, as well as of other functions, depends on the depth of injury, the object, and the conditions discussed below. Figures 9, 10, and 13A-D (curves 1) show the dependence of the time of

CELL REPARATION OF NON-DNA INJURY

24 1

h

16C

I2C

80

1

39

41

I

43

t

I

45

:

"C

FIG.9 . Dependence of resumption time of protoplasmic streaming in epidermal cells of Commelinu ufricunu leaves on temperature of 30-minute heating arresting the motion. Abscissa: temperature of heating; ordinate: time required for complete reactivation of protoplasmic streaming in minutes. Curve 1: water-saturated leaves; curve 2: water-deficient leaves. (From Kappen and Lange, 1968.)

resumption of protoplasmic streaming on the heating of different intensity for different objects. The higher the heating temperature at a given time of exposure the lower the reparation rate. The latter may vary considerably for different objects. Thus after 5 minutes of heating at a temperature which exceeds by 3°C the minimum temperature at which it is arrested, protoplasmic streaming resumes 14 to 15 hours, 39 hours, and 25 minutes later in the cells of C. persicifolia andZ. pendula and in the plasmodium, Physarum polycephalum, respectively. Figure 10 illustrates the reactivation of two forms of movement in the plasmodium, P. polycephalurn. Curve 2 describes unordered protoplasmic streaming, and curve 1, shuttle movement. A 10-minute heating stops the first type of movement at 38°C and the second, at 37.5"C. Reactivation of movement after heating at temperatures exceeding these by 3"C, in the case of unordered movement, occurs in 25 minutes and, in the case of shuttle movement, in 40 minutes

242

V. YA. ALEXANDROV

min

4)

110.

90

-

70

-

50-

30

/

10-

/'

FIG. 10. Dependence of resumption time of two different forms of protoplasmic streaming in Physarum polycephalum arrested by 10-minute heating on temperature of heating. Abscissa: temperature of heating; ordinate: time required for reactivation of motion in minutes. (1) Shuttle-like motion; (2) disordered motion. (From Logamin, 1975.)

(Lomagin, 1975). The difference between the reactivation rates of protoplasmic streaming and photosynthesis is even greater in the same parenchymatous cells of the C. persicifolia leaf. Despite the fact that after 5 minutes of heating photosynthesis is stopped completely at 47"C, and protoplasmic streaming at 45°C photosynthesis returns to the control level only 4 days later while protoplasmic streaming, after the same heating, regains its normal level within 24 hours (Lutova, 1962). Thus we see that recovery of different functions proceeds asynchronously. As demonstrated by the findings of Gorban (1964b) the rate of reactivation of protoplasmic streaming in epidermal cells of Zebrina and Tradescantia leaves depends on the cell age. This rate is higher in growing cells than in young ones

243

CELL REPARATION OF NON-DNA INJURY

that have completed their growth, whereas it is higher in the latter than in older cells. E. CONDITIONS AFFECTING REPAIROF HEATINJURY A number of factors affect the rate and completeness of the recovery of function as well as the extent of the repair zone. The most essential factor is temperature, which plays an important role both during recovery and in the period preceding injury of the object. According to I. S. Gorban (unpublished data) the value for reactivation of protoplasmic streaming in Zebrina cells arrested by a 5-minute heating at 47" or 48°C was 1.6 to 1.7 during recovery in the temperature range 15" to 25°C. This question was specifically investigated by Barabalchuk and Kasapov (1978) on epidermal cells of the Tradescantia leaf (Fig. 11). As may be realized by observing the curves for recovery speed after a 5-minute heating of different intensity, in all cases studied reactivation of protoplasmic streaming was highest at a temperature of about 28°C. At 5°C protoplasmic streaming resumed only in part of the cells. The reparation temperature

el,,

I/h 0.4

0.3

0.2

0.I

5

10

15

20

25

30

35

.

40 'C

FIG. 1 1 . Dependence of speed of recovery of protoplasmic streaming in epidermal cells of Tradescantiafluminensis arrested by 5-minute heating on temperature of repair. Abscissa: temperature during repair; ordinate: recovery speed, I/time (hours), required for recovery of motion. Values above the curves are temperatures of 5-minute heating after which repair was observed. [According to the data of Barabalchuk and Kasapov (1978) most kindly presented at our disposal.]

244

V. YA. ALEXANDROV

influenced the extent of the repair zone, which reached its maximum at 23" to 30°C (Fig. 12). Similar data were also obtained for other organisms. A few hours after heat shock, the yeast Saccharomyces cerevisae maintained at 30°C regained its normal state. At O"C, no recovery was observed (Van Halteren, 1950). Heatinjured Staphylococcus was best reactivated at 37"C, which is the optimum temperature for its growth. At 3°C no repair was noted (Iandolo and Ordal, 1966). Heated spores of Penicillium expansurn were reactivated at 23"C, but not at 0°C. Reactivation did not occur under anaerobic conditions (Baldy et af., 1970). At 20" and 37°C Escherichia coli repaired heat injury but there was no reactivation at 0.5"C (Mukherjee and Bhattacharjee, 1970). At 3" to 4°C the leakage rate of 3sCl in the tobacco leaf elevated by a 2-minute heating did not reduce to the norm. At room temperature, the normal level was reached in 24 hours (Benzioni and Itai, 1973). The rather trivial dependence of recovery on the temperature at which it proceeds indicates that reparation involves metabolic processes that require temperatures close to the optimum ones for the given type of cells. According to Lomagin and Zheleznyak (1973) the plasmodium, P. polycephalum, eliminates heat injury most rapidly at pH 2 to 4. At pH 5 to 6

OC

t

8-

76 -

54-

32-

't I

a

I

I

a

I

1

I

5

10

15

20

25

I

30

I

35

-

I

40

L

OC

FIG. 12. Dependence of extent of repair zone of protoplasmic streaming in epidermal cells of Tradescanria fluminensis arrested by 5-minute heating on temperature of repair. Abscissa: temperature of repair; ordinate: extent of reparatory zone in degrees. (According to Barabalchuk and Kasapov, 1978.)

CELL REPARATION OF NON-DNA INJURY

245

reparation is delayed 1.5 times, and at pH 7, more than 2 times. The optimum pH value for the plasmodium is 3 to 4. Kappen and Lange (1968) revealed an opposite effect of water deficiency on the speed of recovery of protoplasmic streaming arrested by weak and more intensive heat treatment in Commelina africana leaves (Fig. 9). Rakhimov (1977) demonstrated that in the leaves of Kochia prosfafa and Carotoides eversmanniana water deficiency improves the recovery of Hill reaction inhibited by heating. Christophersen and Kaufmann (1952) managed to increase reactivation of the heat-injured yeast Torulopsis kefyr by immersing it for a time into glycocol and phenol solutions. There are contradictory data on the effect of light on the recovery of photosynthesis repressed by heating. According to the evidence of Benzioni and Itai (1973) for tobacco leaves and the findings of Semikhatova and Egorova (1976) for maize leaves, light accelerates the repair of heat-inhibited photosynthesis. In contrast, Lomagin and Antropova (1966) observed the inhibitory action of light on the recovery of photosynthesis and other cell functions stopped by heating. Kislyuk (1979) explains this descrepancy by the fact that light exerts an opposite effect on the recovery of the photosynthesis apparatus depending on the depth of its injury. She demonstrated on Tradescantia leaves that light speeds up repair when the photosynthesis is not completely inhibited. In cases in which photosynthesis is completely repressed, light causes photodynamic oxidation that aggravates the damage. As demonstrated by a series of experiments on bacteria, nutritive requirements increase in cells reversibly injured by heat or cold. Heated bacteria cannot grow on minimum media, which maintain the growth of control cells. Therefore, more complex media containing additional ingredients are required for heat-injured cells to reveal their reparability (Heinmets el al., 1954; Arpai, 1964; Iandalo and Ordal, 1966; Clark and Ordal, 1969; Dabbah et a l . , 1969). Strikingly opposite results were obtained by Gomez and his co-workers (Gomez and Sinskey, 1973, 1975; Gomez et al., 1973). In their experiments on Salmonella typhimurium a number of surviving cells, when placed in a rich medium (trypticase soy + yeast extract), were reduced drastically after a 15minute heating at 50°C. No sharp decrease was observed when the cells were placed in a minimal glucose salt medium. The authors showed that singlestranded DNA breaks arise in heated bacteria kept in a rich medium. Additionally, Woodcock and Grigg (1972) observed double-stranded DNA breaks in Escherichia coli. Immediately after heat treatment the breaks may be lacking. No breaks occurred in the case of minimum medium, but they appeared in a rich medium. When heated bacteria were incubated in a minimum medium or in distilled water for a few hours before transfer to a rich medium they lost their ability to form DNA breaks under the influence of the rich medium.

246

V . YA. ALEXANDROV

In any event, both experimental series indicate the important role of the nutrient medium composition in reactivation following heat injury. F. STIMULATION OF CELLREPARABILITY In our laboratory we succeeded in obtaining evidence that demonstrates the potentialities of a rather significant intensification of cell reparability . This may be attained in two different ways: first, increase of recovery speed and extent of the repair zone by repeated reparations; second, enhancement of the reparation capacity by increasing the temperature at which plants or leaves are kept before heat injury. In all the experiments presented in this section the process of reactivation was brought about at the same room temperature. 1. The First Series of Experiments The experiments of Gorban (1975) carried out on the leaves of C. persicifolia, Podophyllum peltatum, and Z . pendula showed that epidermal cells which resumed protoplasmic streaming arrested by heating reactivated more rapidly after repeated heat treatments. Furthermore, in this case the temperature zone within which heat treatment arrests protoplasmic streaming reversibly is extended. Figure 13A illustrates one of the experiments performed on C. persicifolia leaves. Curve 1 indicates the time of resumption of protoplasmic streaming in epidermal cells after a 5-minute heating at different temperatures. The beginning of the upper horizontal line above the abscissa indicates the minimum temperature at which 5-minute heating arrests protoplasmic streaming (43.1"C). The end of the horizontal line corresponds to the temperature of maximum heating after which the cell still retains its ability to reactivate protoplasmic streaming (49.5"C). The repair zone is 6.4"C. The higher the heating temperature the slower the repair and the more time required for resumption of the movement. Curve 2 indicates the same dependence for the leaves which once reactivated protoplasmic streaming arrested by a 5-minute heating at 45°C a day before exposure to different temperatures. As seen from curve 2, the repeated reparation differs from the primary one in three aspects. (i) After recovery of an injury inflicted by heating at 45"C, the heat resistance of cells increases. The minimum temperature of a 5-minute heating arresting protoplasmic streaming increases by 1.9"C (45.0"C). Such a phenomenon of heat hardening was described in detail in a review by Alexandrov et al. (1970). (ii) Injuries induced by temperatures exceeding by the same value the minimum temperature that arrests protoplasmic streaming are eliminated during the second reparation much more rapidly than during the fist. To illustrate this conclusion one must shift curve 2 so much to the left that the heat resistance of cells is enhanced after the first repair (in this case by 1.9"C). For example, while comparing curve 1 and curve 2a one may clearly see that after heating at a

CELL REPARATION OF NON-DNA INJURY

247

temperature exceeding by 6°C the minimum temperature that arrests protoplasmic streaming, the second reparation is accomplished about 10 times more rapidly. (iii) After the first repair intensification of cell reparability not only enhances drastically the recovery speed of the disturbed function but also extends the repair zone. In Fig. 13A the lower horizontal line demonstrates the repair zone of leaves which recovered after a 5-minute heating at 45°C. The width of the zone is 9.0"C. After the first reparation the width of the reparation zone is 6.4"C, leaving a difference of 2.6"C. Similar data were obtained for 2. pendula (Fig. 13B) and P . peltaturn leaves. Stimulation of reparability may be observed after less intensive preliminary heat treatments that do not arrest protoplasmic streaming (Fig. 13C). In this instance no heat hardening occurred and the effect of stimulation was considerably less. In Tradescantia, a 3-hour incubation of the leaves in a moist chamber at 38°C led to such heat hardening and to a sharp acceleration of recovery from the second heat injury. The repair zone was not extended. A 3-hour exposure to 42°C increased the resistance of thermostable functions (ability for plasmolysis, respiration), but inhibited the reparability of the cell-the recovery speed slowed down drastically, and the repair zone was reduced (Alexandrov and Barabalchuk, 1972). Thus, stimulation or repression of reparability may or may not go together with heat hardening of the cells. This supports the fact that the nature of repair and hardening is different. Intensification of cell reparability in the process of recovery was also attained for the unicellular green alga C. eugametos (Alexandrov and Luknitskaya, 1978). Figure 13D shows curves for speed of recovery of cell movement after a 5-minute heating at temperatures indicated on the abscissa. The minimum temperature arresting movement after such heating was about 34°C. Curve 1 indicates the speed of recovery of movement after the first exposure to heat at different temperatures. Curve 2 represents cells after the second exposure to different temperatures, 24 hours after the first 5-minute heating at 41°C. Within this time the mobility of cells arrested by the first heat treatment resumed completely and returned to norm. The primary resistance of cell movement to heating did not differ from the control as well. Comparison of the curves shows that the rate of repeated reactivation exceeds that of primary recovery by several times. Moreover, on repeated reactivation the cells are capable of resuming their mobility after heat treatment, which causes irreversible damage in cells that underwent no thermal stress (43-45°C). The repair zone in repeatedly recovering cells was wider by 3°C. Falkova et al. (1978) observed an increase in the speed of recovery of protoplasmic streaming in leaf epidermal cells of Saintpaulia ionantha after repeated heating, but the repair zone was not extended. Similar phenomena were described by Benzioni and Itai (1975) for tobacco and bean leaves. The authors compared the completeness of recovery of two

20

2

FIG.13. Intensification of cell reparability after recovery of heat injury. Abscissa: temperature of 5-minute heating; ordinate: time of recovery of protoplasmic streaming (A-C) or of cell movement (D)arrested by heating, logarithmic scale. Curves 1: repair after fist heating; curves 2 repair after repeated heating. Horizontal lines at the bottom: corresponding reparatory zone. Onset of the horizontal lines corresponds to the minimum temperature of 5-minute heating arresting the protoplasmic streaming. (A) Epidermal cells of Campanula persicifilia leaves. Curve 2: 1 day before exposure to heat at different temperatures the leaves were heated for 5 minutes at 45°C. (B) Epidermal cells of Zebrina pendula leaf. Curve 2: 1 day before exposure to heat at different temperatures the leaves were 248

.

I

44

45

46

47

40

49

50

OC

51

2

i I II

3,000-

I ,000.

300-

100.

50. L . 34 35 36 37 38 39 40 41 42 43 44 45

OC

heated for 5 minutes at 46°C (A,B). Curves 2a correspond to curves 2 shifted toward a lesser temperature by a value corresponding to the increase of the primary heat resistance of the cell after the first exposure to heat. (C) Epidermal cells of Campanula persicifolia leaf. Curve 2: 1 day before exposure to heat at different temperatures the leaves were heated for 5 minutes at 39°C. (D) Cells of Chlamydomonas eugametos. Curve 2: 1 day before exposure to heat treatment at different temperatures the cells were heated for 5 minutes at 41°C. 249

250

V. YA. ALEXANDROV

functions-fixation of C02 and leucine incorporation into proteins-24 hours after a 2-minute heating at 47.5" in leaves injured once and in leaves heated twice at a day's interval. In the latter case, the recovery was more complete. They also observed an increased reactivation after treatment of the leaves with NaCl solution (3 gdliter). For further analysis of the above phenomenon one must determine to what extent the stimulation of reparability detected after elimination of the first heat injury is reversible. As has been shown for Chlamydomonas the speed of recovery from the second injury falls gradually during the days following the first heat treatment, and 4 days later it cannot be distinguished from the speed of recovery of cells repairing the first damage (Fig. 14). The preliminary experiments of min

70C

60C

50C

40C

300'

I

2

3

4

5

days

FIG.14. Dependence of repeated recovery speed of Chlamydomonas eugametos movement on time elapsing after first heating. The first and repeated exposures 5 minute heating at 41°C. Abscissa: days after first heating; ordinate: time for recovery of cell movement. Curve 1 : repair after first heating; curve 2: repair after repeated heating.

CELL REPARATION OF NON-DNA INJURY

25 1

Gorban (1975) on C. persicifolia leaf cells created an impression that in these nondividing cells a decrease in the stimulated reparability proceeds at a slower rate than in Chlamydomonas. According to Gorban (1978), Tradescantia cells, in which reparability was stimulated after elimination of heat injury, recover more quickly after poisoning with ethanol than cells that underwent no heat trauma. It would be of interest to compare data obtained in our laboratory from the experiments on stimulation of cell reparability in higher plants and unicellular algae with the evidence of inducibility of DNA repair in bacteria and animal tissue cells. Ultraviolet light injures DNA and induces in E. coli the formation of an additional reparatory mechanism, which is repressed in uninjured cells (for the review, see Witkin, 1976). Owing to this, on repeated irradiation with uv light, injuries that cannot be eliminated after the first exposure are repaired (so-called SOS repair). D'Ambrosio and Setlow (1976) revealed, in a tissue culture of Chinese hamster cells, two repair mechanisms, the first (constitutive) of which does not require protein synthesis, and the second of which is inhibited with cycloheximide. Induction of the second postreplication may be attained by weak uv irradiation or by the action of a low concentration of a chemical mutagen, N-acetoxyacetylaminofluorine. 2. The Second Series of Experiments Experiments performed on both isolated leaves and whole plants show that a temperature treatment causing no trauma may also stimulate the reparability of cells. In Zebrina and Podophyllum leaves placed in a moist chamber at 30" to 32°C for 20 hours the speed of recovery increases sharply and the repair zone is extended as compared to leaves kept at 20°C within the same period of time. The maintenance of the leaves at 30" to 32°C for 20 hours did not inhibit protoplasmic streaming and did not change its thermostability (Gorban, 1974). In these experiments one may suggest the presence of some unheeded injury, but Lutova and Zavadsksys (1966) showed the dependence of reparability on the cultivation temperature of plants which does not exceed the tolerant zone. Tradescantia was grown by these investigators at lo", 20", and 28°C. The minimum temperature arresting protoplasmic streaming in the leaf epidermal cells was similar in these plants or even higher in plants kept under cooler conditions. The higher the cultivation temperature, however, the wider the repair zone. According to Gorban (1975), in Zebrina cells kept in the greenhouse in winter at 13" to 15°C a 5-minute heating at 43" and 46°C led to intensification of cell reparability. They eliminated more rapidly the repeated heat injury, and their repair zone was more extended than that in cells which underwent no reparatory cycle. In June, when the temperature in the greenhouse occasionally rose to 30" to 36°C and the primary heat resistance of protoplasmic streaming increased by 1°C (heat hardening), reparability increased drastically as compared with the

252

V. YA. ALEXANDROV

winter season. In summer the recovery speed in control cells was no less than in winter cells stimulated by a preliminary 5-minute heating at 46°C. Essentially, the “summer” cells did not increase reparability further in response to a 5-minute reversible heat injury. It is not unlikely that their capacity to increase reparability was exhausted. If Zebrina is transferred from the greenhouse to a room maintained at a temperature of 18” to 20°C the reparability of cells decreases and they regain their ability to enhance their reactivation capacity in response to a heat stress. The repair zone of epidermal cells of C. persicifolia leaves taken in winter from under the snow and heated for 5 minutes was 2.5”C smaller than that of cells of plants kept in the greenhouse at a temperature of 8” to 10°C for the same time period (Fig. 8C).

V. Mechanism of Heat Injurious Action To gain an idea of the mechanism of repair of heat injury one must determine what in the cell is injured by heat and how. Unfortunately, our knowledge of this subject is less complete and more vague than our knowledge of DNA damage caused by radiation or chemical mutagens. We have already mentioned relevant data on DNA degradation after moderate heating of S . typhirnuriurn and E . coli. Many authors considered this to be the cause of cell lethality. In compliance with the opinions of all the investigators this phenomenon of DNA breaks is not the result of a direct action of heating. These breaks arise secondarily in an enzymatic way. Some of the authors regard them as a consequence of nuclease activation, while other investigators suggest that heat increases the number of local fluctuational openings of the DNA double helix vulnerable to nucleases. It is not inconceivable that elimination of DNA breaks after heating involves some enzymes in the systems that repair radiation lesions. This accounts for the parallelism observed by a number of the authors between the sensitivity of bacteria to moderate heating and ionizing radiation (Bridges et al, 1969; Matsumoto and Kagami-Ishi, 1970; Samoilenko and Ivanov, 1972; Woodcock and Grigg, 1972). The parallelism is not universal, and there occur many exceptions (Haynes, 1964; Bridges el al., 1969). Breaks of DNA threads in cells exposed to heating are not to be regarded as a general cause of heat injury and cell death (Sedgwick and Bridges, 1972). When analyzing causes of a deleterious effect of heating on the cell we must first take into consideration the great significance of different weak bonds along with covalent bonds, in the organization of a cell system. The strength of these weak bonds is comparable to the kinetic energy of heat movement of molecules at physiological temperatures. Therefore, an increase in temperature may result in a rupture of weak bonds and, as a consequence, disturbances of structures in

CELL REPARATION OF NON-DNA INJURY

253

the maintenance of which weak bonds play an important role. These structures are various supermolecular complexes of membrane and fibril types consisting of components fastened through weak bonds. They also include protein macromolecules, the secondary, tertiary, and quaternary structures of which are supported by weak bonds. Breaks of these bonds lead to breakdown of the spatial structure of the protein macromolecule and its transition to a more or less disorganized functionally inactive denatured state. In the literature, we find a great body of evidence indicating that even reversible heat injury of cells may cause denaturation changes in protein. (i) In heat-injured cells one may observe alterations typical of heat denaturation of protein solutions in vitro. After heat injury, as a result of disorganization of the spatial structure of protein, many groups of macromolecules, which are sterically inaccessible in the native molecule, acquire the ability to interact with dissolved substances. The increase in nonspecific reactivity of proteins after heat denaturation is manifested, among other things, in enhancement of binding of a number of dyes, both acidic and basic (Alexandrov and Nasonov, 1939; Braun, 1948, 1949; Singer and Morrison, 1948; Oster and Grimsson, 1949; Haurowitz et al., 1952; Belicer and Varetskaya, 1959; Prokopova and Munk, 1963). In accordance with this, an increase in the affinity of the cytoplasm and nuclei to acidic and basic dyes was reported for a number of objects after reversible heat injury of cells as well as after heating that induced irreversible changes approaching minimal lethal disturbances (Nassonov, 1932; Aisenberg, 1934; Mescherskaya, 1935; Krasilnikova, 1954; Nemchinskaya and Braun, 1962; Vinogradova and Dzhamusova, 1963; Suzdalskaya and Zander, 1963; Suzdalskaya, 1966). Exposure of groups earlier hidden in the globule is also responsible for the fact that heat denaturation of protein solutions at pH values not far from the isoelectric point is commonly accompanied by formation of intermolecular complexes. In the case of such aggregation the viscosity and light scattering of the solution increase. An increase in viscosity after reversible heat injury was recorded by a number of authors for both animal and plant cells (Heilbronn, 1914, 1922; Heilbrunn, 1925, 1928; Port, 1927; BtSlehradek and Melichar, 1930; Zavadskaya, 1963). The increase in light scattering and in the appearance of granular aggregates in the nucleus and cytoplasm was observed repeatedly in the protoplasm of various cells after heat injury (Lepeschkin, 1923; Heilbrunn, 1925; Van Herwerden, 1927; Nassonov, 1932; Aisenberg, 1934; Ushkov and Vasilyeva, 1965). (ii) It is known that the temperature coefficient of heat injury of cells is remarkably high. It amounts to hundreds and thousands. This concerns not only the temperature dependence of the heat death but also the reversible suppression of cell functions. Table 11 gives the Qlo value for a reversible cessation of protoplasmic streaming after heat treatment of cells of different plants. These data are in good agreement with the high activation energy characteristic of heat

(el,,)

254

V . YA. ALEXANDROV TABLE I1 TEMPERATURE COEFFICIENTS OF HEATCESSATION OF PROTOPLASMIC STREAMING IN LEAFEPIDERMAL CELLS Object

Q.0

Dactylis glomerata Panicum miliaceum Triticum vulgare Campanula persicifolia Phragmites communis Elymus arenarius Echeveria secunda Tradescaniia fluminensis Zebrina pendula

2150 820 620 580 550 470 280 120 60

denaturation of proteins. The Qlo value of this process is also very large, differing greatly from the Qlo of common chemical reactions which usually lies in the range of 2 to 3. (iii) Comparison of the heat resistance of the same cells of closely related species shows that cells of a more thermophilic species are, as a rule, more thermostable. This difference corresponds to an appropriate variation in the thermostability of their proteins. After superoptimal heat treatment of plant cells we observe a temporary increase in their thermostability, i.e., heat hardening. The latter is accompanied by an increase in the heat resistance of some proteins (for reference, see Alexandrov, 1977). These facts are consistent with the assumption that cell thermostability is determined by the thermostability of cell proteins. Some authors are prone to ascribe the heat injury of cells to melting of cell lipids which may lead to a breakdown of lipoprotein complexes (B6lehradek, 1935; Heilbrunn, 1956; Evans and Bowler, 1973). Certain facts, however, do not permit explanation in such a way of the deleterious effect of relatively short-term heating on the cells. Thus, changes in the content of unsaturated fatty acids in the muscles of frogs kept at different temperatures, as well as during spawning, did not correlate with the thermostability level of muscles (Ushakov and Glushankova, 1961). Amosova (1963) cultivated seven generations of the fly Calliphora erythrocephala at 17” and 24°C. She found that the heat resistance of the fly muscle tissue (determined by the time of loss of excitability at 42°C) in “warm” and “cold” larvae remained unchanged. The degree of saturation of fatty acids in muscle lipids was higher in the muscles of “warm” larvae. Introduction of butter or sunflower oil to the rat diet did not affect muscle heat resistance, although the degree of saturation of fatty acids shifted accordingly (Glushankova, 1963). I don’t think there is any reason to ascribe the inhibition of the cell

CELL REPARATION OF NON-DNA INJURY

255

locomotory function at short-term heating to injuries of nucleic acids. Goldman et al. (1973) showed that in fibroblasts in a tissue culture denucleated by cytocholasin treatment the locomotory function is wholly retained. (iv) It is known that some substances increase considerably the resistance of protein macromolecules to various denaturants, and to heat in particular. Some of them stabilize only definite proteins, such as enzyme substrates, coenzymes, and allosteric effectors. Other agents are able to increase the resistance of many various proteins. Such antidenaturators of a wide range of activity include some sugars, glycerol, Ca ions, some amino acids, and D20. High hydrostatic pressure of moderate intensity also shows an antidenaturing effect. It is most essential that all these agents, provided the dose is properly chosen, also increase the resistance of animal, plant, and bacterial cells to heat (for references, see Alexandrov, 1977). By way of illustration let us consider two of a great many relevant examples. Replacement of H 2 0 with D20induces an increase in the heat resistance of ribonucleases, lysozyme, phycocyanin, collagen, and glycerinated models of frog ciliary epithelium cells. At the same time, D20 increases the thermostability of various cells of the palate ciliary epithelium and muscle fibers of the frog; the epidermis of C.persicifolia and T . jluminensis leaves; E . coli; and Sraphylococcus phage (for reference, see Denko, 1970). As another example, a hydrostatic pressure of about several thousand atmospheres denatures proteins. A pressure within 100 to 1000 atm exerts an opposite antidenaturing action. A stabilizing effect resulting from the application of hydrostatic pressure during heat exposure was observed for: tobacco mosaic v i m protein (Johnson et al., 1948b); the protein of bacteriophages TI, T2,and T5 (Foster et al., 1949); pyrophosphatase; malate dehydrogenase of B . stearothermophilus (Morita and Haight, 1962; Morita and Mathemeier, 1964); E . coli aspartase (Haight and Morita, 1962); yeast invertase (Johnson et al., 1948a); Achromobacter jischeri lucipherase (Strehler and Johnson, 1954); Electrophus electricus acetylcholinesterase (Millar er al., 1974); and glycerinated models of the frog oviduct ciliary epithelium (Arronet and Denko, 1973). Hydrostatic pressure of about the same intensity (100-700 atm) increases the heat resistance of living cells, B . subtilis spores (Johnson and Zobell, 1949), E . coli cells (Johnson and Lewin, 1946), Photobacterium phosphoreum (Johnson, 1957), epidermal cells of the Tradescantiajluminensis leaves, and cells of the frog oviduct ciliary epithelium (Arronet and Denko, 1973). (v) It is evident that the above arguments, in favor of the assumption that intensive short-term heating in doses reversibly injuring cells inactivates proteins, are indirect. Some of them only indicate that an injuring heat treatment induces breaks of many weak bonds. Such evidence also does not permit one to say with certainty whether these disturbances are confined to intermolecular bonds or they lead to destruction in the inner native structure of protein macromolecules. In any event, it was demonstrated (Makarov et al., 1976; Makarov,

256

V. YA. ALEXANDROV

1977) on crystals of some globular proteins (pepsin, trypsin, catalase, leghemoglobin, aspartate transaminase, myoglobin, ribonuclease) that heat destruction of crystals consists of denaturing protein molecues inside the crystal. This must also be true for other supermolecular protein systems. It is difficult to suggest that intra- and intermolecular weak bonds differ significantly in their strength. Unfortunately, there is little direct evidence in favor of reversible inactivation of intracellular proteins after heating. In this context, of primary interest are the publications of Tomlins ef al. (1971) and Airas (1972, 1976a,b). Tomlins and his co-workers exposed S?aphylococcus aureus to a 15-minute heating at 52°C. One of the symptoms of injury was the loss of tolerance to 7.0% NaCl. Three to four hours after heating, a normal attitude toward the salt was restored. Immediately after heating the authors analyzed the catalytic activity of 15 enzymes in the cell extract. They determined that the activity was inhibited considerably only in three enzymes-oxyglutarate dehydrogenase, malate dehydrogenase, and lactate dehydrogenase. When the extract was prepared 3 hours after heat treatment of the cells, the activity was restored mainly in malate dehydrogenase and lactate dehydrogenase. During that time no oxyglutarate dehydrogenase reaction was observed. Airas used Pseudomona jluorescens for his experiments. The bacteria were cultivated at 25°C. One-hour exposure to heat at 40°C led to inactivation of pantotenate hydrolase in cells, while in a suspension of cells destroyed by freeze-thawing there was no pantotenase activity. When cells were retransferred to 25°C after heating the enzyme activity was restored gradually and the whole process was accomplished about 4 hours after heating. After in vitro inactivation of the enzyme in the presence of certain substances (oxaloacetate, pyruvate) some reactivation was found to occur in 3 hours, but it was much less pronounced than after in vivo inactivation. It could be traced only after partial inactivation of the enzyme following a 10-minute heating at 28" to 37°C. Ron and his co-workers showed that cessation of growth of E. coli upon transfer of the culture from 37" to 45°C is due to inactivation of a thermolabile enzyme of homoserine transsuccinylase. When returned to 37°C the enzyme immediately regains its activity (Ron and Davis, 1971; Ron and Shani, 1971). Reversible inactivation of contractile proteins in ciliary epithelium cells was demonstrated by the experiments of Arronet (1968) with glycerinated models. Glycerinated models of the ciliary epithelium were first produced in 1956 by Alexandrov and Arronet. These are 45% glycerol-killed cells from which all water-soluble substances are extracted and in which the selective permeability of membranes is completely destroyed. The protein contractile apparatus, however, retains its efficiency. After removal of glycerol the cilia are able to respond by vigorous beating to the addition of ATP to the solution in the presence of magnesium ions. When pieces of the frog oviduct mucosa covered with ciliary epithelium are placed into a ringer solution with an energy metabolism inhibitor,

CELL REPARATION OF NON-DNA INJURY

257

NaF, beating of the cilia stops after a short time. When the pieces are placed into glycerol and then washed out with a salt solution also containing NaF, the addition of Mg2+ and ATP results in beating of the cilia despite the presence of the inhibitor. This indicates that NaF stops the movement of cilia in living cells without affecting the ciliary contractile apparatus. On the other hand, if beating in living cells is stopped by a 5-minute heating at 42.2"C and then the cells are immediately transferred to a 45% glycerol solution they will never make working models. But, if heat-treated pieces of tissue are placed into a Ringer solution at room temperature, movement of the cilia is resumed 20 hours afterward and the cells may turn into efficient glycerinated models. In models made of cells just after heating, the ability to respond to the addition of ATP by beating the cilia is not resumed (the experiments of Arronet). These experiments show that the heat-inactivated contractile apparatus is restored in the cell system, while in the model no inflicted heat injury is eliminated. We have no doubt that injurious heat treatment affects various cell components, nucleic acids and lipids inclusive. Indirect and direct evidence presented in this section shows convincingly, however, that reversible injuries of some cell functions induced by short-term intensive heating are underlain by inactivation of certain cell proteins. Specifically, when the question concerns the cessation of the cell motor function the effect is not to be ascribed to any DNA or RNA damage. It is also clear that there is no correlation between the primary heat resistance of the cell contractile activity and the degree of saturation of fatty acids.

VI. Resynthesis or Reactivation? The above arguments permit the conclusion that after restoration of some heat-inhibited functions, and of motor functions in particular, injured proteins are either reactivated in the cell or replaced for newly synthesized ones. Potentialities of the recovery of disturbed functions via reactivation of injured proteins may be estimated by the effect of switch off of the protein synthesis on the repair process. Unfortunately, few relevant investigations are being performed and they are almost exclusively on bacteria. A most detailed study has been made of the influence of metabolic inhibitors on the reparation of heat injury in S.aureus and S . typhimurium. After reversible injury (15 minutes at 55°C) the following lesions were registered in S. aureus cells: damage of the cell membrane manifested in the loss of tolerance to 7.5% NaCl and excretion of some substances from the cell; destruction of ribosomal 30 S subunits; degradation of 16 S rRNA; and enchancement of 23 S rRNA sensitivity to ribonuclease. Four hours after the injured cells are transferred to an appropriate medium they return to their normal state. The repair

258

V. YA. ALEXANDROV

is not prevented by inhibition of protein synthesis with chloramphenicol, 5 methyltryptophan, and cycloserine, nor by repression of mucoprotein synthesis with vancoleucine, inhibition of membrane growth with penicillin, and oxidative phosphorylation with 2,Cdinitrophenol. Reparation may proceed in the absence of cell growth or reproduction, but RNA synthesis is indispensable if the recovery is to occur. Therefore, it is completely inhibited by actinomycin (Stiles and Witter, 1965; Iandolo and Ordal, 1966; Sogin and Ordal, 1967; Rosenthal et a l . , 1972; Iandolo, 1974). The results obtained from analogous experiments with S . typhimurium were different. Reversibly heat-injured Salmonella cells (30 minutes at 48°C) also exhibit a loss of tolerance to salt, an increase in sensitivity to Ph changes and to methylene blue, and an enhanced efflux of substances. This indicates disturbances in the cell membrane. In addition, 30 S subunits of ribosomes, 16 S rRNA, and partially 23 S rRNA are degraded and the weight of 50 S subparticles decreases to 47. Compared to reparation in S . aureus, in this bacteria reparation is repressed as a result of inhibition of protein synthesis with chloramphenicol or chlortetracycline. 2,CDinitrophenol produces the same effect. To eliminate heat injury in Salmonella, just as in Staphylococcus, RNA synthesis is required, and no repair occurs in the presence of rifamicin (Clark and Ordal, 1969; Tomlins and Ordal, 1971a,b; Pierson et a l . , 1971; Gomez et al., 1976). According to Mukherjee and Bhattacharjee (1970) heat-injured E. coli do not reactivate in the presence of chloramphenicol. On the other hand, inhibition of protein synthesis with chloramphenicol does not affect the repair of injury in Streptococcus falcalis (Clark et a l . , 1968) and in P . fluorescens (Gray et a l . , 1973). In the spores of Penicillium expansum repression of protein synthesis with cycloheximide also does not inhibit reparation of thermal damage (Baldy et al., 1970). The above examples show that the effect of switch off of protein synthesis on the reactivation of heat injuries varies in different species. Moreover, it is likely that the possibility of cell recovery after thermal trauma in the absence of protein synthesis depends on the depth of heat injury. The investigations of Bernstam and Arndt (1974) carried out in our laboratory on the plasmodium, P . polycephalum, testify in favor of this supposition. In the plasmodium, protoplasmic streaming stops after a 10-minute heating at 38°C. After a 10-minute heating the repair zone extends up to 45°C. After heating up to 42°C and above 43"C, however, reactivation of protoplasmic streaming proceeds in different ways. In the case in which the plasmodium is heated at 43" to 45"C, protoplasmic streaming is reactivated only after formation of a sclerotium and its germination. After heat treatment at 38" to 42"C, movement in the plasmodium is restored directly without preliminary transformation of the plasmodium into a sclerotium. Poisoning of the plasmodium with cycloheximide, which completely stops protein synthesis, does not prevent a renewal of motor functions after heating at 38" to 42°C. But, after more intensive heating, when reactivation of protoplasmic streaming occurs via formation of a sclerotium, protein synthesis is indispensa-

CELL REPARATION OF NON-DNA INJURY

259

ble. After such heat treatment, in the presence of cycloheximide, the plasmodium is fragmented without formation of a sclerotium and recovery of protoplasmic streaming. It is quite obvious that in many cases the cells are capable of reactivating heat injury in the absence of protein synthesis. Thus, Tomlins et al. (1971) demonstrated that heat-inactivated malate dehydrogenase and lactate dehydrogenase of S. aureus resume their activity in cells poisoned with chloramphenicol. Analogous results were obtained by Airas (1972) for pantothenase of P. Juorescens. In this case also the heat-inactivated enzyme reactivated in a cell in which protein synthesis was completely inhibited with chloramphenicol. Facts presented in this section show that heat-injured proteins in the cell may regain their normal state via reactivation. It is most likely that, in the process of repair, cells not only reactivate injured proteins but also replace them with newly synthesized ones. This mode of repair is the most probable in cases when recovery of function is delayed for a long time. It is evident that to answer the question asked at the beginning of this section, further research is needed, particularly on cells of higher organisms.

VII. Some Evidence of Protein Renativation What kind of physical evidence enables us to assume the recovery in the cell of a native structure of heat-denatured protein macromolecules? Studies on the renativation of proteins after various breaks in their spatial structure have a long history. At present, there is a vast amount of literature considering the problem from different viewpoints (for reference, see Nasonov and Alexandrov, 1940; Neurath et al., 1944; Tongur, 1951; Joly, 1965, Tanford, 1968; Wetlaufer and Ristow, 1973; Anfinse and Scheraga, 1975; Kushner, 1977). Some of the authors believe that the primary protein structure comprises sufficient information to restore their secondary, tertiary, and quaternary structures after destruction by a denaturant. This supports the reversibility of any denaturation which does not break the primary protein structure (Anfinsen, 1962, 1972; Friedland and Hastings, 1967). In practice, however, the completeness and rate of protein renativation depend directly on the nature of both the protein and the denaturant as well as on the renativation conditions. In some instances, for recovering the native state and inducing spontaneous renaturation it is sufficient to remove a denaturing agent, while in others, this would not be enough, and, besides elimination of the denaturant, additional treatment of the denatured protein is required. Protein renativation is readily obtained after denaturing with urea and guanidine chloride. These substances induce a most complete unfolding of protein macromolecules preventing, as a rule, secondary intermolecular interactions. If the tertiary protein structure is supported by disulfide bonds, urea, and guanidine

260

V. YA. ALEXANDROV

chloride, denaturation is carried out in the presence of reagents which break these bonds via reduction of the sulfur atom (e.g., P-mercaptoethanol). To provoke renativation denaturants are removed by dialysis or strong dilution of the reaction mixture. Heat denaturation usually involves a less complete unfolding of the protein macromolecule and denaturation products may contain regions retaining their native structure (Tanford, 1968; Zimmerman and Coleman, 1971). A product obtained after heating is commonly not as readily renatured as products resulting from denaturation with urea and guanidine chloride. Some of the proteins heated in solutions with a pH value not far from the isoelectric point form intermolecular complexes and precipitate. In the overwhelming majority of similar cases cessation of heating does not lead to renativation of protein in the precipitate. Renativation, however, may be induced in many proteins by preliminary dilution of the sediment in urea or guanidine chloride with subsequent elimination of these substances from the solution. The same result is obtained when a heatprecipitated protein is dissolved in an acid or alkaline solution, which are then gradually neurtralized. Thus bacterial luciferase is denatured at 46°C after a few minutes, yielding a white precipitate. If the precipitate is dissolved in 5 or 8 M urea and then the urea concentration is reduced by strong dilution of the solution, the enzyme activity is restored (Friedland and Hastings, 1967). As another example, sperm whale myoglobin at a pH ranging from 5 to 9 is denatured immediately during heating and then precipitates. After cooling the protein native state may be restored by shifting the pH beyond this boundary (Acampora and Hermans, 1967). In the case in which protein is heated in a region of pH far from the isoelectric point, a true equilibrium may be established between the native and denaturated states of protein macromolecules. Many proteins behave in such a way, for example, chymotrysinogen, chymotrypsin, ribonuclease (Brandts, 1967, 1969), myoglobin (Hermans and Acampora, 1967), deoxyribonuclease (Zimmerman and Coleman, 1971), and others. Some proteins show a spontaneous reversibility of heat denaturation also at a pH close to the isoelectric point. Dupont (1965) identified the reversibility of the first stage of heat denaturation of bovine /3-lactoglobulin, at pH 6.85, by changes in its optical activity. When heated in salt solutions, ferrimyoglobin of the mollusk Aplysia is denatured without precipitating, and it rapidly regains its native state as soon as the solution is cooled. In salt-free solutions the protein precipitates after heating, but even in this state the protein is renativated as soon as the heating is over (Brunori et al., 1968, 1972). Many heat-denatured proteins exhibit spontaneous reversibility: vegetable peroxidase (Herrlinger and Kiermeier, 1944; Tamura and Morita, 1975), trypsin (Terminiello et al., 1958), a-amylase, of B. subtilis (Yamamoto et al., 1964), diphosphopyridine nucleotide pyrophosphatase (Swartz et al., 1958), etc. Ron and Shani (1971) ob-

CELL REPARATION OF NON-DNA INJURY

26 1

served a most rapid renativation of homoserine transsuccinylase after heating in both purified preparations and cells. The degree of reversibility depends on the intensity and length of heating (Citri and Zyk, 1972; Tamura and Morita, 1975; Airas, 1976b), as well as on the protein structure. Zymogen of Streptococcus proteinase is reactivated after denaturation with guanidine chloride and heat treatment. Elimination of 100 amino acid residues from the NH2 end of the zymogen renders it incapable of forming an active structure on the unfolded chain after a similar treatment. The authors suggest that this region of the zymogen is indispensable for the formation of native conformation. The ability of zymogen to coil properly after removal of a denaturant, in contrast to mature enzymes, has been recorded for pepsinogen, proinsulin, and chymotrypsinogen. The rate and completeness of renativation of a protein are sometimes influenced by its concentration in the solution. Depending on the nature of the protein, an increase in its concentration stimulates or prevents renativation, but in some cases it exerts no effect (Yutani et al., 1967, 1969; Teipel and Koshland, 1971; Zimmerman and Coleman, 1971; Teipel, 1972; Carlsson et al., 1973; Chan et al., 1973; Waley, 1973; London et al., 1974). The investigations of this trend are concerned primarily with urea and guanidine chloride denaturation. Renativation is frequently much dependent on the presence of certain substances in the medium. In some cases the effect of such substances is strictly specific. Some substances affect similarly the renativation of different proteins. Thus, according to Michaeli et al. (1967), acethylcholinesterase of bovine erythrocytes after heat treatment (60°C) yields an irreversibly inactivated product. If antibodies are added just after heating or 24 hours later, the enzyme reactivates but only partially. PNA polymerase from E. coli, denatured with urea, does not reactivate after its elimination. If a DNA from the salmon sperm is present during reactivation 55% of the enzyme activity is restored (Lill and Hartmann, 1970). Phosphorylase b kinase denatures partially at 30°C. It reactivates upon addition of Mg2+ and ATP. The process is accelerated by cyclic AMP (Vandenheede et al., 1977). Alpers et al. (1971) found that ATP catalyzes the conformation transition of phosphofructokinase from a nonactive state induced by a low pH value into an active state, due to alkalinization of the solution. There are some other publications with references to the improvement of reactivation of some denatured enzymes by their substrates and cofactors (Steiner, 1961; Deal, 1969; Lill and Hartmann, 1970; Levi and Kaplan, 1971). a-Asparaginase heated at 82°C is not able to reactivate spontaneously. At 4"C, in the presence of substrates (L-asparagine, D-asparagine), it regains its activity (Citry and Zyk, 1972). This is not a common rule, however, and there occur enzymes which are not affected positively by substrates and cofactors (Chilson et al., 1966; Kohn, 1970; Levi and Kaplan, 1971). Not infrequently protein reactivation is stimulated by substances that increase the resistance of its native macromolecules (Chilson et al.,

262

V . YA. ALEXANDROV

1966; Andersson, 1969; Ogasahara et al., 1970; Citri and Zyk, 1972; Airas, 1976a,b). It is known, for example, that Ca2+ stabilizes a-amylase, and its presence promotes a more complete and quicker reactivation of the enzyme (Yutani et al., 1967). In 1963, in laboratories headed by Anfinsen (Goldberger et al., 1963, 1964; Epstein and Goldberger, 1963; Steiner et al., 1965; Fuchs et al., 1967; Anfinsen, 1972) and Straub (Venetianer and Straub, 1963a,b, 1964, 1965; Krause et al., 1965), an enzyme accelerating reactivation of proteins, which contains S-S bonds, was obtained. In the presence of this enzyme, after elimination of reagents breaking S-S bonds and inducing unfolding of the polypeptide chain, ribonuclease, lysozyme, and soybean trypsin inhibitor reconstitute their native structure quicker and fuller than during spontaneous reactivation in the absence of the enzyme. The latter was obtained from the liver and pancreas microsomal fraction of various animals. Its molecular weight is 42,000. It contains three cysteine residues two of which form an S-S bond. According to the abovementioned authors, the spontaneous renativation of proteins with the tertiary structure supported by S-S bonds is perplexing due to the fact that random S-S bonds uncharacteristic of the native molecule are formed as a result of SH group oxidation. However, the bonds are gradually rearranged and during this process regular S-S bridges arise in some of the molecules which regain their efficiency. The discovered enzyme is considered to catalyze the rearrangement of random irregular pair combinations of cysteine residues. Therefore, the efficiency of the “disulfide interchange enzyme” is a function of the number of S-S bonds in the substrate protein. The work of Yutani et al. (1967) showed, however, that the question is probably more complicated. These authors noticed that to induce this effect very large concentrations of the enzyme are required as compared to ordinary enzymatic reactions. Moreover, in the presence of the “disulfide interchange enzyme” they observed a sharp increase in reactivation of a-amylase previously denatured with urea, despite the fact that bacterial a-amylase contains no S-S bonds. An increase in the speed of renativation of a-amylase was also caused by some other proteins, for instance, bovine serum albumin and casein. The positive effect of serum albumin on the renativation of a-amylase and some other proteins has been repeatedly reported (Ogasahara et al., 1970; Bertagnolio et al., 1970; Kohn, 1970; Chan et al., 1973). However, it does not affect the recovery of all the proteins (Waley, 1973). A favorable effect of introduction of alien proteins or of enhancement of denatured protein concentration on recovery indicates that, sometimes, protein-protein interactions are required for normalization of the protein state. The data presented in this section show that proteins denatured by various agents, heating inclusive, can reconstitute their native spatial structure. The similarity of the renativated protein to the original one was in many cases demon-

CELL REPARATION OF NON-DNA INJURY

263

strated by various physicochemical, immunological, and functional indices. Sometimes the renativation product was not fully identical to the native protein. After cessation of heating, some of the proteins more or less rapidly regain their original state according to the scheme; native protein denaturated protein. Spontaneous transition does not rule out intermediate states between the native and denaturated form: H g D ’ e D . For renativating other proteins it is not sufficient to stop heating, and an additional treatment of the denaturated product is needed. The potentialities and mechanism for the protein reactivation in this case are dictated not only by the nature of the protein but also by the composition of the renativation medium. A number of substances exert a considerable effect on the speed and completeness of protein renativation. Sometimes this effect is highly specific, while in other cases, one and the same substance may aid renativation of various proteins. In separate ceses possibilities for catalytic speeding of protein renativation have been demonstrated.

*

VIII. What Happens in the Cell? Facts presented in this review force us to assume that cells are capable of repairing injuries of protein components of the protoplasm. This is likely to be accomplished in two ways-via renativation of injured protein macromolecules or substitution of the latter with newly synthesized macromolecules. Our knowledge in this field is too poor to answer the above question unambiguously when each of the two types of repair is used. Possibilities for renativation of protein in vitro were demonstrated in a great number of publications described in Section VII. Section VI contains evidence that renativation of heat-inactivated proteins proceeds in living cells as well. Let us assume that the above-described reparation of heat injury in cells of higher plants and in the unicellular alga, C. eugameros, is brought about through reactivation of injured proteins. Then the question arises whether such renativation occurs spontaneously without being actively aided by the cell, or does the cell make use of a mechanism promoting reactivation? The answer should be derived from a stimulation of cell reparability reported in Section IV,F. The experiment described there showed that in the process of heat injury elimination cells increase their reparability. Due to this the secondary heat injury may be eliminated more rapidly and after more intensive heating than lesions in cells that underwent no heat trauma. Stimulation of cell reparability is reversible, and the stimulated state vanishes in a few days. Experiments on stimulation of cell reparability indicate that the reactivation of heat-injured proteins involves metabolic activity of the cell. In fact, if the protein was denatured in vitro and spontaneously restored to its initial native conformation after removal of a denaturing agent, one must not expect, from the standpoint of physics, that after

264

V. YA. ALEXANDROV

repeated denaturation it would renativate more quickly. Therefore, stimulation of reparability in cells that recovered after trauma suggests a biochemical mechanism, which brings or aids in bringing the cell back to the normal state. Thus it may be assumed that, due to the positive feedback, functional loading of this mechanisms increases its working efficiency. The experiments of the second category presented in Section IV,F show that, in cells of higher plants, the higher the reparability the higher the cultivation temperature. This evidence allows the assumption that the same reparatory mechanism operates during the cell’s everyday life and that upon an increase in the temperature within the tolerant zone its loading enhances. It may be suggested that in the normal cell there occur, although few, spontaneous breaks of protein macromolecules, which are to be corrected. With increase in temperature the frequency of these disturbances must grow depending on labilization of protein macromolecules (Alexandrov, 1977). In case the assumption reflects what happens in the cell, the hypothetic reparatory mechanism must be acknowledged not only as one of the elements that determine resistance of the cell to the action of deleterious agents but also as an element providing for reliability of the cell system maintaining its integrity under normal conditions. Unfortunately, at our disposal there is no evidence of the effect of cut off of protein synthesis on the repair of heat injury for objects in which the reparability was stimulated. Therefore, another assumption may be made that cells reactivate by replacing injured proteins with newly synthesized ones. Then stimulation of the reparatory ability should be regarded as resulting from the activation of protein synthesis in response to heat injury. Such activation may be suggested to be based on derepression of appropriate genes or intensification of some other chains in the protein synthetic apparatus. It is obvious that the question “What happens in the cell?” will be answered only after further studies by cytophysiologists and biochemists. Although the data presented in this review do not solve the question, they nonetheless seem reliable enough to draw the attention of specialists in these fields to the problem of injury and repair of protein components of the protoplasm.

REFERENCES Acampora, G . , and Hermans, J . (1967). J . Am. Chem. SOC. 89, 1543. Airas, R. K . (1972). Biochem. J . 130, 1 1 I . Airas, R. K . (1976a). Biochim. Biophys. Actu 452, 193. Airas, R. K . (1976b). Biochim. Biophys. Acru 452, 201. Aisenberg, E. (1934). Arch. Anar. Histol. Embryol. 13, 115. Alexandrov, V. Ya. (1948). Tr. Inst. Cyrol., Gisrol, Embriol., Akud. Nuuk SSSR 3 , 3 . Alexandrov, V. Ya. (1955). Tr. Bot. Insr. Akad. Nuuk SSSR, ser. 4, Exp. Bor. 10, 309. Alexandrov, V . Ya. (1956). Bot. Zh. (Leningrad) 41, 939.

CELL REPARATION O F NON-DNA lNJURY

265

Alexandrov, V. Ya. (1964). Tsitologiya 6, 133. Alexandrov, V. Ya. (1965). Tsitologiya 7, 447. Alexandrov. V. Ya. (1977). “Cells, Molecules, and Temperature. ” Springer-Verlag, Berlin and New York. Alexandrov, V. Ya., and Arronet, N. I. (1956). Dokl. Akad. Nauk SSSR 110, 457. Alexandrov, V. Ya., and Barabalchuk, K. A. (1972). Tsitologiya 14, 1328. Alexandrov. V. Ya., and Dzhanumov, D. A. (1972). Tsitologiya 14, 713. Alexandrov, V. Ya., and Luknitskaya, A. F. (1978). Dokl. Acad. Nauk SSSR 243, 217. Alexandrov, V. Ya., and Nasonov, D. N. (1939). Arch. Anat. Hisrol. Embryol. 22, 11. Alexandrov, V. Ya., Khachaturov, E. N., and Shukhtina, H. G. (1963). Tsitologiya 5 , 159. Alexandrov, V. Ya., Lomagin, A. G., and Feldman, N. L. (1970). Protoplasma 69, 417. Alpers, J. B., Panlus, H., and Bazylewicz, G. A. (1971). Proc. Natl. Acud. Sci. U.S.A. 68, 2937. Amosova, I. S. (1963). I n “Problems of Cytoecology of Animals” (B. P. Ushakov, ed.), p. 102. Nauka, Moscow-Leningrad. Andersson, L.-0. (1969). Arch. Biochem. Biophys. 133, 277. Anfinsen, C. B. (1962). Brookhaven Symp. Biol. 15, 184. Anfinsen, C. B. (1972). Biochem. J. 128, 737. Anfinsen, C. B., and Scheraga, H. A. (1975). Adv. Protein Chem. 29, 205. Arpai, J. (1964). Z. Allg. Mikrobiol. 4, 105. Arronet, N. 1. (1963). Tsitologiya 5 , 222. Arronet, N. I. (1968). Dokl. Akad. Nauk SSSR 179, 967. Arronet, N. I., and Denko, E. I. (1973). Dokl. Akad. Nauk SSSR 211, 461. Baldy, R. W., Sommer, N. F., and Buckley, P. M. (1970). J . Bacteriol. 102, 514. Barabalchuk, K. A., and Kasapov, P. S. (1978). Tsitologiya 20, 209. Bauer, H. (1970). Ph.D. Dissertation, Leopld-Franzens-Universitat, Innsbruck. Bauer, H. (1972). Photosynthetica 6, 424. Bauer, H., Larcher, W., and Walker, R. B. (1975). In “Photosynthesis and Productivity in Different Environments” (J. P. Cooper, ed.), p. 557. Cambridge Univ. Press, London and New York. Bglehrhdek, J. (1935). Protoplasma Monogr.. Berlin 8. Bglehrhdek, J., and Melichar, J. (1930). Biol. Gen. 6, 109. Belicer, V. A,, and Varetskaya, T. V. (1959). Ukr. Biokhim. Zh. 3, 171. Belikov, P. S., and Melechov, E. I. (1975). Fiziol. Rast. 22, 466. Benzioni, A., and Itai, C. (1972). Physiol. Plant. 27, 216. Benzioni, A., and Itai, C. (1973). Physiol. Plant. 28, 493. Benzioni, A,, and Itai, C. (1975). Physiol. Plant. 35, 80. Bernstam, V. A., and Amdt, S. (1974). Arch. Mikrobiol. 95, 357. Bertagnolio, G., Got, R., and Colobert, L. (1970). Experientia 26, 140. Beuchat, L. R., and Lechowich, R. V. (1968). Appl. Microbiol. 16, 772. Brandts, J. F. (1967). In “Therrnobiology” (A. H. Rose, ed.), p. 25. Academic Press, New York. Brandts, J. F. (1969). In “Structure and Stability of Biological Macromolecules” (S. N. Timasheff and G. D. Fasman, eds.), p. 213. New York. Braun, A. D. (1948). Dokl. Akad. Nauk SSSR 62, 2. Braun, A. D. (1949). “The Interaction of the Native and Denatured Proteins with Dye-stuff” (Author’s abstract of disseration). Inst. Exp. Med., Leningrad. Bridges, B. A,, Ashwood-Smith, M. J., and Munson, R. J. (1969). J . Gen. Microbiol. 58, 115. Brunori, M., Antonini, E., Fasella, P., Wyman, J., and Rossi Fanelli, A. (1968). J . Mol. Biol. 34, 497. Brunori, M., Giacometti, G. M., Antonini, E., and Wyman, J. (1972). J . Mol. Biol. 63, 139. Carlsson, U., Henderson, L. E., and Lindskog, S. (1973). Eiochirn. Biophys. Acta 310, 376.

266

V. YA. ALEXANDROV

Chan, W. W. C., Mort, J. S., Chong, D. K. K.,and Macdonald, P. D. M. (1973). J . Eiol. Chem. 248, 2778. Chilson, 0. P., Kitto, G. B., Pudles, T., and Kaplan, N. 0. (1966). J. Eiol. Chem. 241, 2431. Christophersen, J . , and Kaufmann, W. (1952). Narurwissenschafren 39, 67. Citri, N., and Zyk, N. (1972). Biochemistry 11, 2103. Clark, C. W., and Ordal, Z. J. (1969). Appl. Microbiol. 18, 332. Clark, C. W., Witter, L. D., and Ordal, Z. J. (1968). Appl. Microbiol. 16, 1764. Dabbah, R., Moats, W. A., and Mottick, J. F. (1969). J . Dairy Sci. 52, 608. D’Ambrosio, S. M., and Setlow, R. B. (1976). Proc. Natl. Acad. Sci. U.S.A. 73, 2396. Daniell, J. W., Chappell, W. E., and Couch, H. B. (1969). Plant Physiol. 44, 1684. David, H. L. (1976). Ann. Microbiol. 1278, 338. Davies, M. E., and Exworth, C. P. (1973). Eiochem. Eiophys. Res. Commun. 50, 1075. Deal, W. C., Jr. (1969). Biochemistry 8, 2795. Denko, E. I. (1970). Vsp. Sovrem. Eiol. 70, 41. Dupont, M. (1965). Eiochim. Eiophys. Acta 102, 500. Egorova, L. I. (1975). Eor. Zh. 60, 1OOO. Egorova, L. I. (1976). Eot. Zh. 61, 945. Epstein, C. J . , and Goldberger, R. F. (1963). J . Eiol. Chem. 238, 1380. Errera, M. (1961). J. Cell. Comp. Physiol., Suppl. 1, 209. Evans, P. R., and Bowler, K. (1973). Sub.-Cell. Eiochem. 2, 91. Falkova, T. V., Kotovshchikova, N. I., and Chernyak, V. V. (1978). Tsitologiya 20, 441. Falymva, S.,and Sagatov, Z. (1977). Uzb. Eiol. Zh. 2, 19. Favard, A. (1963). Rev. Gen. Eor. 70, 661. Foster, R. A. C., Johnson, F. H., and Miller, V. K. (1949). J . Gen. Physiol. 33, 1. Friedland, J., and Hastings, J. W. (1967). Eiochemisrry 6, 2893. Fuchs, S., De Lorenzo, F., and Anfinsen, C. B. (1967). J . Eiol. Chem. 242, 398. Glushankova, M. A. (1963). In “Problems of Cytoecology of Animals” (B. P. Ushakov, ed.), p. 208. Nauka, Moscow-Leningrad (in Russian). Goldberger, R. F., Epstein, C. J., and Anfinsen, C. B. (1963). J . Eiol. Chem. 238, 628. Goldberger, R. F., Epstein, C. J., and Anfinsen, C. B. (1964). J . Eiol. Chem. 239, 1406. Goldman, R. D., Pollack, R., and Hopkins, N. H. (1973). Proc. Nafl. Acad. Sci. U.S.A. 70, 750. Gomez, R. F., and Sinskey, A. J. (1973). J. Eacteriol. 115, 522. Gomez, R. F., and Sinskey, A. J. (1975). J . Eacteriol. 122, 106. Gomez, R. F., Sinskey, A. J., Davies, R., and Labuza, T. P. (1973). J. Gen. Microbiol. 74, 267. Gomez, R. F., Blais, K. D., Herrero, A., and Sinskey, A. J. (1976). J . Gen. Microbiol. 97, 12. Gorban, I. S. (1964a). In “Cytological Aspects of Adaptation of Plants to the Environmental Factors” (V. Ya. Alexandrov et al., eds.), p. 60. Nauka, Moscow-Leningrad. Gorban, I. S. (1964b). In “The Cell and Environmental Temperature” (A. S. Troshin, ed.), p. 197. Nauka, Moscow-Leningrad. Gorban, 1. S. (1974). Tsitologiya 16, 1 1 1 1 . Gorban, I. S. (1975). Tsitologiya 17, 936. Gorban, I. S. (1980). Tsitologiya (in press). Gray, R. J. H., Witter, L. D., and Ordal, Z. J. (1973). Appl. Microbiol. 26, 78. Haight, R. D., and Morita, R. Y. (1962). J . Eacteriol. 83, 112. Haurowitz, F., Di Moia, F., and Tekman, Sh. (1952). J. Am. Chem. SOC. 74, 2265. Haynes, R. H. (1964). Photochem. Photobiol. 3, 429. Heilbronn, A. (1914). Jahrb. Wiss. Eot. 54, 357. Heilbronn, A. (1922). Jahrb. Wiss.Eot. 61, 284. Heilbrunn, L. (1925). J . Exp. Zool. 41, 243. Heilbrunn, L. (1928). Protoplasma-Monogr. Berlin 1.

CELL REPARATION OF NON-DNA INJURY

267

Heilbrunn, L. V. (1956). “The Dynamics of Living Protoplasm.” Academic Press, New York. Heine, U., Sverak, L., Kondratick, J., and Bonar, R. A. )1971). J . Ultrastruct. Res. 34, 375. Heinmets, F., Taylor, W. W., and k h a n , J. J. (1954). J. Bacreriol. 67, 5 . Hermans, J., Jr., and Acampora, G. (1967). J. Am. Chem. SOC.89, 1547. Herrlinger, F., and Kiermeier, F. (1944). Biochem. Z . 317, 1. Hurwitz, C., Rosano, C. L., and Blattberg, B. (1957). J. Bacreriol. 73, 743. Iandolo, J. J. (1974). Ann. N.Y. Acad. Sci. 236, 160. Iandolo, J. J., and Ordal, 2. J. (1966). J . Bacreriol. 91, 134. Ilyinskaya, N. B. (1960). Dokl. Akad. Nauk SSSR 133, 722. Johnson, F. H. (1957). Symp. Soc. Gen. Microbiol. 7, 134. Johnson, F. H., and Lewin, I. (1946). J. Cell. Comp. Physiol. 28, 23. Johnson, F. H., and Zobell, C. E. (1949). J. Bacreriol. 57, 353. Johnson, F. H., Kauzmann, W. J., and Gender, R. L. (1948a). Arch. Biochem. 19, 229. Johnson, F. H., Baylor, M. B., and Fraser, D. (1948b). Arch. Biochem. 19, 237. Joly, M. (1965). “A Physico-Chemical Approach to the Denaturation of Proteins.” Academic Press, New York. Kappen, L., and Lange, 0. L. (1968). Protoplasma 65, 119. Kislyuk, I. M. (1979). Photosynthetica (in press). Kohn, L. D. (1970). J . Biol. Chem. 245, 3850. Krasilnikova, V. I. (1954). Fiziol. Zh. SSSR im. 1. M . Sechenova 40, 476. Krause, E. G., Venetianer, P., and Straub, F. B. (1965). Acra Physiol. Acad. Sci. Hung. 27, 295. Kuhne, W. (1864). “Untersuchungen uber das Protoplasma und die Contractilitat.” Engelmann, Leipzig. Kushner, V. P. ( 1 977). “Conformational Variability and Denatutation of Biopolymers. ” Nauka, Moscow. Lange, 0. L. (1965). PIanta 64, 1. Lepeschkin, W. (1923). “The Constancy of the Living Substance,” Studies from the Plant Physiol. Labor., Vol. 1, p. 6 . Charles University, Prague. Levi, A. S., and Kaplan, N. 0. (1971). J. Biol. Chem. 246, 6409. Lill, U. I . , and Hartmann, G. R. (1970). Biochem. Biophys. Res. Commun. 39, 930. Lomagin, A. G. (1975). Tsirologiya 17, 1273. Lomagin, A. G. (1978). Planra 142, 123. Lomagin, A. G., and Antropova, T. A. (1966). Planta 68, 297. Lomagin, A. G., and Zheleznyak, L. I. (1973). Tsirologiya 15, 524. Lomagin, A. G., Antropova, T. A,, and Semenikhina, L. V. (1966). Planfa 71, 119. Lomagin, A. G., Antropova, T. A., and Semenikhina, L. V . (1967). Fiziol. Rast. 14, 1071. London, J., Skrzynia, C., and Goldberg, M. E. (1974). Eur. J . Biochem. 47, 409. Lutova, M. I. (1962). Bor. Zh. 47, 1761. Lutova, M. I., and Zavadskaya, I. G. (1966). Tsirologiya 8, 484. Makarov, A. A . (1977). “Some Physical Aspects of Protein Thermostability” (Author’s abstract of dissertation), Physical Dept., Moscow University. Makarov, A. A,, Esipova, N. G., Monoselidze, D. R., Mgeladze, G . N., Modzhagaladze, G. K., Kuranova, I. R., and Grebenko, A. I. (1976). Biochem. Biophys. Acra 434, 286. Matsumoto, S., and Kagami-Ishi, Y. (1970). Jpn. J . Genet. 45, 153. Meshcherskaya, K. (1935). Arkh. Biol. Nauk, Ser. B 37, 827. Michaeli, D., Pinto, J. D., and Benjamini, E. (1967). Nature (London) 213, 77. Millar, D. B., Grafins, M. A,, Wild, J. R., and Palmer, D. A. (1974). Biophys. Chem. 2, 189. Monfort, C., Ried, A., and Ried, J. (1955). Beitr. Biol. PJlanz. 31, 349. Monfort, C., Ried, A , , and Ried, J. (1957). Biol. Zenrralbl. 76, 257. Morita, R. Y., and Haight, R. D. (1962). J. Bacreriol. 83, 1341.

268

V. YA. ALEXANDROV

Morita, R. Y., and Mathemeier, P. F. (1964). J. Eacreriol. 88, 1667. Mukhejee, P., and Bhattacharjee, S. B. (1970). J . Gen. Microbiol. 60, 233. Nasonov, D. N. (1959). “Local Reactions of Protoplasm and the Spreading Irritation.” I d . Akad. Nauk SSSR, Moscow-Leningrad. Nasonov, D. N., and Alexandrov, V. Ya. (1940). “Reaction of Live Matter to External Influences.” Izd. Akad. Nauk SSSR, Moscow-Leningrad. Nassonov, D. (1932). Protoplasma 15, 239. Neale, E. K . , and Chapman, G. B. (1970). J. Eacreriol. 104, 518. Nemchinskaya, V. L., and Braun, A. D. (1962). Tsitologiya 4, 409. Neurath, H., Greenstein, J. P., Putnam, F. W., and Erickson, J. 0. (1944). Chem. Rev. 34, 157. Ogasahara, K . , Yutani, K., Imanishi, A., and Isemura, T. (1970). J. Eiochem. (Tokyo) 67, 83. Olejnikova, T. V. (1964). In “Cytological Aspects of Adaptation of Plants to the Environmental Factors” (V. Ya. Alexandrov et al., eds.), p. 70. Nauka, Moscow-Leningrad (in Russian). Oster, G., and Grimsson, H. (1949). Arch. Eiochem. 24, 119. Pierson, M. D., Tomlins, R. J., and Ordal, Z. J. (1971). J . Eacreriol. 105, 1234. Pierson, M. D., Payne, S. L., and Ades, G. L. (1974). Appl. Microbiol. 27, 425. Port, J. (1927). Protoplasma 1, 566. Preobrazhenskaya, T. A., and Shnoll, S. E. (1965). In “The Biophysics of the Cell” (G. M. Frank, ed.), p. 68. Nauka, Moscow (in Russian). Prokopova, E., and Munk, P. (1963). Collect. Czech. Chem. Commun. 28, 957. Rakhimov, G. T. (1977). Uzb. Eiol. Zh. 2, 42. Reginier, J . , and Bazin, S. (1940). C.R. Seances SOC.Eiol. Ses Fil. 133, 648. Roberts, C. T., and Orias, E. (1974). J . Cell Eiol. 62, 707. Ron, E. Z., and Davis, B. D. (1971). J. Eacreriol. 107, 391. Ron, E. Z., and Shani, M. (1971). J . Bacreriol. 107, 397. Rosenthal, L. J., Martin, S. E., Pariza, M., and Iandolo, J. J. (1972). J. Eacreriol. 109, 243. Sachs, T. (1864). Flora (Jena) 47, 5-12; 33-39, and 65. Samoilenko, I. I., and Ivanov, K . K. (1972). Zh. Mikrobiol., Epidemiol. Imrnunobiol. 49, 113. Sedgwick, S. G., and Bridges, B. A. (1972). J. Gen. Microbiol. 71, 191. Semikhatova, 0. A. (1964). In “Cytological Aspects of Adaptation of Plants to the Environmental Factors” (V. Ya. Alexandrov er al., eds.), p. 5 5 . Nauka, Moscow-Leningrad (in Russian). Semikhatova, 0. A., and Egorova, L. I. (1976). Eor. Zh. 61, 313. Semikhatova, 0. A,, Saakov, V. S., and Gorbacheva, G. I. (1962). Tr. Eor. Insr., Akad. Nauk SSSR, Ser. 4 15, 25. Simard, R., and Bemhard, W. (1967). J . Cell Eiol. 34, 61. Simard, R., Amalric, F., and Zalta, J.-P. (1969). Exp. Cell Res. 55, 359. Singer, M., and Morrison, R. L. (1948). J. Eiol. Chem. 175, 133. Skholl, E. D. (1963). In “Problems of Cytoecology of Animals’’ (B. P. Ushakov, ed.), p. 220. Nauka, Moscow-Leningrad (in Russian). Skholl, E. D. (1965). In “Heat Resistance of Cells of Animals” (B. P. Ushakov er al., eds.), p. 106. Nauka, Moscow-Leningrad (in Russian). Sogin, S. J., and Ordal, Z. J. (1967). J. Eacreriol. 94, 1082. Stabrovskaya, V. I. (1967). Tsitologiya. 9, 719. Steiner, D. F. (1961). Eiochim. Eiophys. Acru 54, 206. Steiner, R. F., De Lorenzo, F., and Anfinsen, C. B. (1965). J . Eiol. Chem. 240, 4648. Stiles, M. E., and Witter, L. D. (1965). J. Dairy Sci. 48, 677. Strehler, B. L . , and Johnson, F. H. (1954). Proc. Narl. Acad. Sci. U . S . A . 40, 606. Suzdalskaya, I. P. (1966). Tsitologiya 8, 621. Suzdalskaya, I. P., and Zander, N. V. (1963). Fiziol. Zh. SSSR im. I . M . Sechenova 49, 249. Swartz, M. N., Kaplan, N. O., and Lamborg, M. F. (1958). J . Eiol. Chem. 232, 1051.

CELL REPARATION OF NON-DNA INJURY

269

Tamura, Y., and Morita, Y. (1975). J. Eiochem. (Tokyo) 78, 561. Tanford, C. (1968). Adv. Protein Chem. 23, 187. Teipel, J . W. (1972). Eiochemisrry 11, 4100. Teipel, J . W., and Koshland, D. E. (1971). Biochemistry 10, 792. Terminiello, L., Bier, M., and Nord, F. F. (1958). Arch. Biochem. Eiophys. 73, 171. Thomer, W. (1919). Z . Alleg. Physiof. 18, 226. Tomlins, R. I . , and Ordal, Z. J. (1971a). J . Eacteriol. 105, 512. Tomlins, R. I . , and Ordal, 2. J. (1971b). J . Eacteriol. 107, 134. Tomlins, R. I . , Pierson, M. D., and Ordal, Z. J. (1971). Can. J. Microbiol. 17, 759. Tongur, V. S. (1951). Usp. Sovrem. Biol. 31, 391. Ushakov, B. P., and Glushankova, M. A. (1961). Tsitologiya 3, 707. Ushakov, V. B., and Vasilyeva, V. V. (1965). In “The Biophysics of the Cell” (G. M. Frank, ed.), p. 13 1. Nauka, Moscow. Vandenheede, J. R., Keppens, S., and de Wulf, H. (1977). Eiochim. Eiophys. Acra 481, 463. Van Halteren, P. (1950). Bull. SOC.Chim. Eiol. 32, 458. Van-Herwerden, M. (1927). Protopfasma 2, 271. Venetianer, P., and Straub, F. B. (1963a). Eiochim. Eiophys. Acra 67, 166. Venetianer, P., and Straub, F. B. (1963b). Acta Physiof. Arad. Sci. Hung. 24, 41. Venetianer, P., and Straub, F. B. (1964). Eiochim. Eiophys. Acta 89, 189. Venetianer, P., and Straub, F. B. (1965). Acta Physiol. Acad. Sci. Hung. 27, 303. Vinogradova, A. N., and Dzhamusova, T. A. (1963). Tsirologiya 5, 279. Waley, S. G. (1973). Eiochem. J . 135, 165. Wetlaufer, D. B., and Ristow, S . (1973). Annu. Rev. Eiochem. 42, 135. Witkin, E. M. (1976). Eacreriol. Rev. 40, 869. Woodcock, E., and Grigg, G. W. (1972). Nature (London), New Eiol. 237, 76. Wunderlich, F., and Peyk, D. (1969). Exp. Cell Res. 57, 142. Yamamoto, T., Nishida, A., and Fukumoto, J. (1964). Agric. Eiol. Chem. 28, 656. Yutani, K., Yutani, A,, and Isemura, T. (1967). J. Eiochem. (Tokyo) 62, 576. Yutani, K . , Yutani, A,, and Isemura, T. (1969). J . Eiochem. (Tokyo) 66, 823. Zavadskaya, I. G. (1963). Tsitologiya 5, 151. Zavadskaya, I. G., and Antropova, T. A. (1979). Eor. Zh. 64, 950. Zimmerman, L. B.. and Coleman, N. F. (1971). J. Eiol. Chem. 246, 309.

This Page Intentionally Left Blank

.

INTERNATIONAL REVIEW OF CYTOLOGY VOL . 60

Ultrastructure of the Carotid Body in the Mammals ALAINVERNA Laboratory of Cytology. University of Bordeaux II. Talence. France I . Introduction . . . . . . . . . . . . . . . . . . . . I1 . Histological Features . . . . . . . . . . . . . . . . . A . Cell Clusters . . . . . . . . . . . . . . . . . . B . Blood Vessels . . . . . . . . . . . . . . . . . . C . Nerve Fibers . . . . . . . . . . . . . . . . . . D . Ganglion Cells . . . . . . . . . . . . . . . . . E . Other Cellular Elements . . . . . . . . . . . . . . 111. Ultrastructure of Type I and Type I1 Cells . . . . . . . . . A . Type I Cells . . . . . . . . . . . . . . . . . . B . TypeIICells . . . . . . . . . . . . . . . . . . IV . Type I Cell Innervation . . . . . . . . . . . . . . . . A . Origin of Nerve Endings on Type I Cells . . . . . . . . B . Ultrastructure of Nerve Endings on Type I Cells . . . . . C . Ultrastructure of Junctions between Nerve Endings and Type 1 Cells . . . . . . . . . . . . . . . . . . . . . D . Effects of Denervation upon Type ]/Type I1 Cells . . . . . E . Functional Interpretations of Type 1 Cell-Nerve Ending Relationships . . . . . . . . . . . . . . . . . . F . Noninnervated Type I Cells . . . . . . . . . . . . . V . Vascular Innervation and Efferent Inhibition . . . . . . . . A . Vasomotor Innervation . . . . . . . . . . . . . . . B . Barosensory Innervation . . . . . . . . . . . . . . C . The Problem of the Efferent Inhibition . . . . . . . . . . VI . Ultrastructural Changes after Stimulation of Chemoreceptor and Pathology . . . . . . . . . . . . . . . . . . . . . A . Changes in Type I Cells . . . . . . . . . . . . . . B . Changes in Nerve Endings . . . . . . . . . . . . . C . Pathology . . . . . . . . . . . . . . . . . . . VII . Embryology and Development . . . . . . . . . . . . . VIII . Concluding Remarks . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . .

295 300 301 307 307 308 312 313 315 315 318 318 320 322 323

.

I Introduction The carotid body of mammals is a very small organ that lies dorsal to the common carotid artery bifurcation . Its maximum dimension averages approximately 3 mm in man and 1.5 mm in the cat or dog . It receives arterial blood from a short vessel but its exact location varies greatly not only from one species to 27 1

Copyright 0 1979 by Academic Press. Inc . All rights of reproduction in any form reserved . ISBN 0-12-3643WO

272

ALAIN VERNA

another but also from one individual of a given species to another (see Adams, 1958; Seidl, 1975). The carotid body is innervated by a branch of the glossopharyngeal nerve: Hering’s nerve or the sinus nerve. The latter denomination is due to the fact that the carotid body lies close to the carotid sinus and that both structures are innervated by the same nerve. The carotid body is also connected to the superior cervical ganglion by one or several nerves: the ganglioglomerular nerve(s) (Fig. 1). Since its discovery, the carotid body has been the subject of many controversies. It was considered as a gland by Luschka (1862) and then, as a vascular organ by Arnold (1865). The occurrence of chromaffin cells, described by Kohn (1900), led him to consider the carotid body as a paraganglion, a concept which was corroborated by the existence of a sympathetic innervation. However, some authors underlined that chromaffin cells are rather few in the carotid body (Monckeberg, 1905). What is more, de Castro (1926) demonstrated that carotid body cells are principally innervated by the glossopharyngeal nerve. Until then, this innervation was thought to be efferent (centrifugal), but in 1928, de Castro showed that transection of the glossopharyngeal nerve root (between its sensory ganglions and the medulla) does not affect the innervation of carotid body cells. This experiment demonstrated that this innervation is afferent (sensory) from IX

la

ica

SI

cca

cca

A

B

FIG.1. Schematic representation of carotid bifurcation in the (A) cat and the (B) rabbit (ventral view of the left carotid). ap, ascending pharyngeal artery; cb, carotid body; cca, common carotid artery; cs, carotid sinus; eca, external carotid artery; ica, internal carotid artery; ggn, ganglioglomerular nerve; la, lingual artery; oa, occipital artery; scg, superior cervical ganglion; sla, superior laryngeal artery; sn, sinus nerve; IX, glossopharyngeal nerve.

ULTRASTRUCTURE OF THE CAROTID BODY

273

which de Castro (1928) deduced that the carotid body is a sensory organ which “tastes” the blood. This prophetical interpretation was confirmed physiologically by Heymans and co-workers (Heymans and Bouckaert, 1930; Heymans et al., 1933), who found that the carotid body is an arterial chemoreceptor which is sensitive to the chemical composition of the blood and which is particularly stimulated by hypoxia, hypercapnia, and acidemia. Its excitation leads to respiratory and cardiovascular reflexes (see Heymans and Neil, 1958; Angell-James and Daly, 1969; Biscoe, 1971; Howe and Neil, 1971). Arterial chemoreceptors are not only restricted to the carotid body. They are also present in the aortic arch region where there are several groups of cells referred to as “aortic bodies.” Their chemosensory function is well known (see Howe and Neil, 1971, for references) and it has been recently demonstrated that their ultrastructure is identical to that of the carotid body (Abbott and Howe, 1972). Moreover, “miniglomera’ ’ have been found in the carotid bifurcation region at some distance from the carotid body itself (de Castro, 1962) and even around the common carotid artery of the cat (Matsuura, 1973). In the latter case, both physiological and ultrastructural features of “miniglomera” were found to be identical to those of the carotid body. It seems, consequently, legitimate to extrapolate results concerning the carotid body to the whole arterial chemoreceptor system. The carotid body ultrastructure was studied as early as 1957 and the first works demonstrated the occurrence of osmiophilic vesicles in a category of cells (Lever and Boyd, 1957; Lever et al., 1959). The electron microscope also demonstrated nerve endings on these cells (type I cells) but, at the same time, showed that these terminals do not penetrate the type I cell cytoplasm as suggested by de Castro (1951). Furthermore, Lever et al. (1959) observed synaptic-like vesicles in these nerve endings. This observation was surprising since synaptic vesicles are, classically, located in the presynaptic part of the junctions, whereas, according to de Castro (1928), the nerve endings in question should be postsynaptic (the presynaptic element being the type I cell). The occurrence of synaptic-like vesicles in nerve endings on type I cells has been repeatedly confirmed by many authors and that is probably what led Biscoe and Stehbens (1967) to express doubts about de Castro’s conclusions and, then, Biscoe et al. (1970) to repeat his degeneration experiment. The results of Biscoe et al. (1970) were diametrically opposed to those of de Castro (1928) and, consequently, these authors asserted that type I cell innervation is efferent, a conclusion which implies that type I cells are not sensory. However, some authors did not accept this conclusion (Eyzaguirre et al., 1972); others postulated a double innervation of type I cells, efferent and afferent (Kobayashi, 1971b). Finally, more recent experiments have demonstrated that de Castro (1928) was right in asserting the afferent (sensory) nature of type I cell innervation (Fidone et al., 1975, 1977; Smith and Mills, 1976).

274

ALAIN VERNA

However, the interpretation of the relations between type I cells and sensory nerve terminals is still in dispute and the transducer element cannot be considerbd as identified. Nevertheless, the ultrastructural descriptions are relatively concordant despite these different interpretations. This is not the case for the effects of experimental stimulation of chemoreceptors on carotid body ultrastructure. For example, some authors report that chronic hypoxia increases the number of osmiophilic vesicles in type I cells (Mgller et al., 1974), while others describe the inverse effect (Blessing and Kaldeweide, 1975; Laidler and Kay, 1978). The embryology of the carotid body also has been a subject of great controversy, particularly with respect to the origin of the type I cells, but this problem has been solved recently (Le Douarin et al., 1972; Pearse er al., 1973). Thus some very important results have been obtained in the course of the last 6 years. The object of this review is to present the current status of our knowledge concerning the structure of the carotid body and to discuss the possible functional implications of morphological observations.

11. Histological Features Histologically, the carotid body is characterized by an association between cell clusters and capillaries. These elements are invested by a collagenous connective tissue, more or less abundant according to the species. Many myelinated and unmyelinated nerve fibers travel in the connective tissue. Where the connective tissue is abundant the cell clusters are dispersed and the organ is described as “diffuse” or “disseminated” (Kohn, 1900; Watzka, 1943). This is the case in the rabbit, for example (Fig. 2). In other species, the interstitial tissue is minimal and the carotid body is described as “compact” (cat). However, this distinction seems to be of minor importance and, nevertheless, the amount of connecthe tissue increases with age (Watzka, 1943). Frequently, several cell clusters constitute a lobule which is supplied by one and the same arteriole (Adams, 1958). The cell cluster and its associated capillary are considered by Seidl(l975) as the basic functional unit or “glomoid. ”

A. CELLCLUSTERS

The cell clusters are made of a varying number of cells. The aspect of their nuclei enabled Gomez (1908) to describe two kinds of cells: type I and type 11. Type I cells exhibit a spherical or ovoid nucleus having little affinity for stains. Type I1 cells show a smaller nucleus which is frequently flattened or reniforrn and contains a condensed chromatin; these cells are less numerous than type I cells and are situated at the periphery of the clusters (Fig. 2). This classification

ULTRASTRUCTURE OF THE CAROTlD BODY

215

FIG. 2. Semithin section of rabbit carotid body showing numerous capillaries (Ca), several cell clusters (arrows), and nerve fibers, either in small bundles (Bnf) or isolated (Nf). The collagenous connective tissue (Col) is abundant in this species. Bar = 20 p m . Inset: Higher magnification showing the aspect of the nuclei of type 1 cells (NI) and type I1 cells (NII). Bar = 5 pm.

276

ALAIN VERNA

of carotid body cells was used again by de Kock and Dunn (1966) and also by Biscoe and Stehbens (1966) in the course of ultrastructural studies which revealed other important differences between these cells (see Section 111). However, other denominations are currently in use, e.g., type I and type I1 cells are called “glomus cells” and “sustentacular cells, respectively (Nishi and Stensaas, 1974), or “glomus cells” and “sheath cells” (McDonald and Mitchell, 1975a,b). Other names were used in the past (a list can be found in Biscoe, 1971). The cell islands are always close to a blood vessel and, by means of three-dimensional reconstructions, Seidl (1975) has shown that the cells are clustered on one side of the vessel. The distance between cell clusters and capillaries is variable and has been the subject of many contradictory estimations. This is not surprising since this distance is frequently below the resolving power of the light microscope, as shown later with the electron microscope. A statistical analysis of this problem has been recently undertaken (Lubbers et a l . , 1977). ”

B. BLOODVESSELS 1 . Arteries The carotid body is usually supplied by one artery originating from the carotid bifurcation area or from the occipitopharyngeal trunk in the cat (the internal carotid artery is vestigial in this species). However, considerable variations occur, even in a given species, e.g., Seidl (1975) has observed cat carotid bodies supplied by more than three arteries. Moreover, half of the carotid bodies studied by this author were supplied by arteries originating from the external carotid artery, occipital artery, ascendant pharyngeal artery, or common carotid artery. Frequently, the glomic artery (arteries), after providing branches to the carotid body, leaves the organ to supply neighboring structures such as the wall of the carotid sinus or the superior cervical ganglion (Chungcharoen et a l . , 1952). The histological structure of glomic arteries is unusual. These vessels have a very flexible wall, made of concentric elastic laminae surrounded by only a few layers of smooth muscle fibers (de Castro and Rubio, 1968; Heath and Edwards, 1971). According to de Castro (1940) the carotid body arteries are innervated by barosensory nerve fibers (see Section IV,B).

2. Capillaries The capillary network of the carotid body has been clearly demonstrated by means of gelatin injections (de Castro, 1940, 1951) and, more recently, by scanning electron micrographs of cast preparations (Keller et a l . , 1972; Seidl, 1975). It must be underlined that the vessels in question are not sinusoids but are true capillaries with a continuous (although fenestrated) endothelium. De Castro and Rubio (1968) have described two kinds of capillaries: Type I capillaries are

ULTRASTRUCTURE OF THE CAROTID BODY

277

very convoluted vessels 14 to 28 pm in diameter; type I1 capillaries are less numerous, have a diameter of 6 to 12 pm,and tend to constitute bridges between type I capillaries. This distinction has not been confirmed by other workers. On the other hand, de Castro and Rubio (1968) have noted that the communication between capillaries and venules is achieved through narrow postcapillary channels, and Seidl (1975) also described “vessels having a large lumen suddenly narrow prior to merging into the sinus. This situation may involve great variations in the blood velocity. Pericytes are usually present around carotid body capillaries. These cells have been frequently misinterpreted in the past (see Adams, 1958). ”

3. Arteriovenous Anastomoses The occurrence of arteriovenous anastomoses in the carotid body has been postulated by many authors (de Castro, 1940, 1951; de Boissezon, 1943, Celestino de Costa, 1944; Serafini-Fracassini and Volpin, 1966; de Castro and Rubio, 1968; Schafer et al., 1973; Seidl, 1975). According to de Castro and Rubio (1968), the arterial segment of these anastomoses is provided with baroreceptor nerve endings and these authors suggest that arteriovenous shunts regulate the blood flow through the capillaries by bypassing the blood directly to the veins under certain circumstances. Schafer et al. (1973) and Seidl(1975) also describe shunt vessels outside the carotid body, at its arterial pole. They distinguish two kinds of arteriovenous anastomoses (bridge anastomoses and spiral anastomoses). Unfortunately, in a more recent study, Seidl (1976) was unable to verify the presence of these structures. However, this author observed several cases where the artery and vein approached each other up to within a few micrometers. It must be added that other authors were not convinced that arteriovenous anastomoses actually exist in or around the carotid body (Hollinshead, 1942; Edwards, discussion on the paper of de Castro and Rubio, 1968). 4. Veins The carotid body venules form a superficial plexus from which several veins leave the organ to join neighboring venous trunks. The histology of this venous system has not been the object of special comments, except the possible occurrence of baroreceptor nerve endings in the wall of the veins described by Abraham (1 968). C. NERVEFIBERS The abundance of nerve fibers is the third histological characteristic of the carotid body. .These fibers originate from the sinus nerve, the ganglioglomerular nerve(s), or intrinsic ganglion cells.

278

ALAIN VERNA

1 . Fiber Content of Nerves

a. The sinus nerve of the cat contains about 600 to 700 myelinated nerve fibers and a greater number of unmyelinated ones (de Castro, 1951). According to Eyzaguirre and Uchizono (1961) the ratio of unmyelinated to myelinated fibers varies along the nerve: It is greater near the carotid body than at the middle of the nerve. The authors suggest that a branch having a larger proportion of myelinated axons leaves the sinus nerve some distance before the carotid body to innervate the carotid sinus. Similar observations are made by Laurent and Banes (1964) who do not find unmyelinated fibers in the rabbit sinus nerve near its junction with the glossopharyngeal nerve, whereas such fibers appear numerous near the carotid body. However, these authors propose another explanation, namely, that unmyelinated fibers in the sinus nerve are terminal parts of myelinated ones. The diameters of the myelinated fibers range approximately from 1 to 10 pm in the cat and the rabbit (Eyzaguirre and Uchizono, 1961; Laurent and Banes, 1964). In the latter species, the diameter distribution curve is bimodal with a peak around 2 to 3 p m and another between 5 and 8 pm (Laurent and Barres, 1964; A. Verna, unpublished observations). Electrophysiological methods enabled Fidone and Sat0 (1969) to estimate that, among these myelinated fibers, approximately two-thirds are chemoreceptor afferents and one-third are baroreceptor afferents. The unmyelinated fibers of the sinus nerve have diameters ranging from 0.1 to 1.3 pm with a unimodal distribution (Eyzaguirre and Uchizono, 1961). According to Fidone and Sat0 (1969) about half of these fibers are sensory, two-thirds being baroreceptor afferents and one-third chemoreceptor afferents. The other half are efferent fibers of sympathetic origin or of central origin (see Section V). b. The ganglioglomerular nerve is essentially made of unmyelinated nerve fibers and contains only a few myelinated axons. The unmyelinated fibers have diameters between 0.1 and 2.0 p m and a possible bimodal distribution has been suggested (Eyzaguirre and Uchizono, 1961). 2. Nerve Fibers in the Carotid Body The nerve fibers are grouped in bundles at the organ periphery; they are more dispersed in the central regions and unmyelinated fibers are much more numerous than myelinated ones. Occasionally, myelinated nerve fibers loosing their myelin sheath have been observed in the carotid body (Kondo, 1971; Verna, 1975). The nerve fibers are accompanied by Schwann cells. However, when nerve fibers terminate on type I cells, Schwann cells are relayed by type I1 cells and it is frequently difficult to distinguish these two types of cells. With methylene blue or silver impregnation techniques, de Castro (1940) showed that a single fiber may innervate several cell clusters. Conversely, some type I cells receive terminals from different fibers (see Section IV). Other fibers go to the blood vessels and will be considered in Section V.

ULTRASTRUCTURE OF THE CAROTID BODY

279

D. GANGLION CELLS The occurrence of some ganglion cells in the carotid body has been known since the first histological studies (Kohn, 1900). These neurons are usually few and are located at the organ periphery. According to Biscoe and Silver (1966), there are three to five ganglion cells per carotid body in the cat. However, these cells seem to be more numerous in some mammals (as the hedgehog) than in others (Adams, 1958). They are generally multipolar neurons, of greater size than type I cells, frequently isolated or sometimes grouped in microganglia. Cell clusters made of type Ihype I1 cells and ganglion cells have been described in the “pilot whale” (de Kock, 1956) and in the rat (Kondo, 1976) but his arrangement appears rather exceptional. It has been suggested that carotid body ganglion cells may be of two kinds: orthosympathetic and parasympathetic (Smith, 1924; Watzka, 1943; McDonald and Mitchell, 1975b). Their function is probably related to the vasomotor innervation and will be considered in Section V.

E. OTHERCELLULAR ELEMENTS 1. Connective Cells Connective cells are numerous in the carotid body, particularly in its disseminated type. These cells form a thin capsule (sometimes inconspicuous) around the organ and its interstitial stroma. The latter contains many collagenous fibers but practically no elastic fibers. 2. Must Cells Mention must be made of the frequent occurrence of mast cells in the carotid body. These cells may be responsible, at least in part, for the serotonin content reported by some authors (Chiocchio et a l . , 1967, 1971a). 111. Ultrastructure of Type I and Type I1 Cells

The ultrastructure of carotid body-specific cells has been the object of many studies and a list of works prior to 1969 can be found in Biscoe (1971) and Kobayashi (1971b). Many species were used during these studies, in particular the cat. However, it seems there are no remarkable species differences with regard to the ultrastructure of type I and type I1 cells. A. TYPEI CELLS 1. General Description

Type I cells are enveloped by type I1 cells except in some places where the type I cell membrane is separated from extracellular spaces by a basement mem-

280

ALAIN VERNA

brane only. Type I cells are of complex chape and a section in a cell cluster shows very intricate cytoplasmic portions (Fig. 3). These cells frequently send cytoplasmic processes toward other cell clusters or toward capillaries (Nishi, 1976). Because of this complex shape, the size of type I cells is difficult to determine; it can be said that the perikaryon measures about 10 pm. The nucleus is usually spherical or ovoid and contains a dispersed chromatin of low electron density. The nuclear envelope shows numerous pores. The cytoplasmic organelles are evenly distributed and, consequently, the cell shows no polarity. There are many mitochondria with transversal cristae. However, according to Seidl et al. (1977), the percentage of the cytoplasm volume occupied by mitochondria is considerably smaller in type I cells than in liver cells. This is surprising since the oxygen consumption of the carotid body is high (Daly et al., 1954; Leitner and Liaubet, 1971; h r v e s , 1970). The Golgi apparatus is well developed and different kinds of vesicles are present around it: clear smooth vesicles and coated vesicles, some of which contain an electrondense material (Fig. 4). The granular endoplasmic reticulum is usually dispersed in the cytoplasm but sometimes has an arrangement similar to that of the Nissl body of neurons. Free ribosomes, singly or associated in polysomes, are numerous. The centrioles show the usual structure and frequently give rise to a cilium of the 9 + 0 pattern. This kind of cilium has been observed not only in carotid body type I cells (Hess, 1968; Kobayashi, 1968; Kondo, 1971) and type I1 cells (Hess, 1968) but also in many other tissues. These cilia appear, consequently, nonspecific to carotid body cells and their function, if any, is probably not related to the chemoreception. Many microtubules are also present in the cytoplasm. They are intermingled in the perikaryon but show a parallel arrangement in the fingerlike processes; this often makes it difficult to distinguish between type I cell processes and nerve fibers. Other organelles and inclusions are occasionally visible in the cytoplasm: multivesicular bodies, lysosomes, pinocytotic vesicles, lipid droplets, a few glycogen particles. Mention must be made also of some small clear vesicles, similar to synaptic vesicles. These vesicles (about 60 nm in size) are preferentially located below the plasma membrane when the latter is exposed to extracellular spaces (without a type I1 cell cover) (Fig. 5 ) ; they are also involved in some junctions between type I cells and nerve endings (see Section IV,C,2). The content of these synaptic-like vesicles is unknown, but McDonald and Mitchell (1975a) have shown that about 40% of the synaptic-like vesicles in type I cells have a dense core after 5-hydroxydopamine administration. This result suggests that some of these vesicles may store a catecholamine. Finally, the most prominent feature of type I cells is the presence of numerous electron-dense membrane-bound granules, the so-called dense-cored vesicles. These vesicles, described for the first time in the carotid body by Lever and Boyd

ULTRASTRUCTURE OF THE CAROTID BODY

28 1

FIG. 3 . Electron micrograph of rabbit carotid body. This section of a cell cluster shows several type I cells (CI) and type I1 cells (CII). Dense-cored vesicles are abundant in certain regions of the type I cell cytoplasm (arrows). The cell cluster is very close to a capillary (Ca). Col, Collagen; Nf, unmyelinated nerve fiber; NE, nerve ending. Bar = 2 pm.

282

ALAIN VERNA

FIG.4. Type I cell cytoplasm (rabbit carotid body). Section at the level of the Golgi apparatus ( G ) showing many cytoplasmic components: M, mitochondria; R, granular endoplasmic reticulum; ri, polyribosomes; mv, multivesicular bodies; ly , lysosome; dcv, dense-cored vesicles, mt, microtubules; PI,plasma membrane. Arrowheads indicate coated vesicles, either “empty” or containing a dense material. Bar = 1 pm.

ULTRASTRUCTURE OF THE CAROTID BODY

283

FIG.5 . A type I cell process (CI) is only separated from collagenous connective tissue (Col) by a basement membrane (Bm) (rabbit carotid body). At this level, numerous small synaptic-like vesicles (V) accumulate beneath the plasma membrane. Structures similar to presynaptic dense projections are also present (arrowheads). M, mitochondria. Bar = 0.5 p m .

(1957), are usually spherical, more rarely irregular, in shape; their size is very variable (from about 60 to 300 or 400 nm) not only from cell to cell but also within a given cell. There is always an electron-lucent space between the dense core and the membrane; this space is of variable size and, thus, the vesicles appear more or less “full.” The dense core is of variable electron opacity and, if sufficiently light, shows a faintly granular substructure. An extensive descriptive study of the dense-cored vesicle morphology has been published by Matthiessen et al. (1973) with respect to fixation conditions. The amount of dense-cored vesicles varies considerably from cell to cell. Usually, dense-cored vesicles are evenly distributed in the cytoplasm; however, they are sometimes (particularly in the mouse carotid body) more abundant just below the plasma membrane. Some authors have published exocytosis pictures (Bliimcke et al., 1967a; Bock et al., 1970; Bock and Gorgas, 1976b), but it seems that such pictures are very uncommon for the carotid body.

2 . The Content of Dense-Cored Vesicles As early as their discovery, type I cells dense-cored vesicles were compared to those of the adrenal medulla by Lever and Boyd (1957), and some years later,

284

ALAIN VERNA

Lever et al. (1959) wrote: “possibly, the electron-dense membrane bound granular bodies. . . represent (as in the adrenal medulla) a stored form of some catecholamine. This conclusion was based on the demonstration (with the light microscope) of a faint positive chromaffin reaction in type I cells and on the effects of reserpine which causes the dense cores of vesicles to disappear. However, this result was, in part, an artifact since the aspect of dense-cored vesicles depends on the fixation method, as shown later by Chen et al. (1969): After reserpine, most dense-cored vesicles appear “empty ” after osmium tetroxide fixation (method used by Lever et al., 1959) but show a normal dense content after glutaraldehyde-osmium tetroxide fixation. Nevertheless, type I cells contain catecholamines andor indoleamines as shown unequivocally by the formolinduced fluorescence method (for references, see Kobayashi, 1971a; Bock and Gorgas, 1976b), but the intracellular localization of monoamines cannot be studied with the light microscope. The first evidence of the presence of catecholamines in type I cell dense-cored vesicles was given by Chen and Yates (1969). These authors demonstrated, by high-resolution autoradiography, that after administration of labeled catecholamine precursors there are many silver grains over type I cells, associated mainly with dense-cored vesicles. The same authors also demonstrated a positive reaction in dense cores after glutaraldehyde-potassium dichromate incubation (method of Wood and Barnett, 1964), i.e., a chromaffin reaction at the dense-cored vesicle level. A more direct argument has been given by Lishajko (1970) who has shown that isolated densecored vesicles (from a human carotid body tumor) release and take up dopamine. Finally, type I cell dense-cored vesicles seem to incorporate “false precursors of catecholamines such as 5-hydroxydopa (Hellstrom, 1975b) and 6-hydroxydopamine (Hess, 1976), and they undergo some shrinkage after catecholamine depletion by reserpine (Hess, 1977b). Obviously, these results do not prove that all type I cell dense-cored vesicles contain catecholamines. It is well known that many different secretion products are stored in similar dense-cored vesicles and it has been demonstrated that, even in a central monoaminergic neuron, some dense-cored vesicles contain substances other than monoamines, such as acid phosphatase for example (Sotelo, 1971). It is possible that, in the type I cells as well, some vesicles do not contain catecholamines. However, they, probably, are few according to the results of ultracytochemical studies. On the other hand, dense-cored vesicles contain not only catecholamines, but also proteins, ATP, and calcium. The proteins are probably responsible for the electron opacity of the dense cores; this explains why, after reserpine, catecholamine-depleted vesicles may have an electrondense content. The presence of proteins in type I cell vesicles has been demonstrated by means of Pronase digestion (Chen et al., 1969). This protein content may be considered as a binding substance for catecholamines but, for some authors, it is a secretion product per se (Capella and Solcia, 1971; Pearse, 1969). ”

”

285

ULTRASTRUCTURE OF THE CAROTID BODY

However, this hypothetical secretory polypeptide (called ‘‘glomin by Pearse, 1969) has not been identified yet. Furthermore, an autoradiographic study using tritiated leucine, monoamines, and ATP has shown that the turnover of these products in type I cells is slow by comparison with endocrine cells of the adrenal medulla or of the gut (Kobayashi, 1976, 1977). Another problem is to specify what kind of monoamine is present in a given type I cell. Unfortunately, the cytochemical identification of monoamines is not easy and species differences are an additional confusing factor. To summarize, both cytochemical and biochemical studies have demonstrated the presence of large amounts of dopamine and, to a lesser extent, of noradrenaline in the carotid body of many mammals (Dearnaley et al., 1968; Knoche et al., 1969; Zapata et al., 1969; Lishajko, 1970; Chiocchio et al., 1971a,b; Hellstrom and Koslow, 1976). On the other hand, results concerning adrenaline and serotonin are more conflicting. Thus, we question if these different monoamines are stored in different cells, and this leads us to consider a possible classification of type I cells. ”

3 . Different Kinds of Type I Cells Until now, we have considered type I cells as a homogenous population. However, several authors have attempted to classify these cells on the basis of ultrastructural or cytochemical features or both. a. Light and Dark Cells. In early studies (Garner and Duncan, 1958; Lever et al., 1959) light-dark variations of the electron opacity of the type I cell cytoplasm were noticed in the rabbit and, to a lesser extent, in the cat carotid body. This observation has been confirmed for other species such as man (Grimley and Glenner, 1968), horse (Hoglund, 1967), monkey (Al-Lami and Murray, 1968b), and, again, cat (Morita et al., 1969) and rabbit (A. Verna, unpublished observations). However, only electron-dense cells have been found in rat and mouse carotid bodies (Bock and Gorgas, 1976b). It must be added that the ‘‘light’’ or “dark” appearance of cytoplasms is not restricted to carotid body cells but has also been observed in many other cells, e.g., the adrenal medulla. However, the significance of this background density variation is obscure. For many authors it is a fixation artifact (Benedeczky and Smith, 1972; Wacker and Forssmann, 1972); for others it corresponds to different levels of activity (Garner and Duncan, 1958). Nevertheless, this criterion seems to be insufficient to subdivide type I cells. b. ChromafSin and Nonchromaffin Type I Cells. Chromaffinity in the carotid body has been a matter of controversy for years because of the contradictory results reported by many authors (see Adams, 1958; Kobayashi, 1971a). Nowadays, these controversies are only of historical interest since there is no doubt that type I cells contain catecholamines, despite the fact that they are generally chromaffin negative. However, it appears that a few cells react positively to the chromaffin reaction, at least in the carotid body of the dog (Kobayashi, 1968)

286

ALAIN VERNA

and man (Kobayashi, 1971a). At first sight, it would seem that nonchromaffin cells contain monoamines in amounts too small to react positively; in this interpretation it is not necessary to suppose that chromaffin and nonchromaffin cells are distinct cell types. But it is also possible that chromaffin and nonchromaffin cells contain different monoamines and, in this case, they may be considered to be different cells. This interpretation has been postulated by Kobayashi (1968) who described, in the dog carotid body, very distinct chromaffin and nonchromaffin cells and supposed that (but without experimental support) they contain noradrenaline and adrenaline, respectively. These cells also showed other ultrastructural differences, particularly with respect to their dense-cored vesicles: Those of nonchromaffin cells were of moderate electron density and measured about 100 to 200 nm, whereas chromaffin cell dense-cored vesicles were very electron opaque and measured up to 300 nm. It must be mentioned here that, in the rabbit carotid body as well, a few cells are characterized by very large and very electron-opaque dense-cored vesicles; now these cells take up tritiated noradrenaline whereas the other cells do not (A. Verna, unpublished observations). This observation is therefore consistent with the suggestion of Kobayashi on the possible presence of noradrenaline in chromaffin type I cells. Furthermore, dopamine P-hydroxylase (the enzyme that catalyzes the conversion of dopamine to noradrenaline) has been identified in the denervated cat carotid body by Belmonte et al. (1977). This result is consistent with the observations of Morita et al. (1969) who have described, in the cat carotid body, a few cells characterized by their very large (170 to 400 nm) and very opaque dense-cored vesicles; the possibility that these cells may contain noradrenaline was evoked by these authors and confirmed later by Bock and Gorgas (1976b). The scarcity of such cells is consistent with the relatively low levels of dopamine P-hydroxylase found by Belmonte et al. (1977). Thus, it may be considered, as a working hypothesis, that most type I cells contain dopamine (which does not give a positive chromaffin reaction under the light microscope), whereas a few, characterized by larger dense-cored vesicles, contain noradrenaline; the latter cells are chromaffin positive but their irregular occurrence would explain the contradictory results obtained by different authors on different species. Are chromaffin and nonchromaffin cells really distinct cell types? It is difficult to answer this question since it has been suggested, on cytochemical grounds, that intermediate forms (containing a mixture of noradrenaline and dopamine) also exist in the cat carotid body (Chiocchio et al., 1971b; Bock and Gorgas, 1976b). It may be useful to consider here the morphometric analysis of dense-cored vesicles since chromaffin and nonchromaffin cells also differ by the size of their dense-cored vesicles. c. Morphometric Analysis of Dense-Cored Vesicles. The first attempt was that of Morita et al. (1969) who have described, in the cat carotid body, “light” and “dark” type I cells, the latter group being further divided into three sub-

ULTRASTRUCTURE OF THE CAROTID BODY

287

types, according to the mean diameter of their dense-cored vesicles. Unfortunately, the authors give neither the number of cells nor the number of vesicles which have been used to calculate each mean. The results of Hellstrom (1975a) and McDonald and Mitchell (1975a,b) appear more convincing at fiist sight. They describe, in the rat carotid body, two kinds of type I cells which differ significantly in the mean diameter of their dense-cored vesicles. According to Hellstrom (1975a) “small-vesicle cells” and “large-vesicle cells” have densecored vesicles of 47.4 and 63.0 nm mean profile, respectively. The author gives the frequency histograms of the mean diameters for 33 cells from one carotid body and 30 cells from another: These histograms are bimodal but, here again, we do not know the exact number of measured vesicles in each cell. Once distinguished by the size of their dense-cored vesicles, the two kinds of cells were found to have other differences: The dense-cored vesicles are almost twice as abundant in large-vesicle cells as in small-vesicle cells; the volume density of the mitochondria is slightly larger and the volume density of the nuclei is slightly smaller in large-vesicle cells than in small-vesicle cells. Finally, large-vesicle cells are about 1.5 times as numerous as the other cells. McDonald and Mitchell (1975a,b) also described two kinds of type I cells which differ in the mean diameter of their dense-cored vesicles: 90 versus 116 nm. These numbers are very different from those of Hellstrom (1975a) although they concern the same species, rat (probably owing to some methodological factor). McDonald and Mitchell (1975a,b) also remarked, as Hellstrom (1975a) did, that dense-cored vesicles are more abundant in cells with larger vesicles but found the two types of cells in nearly equal proportions. Furthermore, they added that only a few small-vesicle cells (called “B cells” by these authors) are in contact with nerve fibers. McDonald and Mitchell (1975a,b) give the number of investigated cells and the total number of measured vesicles. It should be pointed out that an average of only 34 measured vesicles per cell has been considered as sufficient to characterize 10 cells as “B cells.” One may wonder if such a small sample is representative of a population of several thousands of vesicles. Another critical factor is the way in which cells are selected to be studied. It is absolutely necessary to avoid an a priori selection and it is regrettable that McDonald and Mitchell (1975a,b) did not say if the cells were selected at random or not. In a preliminary study, I considered the dense-cored vesicles in 30 type I cells, selected at random, from a rabbit carotid body (Verna, 1977). An average of 250 vesicles (140 < n < 434) were measured for each cell. The results showed, first, great variations in the size of dense-cored vesicles (extreme values: 50 to 250 nm) not only from cell to cell, but also within a given cell. The mean diameter was calculated for each cell, and the extreme values were 93 and 146 nm. Thus, there were small- and large-vesicle cells in the rabbit carotid body as well but these cells were the extremes of a continuous series; this was shown by the unimodal distribution of the 30 mean diameters, most cells having dense-cored

288

ALAlN VERNA

vesicles about 115 nm in mean diameter. In conclusion, it seems impossible to distinguish clear-cut categories of type I cells in the rabbit carotid body from a morphometric study of their dense-cored vesicles. d. Significance of the Various Aspects of Type I Cells. As we have seen, there exist very different type I cells, either from a cytochemical or from a cytological point of view, and there probably exist intermediate forms as well. Unfortunately, there has been no functional interpretation of these findings until now. Does this diversity reflect different stages of evolution of a single cell type? We must not forget that the electron microscope gives us a static picture. However, it is possible that type I cells undergo some changes during their life. Mitotic type I cells have been observed, although rarely (Kondo, 1971; Verna, 1977; see also Adams, 1958); some degenerating type I cells, characterized by very large and opaque dense-cored vesicles and pycnotic nuclei, are often present in the rabbit carotid body (Verna, 1977). So an evolution of type I cells with correlative morphological (and probably biochemical) changes is not inconceivable. It may be noteworthy to recall here that Celestino da Costa (1944) considered type I cells as undifferentiated or immature cells (“metaneurogonia”) of sympathetic origin, like chormaffin cells of the adrenal medula. This interpretation is compatible with the recent demonstration of the embryological origin of type I cells (see Section VII) and has received additional support from the work of Korkala er al. (1973): These authors have shown that glucocorticoids increase the storage of catecholamines in type I cells of the adult rat, an effect which is usually observed only in the catecholamine-storing cells (i.e., SIF cells) of newborn animals (Eranko and Eranko, 1972). Korkala er al. (1973) suggested that carotid body type I cells remain, in adult rats, at a relatively primitive stage. From this point of view, carotid body chromaffin cells, which look like adrenal chromaffin cells, may be considered as aberrantly differentiated type I cells. It may be recalled here that a small adrenal medulla has been observed inside the carotid body by de Castro (1926). It would be interesting to see if carotid body chromaffin cells are innervated by glossopharyngeal afferent nerve fibers (as any type I cell; see Section IV) or by sympathetic preganglionic fibers. McDonald and Mitchell (1975b) say they have unpublished data in favor of the latter possibility. 4. Connections between Type I Cells

Type I cells in a cluster are separated from each other (and from type I1 cells and nerve endings) by an intercellular space of about 20 nm. It has been shown that all these intercellular spaces are accessible to vascularly injected horseradish peroxidase (Woods, 1975). Adjacent type I cells sometimes present symmetrical junctions of the zonula adherens type. These junctions have been observed by many authors working on different species (Biscoe and Stehbens, 1966; Hess, 1968; Kondo, 1971) but seem to be more frequent in the rat carotid body (Hess, 1975a; McDonald and Mitchell, 1975b; Morgan et a l . , 1975). Another kind of

ULTRASTRUCTURE OF THE CAROTID BODY

289

junction seems to be frequent in the latter species; it is characterized by an asymmetrical structure with an accumulation of dense-cored vesicles and, sometimes, of synaptic-like vesicles on only one side of the junction (Hess, 1975a; McDonald and Mitchell, 1975b; Morgan et al., 1975). These junctions are considered as “synapses” by McDonald and Mitchell (1975b) but without any evidence other than a morphological one. A junction of this kind has been reported between a type I and a type I1 cell (Hess, 1975a); similar accumulations of vesicles also occur at the level of membrane differentiations facing interstitial connective spaces (see Section IV,E,l, and Fig. 5). In the latter case, it is difficult to consider this structure as a synapse. Finally, a few narrower junctions (tight junctions?) have been reported between type I cells (Al-Lami and Murray, 1968a; Hess, 1975a) but we do not know their exact nature for lack of tracer studies. There has been no physiological evidence until now (because of technical difficulties) concerning a possible coupling between type I cells. B. TYPEI1 CELLS Type I1 cells are much less numerous than type I cells: According to Biscoe and Pallot (1972) there are about four or five times as many type I cells as type I1 cells in the cat carotid body; in the rat, type I cells outnumber type I1 cells by a factor of 3 to 5 (McDonald and Mitchell, 1975b). Type I1 cells are located at the periphery of type I cell clusters (Fig. 3) and surround the major part of these cells with a thin cytoplasmic layer, sometimes only 0.2 p m thick. Type I1 cells also send cytoplasmic sheets between neighboring type I cells, toward the cluster center, but this occurs more frequently in the cat than in the rat carotid body (Hess, 1975a). The type I1 cell nucleus is generally flattened, sometimes lobulated, and more electron dense than the type I cell nucleus. The cytoplasm contains the usual organelles: few mitochondria, a Golgi apparatus, endoplasmic reticulum, centrioles, sometimes a cilium of the 9 + 0 pattern, microtubules, and microfilaments. Type I1 cells do not contain dense-cored vesicles; they do not take up tritiated monoamines. Type I1 cells envelop nerve fibers and nerve endings like Schwann cells but, as for type I cells, a certain proportion of nerve ending membrane is not covered and, thus, is exposed to surrounding connective spaces through a basement membrane. Type I1 cells also have a basement membrane.

IV. Type I Cell Innervation As pointed out in the Introduction, the interpretation of the relationship between type I cells and nerve fibers is one of the most disputed problems concerning the carotid body. Between 1970 and 1975, the great question was whether

290

ALAIN VERNA

type I cell innervation is sensory in nature or not. Now it has been demonstrated that many, if not all, nerve endings on type I cells are afferent (sensory) as previously suggested by de Castro (1928). However, there is still a controversy about the part played by the type I cells in the sensory process: For some authors, these cells are the transducing elements which stimulate the nerve endings by release of an excitatory transmitter; for others, the transducing element is the nerve ending itself, but its activity is modulated by type I cells by release of an inhibitory transmitter. In fact, these very different conceptions arise principally from contradictory results of physiological experiments or from speculative interpretations of morphological observations. However, the latter are relatively concordant. We shall first consider the origin of nerve endings on type I cells, then their cytological features and the ultrastructure of their junctions with type I cells. Finally, current hypotheses as to the meaning of these structures will be briefly considered. A.

ORIGIN OF

NERVEENDINGS ON TYPEI CELLS

Two methods have been used up to now to determine the source of nerve endings in the carotid body: One of them is the well-known degeneration method; the other is the more recent tracer method which uses the axoplasmic flow. That it does not provoke pathological changes which are, sometimes, difficult to interpret is the great advantage of this method. 1.. Degeneration Studies The nerve fibers ending on type I cells may originate from (1) intracranial neurons, (2) neurons located in the sensory ganglions of the glossopharyngeal nerve, (3) neurons located in the sympathetic superior cervical ganglion, and (4) neurons located in the carotid body itself. In 1926, de Castro showed, with the light microscope, that transection of the glossopharyngeal nerve leads to degenerative changes in nerve fibers and nerve endings associated with type I cells. This result was confirmed in electron microscopy by Biscoe and Stehbens (1967), Hess (1968), and Hess and Zapata (1972). However, it is more difficult to determine if all the nerve endings on type I cells arise from glossopharyngeal nerve fibers (via the sinus nerve). In fact, Biscoe and Stehbens (1967) have observed a few apparently normal nerve endings which persist 3 months after cutting the sinus nerve. Hess and Zapata (1972) also admit that 20 days after severance of the IXth nerve (in cats) “an occasional synapse, perhaps normal in appearance, can be found,” and more recently, McDonald and Mitchell (1975a,b) assert that, in the rat carotid body, about 5% of the nerve endings on type I cells remain unaffected 25 days after glossopharyngeal nerve section. From results of further degeneration studies (after

ULTRASTRUCTURE OF THE CAROTID BODY

29 1

removal of the superior cervical ganglion or section of the sympathetic trunk) it was concluded by McDonald and Mitchell (1975a,b) that the few endings which did not degenerate after sinus nerve section were terminals from sympathetic preganglionic fibers which reach the carotid body by the ganglioglomerular nerves. In a serial ultrathin section analysis of the rat carotid body, Kondo (1976) also described a preganglionic efferent fiber in direct contact with a type I cell but he added: “The latter pathway seems to be of minor importance because of its rare occurrence.” This opinion is shared by Hess (1977a) who, studying chronically denervated rat carotid bodies, observed no nerve terminal on type I cells after section of the glossopharyngeal nerve and concluded: “The glomus cells do not receive any significant autonomic innervation. ” Furthermore, Hess (1977a) called attention to the risk of misinterpreting a type I cell process as a nerve fiber. On the other hand, it is also possible that a few type I cells, located at the organ periphery, may be related to the dendrites of intrinsic ganglion cells as demonstrated by Kondo (1976). However, these junctions are obviously of low frequency due to the well-known paucity of ganglion cells in the carotid body. To summarize, and without minimizing the functional importance of the abovementioned unusual autonomic fibers, it can be said that most nerve endings on type I cells are sinus nerve fiber terminals. The problem now is to determine if this innervation is sensory or not. One way to do this is to localize the neuronal cell bodies: If they are in the glossopharyngeal sensory ganglions it will be possible to admit their sensory nature. This attempt was first made by de Castro (1928) by way of intracranial section of the glossopharyngeal nerve roots, above (central to) the sensory ganglions. He reported that no change could be detected in type I cell innervation 12 days after operation. This study was, of course, a light microscope study after silver staining but the results were confirmed later in electron microscopy by de Castro and Rubio (1968), who concluded that the fibers innervating type I cells have their cell bodies in the glossopharyngeal nerve sensory ganglions and, consequently, are afferent (sensory) in nature. Unfortunately, the same experiment was repeated by Biscoe et al. (1970) but with diametrically opposed conclusions. These authors reported that about 60% of nerve endings on type I cells degenerate 128 days after cutting the glossopharyngeal nerve intracranially and they concluded in favor of the efferent nature of the type I innervation. This interpretation was also supported by the occurrence of synaptic-like microvesicles in the nerve endings which make them similar to efferent or motor terminals in other areas. To resolve the apparent discrepancy between de Castro’s and Biscoe ’s works, the same experiment was repeated by Hess and Zapata (1972), Nishi and Stensaas (1974), and McDonald and Mitchell (1975a,b) but with results supporting the original conclusions of de Castro (1928) and not those of Biscoe’s group. It is

292

ALAIN VERNA

possible, as pointed out by Hess and Zapata (1972), that Biscoe et al. (1970) performed their operation of intracranial severance of the IXth nerve too closely to the petrous ganglion, causing injury and retrograde degeneration of the sensory neurons, This would also explain the slow rate of degeneration they reported. Finally, the problem was investigated using more physiological methods. 2. Autoradiographic Studies It has been well demonstrated that after perikaryal uptake of a tracer, it migrates with the axoplasmic flow and accumulates in the axon terminals. This phenomenon makes it possible to trace neuronal projections (Lasek et al., 1968) and has been used recently by Fidone et al. (1975, 1977) and also by Smith and Mills (1976). After administration of tritiated amino acids to the petrosal ganglion these authors observed autoradiographically that most nerve endings on type I cells were labeled. Consequently, it appears that these nerve endings arise from neurons in the petrosal ganglion. Since this ganglion seems to consist entirely of sensory unipolar neurons (Stensaas and Fidone, 1977) we can consider the sensory nature of most terminals on type I cells as demonstrated. It must be added that many of the labeled sensory nerve endings observed by Fidone et al, (1 975, 1977) and by Smith and Mills (1976) exhibit clear synaptic-like microvesicles. This point will be discussed later. Finally, the use of this new method leads to the same conclusions as those expressed by de Castro (1928) nearly half a century ago. B. ULTRASTRUCTURE OF NERVE ENDINGS ON TYPEI CELLS 1 . Identification The nerve endings on type I cells are characterized by a local accumulation of organelles such as mitochondria and clear microvesicles. They may also contain some dense-cored vesicles and glycogen particles but are often devoid of neurotubules and microfilaments. Their mitochondria are generally of smaller diameter and have a more electron-dense matrix than those of type I cells. However, this distinction is impossible in the case of type I cell processes whose mitochondria are frequently similar to nervous mitochondria. It should be emphasized that it is sometimes difficult to distinguish between nerve terminals and type I cell processes without the help of serial sections: As previously mentioned, type I cell processes may also contain clear microvesicles and a few dense-cored vesicles. A good criterion to identify type I cell processes is the presence of ribosomes. Unfortunately, ribosomes are scarcely distributed in type I cell extensions, so the probability of their occurring in a single section is very low.

2 . Size and Shape The size of nerve endings on type I cells varies considerably (roughly from 1 to 10 pm), but this notion of size is of very little significance because of the very

ULTRASTRUCTURE OF THE CAROTID BODY

293

complicated shape of many terminals (see three-dimensional reconstructions in Nishi and Stensaas, 1974). It would be better to quantify the surface of contact between nerve endings and type I cells or the volume of the terminals. Based on morphological criteria, several attempts have been carried out in the past to separate various kinds of nerve terminals, such as “basket” and “bulbous endings” (Al-Lami and Murray, 1968a), “calyciform” endings (Eyzaguirre et al., 1972; Nishi and Stensaas, 1974), “bouton-like” endings (Eyzaguirre et al., 1972; Vema, 1971), and “en passant” endings (Hoglund, 1967; Abbot et al., 1972; Vema, 1971), etc. In fact, later studies have shown that nerve terminals on type I cells are very polymorphous: So, one and the same nerve fiber may innervate several type I cells, after branching, by means of terminals of very different size and shape. This conclusion results from both light microscope studies (examination of de Castro’s original slides by Eyzaguirre and Gallego, 1975) and serial ultrathin section analysis (Biscoe and Pallot, 1972; Nishi and Stensaas, 1974; Kondo, 1976). This polymorphism explains the very different aspects previously described in single ultrathin sections and is consistent with the great variation in size observed along the course of one terminal thanks to a particularly favorable plane of section (Vema, 1973). In conclusion, it is very difficult to distinguish different kinds of nerve terminals on type I cells based on morphological criteria, especially in single section studies. However it must be added that Nishi and Stensaas (1974) describe both large and small calyciform endings after three-dimensional reconstruction from serial sections. The significance of this distinction is obscure, particularly because Nishi and Stensaas (1974) demonstrate that both kinds of terminals are glossopharyngeal afferents which do not degenerate after intracranial section of nerve roots. Another interesting result given by the same authors is the fact that large calyciform endings frequently send short processes to neighboring type I cells. 3. Organelle Content

a. Mitochondria. Mitochondria are abundant in nerve endings on type I cells. They frequently show longitudinally oriented cristae and a very electrondense matrix. However, these features depend on the fixation conditions and also on the species. The mitochondria are sometimes so numerous that they completely fill the neuroplasm (Fig. 6), leading Bock et al. (1970) to call these nerve endings “Mitochondriensacke. These aspects have been observed not only in different mammal species such as man (Bock et al., 1970), guinea pig (Kondo, 1971), and rabbit (Vema, 1971, 1973), but also in the avian carotid body (King et al., 1975) and the amphibian carotid labyrinth (Kobayashi, 1971b). It may be fascinating to observe a great amount of mitochondria in the sensory nerve endings of an organ which respond to the level of arterial oxygen partial pressure. However, it must be added that accumulation of mitochondria in nerve terminals is not a feature of the carotid body. It has been observed also in nerve terminals ”

294

ALAIN VERNA

FIG.6 . Nerve ending (NE) showing an unusual accumulation of slender mitochrondria (rabbit carotid body). CI, Type I cell; CII, type I1 cell. Bar = 1 p m .

or fibers in the carotid sinus (Rees, 1967a; Chiba, 1972; Knoche and Addicks, 1976), in the heart (Chiba and Yamauchi, 1970), in the brain (Sotelo and Palay, 1968; Bouchaud, 1974), and in the spinal cord (Leonhardt, 1976). The problem is to determine the significance of these accumulations of mitochondria in nerve endings. For some authors it has been considered as a characteristic feature of sensory nerve endings (Munger, 1971; King et al., 1974) but for others it is a pathological change, perhaps linked to a process of aging (Seitelberger, 1971; Bouchaud, 1974; Leonhardt, 1976). In favor of this interpretation is the fact that many mitochondria-rich nerve endings exhibit a noticeable proportion of altered mitochondria, lamellar bodies and dense bodies. On the other hand, large dendritic varicosities observed in the brain and filled with mitochondria have been interpreted as growing dendritic tips by Sotelo and Palay (1968). Maybe both phenomena (degeneration and regeneration) occur in the carotid body, leading to some renewal of nerve endings as suggested for the carotid sinus (Knoche and Addicks, 1976) and the olfactory system (Graziadei, 1973). It must be underlined that mitochondria accumulation does not characterize a special kind of nerve terminal in the carotid body: Both serial section (Biscoe and Pallot, 1972) and single section studies (Verna, 1973) have shown that the amount of mitochondria varies greatly from one region of a given nerve ending to another.

ULTRASTRUCTURE OF THE CAROTID BODY

295

b. Dense-Cored Vesicles. Nerve endings terminating on type I cells usually contain a few dense-cored vesicles (about 90 to 110 nm in diameter) but these vesicles are never associated with the synaptic-like membrane differentiations. From the work of McDonald and Mitchell (1975a,b) it appears that dense-cored vesicles are less numerous in the sinus nerve afferent terminals than in the preganglionic sympathetic efferent endings. In both cases, we do not know the content of these vesicles. c. Synaptic-like Microvesicles. Electron-lucent vesicles about 60 nm in diameter are almost always present in nerve endings on type I cells but are very irregularly distributed: They are often dispersed in the neuroplasm, sometimes densely packed in clusters. It has been suggested by Nishi and Stensaas (1974) and McDonald (1977b) that synaptic-like vesicles are more abundant in small than in large nerve endings. According to McDonald (1977b) the number and distribution of these vesicles vary with the activity of nerve terminals (see Section VI,B). In a quantitative study of the rat carotid body, McDonald and Mitchell (1975a,b) measured the clear vesicles and found a mean diameter of 61 nm in sinus nerve afferent terminals and of 53 nm in the few preganglionic sympathetic efferent terminals (previously identified by degeneration experiments). The significance of synaptic-like vesicles in sensory terminals will be discussed below, but it already can be said that we have no evidence about their content. We presume they contain no acetylcholine, since the acetylcholine contents of normal and denervated carotid bodies are not significantly different from each other (Fidone et a l . , 1976). d. Other Nerve Terminal Components. The nerve terminals on type I cells, like terminals anywhere in the nervous system, occasionally show pinocytotic vesicles, lysosomes, and smooth reticulum profiles. Glycogen particles are also frequently present and sometimes accumulate in very large amounts (Verna, 1973). This is not peculiar to the carotid body and has been observed in many other nerve endings such as, for example, the carotid sinus barosensory terminals (Chiba, 1972; Bock and Gorgas, 1976a; Knoche and Addicks, 1976). C. ULTRASTRUCTURE OF JUNCTIONS BETWEEN NERVEENDINGS A N D TYPEI CELLS Nerve terminals and type I cells are always separated by an extracellular space approximately 20 nm wide. From place to place, the juxtaposed membranes show local differentiations characterized by an increased electron density. At this level, an electron-dense material is present in both the nerve terminal and the type I cell adjoining cytoplasm. The intercellular gap also contains a material of slight electron density appearing in cross section as a dark line equidistant from the membranes. Generally these specialized zones do not exceed 0.5 p m in length but several of them may follow one another along the area of apposition

296

ALAlN VERNA

between nerve ending and type I cell. The junctional zones may be symmetrical without any associated vesicles; they are consequently similar to the “puncta adherentia” well known in the nervous system (see Peters et al., 1970). In other cases, different kinds of vesicles are clearly related to the junctional dense material: These junctions are therefore comparable to chemical synapses and the association between the vesicles and the junctional dense material may be referred to as a “synaptic complex” by analogy with the structure of the nervous system synapses (Palay, 1958). However, the junctions between carotid body type I cells and nerve endings are characterized by their polymorphism since different kinds of vesicles may be observed in variable proportions and, moreover, vesicles may be present on one side of the junction or on the other and sometimes on both. In this review, the term junction will be used in preference to synapse because the synaptic nature of the relationship between type I cells and nerve endings cannot be considered as having been demonstrated. 1 . Junctions Characterized by an Accumulation of Vesicles on the Neural Side The synaptic-like clear vesicles which are numerous in the nerve endings on type I cells sometimes accumulate at the level of a membrane area characterized by increased electron density and the presence of some cone-shaped patches of opaque cytoplasmic material next to it. These structures are identical to the “dense projections” described by Gray (1963) in the central nervous system synapses. On the type I cell side, the facing membrane also shows an increased electron density and an associated cytoplasmic dense material but the latter is of more uniform thickness and sometimes is inconspicuous. There is no accumulation of vesicles (Fig. 7). This kind of junction was the first to be described in the ultrastructural study of the carotid body (Biscoe and Stehbens, 1966) and has been observed regularly since those early studies, involving nerve endings of very different aspects such as “knobs” of medium size and large terminals filled with mitochondria. Comparative studies have demonstrated its occurrence in the carotid body (or homologous organ) of many vertebrate species (see Kobayashi, 197 1 b). However, if these descriptions were concordant, it would be surprising to find such junctions in a sensory receptor since the structure of these junctions suggests a transmission from nerve endings toward type I cells: Synaptic vesicles and dense projections are indeed classically thought to be presynaptic in the nervous system synapses (see Peters et al., 1970; Akert et a l . , 1972). This is one of the reasons which has lead some authors to express doubts about the afferent nature of the type I cell innervation (an edifying list of papers and author’s conclusions on this problem can be found in Kobayashi, 197 lb). Fortunately, the above-mentioned autoradiographic studies have demonstrated that efferent-like junctions actually occur between sensory sinus nerve afferent terminals and type I cells (see Fig. 2 in Smith and Mills, 1976, for example). One consequence of this is the difficulty

ULTRASTRUCTURE OF THE CAROTID BODY

297

FIG.7. Junction between a type I cell (CI) and a nerve ending (NE) (rabbit carotid body). Synaptic-like vesicles (V) accumulate in the nerve ending at the level of membrane differentiations whereas there is no vesicle on the type I cell side. Note the nerve ending area (arrow) which is not covered by the type I1 cell (CII). Col, Collagen; M, mitochondria; Nu, nucleus. Bar = 0.5 pm.

to distinguish morphologically efferent-like ‘‘synaptic complexes” in sensory terminals from true efferent junctions involving possible autonomic fibers on type I cells. However, McDonald and Mitchell (1975a,b) assert that it is possible in the rat carotid body. They specify that in sympathetic efferent terminals, clear vesicles are slightly smaller and more closely packed and dense-cored vesicles more numerous than in sinus nerve afferent terminals. 2. Junctions Characterized by an Accumulation of Vesicles on the Type I Cell Side In these junctions, different kinds of vesicles accumulate near the type I cell junctional membrane which shows associated dense projections, whereas, in the nerve ending, clear vesicles are located away from the junctional zone. The neuroplasmic junctional material is very marked and shows a uniform thickness and a filamentous structure. This disposition is therefore compatible with the “classical” interpretation of de Castro (1928) since it suggests a transmission from type I cell to nerve ending. However, it was only in 1970 that the occurrence of vesicles associated with the junctional zones of the type I cell membrane were first noticed in the mouse carotid body (Kobayashi and Uehara, 1970). It seems that this kind of junction is more frequent in the mouse and rat than in the

298

ALAlN VERNA

cat carotid body. This could explain why these junctions were discovered so late since the cat was the most frequently used species in carotid body studies. Later, similar observations were made on other mammals such as the rabbit (Verna, 1971, 1973), the guinea pig (Kobayashi, 1971b), the rat (Hess, 1975a; McDonald and Mitchell, 1975a,b; Morgan et al., 1975; Kondo, 1976), and, finally, the cat (Smith and Mills, 1976). Identical junctions also occur in the carotid body of birds, such as the swallow (Kobayashi, 1971b), the domestic fowl (King et al., 1975), and the duck (Osborne and Butler, 1975). However, these descriptions revealed some minor differences possibly related to the species considered. In the mouse carotid body, the “afferent synaptic complexes” described in type I cells by Kobayashi and Uehara (1970) are characterized by small vesicles about 30 to 40 nm in diameter and filled with an electron-dense granular substance. In the rabbit carotid body different kinds of vesicles may be observed in type I cell “synaptic complexes,” namely: small vesicles (about 60 nm), occasionally small granular vesicles (about 70 nm), and dense-cored vesicles (about 90 nm) (Verna, 1973, 1975). The latter are consequently somewhat smaller than the mean diameter of the whole dense-cored vesicle population (about 115 nm in this species; Verna, 1977). In some junctions only small vesicles (the majority of them, if not all, having a clear content) take part in the synaptic complex; in others, there are only a few large dense-cored vesicles (Fig. 8), but the most frequent occurrence is a mixture of the different kinds of vesicles. In the rat carotid body, both small clear vesicles and large dense-cored vesicles participate in the junctions with nerve endings (McDonald and Mitchell, 1975a,b; Morgan et al., 1975; Kondo, 1976). According to McDonald and Mitchell (1975b) the ratio of small clear vesicles to large dense-cored vesicles is about 111 except at the junctions between nerve endings and type I cell processes where small clear vesicles are much more numerous than large dense-cored vesicles (average ratio: 25/1). Both McDonald and Mitchell (1975b) and Kondo (1976) assert that large dense-cored vesicles involved in the junctions are not different from those located elsewhere in the cell. 3 . So-called “Reciprocal Synapses”

The two kinds of junctions described above may occur side by side at the level of one and the same nerve terminal. This arrangement has been described in the rat carotid body as “reciprocal synapse” by McDonald and Mitchell (1975a,b). It has also been observed by Morgan et al. (1975) in the same species and by Smith and Mills (1976) in the cat. The rabbit carotid body also shows similar pictures but they are exceptional in this species (Fig. 9). Lastly, “reciprocal synapses” have also been described in the domestic fowl (King et al., 1975) and the duck (Osborne and Butler, 1975) carotid body. According to McDonald and Mitchell (1975a,b) only 5% of sinus nerve afferent terminals show “reciprocal synapses” but this evaluation may be underestimated since it is based on single

ULTRASTRUCTURE OF THE CAROTID BODY

299

FIG.8. Junctions between type I cells (CI) and nerve endings (NE)(rabbit carotid body). (A) In this junction, synaptic-like vesicles accumulate on the type I cell side. Most of these vesicles have clear contents but a few exhibit a dense granule (arrow). Note the unsymmetric arrangement of the junctional dense material. Nu, Nucleus. Bar = 0.5 p m . (B) In this case, large dense-cored vesicles (arrow) accumulate in the type I cell at the level of the junctional dense material. CII, Type I1 cell. Bar = 0.5 p m .

300

ALAIN VERNA

FIG.9. Junctions similar to those in Fig. 7 (fine arrow) and Fig. 8B (thick arrow) sometimes occur side by side, as in this example (rabbit carotid body). This arrangement is described as “reciprocal synapse” by some authors (see text). CI, Type I cell; M, mitochondria; NE, nerve ending; Nu, nucleus. Bar = 0.5 pm.

section studies. The interval between both members of a “reciprocal synapse” may be as small as 100 nm. It must be added that “reciprocal synapses” are described as paired junctions of opposite polarity. However, we do not know (for lack of statistical studies on serial sections) if this association in pairs occurs at random or not.

D. EFFECTSOF DENERVATION UPON TYPEI/TYPE11 CELLS It is well known that taste buds which, like the carotid body, are innervated by the IXth nerve, degenerate rapidly after denervation (see Guth, 1971). It was consequently legitimate to expect similar behaviour in carotid body cells: Therefore, short-term and long-term consequences of denervation upon carotid body ultrastructure have been explored by several authors. Biscoe and Stehbens (1967) did not find any alteration in the ultrastructure of type I cell nor of type I/type I1 cell relations up to 3 months postoperatively . Hess (1968) has noted that, 7 days after denervation, type I1 cells show proliferative cytoplasmic extensions whereas type I cells do not exhibit any striking changes. Abbott et al. (1972) have studied more precisely the early ultrastructural changes after sinus nerve section. These authors write: “The majority of glomus cells appear to be unaffected by loss of their nerve supply. However, they add ”

ULTRASTRUCTURE OF THE CAROTID BODY

30 1

that lysosomes and lipid bodies are slightly more numerous in type I and type I1 cells between 1 and 10 days after denervation. They also describe, at the 21-day stage, some hypertrophy of type I cells which leads to cytoplasmic processes filling the spaces presumably left by the disappearance of nerve endings. In a more recent study, Hess (1977a) underlined that, “after denervation, the early changes seen in the carotid body reside in the capsule cell.” According to this author, these cells (type I1 cells), like Schwann cells, hypertrophy and ensheath type I cells with more numerous layers and in a more irregular manner for about 2 or 3 months. On the contrary, type I cells do not show any morphological change even up to 13 months after denervation. It is consequently clear that carotid body type I cells, contrary to gustatory cells, are not morphologically affected by denervation. However, this conclusion does not mean that denervation does not elicit any effect at all on type I cells: It has been demonstrated that, after degeneration of their sensory terminals, these cells become hyposensitive to reserpine which can no longer totally deplete their catecholamine stores (Hess, 1975b). From this observation Hess deduced that intact sensory nerve endings exert a retrograde effect on type I cells, which he considered as a trophic influence. It is not known if the carotid body development is dependent on its sensory nerve supply as suggested for other receptors such as avian Herbst and Grandry corpuscules (Saxod, 1972) and mammal mechanoreceptors (Zelena, 1976).

E. FUNCTIONAL INTERPRETATIONS OF TYPEI CELL-NERVE ENDING RELATIONSHIPS In summary, the studies reported in this chapter have demonstrated that:

i. Type I cells are innervated by sensory nerve fibers from the glossopharyngeal nerve. ii. A few efferent autonomic fibers may also end on some type I cells. iii. The structure of the junctions between type I cells and nerve endings is similar to the structure of chemical synapses in the nervous system. iv. Two kinds of junction may be distinguished according to the situation of vesicles (on type I cell side or on nerve ending side). v. These two kinds of junction may coexist side by side at the level of the same nerve ending. Several attempts have been made to conciliate ultrastructural and physiological findings. Some of them, based on the assumption of a type I cell efferenr innervation (Biscoe, 1971), are no longer defensible. However, among the hypotheses which take the sensory nature of type I cells into account, a controversy still remains as to the identity of the transducer structure. Basically, there are two opposite views: For some authors the type I cell is the receptor, for others it is not, the transducer being the sensory nerve terminal itself.

302

ALAlN VERNA

1 . Type I Cell as a Transducer In this interpretation the type I cell is responsible for the transduction of stimuli and the subsequent depolarization of the afferent nerve terminals, possibly by release of a chemical transmitter. This is the “classical” interpretation of de Castro (1928), more recently defended by Eyzaguirre et al. (1972). The structure of the junctions characterized by the occurrence of vesicles on the type I cell side supports this view. Furthermore, the participation of a chemical transmitter is also supported by the experiments of Eyzaguirre e f al. (1965): These authors have shown that stimulation in vitro of a superfused carotid body induces the release of a substance which increases the sensory discharge of a second carotid body, situated downstream (Loewi effect). Some years later, Eyzaguirre and Zapata (1968) demonstrated that this Loewi effect persists even if the stimulated carotid body is totally denervated 4 days previously. This result suggests that the presumed transmitter is located in carotid body cells and not in nerve terminals. Unfortunately, as we have seen, different kinds of vesicles are involved in junctions, and we know nothing about their content. It is possible that the large dense-cored vesicles (diameter about 90 nm) contain a catecholamine as do the other dense-cored vesicles located elsewhere in the cell, but we have no information about the small-vesicle contents. Do these different vesicles contain different substances? This possibility cannot be ruled out. Furthermore, the accumulation of small vesicles near a particular zone of the type I cell membrane is not restricted to the occurrence of a nerve ending: Small vesicles (most of them having a clear content) may accumulate at the level of a membrane area covered by a basement membrane only and facing the extracellular spaces without any interposed type I1 cell cytoplasm. At these sites, a dense material similar to the synaptic dense projections occurs beneath the type I cell membrane (Fig. 5). This observation suggests the release of the contents of the microvesicles at nonsynaptic sites, since the target element is lacking. It must be recalled that similar observations have been made by Taxi er al. (1969) describing nonsynaptic accumulation of vesicles in ganglion cell dendrites and also in chromaffin cells (SIF cells) of the rat sympathetic superior cervical ganglion. Several interpretations could be proposed for these observations: Small vesicles actually contain a chemical transmitter which is released not only at the nerve ending level but also at other sites around the cell. In this way, we can presume like Torrance (1968) that nerve endings “respond to a general concentration of a transmitter rather than to a localized high concentration within a narrow synaptic cleft. This interpretation could also explain why, during reinnervation experiments, it is possible to record a chemoafferent activity when nerve terminals are close to type I cells whereas “synaptic” junctions are still few (Zapata et a l . , 1969). It is also possible that microvesicles have nothing to do with nerve depolarization. In this way, type I cell microvesicles can be compared with the apparently ”

ULTRASTRUCTURE OF THE CAROTID BODY

303

identical vesicles which occur in neurosecretory processes where their function, although unknown, is not synaptic (see Douglas et a l . , 1971). Many substances have been proposed as possible chemical transmitters in the chemosensory process (see Torrance, 1968; Joels and Neil, 1968; Eyzaguirre and Zapata, 1968; Biscoe, 1971; Howe and Neil, 1971). Although this problem is not directly related to this review, it must be said that none of the proposed substances fully accounts for the whole body of experimental results. Another confusing factor is the possible occurrence of species differences: It seems that dopamine inhibits the chemoafferent activity in the cat (Black et a l . , 1972; Sampson, 1972; Zapata, 1975; Nishi, 1977) but may be excitatory in the dog (Black et a l . , 1972) and in the rabbit (Monti-Bloch and Eyzaguirre, 1977). The identification of the transmitterts), therefore, needs further investigations. Nevertheless, type I and/or type I1 cells appear necessary to the generation of chemoafferent impulses since after their destruction, the regenerating nerve fibers do not exhibit any specific chemosensitivity (Verna et a l . , 1975). Furthermore, it has been shown by Zapata et al. (1969) that it is possible to reinnervate the carotid body with mechanoreceptor fibers and to obtain, after several months, chemosensory responses to the usual stimuli. More recently, Zapata et al. (1976, 1977) have shown that after a sinus nerve crush there is a good correlation between the reappearance of chemosensory discharges and the reestablishment of contacts between sensory nerves and type I cells. Therefore, these results suggest that type I and/or type I1 cells are responsible for the receptors’ specificity, although, as underlined by Eyzaguirre et al. (1977), they do not indicate whether the carotid body cells are the primary transducer or if they condition nerve endings to become chemosensitive. However, if we admit that information is conveyed from type I cells to their afferent nerve endings, it is still necessary to account for the occurrence of synaptic-like vesicles in these nerve endings. Hess and Zapata (1972) have suggested that sensory terminals on type I cells could have, in addition to their afferent functions, efferent functional effects. According to these authors, such a dual role may provide the morphological basis for the efferent inhibition described in the carotid body (see Section V). Nishi and Stensaas (1974) and Vazquez-Nin et al. (1977) envisage such inhibitory effects and suggest an axon reflex mechanism for the release of the contents of the vesicles, assuming that impulses generated in an afferent fiber may antidromically invade collaterals of the same axon. Efferent inhibitory effects are also postulated by some authors who consider the nerve terminal itself as a transducer. 2. Nerve Terminal as a Transducer

The hypothesis that carotid body nerve fibers may be chemosensitive by themselves was first evoked by Biscoe and Taylor (1963) and later developed by

304

ALAlN VERNA

Biscoe (1971) essentially on theoretical grounds. He attributed the transducer function to very small nerve terminals (less than 0.1 p m in diameter) enclosed in type I1 cells, whereas type I cells were thought to be part of an efferent pathway which controls the receptor excitability. However, as we have seen, type I cells receive only minor, if any, efferent innervation, and the very small terminals enclosed in type I1 cells (previously reported by de Kock and Dunn, 1968) do not exist. Nevertheless, the possibility that afferent nerve terminals on type I cells are transducer elements still remains and is particularly supported by McDonald and Mitchell (1975a,b). Their interpretation is principally based on the results of a physiological experiment from Mitchell et al. (1972): These authors have shown that some nerve fibers in a neuroma formed from the cut end of the cat’s sinus nerve show normal chemoreceptive properties, although the neuroma does not contain any type I cell. This possibility is somewhat conflicting with the aforementioned results of Zapata et af. (1969, 1976, 1977) and Verna et al. (1975) which suggest that carotid body cells are necessary for chemoreception. It is possible that the chemoafferent activity recorded by Mitchell et al. (1972) arose from fibers which had reinnervated some type I cells either in the carotid body itself (which was removed in only one of the four studied cats) or in ectopic carotid bodies known to occur along the cat’s common carotid artery (Matsuura, 1973). Mitchell et al. (1972) argue they have “cut all connecting tissue between the carotid bifurcation and the neuroma. However, vascular connections were obviously intact since chemoafferent activities were tested via bloodborne changes; so the possibility that nerve fibers had escaped from the neuroma, along blood vessels toward type I cells elsewhere, cannot be ruled out. Moreover, this possibility was favored by the long delay (12 to 18 months) allowed to nerve regeneration. However, even if we admit with McDonald and Mitchell (1975a,b) that nerve endings on type I cells are chemoreceptive, it is still necessary to envisage some function for type I cells since they do exist. According to McDonald and Mitchell ( 1975a,b) these cells are interneurons modulating the sensitivity of chemoreceptive nerve endings in the following way: When sensory nerve terminals are stimulated, they release an excitatory transmitter; this transmitter induces the cells to release dopamine which, in turn, inhibits the sensory terminals. Thus, type I cells and nerve endings are interconnected by “reciprocal synapses” which form an inhibitory feedback loop. In addition, McDonald and Mitchell (1975a,b) postulate that the few preganglionic sympathetic endings they describe on some type I cells enhance dopamine release from these cells and that synaptic interconnections between type I cells enable them to influence one another. This theory therefore explains most of the ultrastructural and physiological data concerning the carotid body. However, it must be noted that none of the ”

ULTRASTRUCTURE OF THE CAROTID BODY

305

assumptions which support this theory can be considered as having been conclusively demonstrated. As already mentioned, the evidence that regenerating sinus nerve endings are chemosensitive by themselves (Mitchell et al., 1972) is not entirely convincing. The synaptic nature of the vesicles occurring in afferent nerve terminals is not proved at all: We do not know what they contain; they are presumed to be synaptic only by comparison with similar vesicles in other structures. It may be recalled that synaptic-like vesicles occur in nerve endings of the Pacinian corpuscle, for example, where there is no synapse (Zelena, 1978). McDonald (1977b) argues that conditions stimulating chemoafferent fiber activity (hypoxia, hypercapnia, antidromic electrical stimulation) reduce the number of microvesicles in sensory nerve terminals and also lead these vesicles to become more dispersed. This result is interpreted by the author as reflecting transmitter release but this is a very indirect evidence. The many unsuccessful attempts which have been made to correlate acetylcholine release with synaptic vesicle depletion in motor terminals lead us to be very prudent on this subject. It may be added that microvesicles in nerve endings on type I cells are more frequently sparse than packed at the level of junctions. Moreover, it seems that antidromic stimulation of the sinus nerve does not alter the type I cell membrane potential (Goodman and McCloskey , 1972). Furthermore, the concept of “reciprocal synapse, which has been introduced for dendrodendritic junctions in the olfactory bulb (see Reese and Shepherd, 1972, for references), seems to be questionable and it has been suggested recently that the two elements of these junctions interact by a synaptic mechanism in one direction and a nonsynaptic mechanism in the reverse direction (Ramon-Moliner, 1977). It is possible that afferent nerve endings also exert nonsynaptic effects on type I cells. What can this effect be? McDonald and Mitchell (1975a,b) suggest that the substance released by nerve terminals leads type I cells to liberate dopamine, which, in turn, inhibits the terminal activity. This model does not work for those species in which dopamine seems to be excitatory (dog and rabbit, according to Black et al., 1972, and Monti-Bloch and Eyzaguirre, 1977, respectively). Lastly it has been shown by Hellstrom et al. (1976) and Hellstrom (1977) that hypoxia, by itself, induces dopamine depletion which is quite similar in denervated and intact rat carotid bodies. Thus, things are probably more complicated than expected. ”

3. Other Interpretations

Other mechanisms of interaction between nerve fibers and type I/type I1 cells have been proposed by many authors. Some of these will be only briefly reported here, insofar as they have few ultrastructural implications. a. It has been suggested by Paintal (1967, 1968) that the transducer element may be the type I1 cell which is supposed to excite afferent nerve endings by “some physical change. This hypothesis was slightly modified by Jones (1975) according to which stimuli lead type I cells to release acetylcholine which in”

306

ALAIN VERNA

duces type I1 cells to contract, thereby stimulating the nerve endings they enclose. There are several reports suggesting that the carotid body contains (Eyzaguirre et al., 1965; Jones, 1975) and releases acetylcholine (Metz, 1969). More recent works have shown that this acetylcholine content is not affected by denervation (Fidone et al., 1977) and that the sinus nerve contains no more acetylcholine than noncholinergic nerves (Golberg et al., 1978). These results support the assumption that acetylcholine is located in carotid body cells and not in nerve endings. Furthermore, the ultracytochemical localization of choline acetyltransferase(Ballard and Jones, 1972) and of the high-affinity component of choline uptake (Fidone et al., 1977) suggests that type I cells are capable of acetylcholine synthesis. Unfortunately, the fine structural localization of cholinesterasesin the carotid body is not very enlightening: Cholinesterases seem to be present (in small amounts) all around the type I1 cells and the nerve endings covered by these cells (Ballard and Jones, 1971); the localization of acetylcholinesterase is more controversial (see Jones, 1975). However these observations may be interpreted not only in terms of transmitter hydrolysis but also in terms of the intracellular role of acetylcholine (see Koelle, 1969; Welsch and Pearse, 1969; Satler et al., 1974). Thus, the role of acetylcholine in the carotid body, frequently contested as a chemical transmitter (Heymans and Neil, 1958; Paintal, 1969; Sampson, 1971), is still obscure. Nevertheless, there is no convincing argument which could lead us to think that type I1 cells have a more sophisticated function than Schwann cells. Mills (1972) also has considered type N cells as the possible oxygen sensor, containing cytochrome a3 with a low affinity for oxygen. This author postulates the following mechanism: In hypoxia, type I1 cells release potassium ions which could depolarize type I cells, leading these cells to release an excitatory transmitter acting on the afferent terminals. This interpretation was revised later by Mills (1975) who “no longer excludes the possibility that the respiratory chain with low oxygen affinity is in the type I cell.” b. The originality of the theory proposed by Osborne and Butler (1975) is to suppose that the afferent nerve fibers are endogenously active and therefore spontaneously discharge. The proposed mechanism is the following: First, during normoxia, type I cells release dopamine which inhibits the spontaneous afferent discharge. In hypoxia, this release of dopamine is reduced, allowing the afferent terminals to discharge at a higher rate. This increased activity is accompanied by the release of acetylcholine from the terminals which further reduces the rate of dopamine secretion by type I cells. There is, consequently, a positive feedback loop. This theory, although ingenious, does not fit in with the results of Fidone et al. (1976) who suggest that acetylcholine is not located in nerve endings. Second, in a pharmacological study with a dopamine inhibitor, Docherty and McQueen (1977) showed that, in the cat, inhibition of sensory activity by dopamine is not substantial. Third, hypoxia increases dopamine

ULTRASTRUCTURE OF THE CAROTID BODY

307

release, at least in the rat (Hellstrom et al., 1976; Hellstrom, 1977) and the rabbit (Gonzalez and Fidone, 1977) carotid body. c. Torrance (1975) proposes that nerve terminals on type I cells are only sensitive to acidity and that hypoxia acts via a pH decrease in the intercellular space between nerve endings and type Utype 11 cells. This hypothesis has the merit of accounting for the convergence of hypoxic and hypercapnic stimuli (see Torrance, 1976, 1977, for full development of his theory). Torrance (1976) suggests that the pH is manipulated by type I1 cells but does not exclude the possibility that nerve endings, alternatively, may be stimulated by a transmitter (possibly polypeptidic) released from type I cells in response to a change in pH at their surface. F. NONINNERVATED TYPEI CELLS Finally we should mention that some type I cells are not innervated. From statistical data, McDonald and Mitchell (1975a,b) concluded that about 50% of type I cells (in the rat carotid body) are not connected to nerve endings but only to other type I cells. However, this estimation has not been corroborated by serial section analysis: Kondo (1976) reconstructed a cell cluster from the same species (rat) but only found one noninnervated cell among the 19 type I cells of the reconstructed cluster. A serial section analysis was also carried out by Nishi (1976) working on the cat carotid body: He only found 3 noninnervated cells out of 22 type I cells. According to this author, type I cells which are not innervated do not differ in appearance and organelle content from innervated cells.

V. Vascular Innervation and Efferent Inhibition It is well known that the carotid body is innervated not only by the sinus nerve but also by autonomic fibers coming from the superior cervical sympathetic ganglion via the ganglioglomerular nerve(s) (see Adams, 1958, for references). Degeneration experiments have shown that carotid body blood vessels are innervated by sympathetic fibers (Biscoe and Stehbens, 1967) and it has been demonstrated that sympathetic stimulation usually increases the chemoafferent activity in the sinus nerve as a consequence of a decrease in the carotid body blood flow (Floyd and Neil, 1952; Eyzaguirre and Lewin, 1961; Biscoe and Purves, 1967). However, it has been recently suggested that sympathetic efferent fibers may have more variable effects, sometimes leading to inhibitory effects, sometimes to different kinds of excitatory effects, or sometimes to no effect at all upon the chemoafferent discharge (O’Regan, 1977). Autonomic fibers are also present in the sinus nerve and there are both morphological and physiological results suggesting the occurrence of sympa-

308

ALAlN VERNA

thetic and parasympathetic fibers in this nerve. Both kinds of fibers seem to have opposite vasomotor effects. Thus, parasympathetic fibers appear to be vasodilatory, leading to a decrease in chemoafferent activity. However, a new controversy arose a few years ago concerning the possible occurrences of nonvasomotor inhibitory fibers in the sinus nerve. Some authors think that such fibers do exist, but others disagree with this and ascribe all inhibitory effects to vasomotor fibers. Besides, both light and electron microscope studies mention a barosensory nerve supply to carotid body blood vessels. From a morphological point of view, we must consider therefore the courses and destinations of vasomotor and barosensory fibers and look for possible nonvasomotor inhibitory fibers. A. VASOMOTOR INNERVATION

The first ultrastructural demonstration of nerve fibers related to carotid body blood vessels was carried out by Biscoe and Stehbens (1966). They described unmyelinated nerve fibers, sometimes as small as 0.1 pm in diameter, more or less invested by Schwann cells and situated close to blood vessels. Among these fibers, nerve endings were identified by their mitochondria and microvesicles (about 50 nm) and large dense-cored vesicle (65 to 100 nm) contents. However, in this study, the origins of nerve fibers and endings were not determined. Now it appears that carotid body blood vessels receive sympathetic and possibly parasympathetic postganglionic nerve fibers. 1 . Sympathetic Nerve Fibers a. Identification. The above-mentioned results of Biscoe and Stehbens (1966) were enlarged upon in 1967 when they described degenerative changes in the perivascular nerve fibers after cutting the ganglioglomerular nerve(s), i.e., after postganglionic sympathectomy. However, no changes were seen after cutting the preganglionic cervical sympathetic, This observation therefore established, ultrastructurally, the existence of a vascular postganglionic sympathetic innervation in the carotid body. This conclusion has been largely confirmed afterward, using different methods. It is, indeed, very easy to demonstrate postganglionic sympathetic fibers thanks to their ultrastructural and cytochemical feature-. In conventional electron microscopy these fibers are characterized by varicosities containing some large dense-cored vesicles and numerous small vesicles (about 60 nm in diameter) which are either “empty” or contain one dense granule. These different kinds of vesicles are known to contain noradrenaline (Bisby and Fillenz, 1971) and the occurrence of even a limited number of small granulated vesicles is a good criterion to identify noradrenergic neurons (see Taxi, 1969). It is also possible to demonstrate noradrenergic terminals by highresolution autoradiography after administration of tritiated noradrenaline. Finally, aminergic nerve endings can be located with the light microscope using the

ULTRASTRUCTURE OF THE CAROTID BODY

309

formol-induced fluorescence method (Falck et al., 1962). All these methods have been used in carotid body studies and converge to prove the existence of vascular postganglionic sympathetic innervation: Perivascular nerve profiles in the carotid body appear to contain small granulated vesicles (Fig. 10) (Rees, 1967b; Kondo, 1971; McDonald and Mitchell, 1975a,b); a dense network of aminergic varicose nerve fibers has been demonstrated by Falck’s method (Rees, 1967b; Korkala et al., 1974; Verna, 1975), but it must be pointed out that fluorescent fibers appear related not only to blood vessels but also sometimes to type I cells (see Section V,A,c below). Finally, labeled nerve fibers may be seen in the vicinity of blood vessels after tritiated noradrenaline administration (Verna, 1975). Thus, the existence of noradrenergic sympathetic fibers related to carotid body blood vessels is well proved. There are, usually, several perivascular nerve fibers enclosed in a given Schwann cell but these fibers may be of different origin (sympathetic and parasympathetic according to McDonald and Mitchell, 1975b). Noradrenergic nerve endings (or, more exactly, varicosities) are often more or less devoid of Schwann cell sheaths but the distance between noradrenergic endings and smooth muscle cells or pericytes is rarely less than 0.1 pm (“at distance” junction). Although there is, unfortunately, no report about the distribution of nerve fibers along blood vessels, it seems that noradrenergic

FIG. 10. Sympathetic nerve profile (N) next to a capillary (Ca) (rabbit carotid body). Sympathetic endings contain both large dense-cored vesicles (arrowheads) and small vesicles some of which have a dense core (arrows). E, Endothelial cell; F, cytoplasmic processes of fibrocytes; M, mitochondria; Py,cytoplasmic processes of pericytes; S , Schwann cell. Bar = 0.5 pm.

310

ALAIN VERNA

fibers are abundant around carotid body arteries, arterioles, and capillaries, but absent at the venous level. It must be emphasized that there are no smooth muscle cells but only pericytes around capillaries. b. Origin and Courses of Sympathetic Nerve Fibers. i. The ganglioglomerularnerve(s) pathway. As a rule, it can be said that carotid body aminergic fibers are unmyelinated axons of ganglion cells located in the superior cervical ganglion which reach the carotid body by the ganglioglomerular nerve(s). This has been shown by degeneration studies with the electron microscope (Biscoe and Stehbens, 1967; Rees, 1967b; Nishi and Stensaas, 1974; McDonald and Mitchell, 1975b) and with Falck’s method (Rees, 1967b; Verna, 1975). However, there are two possible variants: postganglionic fibers leaving the superior cervical ganglion to reach the carotid body via the sinus nerve and postganglionic fibers arising from sympathetic ganglion cells located in the carotid body itself. ii. The sinus nerve pathway. The occurrence of sympathetic nerve fibers in the cat sinus nerve has been suggested on physiological grounds (Biscoe and Sampson, 1968; Neil and O’Regan, 1971a). According to Biscoe and Sampson (1968) these fibers arise from postganglionic branches of the superior cervical ganglion joining the glossopharyngeal nerve and, then, course down the sinus nerve toward the carotid body. However, there is no histological study as to the destination of these fibers. Biscoe and Sampson (1968) have minimized their physiological importance, for the additional reason that such fibers are not always present. Furthermore, electrophysiological recordings have failed to demonstrate sympathetic fibers in the rabbit sinus nerve (Laurent and Jager-Barres, 1969) and, histologically, section of the rabbit sinus nerve does not seem to significantly affect the abundance of the carotid body noradrenergic innervation (A. Verna, unpublished observations). iii. Intrinsic sympathetic ganglion cells. McDonald and Mitchell (1975b) have observed few postganglionic sympathetic endings which remain intact after sinus nerve section and superior cervical ganglion removal (only three examples in sections from two rat carotid bodies). These nerve endings were indentified by their small granulated vesicles and were supposed to arise from sympathetic ganglion cells in the carotid body. Sympatheticganglion cells are usually located outside innervated organs (contrary to parasympathetic neurons), but their presence within the carotid body may be explained by the intimate connections which occur between the carotid body and the superior cervical ganglion anlagen during embryonic development (see Section VII). It is possible that the few myelinated nerve fibers present in the ganglioglomerularnerve(s) (Eyzaguirre and Uchizono, 1961) are sympathetic preganglionic fibers reaching intrinsic sympathetic neurons in the carotid body. However, we must underline that these neurons are few in number and, consequently, most sympathetic postganglionic nerve fibers in the carotid body arise from ganglion cells in the superior cervical ganglion.

ULTRASTRUCTURE OF THE CAROTID BODY

31 1

C. Functional Significance of Postganglionic Sympathetic Innervation. Sympathetic innervation is classically thought to be vasoconstrictory in the carotid body and therefore to mediate a decrease in blood flow (Daly et a l . , 1954; Purves, 1970). However, the effects of sympathetic excitation are far from clear: It has been suggested that sympathetic nerve fiber activity also tends to diminish the rate of carotid body oxygen consumption (Purves, 1970) but the mechanism of this effect is unknown (direct effect on carotid body cells? redistribution of blood?). More recently, O’Regan (1977) has studied the effects of sympathetic stimulation on chemoafferent activity. This author describes occasional inhibitory effects and two kinds of excitatory effects: The first kind is an early, transient increase whereas the second one is a slowly developing increase. The latter is presumed to be a vasoconstrictor a-adrenergic effect but the first excitatory effect is resistant to a-blockade and, thus, presumed to be nonvasomotor in nature. In this context, it may be worthwhile to recall that aminergic nerve endings have been observed close to type Vtype I1 cells by several authors. Kondo (1971) has shown a nerve terminal containing small granulated vesicles separated from a type I cell by only a slender type I1 cell process (guinea pig carotid body). Noradrenergic nerve profiles invested by thin processes of type I1 cells have been observed, autoradiographically, in the rabbit carotid body (Verna, 1975). McDonald and Mitchell (1975b) also report (but without giving any illustration) they have observed, in the rat carotid body, a few sympathetic nerve endings in contact with type I cells. This assertion has been repeated (but again without illustration) by Knoche and Kienecker (1977) concerning the rabbit carotid body. It seems therefore that carotid body-specific cells are surrounded by noradrenergic nerve endings (as clearly indicated by Falck’s method) which can even be in contact with them. Consequently, it may be legitimate to envisage sympathetic actions not only on carotid body blood vessels but also on the type I cell/type I1 cell/afferent ending complex. Nonvasomotor sympathetic actions have also been envisaged concerning the cochlea (Densert, 1974; Densert and Flock, 1974). However, it must be added that, after some controversies, the effects of noradrenaline upon chemosensory discharges appear to be weak and of short duration, at least in the cat carotid body (Llados and Zapata, 1978). Nevertheless, sympathetic effects exerted at the level of cell clusters, in addition to those on blood vessels, could explain the variable effects observed after sympathetic stimulation (O’Regan, 1977) or noradrenaline administration (Llados and Zapata, 1978).

2 . Parasympathetic Nerve Fibers The existence of a parasympathetic vascular innervation has been suggested by de Castro (1926, 1928, 1951) and de Castro and Rubio (1968). According to these authors there are parasympathetic ganglion cells in the carotid body which

3 12

ALAIN VERNA

are innervated by preganglionic fibers coursing in the sinus nerve. The postganglionic fibers were supposed to innervate smooth muscle cells of carotid body arteries or arterioles. However, this problem has received little attention and it was only in 1975 that ultrastructural evidence was presented on this subject by McDonald and Mitchell (1975b). These authors have described, in the rat carotid body, perivascular nerve endings containing small vesicles without dense cores and therefore, of nonsympathetic nature. These endings did not degenerate after severance of all nervous connection to the carotid body and, consequently, were presumed to arise from parasympathetic ganglion cells in the carotid body. Furthermore, McDonald and Mitchell (1975b) demonstrated that most of carotid body ganglion cells were denervated by cutting the glossopharyngeal nerve. Thus, these results are concordant not only with the light microscope studies of de Castro but also with physiological observations showing the occurrence in the cat sinus nerve of fibers whose activity increases the carotid body blood flow, an effect which is blocked by atropine (Neil and O’Regan, 1969a, 1971a). However, it is surprising that McDonald and Mitchell (1975b) describe parasympathetic vasomotor endings as being abundant since they only found 25 ganglion cells in one carotid body. Moreover, this number may be exceptional since, in the same species (rat), Hess (1977) found a mean number of about 5 ganglion cells per carotid body (actually from 1 to 8 neurons in 12 carotid bodies). Since type I cell processes also contain synaptic-like vesicles without dense cores and, of course, do not degenerate after denervation, there is an obvious possibility of confusion. Alternatively, it is possible that parasympathetic terminals innervate only certain portions of the vascular apparatus and it is regrettable that McDonald and Mitchell (1975b) describe the vasomotor innervation only in terms of endings on “blood vessels.” According to de Castro and Rubio (1968) postganglionic parasympathetic fibers innervate carotid body arteries and arterioles and are the efferent pathway of a reflex which tends to maintain a constant flow in spite of arterial pressure variations. The afferent part of this reflex has been attributed by these authors to baroreceptor endings in the wall of carotid body arteries.

B. BAROSENSORY INNERVATION De Castro (1940, 1951) has described, in silver-impregnated material, baroreceptor nerve endings in the adventitia of carotid body arterioles and in the arterial segment of arteriovenous shunts. Presumed baroreceptor endings were also described, in light microscopy, by Abraham (1968) around carotid body arteries of some species (dog), in the connective tissue of others (horse and bovides), and in vein walls of human carotid body. This question has been recently investigated with the electron microscope by Gorgas and Bock (1977). They describe mitochondria-rich nerve endings located in the tunica adventitia of

ULTRASTRUCTURE OF THE CAROTID BODY

313

small arteries (one to two layers of smooth muscle cells). These endings are branched lanceolate terminals which adhere helically or circularly to the arterial wall. Furthermore, Gorgas and Bock (1977) describe junctions between these presumed baroreceptor endings and slender processes of type I cells. Unfortunately, although their interpretation may be correct, it is absolutely impossible to affirm that a given nerve ending is barosensory or not on morphological criteria alone. This is more especially the case in the carotid body insofar as the ultrastructure of the putative barosensory terminals described by Gorgas and Bock (1977) is identical to the ultrastructure of the afferent nerve endings on type I cells. This criticism may also be applied to the description of axon swellings in the carotid body connective tissue as possible baroreceptor endings (Kobayashi, 1971b). Although such axon swellings which do not contact type Utype I1 cells actually exist (Nishi and Stensaas, 1974) their origin and significance are open to question. C. THEPROBLEM OF THE EFFERENT INHIBITION The problem of so-called efferent inhibition is, primarily, a physiological one. Its origin lies in the electrophysiological observations of Biscoe and Sampson (1968) who recorded spontaneous nervous activities from the central cut end of the sinus nerve (in the cat). This efferent activity was also demonstrated by Laurent and Jager-Barres (1969) in the rabbit and again in the cat by Neil and 0 ’Regan ( 1969b, 1971b) , Possible physiological effects of sinus nerve efferent activity were investigated by Neil and O’Regan (1969a, 1971a) and Fidone and Sat0 (1970). It was shown by these authors that electrical stimulation of the sinus nerve decreases the chemoafferent activity in that nerve. However it is necessary, in this kind of experiment, to distinguish true inhibitory effects from antidromic depression of afferent fibers due to stimulus spread along the nerve. This was done by Fidone and Sat0 (1970) who concluded that true efferent inhibition, mediated by unmyelinated fibers, may be elicited in the sinus nerve but appears to be less potent than antidromic depression. Furthermore, Neil and O’Regan (1969b, 197lb) demonstrated that the efferent activity recorded from slips of otherwise intact sinus nerves markedly increases when the chemoafferent activity is raised by systemic hypoxia or asphyxia. These results suggest a feedback inhibitory mechanism similar to the efferent control of other receptors such as the auditory system (Fex, 1962) or the lateralline organ of fishes (Russell, 1968). However, the results of Neil and O’Regan were severely criticized by Goodman (1973) who attributed their results to antidromic depression and vasomotor effects. There are, indeed, vasomotor nerve fibers present in the cat sinus nerve: Neil and O’Regan themselves (1969a, 197la) have demonstrated that stimulation

314

ALAIN VERNA

of the sinus nerve efferents usually increases the carotid body flow, an effect which can be blocked by atropine. It is well known that chemoafferent activity is inversely related to carotid body blood flow (Daly et al., 1954). Thus, there are three possible mechanisms which can decrease the chemoafferent discharge: (1) antidromic depression, either artifactual in the case of electrical stimulation or physiological if primary afferent depolarization or axon reflexes occur; (2) increase in the carotid body blood flow; (3) efferent inhibition directly exerted on chemoreceptors. According to Goodman (1973) the latter mechanism does not exist and the so-called “efferent inhibition” is only of vasomotor origin. But O’Regan (1975) showed that inhibition can still be observed in the ischemic carotid body, i.e., in the absence of blood flow and, therefore, of vasomotor effects. Furthermore, Willshaw (1975) was able to measure the carotid body blood flow and found that the inhibition is not accompanied by any significant change in blood flow. On the other hand, Belmonte and Eyzaguirre (1974) and McCloskey (1975) did not observe efferent inhibition in the absence of carotid body perfusion. Thus, the influence of vasomotor effects on chemoafferent discharge is not contested but the question remains: Are there other, more direct, inhibitory mechanisms? 1. Efferent Influences on Type I Cells The so-called efferent inhibition being mediated by sinus nerve fibers, we shall not discuss here the possible effects exerted by preganglionic sympathetic nerve endings on type I cells described by some authors and contested by others (see Section IV). This being said, the question now is: Do sinus nerve efferents affect type I cells? As we have seen, such fibers (of parasympathetic nature) may end on ganglion cells but there is no evidence of sinus nerve efferent fibers terminating on type I cells. Nevertheless, it has been shown that centrifugal sinus nerve activity may affect the type I cell content (Yates et al., 1970) and synthesis and release (Mills and Slotkin, 1975; Sampson et al., 1975) of catecholamines. Thus, we have to explain efferent effects without efferent terminals. One possibility is to admit, as we have already mentioned (see Section IV,E,l), that in addition to their afferent activity, sensory nerve terminals may exert efferent effects on type I cells. This phenomenon may occur under artificial circumstances, such as electrical stimulation of the sinus nerve, or under natural circumstances, either by an axon reflex mechanism as suggested by Nishi and Stensaas (1974) and VBzquez-Nin et al. (1977) or by the mechanism proposed by McDonald and Mitchell (1975a,b), i.e., a transmitter release (at the level of so-called reciprocal junctions) concomitant to nerve ending excitation. However, these mechanisms do not involve efferent fibers, which are known to be present in the sinus nerve, so we must consider yet another possibility of nonvasomotor inhibitory effects, namely, efferent control of the afferent nerve endings themselves.

ULTRASTRUCTURE OF THE CAROTID BODY

315

2. Efferent Influences on Afferent Nerve Fibers It has been suggested by some authors (Verna, 1973; King et al., 1975; Morgan et al., 1975) that the activity of the type I cell afferent innervation may be controlled by efferent fibers terminating on the afferent endings themselves. This interpretation was based on electron micrographs showing presumed nerve profiles (containing synaptic-like vesicles) which were presynaptic to nerve endings in contact with type I cells. Unfortunately, none of the authors who have described these junctions between nerve endings has been able to prove the neural nature of the presumed efferent terminal. So, it cannot be excluded, as pointed out by McDonald and Mitchell (1975a,b), that the presumed efferent “fibers” actually were type I cell processes since we know that such processes may contain numerous synaptic-like vesicles (see Section IV,C,2). Another possibility of efferent influences on the chemoafferent activity has been suggested which does not involve efferent fibers but involves interactions between afferent fibers running close together (Willshaw, 1977, discussion of his paper). However, this hypothesis has so far received no experimental support. Thus, there are both physiological and morphological data which explain “efferent inhibition” by activity of vasomotor fibers, whereas there are no conclusive morphological results, only physiological ones, supporting the existence of nonvasomotor “efferent inhibition” mediated by sinus nerve efferent fibers.

VI. Ultrastructural Changes after Stimulation of Chemoreceptor and Pathology Since it has been postulated (de Castro, 1928) and then demonstrated (Heymans and Bouckaert, 1930) that the carotid body is an arterial chemoreceptor, many attempts have been made to demonstrate morphological changes consecutive with stimulation by hypoxia, hypercapnia, asphyxia, and so on. So, Hollinshead ( 1945) has described cytological modifications of carotid body cells after severe hypoxia and similar investigations were also made with the electron microscope as early as 1958 by Hoffman and Birrel. The first studies were principally focused on type I cells but more recent reports also consider changes in nerve ending ultrastructure. A. CHANGES IN TYPEI CELLS The first report of changes in carotid body ultrastructure following anoxia was from Hoffman and Birrel (1958) who described the disappearance of type I cell dense-cored vesicles as a consequence of severe anoxia. However, these results are unreliable because of the poor fixation methods available at that time. Bliimcke et af. (1967a) also have studied the effects of oxygen deficiency on the rat carotid body ultrastructure. They submitted rats to gas mixtures containing

316

ALAIN VERNA

10,5, and 2.5% oxygen for 5 to 20 minutes but only found changes in type I cells after extreme hypoxia (10 minutes in 2.5% 0 2 ) . In these conditions, Blumcke et al. (1967a) observed an increased frequency of exocytosis pictures, suggesting the discharge of type I cell dense-cored vesicles. They also describe other changes in these cells such as swelling of the mitochondria, disintegration of the chromatin in the nucleus, and distension of the perinuclear cleft. Moreover, they show by formol-induced fluorescence microscopy that, after extreme oxygen deficiency (20 minutes in 2.5% 0 2 )the , catecholamine content of type I cells disappears. Blumcke et al. (1967a) conclude from their observations that type I cells release their catecholamines (by exocytosis) in response to hypoxia. Although this conclusion may be right (see Mills and Slotkin, 1975; Hellstrom et af., 1976; Hellstrom, 1977) the results of such extreme stimulations (the authors confess that rats lost consciousness after 1 minute in 2.5% 0,) seem to be doubtful. The swelling of the mitochondria was certainly unspecific and the exocytosis pictures were not convincing due to very marked osmotic artifacts enlarging the intercellular spaces up to 100 nm. In another paper, Blumcke et al. (1967b) considered the effects of hypercapnia on type I cells. Here again they used unphysiological stimuli (up to 75% C02) and, hence, it was not surprising that type I cell cytoplasm showed a “high degree of oedema. The authors also described a disintegration of type I cell dense-cored vesicles leading their dense cores to disappear in the cytoplasm. These aspects were probably due, in fact, to the use of collidine buffer in the fixative (Chen et al., 1969). Using somewhat better fixation methods, Al-Lami and Murray (1968a) reported that hypoxia (9% O2 for 45 minutes) leads to a slight increase in the relative number of type I cell dense-cored vesicles (in the cat carotid body). They concluded from this observation that type I cells do not release the content of their dense-cored vesicles after hypoxia. Such a conclusion is purely speculative since the observation could be also explained by an increased rate of synthesis which compensates an increased rate of discharge. The same remark may be made for the conclusion of Chen et al. (1969), who demonstrated, with a cytochemical method, that reserpine depletes the amine content of type I cell dense-cored vesicles whereas hypoxia does not. However this result does not validate their suggestion that type I cells do not release catecholamines in response to hypoxia. At the same time, Zapata et al. (1969) observed no detectable changes in the carotid body cytology induced by hypoxia (5% O2 in N, for 3 hours in cats) and, more recently, McDonald and Mitchell (1975a) found no striking changes in type I cells after hypoxia, except a slight enlargement of mitochondria and a decrease in the electron density of their matrix. On the other hand, Hellstrom (1977) reported similar changes about type I cell mitochondria but suggested that there is a small decrease in the number of dense-cored vesicles after hypoxia. ”

ULTRASTRUCTURE OF THE CAROTID BODY

317

In the above-reported studies only temporary periods of hypoxia were used (no more than a few hours) but other investigators have considered the consequences of chronically hypoxic conditions on the carotid body. It has been shown that subjects (man and animals) living at high altitudes have enlarged carotid bodies (Arias-Stella, 1969; Edwards et al., 1971b; Barer et al., 1972; Blessing and Wolff, 1973; Laidler and Kay, 1975). According to Blessing and Wolff (1973) and Laidler and Kay (1975) the enlargement of the carotid body in animals submitted to high altitude is due to an increased volume of cellular structures and capillaries. From an ultrastructural point of view, Edwards et al. (1972) reported only some minor changes between sea-level and high-altitude guinea pig carotid bodies. On the other hand, Moller et al. (1974) claims there is a marked increase in the number of dense-cored vesicles and mitochondria in type I cells of hypoxic rabbits. This assertion is only based on subjective grounds since the authors write “it is our impression that the number of dense-cored vesicles and mitochondria in the type I cells is at least doubled in the hypoxic rabbits. Such an “impression” is not conclusive because the amount of dense-cored vesicles in rabbit type I cells varies greatly from cell to cell (in a 1 to 10 range; A. Verna, unpublished observations). There is consequently an obvious problem of sampling. Contrary to Moller et al. (1974), Blessing and Kaldeweide (1975) found a noticeable decrease in the number of dense-cored vesicles following adaptation to a simulated altitude of 7000 m (in rats). Furthermore they observed dilated capillaries and thrombosis leading to type I cell degeneration. However, Laidler and Kay (1978) have made a more precise study: They have used stereological methods to investigate carotid bodies from rats living at a simulated altitude of 4300 m for 4 to 5 weeks. They found a 3-fold increase in the volume of type I cells (due to an increased volume of cytoplasm) in hypoxic rats. On the other hand, no change in the volume proportion of type I cells occupied by mitochondria was reported. The authors deduced from their data that the number of mitochondria in each type I cell was increased, but it must be noted that possible changes in length of mitochondria have not been considered. Dense-cored vesicles were reported to increase in diameter but without significant change in volume proportion of cytoplasm due to a reduction in their mean number per volume unit of cytoplasm. Thus, chronically hypoxic conditions really seem to affect type I cell ultrastructure. However, it is difficult to deduce from these results any conclusion whatsoever. Laidler and Kay (1978) suggest that type I cells react to hypoxia by an increased metabolic activity accompanied by an increased rate of dense-cored vesicle release. However, other interpretations may be proposed as well, as long as no changes are reported in the number of ribosomes. Finally, the effects of acute hemorrhagia were also studied by Korkala and Waris (1975) who have observed (and confirmed with a statistical method) that type I cell dense-cored vesicles move toward nerve endings in response to severe hemorrhagia. The authors concluded that type I cells probably have an inhibitory ”

318

ALAIN VERNA

modulating function, a conclusion which is not, actually, founded on their morphological observations.

B. CHANGES IN NERVE ENDINGS Modifications of nerve ending ultrastructure after chemoreceptor excitation have been reported only by McDonald and Mitchell (1975a, 1976) and McDonald (1977a,b). These authors describe conspicuous changes in afferent nerve terminals on type I cells following increased activity of these terminals (produced by hypoxia, by hypercapnia, or by antidromic electrical stimulation of the sinus nerve). These changes concern mitochondria and synaptic-like vesicles. After 1 minute of severe hypoxia (100% N2) the mitochondria become swollen and exhibit a more electron-lucent matrix than the mitochondria in hyperoxic animals. However, efferent nerve terminals do not show mitochondria1 swelling in response to hypoxia. In addition, the abundance and packing density of the synaptic-like vesicles in sensory endings on type I cells are decreased by stimulation of the chemoafferent activity: After hypoxia produced by ventilating rats for 10 minutes with 10 or 5% O2 in N2, the mean concentration of synaptic-like vesicles is reduced by about 27%. This observation is interpreted by the authors as reflecting a rate of exocytosis which exceeds the rate of vesicle formation, these vesicles being thought to contain a synaptic transmitter (McDonald and Mitchell, 1975a, 1976; McDonald, 1977a,b) (see Section IV,E,2).

C. PATHOLOGY 1. The Carotid Body in Respiratory Diseases and Anemia The structure of the carotid body has been investigated in patients suffering diseases which may affect the chemoreceptors by a correlative hypoxia. For example, Simhrszky and Lapis (1970) have studied the ultrastructure of carotid bodies in individuals suffering bronchial asthma. These carotid bodies were removed for therapeutic purposes (a somewhat surprising practice) but their ultrastructure appeared to be normal. On the other hand, carotid body enlargement has been described as a response to chronic hypoxia due to chronic bronchitis (Heath et al., 1970), emphysema (Edwards et al., 1971a), or cyanotic congenital heart disease (Lack, 1977). The effects of chronic anemia upon carotid body histology have also been investigated. Tramezzani et al. (1971) reported that carotid bodies enlarge in cats made chronically anemic by daily bleeding. The authors suggested, together on experimental grounds, that the carotid body controls erythropoiesis. This asser-

ULTRASTRUCTURE OF THE CAROTID BODY

319

tion led Winson and Heath (1973) to correlate carotid body weight to hemoglobin level but carotid bodies were found not to be heavier in anemia. Furthermore, the experimental data reported by Tramezzani et al. (1971) to support their statement were confirmed neither by Lugliani et al. (1971) nor by Hansen et al. (1973) who rejected the suggestion that the carotid body produces erythropoietin. Finally, there has also been an attempt to relate the carotid body size to the sudden infant death syndrome, but with inconclusive results (Naeye et al., 1976). 2. Carotid Body Tumors Carotid body tumors, although occurring infrequently, have been known for almost a century (Shamblin et al., 1971) and there is a considerable volume of literature on this subject. However, it seems that some confusion exists regarding the terminology of this disease. Carotid body tumors are sometimes referred to as “glomus tumors” but this term more frequently designates glomera of the skin. Lattes (1950) first used the name “nonchromaffin paraganglioma” for tumors of the carotid body and aortic body but also for those of the ganglion nodosum and the jugular body. This term was opposed to “chromaffin paraganglioma” or ‘‘pheochromocytoma” (which designates tumors of the adrenal medulla for example), but, as pointed out by Karnauchow (1963, chromaffinity is not a good criterion to subclassify these tumors. As in the normal carotid body, the chromaffinity of carotid body tumors is negative or doubtful (LeCompte, 1951), or positive only at the level of a few cells (Pryse-Davies et al., 1964). For this reason, Totten (1973) recommends the use of only “paraganglioma” accompanied by the anatomical site of the tumor. But the most usual term is perhaps “chemodectoma,” an appellation introduced by Mulligan (1950). Unfortunately, this term is used not only for carotid and aortic body tumors but also for tumors of other structures whose chemoreceptor function is not established. Only carotid and aortic body tumors are considered here. According to Shamblin et al. (1971) about 500 cases of human carotid body tumors have been reported up to 1970. These tumors are consequently infrequent. Among these cases only a few percent were malignant. The incidence of metastases is very low but some cases have been reported (Fanning et al., 1963; Whimster and Masson, 1970; Hortnagl et al., 1973; Villiaumey et al., 1974). The tumors usually grow slowly but may be bilateral or even multicentric. Familial cases have been described. One report suggests that carotid body tumors are more frequent in people living at high altitude (Saldana et al., 1973). Symptoms and clinical signs are ordinarily minimal and often due to compression of neighboring structures. However, a few cases of catecholamine secretion have been reported (Glenner et al., 1962; Hamberger et al., 1967; Hortnagl et al., 1973).

320

ALAIN VERNA

Carotid and aortic body tumors have also been described in other mammals and more especially in the dog (Parodi and Lekkas, 1970; Cheville, 1972). Whereas carotid body tumors are preponderant over aortic ones in man, the reverse is true for the dog (Parodi and Lekkas, 1970). Histologically, carotid body tumors are characterized by a lobulated and richly vasularized structure. Because of this vascular nature, biopsy is a hazardous procedure. The tumor cells are grouped in clusters and frequently show enlarged nuclei, Nuclear pleomorphism is sometimes considered as indicating a malignant character (Hortnagl ef al., 1973) but difficulty in judging carotid body tumors benign or malignant on histological grounds has been underlined by many authors (LeCompte, 1951; Pryse-Davies et al., 1964). Mitoses are always rare. Despite the inconstant results of the chromaffin reaction, catecholamines have been characterized in carotid body tumors with fluorescence microscopy (Grimley and Glenner, 1967). Nerve endings have been described by some authors (Costero and Barroso-Moguel, 1961; Pryse-Davies et al., 1964) but have not been observed by others (Grimley and Glenner, 1967; Macadam, 1969; Alpert and Bochetto, 1974). Type I and type I1 cells have been characterized in carotid body tumors. However, type I1 cells are usually few or even absent. Type I cells are polygonal and have a “light, ” “dark, or intermediate appearance (Alpert and Bochetto, 1974). Dense-cored vesicles have been described in these cells by many authors (Grimley and Glenner, 1967; Toker, 1967; Macadam, 1969; Capella and Solcia, 1971; Hortnagl ef al., 1973; Alpert and Bochetto, 1974). It has been shown by Lishajko (1970) that isolated dense-cored vesicles from a human carotid body tumor release and take up dopamine. On the other hand, Capella and Solcia (197 1) suggest that some tumoral cells may contain serotonin (the others containing some polypeptidic substance), whereas Hortnagl et al. (1973) characterize noradrenaline and dopamine P-hydroxylase in a liver metastase of a carotid body tumor. Type I cell mitochondria are more or less numerous but frequently swollen due to poor fixation conditions. In conclusion, the study of carotid body tumors seems to add little to our knowledge of the normal carotid body. ”

VII. Embryology and Development The embryology of the carotid body has been very controversial until recent years, particularly with respect to the origin of type I cells (see Adams, 1958, for historical review). According to Rogers (1965) the carotid body appears as a primary condensation of cells on the third aortic arch, but this author concluded that it was impossible to know whether these cells derive from local mesenchyme or migrate from the neural crest. However, this problem has been solved recently, at least for birds, by Le Douarin et al. (1972) and Pearse ef al. (1973).

ULTRASTRUCTURE OF THE CAROTID BODY

32 1

These authors have demonstrated that carotid body type I (and possibly type 11) cells are of neural crest origin. This result was obtained from heterospecific grafting experiments together with histological and histochemical studies making it possible to identify the cells of the graft among the cells of the receiver. Although these studies were made on avian species, there is no reason to postulate a different origin for the mammalian carotid body type I cells. The development of the carotid body anlage has been investigated recently with the light, fluorescence, and electron microscopes. It has been shown with the formol-induced fluorescence method that type I cells of the midterm human fetus contain catecholamines (Hervonen and Korkala, 1972; Korkala and Hervonen, 1973). Furthermore, it was shown by Korkala and Hervonen (1973) that a connecting cell cord links the carotid body to the sympathetic anlage in the 7-week-old human fetus. This process, which is made of small fluorescent cells, disappears after the tenth week of development. The authors suggest a migration of cells from the sympathetic trunk to the carotid body. In the newborn and 2-week-old rat, Korkala et al. (1974) observed a bundle of fluorescent nerve fibers which connect the sympathetic superior cervical ganglion to the carotid body. Occasionally, fluorescent cells, identical to carotid body type I cells, were seen inside or beside the fluorescent nerves. In the mouse embryo, catecholamine-containing cells were identified in the carotid body anlage at 2 weeks of gestation. These cells were able to take up L-dopa which increases their fluorescence but does not modify the number of fluorescent cells (Fontaine, 1974). There are two recent electron microscope studies on the embryonic development of the carotid body. The first one is devoted to the midterm human fetus (Hervonen and Korkala, 1972). The histochemically demonstrated occurrence of catecholamines in cells of this material has been correlated by Hervonen and Korkala (1972) with the occurrence of dense-cored vesicles in many cells, as shown by electron microscopy. These dense-cored-vesicle-containing cells (that is, type I cells) occur in small groups surrounded by processes of other cells which do not contain dense-cored vesicles (type I1 cells). However, type I cells are not always completely enveloped by type I1 cells (as in the adult organ). The dense-cored vesicle size does not differ noticeably from cell to cell (but the authors give no quantitative data). Type I1 cells surround not only type I cell groups but also capillaries and nerve fibers. Some type I1 cells were occasionally found by Hervonen and Korkala (1972) in contact with both endothelial and type I cells. Capillaries are abundant but do not show fenestrations. Type I cells are already richly innervated: Many type I cells are in contact with nerve endings which contain agranular vesicles (40 to 50 nm) and a few larger dense-cored vesicles. Several sites of increased electron density were observed by Hervonen and Korkala (1972) along the course of the longer nerve endings. These authors, in conclusion to their study, underlined that the fetal carotid body looks surpris-

322

ALAIN VERNA

ingly mature and suggest there may be some sort of chemoreceptor activity during pregnancy, as previously described by Biscoe et al. (1969). The second electron microscope study of the embryonic development of the carotid body is from Kondo (1975) and concerns the rat. In the 11-mm embryo (13 days of gestation) this author describes the carotid body anlage as a thickening of the wall of the third branchial artery. This thickening is made of undifferentiated cells and small blood vessels. There are also unmyelinated nerve fibers which contain clear vesicles about 40 to 70 nm in diameter and dense-cored vesicles about 100 nm in diameter. Although some nerve fibers were enveloped by processes of undifferentiated cells, membrane specializations were absent. In the 12-mm embryo (14 days of gestation) dense-cored vesicles appear in some of the undifferentiated cells, particularly those located at the boundaries of the anlage. At this stage, bundles of small unmyelinated nerve fibers are common and sometimes touch undifferentiated cells or cells containing dense-cored vesicles. However there are no junctional specializations. The latter appear in the 17-mm embryo (16.5 days of gestation). Kondo (1975) describes two kinds of junctions with embryonic type I cells, one with membrane densification and vesicles clustered inside the nerve ending, the other with dense material and vesicles inside the type I cell. At this stage, fenestrations of the blood vessels occur. Some type I cells exhibit slender processes as long as 25 pm. Finally, in the 20-mm embryo (17 to 17.5 days of gestation) the carotid body anlage is completely separated from the wall of the internal carotid artery. Undifferentiated cells show a tendency to envelop adjacent type I cells. The number of junctions with nerve endings remains small but the length of the terminals increases and more vesicles are associated with the membrane specializations. These studies are therefore concordant as to the mature appearance of the carotid body before birth. However, the problem of the factors controlling type I cell differentiation is still obscure. Korkala and Hervonen (1973) suggest humoral influences coming from the blood vessels but Kondo (1975) points out that type I cells can be first identified at the anlage periphery and suggests that these cells migrate from a distant site, already differentiated, at least partially. It is obvious that this kind of study comes up against the difficulty involved in identifying undifferentiated cells.

VIII. Concluding Remarks Despite some remaining problems concerning type I cell subdivision or autonomic innervation, for example, it can be said that the carotid body ultrastructure is well known. Conclusive results, recently obtained, have put an end to the confusion about the nature of nerve endings on type I cells. It is clear now that the elementary structure of the carotid body is a dense-cored vesicle-containing cell situated close to a capillary and innervated by a sensory fiber, the cell and the

ULTRASTRUCTURE OF THE CAROTID BODY

323

nerve ending being more or less invested by a satellite cell. However, further data about the three-dimensional structure of the carotid body would be desirable. Such investigations have been undertaken by means of reconstruction from serial sections, a method recently improved by use of computers (see Seidl et al., 1977; Lubbers et al., 1977). Unfortunately, morphological methods do not make it possible to identify the transducer element of the chemosensory process. Morphologists can stimulate chemoreceptors and look for specific changes by comparing the structure of stimulated and unstimulated carotid bodies. However, as we have seen, these changes are subtle and difficult to interpret. On this point, it could even appear somewhat paradoxical to study the structure of unstimulated chemoreceptors after a chemical fixation! Morphologists can also compare the structure of type I cells to the structure of other cells. This procedure leads to conclusions such as: Type I cells are sensory cells, type I cells are secretory cells, type I cells are interneurons, type I cells are paraneurons . . . ! These varied interpretations show with some obviousness that type I cells, actually, have no real specific cytological character. They even look like relatively undifferentiated cells with a high nucleus/cytoplasm ratio. Nevertheless, type I cells (andor type I1 cells) appear necessary to the chemosensory process (Verna et al., 1975; Zapata et al., 1976, 1977) but their exact role, if any, in the mechanism of chemoreceptor excitation has not been elucidated (see Acker and Pietruschka, 1977; Roumy and Leitner, 1977; Torrance, 1977, for current hypotheses). On this subject, intracellular recordings from carotid body slices (Eyzaguirre et al., 1977) or on carotid body cells in culture (Acker and Pietruschka, 1977) look promising. It has been shown by these methods that type I cells have a membrane potential of relatively low value, that they do not generate spikes, and that their membrane potential and input resistance are affected by various chemical and physical stimuli. But, once again, results are difficult to interpret: Among stimuli increasing the chemoafferent discharge, some hyperpolarize (temperature increase), others depolarize (acidity, interuption of superfusion, hyperosmolarity), and others do not seem to affect type I cell membrane potential and input resistance (COz, hypoxia, cyanide, acetylcholine) (Eyzaguirre et al., 1977). These approaches are, nevertheless, at the root of the problem which is, finally, the function(s) of these mysterious cells called type I cells.

REFERENCES Abbott, C. P., and Howe, A . (1972). Acra Anat. 81, 609. Abbott, C. P., Daly, M. de B . , and Howe, A. (1972). Acfu Anar. 83, 161. Abrahbm, A. (1968). In “Arterial Chemoreceptors” (R. W. Torrance, ed.), p. 57. Blackwell, Oxford.

324

ALAIN VERNA

Acker, H., and Pietruschka, F. (1977). In “Chemoreception in the Carotid Body” (H. Acker er a l . , eds.), p. 92. Springer-Verlag, Berlin and New York. Adams, W. E. (1958). “The Comparative Morphology of the Carotid Body and Carotid Sinus.” Thomas, Springfield, Illinois. Akert, K.,Pfenninger, K.,Sandri, C., and Moor, H. (1972). In “Structure and Function of Synapses” (G. D. Pappas and D. P. Purpura, eds.), p. 67. Raven, New York. Al-Lami, F., and Murray, R. G. (1968a). Anar. Rec. 160, 697. AI-Lami, F., and Murray, R. G. (1968b). J . Ulrrastruct. Res. 24, 465. Alpert, L. I., and Bochetto, J. F. (1974). Cancer 34, 564. Angell-James, J. E., and Daly, M. de B. (1969). J. Physiol. (London) 201, 87. Arias-Stella, J. (1969). Am. J . Pathol. 55, 82a. Arnold, J. (1865). Virchows Arch. Pathol. Anat. Physiol. 33, 190. Ballard, K . J . , and Jones, J. V. (1971). J . Physiol. (London) 219, 747. Ballard, K. J . , and Jones, J . V. (1972). J. Physiol. (London) 227, 87. Barer, G. R., Edwards, C. W., and Jolly, A. I. (1976). Clin. Sci. Mol. Med. 50, 311. Belmonte, C., and Eyzaguirre, C. (1974). J . Neurophysiol. 37, 1131. Belmonte, C., GonzAlez, C., and Garcia, A. G. (1977). In “Chemoreception in the Carotid Body” (H. Acker et a / . , eds.), p. 99. Springer-Verlag, Berlin and New York. Benedeczky, I., and Smith, A. (1972). Z . Zellforsch. Mikrosk. Anar. 124, 367. Bisby, M., and Fillenz, M. (1971). J. Physiol. (London) 215, 163. Biscoe, T. J. (1971). Physiol. Rev. 51, 437. Biscoe, T. J., and Pallot, D. (1972). Experientia 28, 33. Biscoe, T. J., and Purves, M. J. (1967). J. Physiol. (London) 190, 389. Biscoe, T. J., and Sampson, S. R. (1968). J . Physiol. (London) 196, 327. Biscoe, T. J., and Silver, A. (1966). J. Physiol. (London) 183, 501. Biscoe, T. J., and Stehbens, W. E. (1966). J . Cell Eiol. 30, 563. Biscoe, T. J., and Stehbens, W. E. (1967). Q. J. Exp. Physiol. Cogn. Med. Sci. 52, 31. Biscoe, T. J., and Taylor, A. (1963). J. Physiol. (London) 168, 332. Biscoe, T. J., Purves, M. J., and Sampson, S. R. (1969). J . Physiol. (London) 202, 1. Biscoe, T. J., Lall, A., and Sampson, S. R. (1970). J . Physiol. (London) 208, 133. Black, A. M. S., Comroe, J. H., Jr., and Jacobs, L. (1972). Am. J. Physiol. 223, 1097. Blessing, M. H., and Kaldeweide, J. (1975). Virchows Arch. E . 18, 315. Blessing, M. H., and Wolff, H. (1973). Virchows Arch. A . 360, 79. Bliimcke, S., Rode, J., and Niedorf, H. R. (1967a). Z. Zellforsch. Mikrosk. Anar. 80, 52. Bliimcke, S., Niedorf, H. R., and Rode, J . (1967b). Z. Zellforsch. Mikrosk. Anat. 81, 557. Bock, P., and Gorgas, K. (1976a). Cell Tissue Res. 170, 95. Bock, P., and Gorgas, K. (1976b). In “Chromaffin, Enterochromaffin and Related Cells’’ (R. E. Coupland and T. Fujita, eds.), p. 355. Elsevier, Amsterdam. Bock, P., Stockinger, L.,and Vyslonzil, E. (1970). Z . ZeNforsch. Mikrosk. Anar. 105, 543. Bouchaud, C. (1974). J. Microsc. (Paris) 20, 215. Capella, C., and Solcia, E. (1971). Virchows Arch. E . 7, 37. Celestino da Costa, A. (1944). Arch. Porr. Sci. Biol. 8, 129. Chen, LLi, and Yates, R. D. (1969). J . Cell. Eiol. 42, 794. Chen, I.Li, Yates, R. D., and Duncan, D. (1969). J. Cell Eiol. 42, 804. Cheville, N. F. (1972). Ver. Pathol. 9, 166. Chiba, T. (1972). J . Electron Microsc. 21, 139. Chiba, T., and Yamauchi, A . (1970). Z. Zellforsch. Mikrosk. Anar. 108, 324. Chiocchio, S. R., Biscardi, A. M., and Tramezzani, H. (1967). Science 158, 790. Chiocchio, S . R., King, M. P., and Angelakos, E. T. (1971a). Hisrochemie 25, 52. Chiocchio, S. R., King, M. P., Carballo, L., and Angelakos, E. T. (1971b). J. Hisrochem. Cyrochem. 19, 621.

ULTRASTRUCTURE OF THE CAROTID BODY

325

Chungcharoen, D., Daly, M. de B., and Schweitzer, A. (1952). J . Physiol. (London) 117, 347. Costero, I., and Barroso-Moguel, R. (1961). Am. J. Pafhol. 38, 127. Daly, M. de B., Lambertsen, C. J., and Schweitzer, A. (1954). J. Physiol. (London) 125, 67. Dearnaley, D. P., Fillenz, M., and Woods, R. I. (1968). Proc. R. SOC.London, Ser. B 170, 195. de Boissezon, P. (1943). Bull. Hisfol. Tech. Microsc. 20, 136. de Castro, F. (1926). Trab. Lab. Invest. Biol. Univ. Madrid 24, 365, de Castro, F. (1928). Trab. Lab. Invest. Biol. Univ. Madrid 25, 331. de Castro, F. (1940). Trab. Inst. Cajal Invest. Biol. 32, 297. de Castro, F. (1951). Acra Physiol. Scand. 22, 14. de Castro, F. (1962). Arch. In?. Pharmacodyn. Ther. 139, 212. de Castro, F., and Rubio, M. (1968). In “Arterial Chemoreceptors” (R. W. Torrance, ed.), p. 267. Blackwell, Oxford. de Kock, L. L. (1956). Nature (London) 177, 1084. de Kock, L. L., and Dunn, A. E. G. (1966). Acra Anar. 64, 163. de Kock, L. L., and Dunn, A. E. G. (1968). In “Arterial Chemoreceptors” (R. W. Torrance, ed.), p. 179, Blackwell, Oxford. Densert, 0. (1974). Acta Oto-Laryngol. 78, 345. Densert, O . , and Flock, d;. (1974). Acra Oro-Laryngol. 77, 185. Docherty, R. J . , and McQueen, D. S. (1977). J. Physiol. (London) 270, 73P. Douglas, W. W., Nagasawa, J., and Schultz, R. (1971). Mem. SOC. Endocrinol. 19, 353. Edwards, C., Heath, D., and Harris, P. (1971a). J . Parhol. 104, 1. Edwards, C., Heath, D., Harris, P., Castillo, Y., Kriiger, H., and Arias-Stella, J. (1971b). J. Pathol. 104, 231. Edwards, C., Heath, D., and Harris, P. (1972). J . Parhol. 107, 131. Eranko, L., and Eranko, 0. (1972). Acra Physiol. Scand. 84, 125. Eyzaguirre, C., and Gallego, A. (1975). In “The Peripheral Arterial Chemoreceptors” (M. J. Purves, ed.), p. 1. Cambridge Univ. Press, London and New York. Eyzaguirre, C., and Lewin, J. (1961). J. Physiol. (London) 159, 251. Eyzaguirre, C., and Uchizono, K. (1961). J. Physiol. (London) 159, 268. Eyzaguirre, C., and Zapata, P. (1968). J . Physiol. (London) 195, 589. Eyzaguirre, C., Koyano, H., and Taylor, J. R. (1965). J . Physiol. (London) 178, 463. Eyzaguirre, C., Nishi, K., and Fidone, S. (1972). Fed. Proc., Fed. Am. SOC. Exp. Biol. 31, 1385. Eyzaguirre, C., Baron, M., and Gallego, R. (1977). In “Tissue Hypoxia and Ischemia” (M. Reivich et a/., eds.), p. 209. Plenum, New York. Falck, B., Hillarp, N.-A., Thieme, G., and Torp, A. (1962). J. Histochem. Cytochem. 10, 348. Fanning, J. P., Woods, F. M., and Christian, H. J. (1963). J. Am. Med. Assoc. 185, 49. Fex, J. (1962). A d a Physiol. Scand. 55, Suppl. 189, 1. Fidone, S. J., and Sato, A. (1969). J. Physiol. (London) 205, 527. Fidone, S. J., and Sato, A. (1970). Brain Res. 22, 181. Fidone, S. J., Stensaas, L. J., and Zapata, P. (1975). J . Neurobiol. 6, 423. Fidone, S. J., Weintraub, S . T., and Stavinoha, W. B. (1976). J . Neurochem. 26, 1047. Fidone, S. J., Zapata, P., and Stensaas, L. J. (1971). Brain Res. 124, 9. Floyd, W. F., and Neil, E. (1952). Arch. Int. Pharmacodyn. Ther. 91, 230. Fontaine, J. (1974). Arch. Anat. Microsc. Morphol. Exp. 63, 217. Garner, C. M., and Duncan, D. (1958). Anat. Rec. 130, 691. Glenner, G. G., Grout, J. R., and Roberts, W. C. (1962). Arch. Pathol. 73, 230. Goldberg, A. M., Lentz, A. P., and Fitzgerald, R.S. (1978). Brain Res. 140, 374. Gomez, L. P. (1908). Am. J. Med. Sci. 136, 98. Gonzalez, C., and Fidone, S. (1977). Neurosci. Left. 6, 95. Goodman, N. W. (1973). J. Physiol. (London) 230, 295. Goodman, N. W., and McCloskey, D. I. (1972). Brain Res. 39, 501.

326

ALAlN VERNA

Gorgas, K., and Bock, P. (1977). In “Chemoreception in the Carotid Body” (H. Acker e t a l . , eds.), p. 55. Springer-Verlag, Berlin and New York. Gray, E. G. (1963). J. Anat. 97, 101. Graziadei, P. P. C. (1973). Tissue & Cell 5, 113. Grimley, P. M., and Glenner, G. G. (1967). Cancer 20, 1473. Grimley, P. M., and Glenner, G. G. (1968). Circulation 37, 648. Guth, L. (1971). In “Handbook of Sensory Physiology” (L. M.Beidler, ed.), Vol. 4, Part 2, p. 63. Springer-Verlag, Berlin and New York. Hamberger, C. A,, Hamberger, C. B., Wersiill, J . , and Siigermark, J. (1967). Acfa Pnrhol. Microbiol. Scand. 69, 489. Hansen, A. J., Fogh, J . , MpllgPrd, K., and Sgrensen, S. C. (1973). Respir. Physiol. 18, 101. Heath, D., and Edwards, C. (1971). Cardiovasc. Res. 5, 303. Heath, D., Edwards, C., and Harris, P. (1970). Thorax 25, 129. Hellstrom, S. (-1975aj.J . Neurocytol. C,‘Y7. Hellstrom, S. (1975b). J. Neurocytol. 4, 439. Hellstrom, S. (1977). In “Chemoreception in the Carotid Body” (H. Acker et al.. eds.), p. 122. Springer-Verlag, Berlin and New York. Hellstrom, S.,and Koslow, S. H. (1976). Brain Res. 102, 245. Hellstrom, S., Hanbauer, I., and Costa, E. (1976). Brain Res. 118, 352. Hervonen, A., and Korkala, 0. (1972). Z . Anat. Entwicklungsgesch. 138, 135. Hess, A. (1968). In “Arterial Chemoreceptors” (R. W. Torrance, ed.), p. 51. Blackwell, Oxford. Hess, A. (1975a). In “The Peripheral Arterial Chemoreceptors” (M. J. Purves, ed.), p. 51. Cambridge Univ. Press, London and New York. Hess, A. (1975b). Brain Res. 98, 348. Hess, A. (1976). Tissue & Cell 8, 381. Hess, A. (1977a). Acta Anat. 97, 307. Hess, A. (1977b). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 201. Springer-Verlag, Berlin and New York. Hess, A., and Zapata, P. (1972). Fed. Proc., Fed. Am. SOC. Exp. Biol. 31, 1365. Heymans, C., and Bouckaert, J. J. (1930). J. Physiol. (London) 69, 254. Heymans, C., and Neil, E. (1958). “Reflexogenic Areas of the Cardiovascular System,” p. 269. Churchill, London. Heymans, C., Bouckaert, J. J., and Regniers, P. (1933). “Le sinus carotidien et la zone homologue cardioaortique,” p. 344. Doin, Paris. Hoffman, H., and Birrel, J. H. W. (1958). Acfa Anat. 32, 297. Hoglund, R. (1967). Z. Zellforsch. Mikrosk. Anat. 76, 568. Hollinshead, W. H. (1942). Anat. Rec. 84, 1. Hollinshead, W. H. (1945). Anat. Rec. 92, 255. Hortnagl, H., Hortnagl, H.,Propst, A., Schwingshackl, H., Weiser, and Winckler, H. (1973). Virchows Arch. B . 12, 330. Howe, A., and Neil, E. (1971). In “Handbook of Sensory Physiology” (E. Neil, ed.), Vol. 3, Part 1, p. 47. Springer-Verlag, Berlin and New York. Joels, N., and Neil, E. (1968). In “Arterial Chemoreceptors” (R.W. Torrance, ed.), p. 153. Blackwell, Oxford. Jones, J. V. (1975). In “The Peripheral Arterial Chemoreceptors” (M. J. Purves, ed.), p. 143. Cambridge Univ. Press, London and New York. Kamauchow, P. N. (1965). Can. Med. Assoc. J . 92, 1298. Keller, H. P., Schafer, D., and Lubbers, D. W. (1972). Z. Naturforsch., Teil E 27, 1118. King, A. S . , McLelland, J., Cook, R. D., King, D. Z., and Walsh, C. (1974). Respir. Physiol. 22, 21.

ULTRASTRUCTURE OF THE CAROTID BODY

327

King, A. S., King, D. Z., Hodges, R. D., and Henry, J . (1975). Cell Tissue Res. 162, 459. Knoche, H., and Addicks, K. (1976). Cell Tissue Res. 173, 77. Knoche, H., and Kienecker, E.-W. (1977). Cell Tissue Res. 184, 103. Knoche, H., Alfes, H., and Mollmann, H. (1969). Experientia 25, 515. Kobayashi, S. (1968). Arch. Hisiol. Jpn. 30, 95. Kobayashi, S. (1971a). Arch. Histol. Jpn. 33, 319. Kobayashi, S. (1971b). Arch. Hisiol. Jpn. 33, 397. Kobayashi, S. (1976). Arch. Histol. Jpn. 39, 295. Kobayashi, S. (1977). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 221. Springer-Verlag, Berlin and New York. Kobayashi, S., and Uehara, M. (1970). Arch. Histol. Jpn. 32, 193. Koelle, G. B. (1969). Fed. Proc., Fed. Am. SOC. Exp. B i d . 28, 95. Kohn, A. (1900). Arch, Mikrosk. Anat. Entwicklungsmech. 56, 81. Kondo, H. (1971). J. Ultrastruci. Res. 37, 544. Kondo, H. (1975). Am. 1. Anat. 144, 275. Kondo, H. (1976). Cell Tissue Res. 173, 1. Korkala, 0.. and Hervonen, A. (1973). Histochemie 37, 287. Korkala, O., and Waris, T. (1975). Cell Tissue Res. 158, 355. Korkala, O., Eranko, O., Partanen, S., Eranko, L., and Hervonen, A. (1973). Histochem. J . 5,479. Korkala, O., Partanen, S., and Eranko, 0. (1974). Z. Anar. Eniwicklungsgesch. 143, 135. Lack, E. E. (1977). Hum. Pathol. 8, 39. Laidler, P., and Kay, J. M. (1975). J. Pathol. 117, 183. Laidler, P., and Kay, J . M. (1978). J. Neurocyiol. 7, 183. Lasek, R. J., Joseph, B. S., and Whitlock, D. G. (1968). Brain Res. 8, 319. Lattes, R. (1950). Cancer 3, 667. Laurent, P., and Barres, M. C. (1964). C.R. Hebd. Seances Acad. Sci. 259, 2694. Laurent, P., and Jager-Banes, M. C. (1969). J. Physiol. (Paris) 61, 403. LeCompte, P. M. (1951). In “Atlas of Tumor Pathology,” Sect. IV, Fasc. 16. Armed Forces Inst. Pathol., Washington, D.C. Le Douarin, N., Le Lievre, C., and Fontaine, J. (1972). C . R. Hebd. Seances Acad. Sci., Ser. D 275, 583. Leitner, L.-M., and Liaubet, M. J . (1971). Pfluegers Arch. 323, 315. Leonhardt, H. (1976). Cell Tissue Res. 174, 99. Lever, J. D., and Boyd, J. D. (1957). Nature (London) 179, 1082. Lever, J. D., Lewis, P. R., and Boyd, J . D. (1959). J . Anat. 93, 478. Lishajko, F. (1970). Acia Physiol. Scand. 79, 533. Llados, F., and Zapata, P. (1978). J. Physiol. (London) 274, 501. Lubbers, D. W., Teckhaus, L., and Seidl, E.(1977). In “Chemoreception in the Carotid Body” (H. Acker ei al., eds.), p. 62. Springer-Verlag, Berlin and New York. Lugliani, R., Whipp, B. J., Winter, B., Tanaka, K. R., and Wasserman, K. (1971). N . Engl. J . Med. 285, 1112. Luschka, H. (1862). Arch. Anat. Physiol. Wiss. Med. 405. Macadam, R. F. (1969). J . Paihol. 99, 101. McCloskey, D. I. (1975). Respir. Physiol. 25, 53. McDonald, D. (1977a). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 207. Springer-Verlag, Berlin and New York. McDonald, D. (1977b). Am. Rev. Respir. Dis. 115, 193. McDonald, D. M., and Mitchell, R. A. (1975a). In “The Peripheral Arterial Chernoreceptors” (M. J. Purves, ed.), p. 101. Cambridge Univ. Press, London and New York. McDonald, D. M., and Mitchell, R. A. (1975b). J . Neurocytol. 4, 177.

328

ALAIN VERNA

McDonald, D. M., and Mitchell, R. A. (1976). In “Morphology and Mechanisms of Chemoreceptors” (A. S. Paintal, ed.), p. 248. Vallabhbhai Patel Chest Inst., Delhi. Matsuura, S. (1973). J. Physiol. (London) 235, 57. Matthiessen, D., Mollmann, H.,and Knoche, H. (1973). Z . Zellforsch. Mikrosk. Anat. 138, 133. Metz, B. (1969). Respir. Physiol. 6, 386. Mills, E. (1972). Fed. Proc., Fed. Am. Soc. Exp. Biol. 31, 1394. Mills, E. (1975). In “The Peripheral Arterial Chemoreceptors” (M. J. Purves, ed.), p. 373. Cambridge Univ. Press, London and New York. Mills, E., and Slotkin, T. A. (1975). Life Sci. 16, 1555. Mitchell, R. A., Sinha, A. K., and McDonald, D. M. (1972). Brain Res. 43, 681. Mqller, M., MQllgArd, K., and Sqrensen, S. C. (1974). J. Physiol. (London) 238, 447. Monckeberg, I. G. (1905). Beitr. Pathol. Anat., Allg. Pathol. 38, 1. Monti-Bloch, L., and Eyzaguirre, C. (1977). Neurosci. Abstr. 3,458. Morgan, M., Pack, R. J., and Howe, A. (1975). Cell Tissue Res. 157, 255. Morita, E., Chiocchio, S. R., and Tramezzani, J. H. (1969). J. Ultrmtruct. Res. 28, 399. Mulligan, R. M. (1950). Am. J. Parhol. 26, 680. Munger, B. L. (1971). In “Handbook of Sensory Physiology” (W. R. Loewenstein, ed.), Vol. 1, p. 523. Springer-Verlag, Berlin and New York. Naeye, R. L., Fisher, R., Ryser, M., and Whalen, P. (1976). Science 191, 567. Neil, E., and O’Regan, R. G. (1969a). J. Physiol. (London) 200, 69P. Neil, E., and O’Regan, R. G. (1969b). J. Physiol. (London) 205, 20P. Neil, E., and O’Regan, R. G. (1971a). J. Physiol. (London) 215, 15. Neil, E., and O’Regan, R. G. (1971b). J. Physiol. (London) 215, 33. Nishi, K . (1976). In “Morphology and Mechanisms of Chemoreceptors” (A. S. Paintal, ed.), p. 1. Vallabhbhai Patel Chest Inst., Delhi. Nishi, K. (1977). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 145. Springer-Verlag, Berlin and New York. Nishi, K., and Stensaas, L. J. (1974). Cell Tissue Res. 154, 303. O’Regan, R. G. (1975). In “The Peripheral Arterial Chemoreceptors” (M. J. Purves, ed.), p. 221. Cambridge Univ. Press, London and New York. O’Regan, R. G. (1977). Ir. J. Med. Sci. 146, 199. Osborne, M. P., and Butler, P. J. (1975). Nature (London) 254, 701. Paintal, A. S. (1967). J. Physiol. (London) 189, 63. Paintal, A. S. (1968). In “Arterial Chemoreceptors” (R. W. Torrance, ed.), p. 253. Blackwell, Oxford. Paintal, A. S. (1969). J. Phpiol. (London) 204, 94P. Palay, S. L. (1958). Exp. Cell Res. 5 , 275. Parodi, A. L., and Lekkas, S. (1970). Bull. Cancer 57, 487. Pearse, A. G. E. (1969). J. Histochem. Cyrochem. 17, 303. Pearse, A . G. E., Polak, J. M., Rost, F. W. D., Fontaine, J., Le Lievre, C., and Le Douarin, N. (1973). Histochemie 34, 191. Peters, A., Palay, S. L., and Webster, H. de F. (1970). “The Fine Structure of the Nervous System. The Cells and their Processes.” Harper (Hoeber), New York. Pryse-Davies, J., Dawson, I. M. P., and Westbury, G . (1964). Cancer 17, 185. Purves, M. J. (1970). J. Physiol. (London) 209, 417. Ramon-Moliner, E. (1977). Brain Res. 128, I . Rees, P. M. (1967a). J. Comp. Neurol. 131, 517. Rees, P. M. (1967b). J. Physiol. (London) 193, 245. Reese, T. S., and Shepherd, G. M. (1972). In “Structure and Function of Synapses” (G. D. Pappas, and D. P. Purpura, eds.), p. 121. Raven, New York.

ULTRASTRUCTURE O F THE CAROTID BODY

329

Rogers, D. C. (1965). J. Anar. 99, 89. Roumy, M., and Leitner, L. M. (1977). In “Chemoreception in the Carotid Body” (H. Acker e t a l . , eds.), p. 257. Springer-Verlag, Berlin and New York. Russell, 1. J. (1968). Nature (London) 219, 177. Saldana, M. J., Salem, L. E., and Travezan, R. (1973). Hum. Pathol. 4, 251. Sampson, S. R. (1971). J. Physiol. (London) 212, 655. Sampson, S. R. (1972). Brain Res. 45, 266. Sampson, S. R., Nicolaysen, G., and Jaffe, R. A. (1975). Brain Res. 85, 437. Satler, J. J., Predan, S., and Brzin, M. (1974). Histochemistry 39, 65. Saxod, R. (1972). J. Embryol. Exp. Morphol. 27, 585. Schafer, D., Seidl, E., Acker, H., and Lubbers, D. W. (1973). Z. Zellforsch. Mikrosk. Anat. 142, 515. Seidl, E. (1975). In “The Peripheral Arterial Chemoreceptors” (M. J. Purves, ed.), p. 293. Cambridge Univ. Press, London and New York. Seidl, E. (1976). Anat. Embryol. 149, 79. Seidl, E., Schafer, D., Zierold, K., Acker, H., and Lubbers, D. W. (1977). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 1. Springer-Verlag. Berlin and New York. Seitelberger, F. (1971). Acra Neuropathol., Suppl. 5 , 17. Serafini-Fracassini, A., and Volpin, D. (1966). Acta Anat. 63, 571. Shamblin, W. R., Remine, W. H., Sheps, S. G., and Harrison, E. G. (1971). A m . J. Surg. 122, 732. Simhszky, J., and Lapis, K. (1970). Acta Morphol. Acad. Sci. Hung. 18, 167. Smith, C. (1924). Am. J. Anat. 34, 87. Smith, P. G., and Mills, E. (1976). Brain Res. 113, 174. Sotelo, C. (1971). J. Ultrasrruct. Res. 36, 824. Sotelo, C., and Palay, S. L. (1968). J. Cell Biol. 36, 151. Stensaas, L. J., and Fidone, S. J. (1977). Brain Res. 124, 29. Taxi, J. (1969). I n “Mechanisms of Synaptic Transmission” (K. Akert and P. G. Waser, eds.), p. 5 . Elsevier, Amsterdam. Taxi, J., Gautron, J., and L’Hermite, P. (1969). C. R. Hebd. Seances Acad. Sci., Ser. D 269, 1281. Toker, C. (1967). Cancer 20, 271. Torrance, R. W. (1968). In “Arterial Chemoreceptors” (R. W. Torrance, ed.), p. 1. Blackwell, Oxford. Torrance, R. W. (1975). J. Physiol. (London) 244, 64. Torrance, R. W. (1976). In “Morphology and Mechanisms of Chemoreceptors” (A. S. Paintal, ed.), p. 131. Vallabhbhai Patel Chest Inst., Delhi. Torrance, R. W. (1977). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 286. Springer-Verlag, Berlin and New York. Totten, R. S. (1973). Hum. Parhol. 4, 453. Tramezzani, J. H., Morita, E., and Chiocchio, S. R. (1971). Proc. Narl. Acad. Sci. U.S.A. 68, 52. Vbzquez-Nin, G., Costero, I., Aguilar, R., and Echeverria, 0. M. (1977). Acta Anat. 98, 233. Verna, A. (1971). J . Microsc. (Paris) 10, 59. Verna, A. (1973). J. Microsc. (Paris) 16, 299. Verna, A. (1975). In “The Peripheral Arterial Chemoreceptors” (M. J. Purves, ed.), p. 75. Cambridge Univ. Press, London and New York. Verna, A. (1977). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 216. Springer-Verlag, Berlin and New York. Verna, A., Roumy, M., and Leitner, L.-M. (1975). Brain Res. 100, 13. Villiaumey, J., Amouroux, J . , Rotterdam, M., Pointud, P., and Delporte, M. P. (1974). Sem. Hop. 50, 854.

330

ALAlN VERNA

Wacker, P., and Forssmann, W. G. (1972). 2.Zellforsch. Mikrosk. Anat. 126, 261. Watzka, M. (1943). I n “Handbuch der mikroskopischen Anatomie des Menschen” (W. von Mollendorff, ed.), Vol. VI, p. 262. Springer-Verlag, Berlin and New York. Welsch, V., and Pearse, A. E. G. (1969). Histochemie 17, 1. Whimster, W. T., and Masson, A. F. (1970). Cancer 26, 239. Willshaw, P. (1975). I n “The Peripheral Arterial Chemoreceptors” (M. J. Purves, ed.), p. 253. Cambridge Univ. Press, London and New York. Willshaw, P. (1977). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 168. Springer-Verlag, Berlin and New York. Winson, M., and Heath, D. (1973). Arch. Pathol. 96, 58. Wood, J. G., and Barmett, R. J. (1964). J . Histochem. Cytochem. 12, 197. Woods, R. I., (1975). I n “The Peripharal Arterial Chemoreceptors” (M.J. Purves, ed.), p. 195. Cambridge Univ. Press, London and New York. Yates, R. D., Chen, I.-Li, and Duncan, D. (1970). J . Cell Biol. 46, 544. Zapata, P. (1975). J . Physiol. (London) 244, 235. Zapata, P., Hess, A., and Eyzaguirre, C. (1969). J. Neurophysiol. 32, 215. Zapata, P., Stensaas, L. J., and Eyzaguirre, C. (1976). Brain Res. 113, 235. Zapata, P., Stensaas, L. J., and Eyzaguirre, C. (1977). In “Chemoreception in the Carotid Body” (H. Acker et al., eds.), p. 44. Springer-Verlag, Berlin and New York. Zelenh, J. (1976). Prog. Brain Res. 43, 59. Zelenh, J. (1978). J . Neurocytol. 7, 71.

INTERNATIONAL REVIEW OF CYTOLOGY. VOL . 60

The Cytology and Cytochemistry of the Wool Follicle DONALD F . G . ORWN Wool Research Organisation of New Zealand. Inc., Christchurch. New Zealand I . Introduction . . . . . . . . . . . . . . . . . . . . General Description of the Follicle . . . . . . . . . . . . I1. Dermal Papilla . . . . . . . . . . . . . . . . . . . III . Bulb . . . . . . . . . . . . . . . . . . . . . . IV . Medulla . . . . . . . . . . . . . . . . . . . . . A . Differentiation . . . . . . . . . . . . . . . . . . B . Cytochemistry . . . . . . . . . . . . . . . . . . V . Cortex . . . . . . . . . . . . . . . . . . . . . . A . Plasma Membrane Differentiation . . . . . . . . . . . B . Nonkeratin Cytoplasmic Components . . . . . . . . . C . Keratin Differentiation . . . . . . . . . . . . . . . D . Ultrastructure of Macrofibrils . . . . . . . . . . . . E . Keratin Synthesis . . . . . . . . . . . . . . . . . VI . Fiber Cuticle . . . . . . . . . . . . . . . . . . . A . Cuticle Shape Differentiation . . . . . . . . . . . . B . Plasma Membrane Differentiation . . . . . . . . . . . C . Cytochemistry of the Cuticle Plasma Membranes . . . . . D . Nonkeratin Cytoplasmic Components . . . . . . . . . E . Keratin Differentiation . . . . . . . . . . . . . . . F. Cytochemistry of Cuticle Keratin . . . . . . . . . . . V n . Inner Root Sheath . . . . . . . . . . . . . . . . . . A . Plasma Membrane Differentiation . . . . . . . . . . . B . Cytochemistry of the Plasma Membrane . . . . . . . . C . Nontrichohyalin Cytoplasmic Components . . . . . . . D . Trichohyalin . . . . . . . . . . . . . . . . . . E . Breakdownof theInner Root Sheath . . . . . . . . . . VIII . Outer Root Sheath . . . . . . . . . . . . . . . . . A . Differentiation . . . . . . . . . . . . . . . . . . B . Cytochemistry . . . . . . . . . . . . . . . . . . IX . Connective Tissue Sheath . . . . . . . . . . . . . . . X . Concluding Remarks . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . .

331 332 334 337 338 339 341 342 342 345 346 350 351 353 354 356 351 358 358 359 360 361 363 363 364 365 366 366 368 369 369 369

.

I Introduction The wool or hair follicle is one of the least-studied biological tissues. mainly because interest is confined to the textile or cosmetic potential of the fiber produced . Nevertheless. the follicle is an exciting and challenging tissue as it 331

.

Copyright @ 1979 by Academic Press. Inc All rights of reproduction in any form nserved. ISBN 0-12-3643@0

332

DONALD F. G . ORWIN

embodies many fundamental biological processes. During the growth of the fiber there appears to be continuous controlled cell division in the bulb of the follicle, with differentiation of several recognizable cell types. Shape changes and relative movements of cells and of the fiber itself occur; proteins responsible for the properties of the fibers are produced; there is controlled cell death and many other features of interest. The morphology of fiber formation has been extensively studied and it is against this background that the cytology and cytochemistry of the wool follicle are reviewed, with reference to the hair follicle where necessary for completeness. GENERAL DESCRIPTION OF THE FOLLICLE Although follicles undergo cycles of activity (anagen) and rest (telogen) in most sheep breeds, a high proportion of the follicles is in the active phase of fiber production (anagen VI of Dry, 1926; Chase et al., 1951) throughout the year. Only this phase will be considered in this review (Fig. 1). The follicle is an epithelium-derived tubular downgrowth into the dermis of the skin. Its dimensions vary according to factors such as breed, but are of the order of 2-3 mm in length and 0.1 mm in diameter. It is surrounded by a connective-tissue sheath comprised mainly of collagen and associated fibroblasts. At the proximal end of the follicle another dermal component, the dermal papilla, is enclosed by the epithelium-derived cells of the bulb (Fig. 1, zone A). The blood supply for the follicle consists of vessels in the connective-tissue sheath and, except in small follicles, the dermal papilla (Ryder, 1956). In the upper part of the follicle there are usually two sebaceous glands which open through the neck of the follicle into the pilary canal. Sudoriferous glands are usually associated with primary follicles (Hardy and Lyne, 1956) and also release their contents into the pilary canal. Attached to the midfollicle wall of primary follicles is the arrector pili muscle (Auber, 1952). In sheep, most follicles are arranged in groups according to their time of development in the fetal skin. A group normally consists of 3 primary follicles (a central follicle forming first, followed by 2 lateral follicles) and a variable number (12-60) of secondary follicles which develop later. Some secondary follicles may form as outgrowths from other secondaries (Hardy and Lyne, 1956). Primary follicles are usually larger than secondary follicles and produce coarser fibers (Ryder and Stephenson, 1968). Fiber formation seems to be similar in all types of follicles. The fiber is formed only from epithelium-derived cells (Fig. 1); apparently no cells are contributed by the dermal components of the follicle. Up to 10 different cell types differentiate from a population of mitotically active bulb cells around the dermal papilla. These are organized into the cell lines of the outer root sheath (2 cell

THE WOOL FOLLICLE

333

FIG. 1. Diagrammatic representations of medullated and nonmedullated wool follicles. See text for description of zones. (Modified from Orwin, 1976a.)

types), the inner root sheath (3), and the fiber (2-5). The fiber is composed of three major cell types, the fiber cuticle, the cortex, and the medulla; the last is usually present in coarse fibers only. As the differentiation of these cell lines follows a regular sequence in the wool follicle, various classification systems have been used based on morphological markers and distances along the follicle (cf. Auber, 1952; Chapman and Gemmell, 1971a). A more precise system has been developed to obviate the difficulty of correlating events in follicles of different lengths and to cope with more

334

DONALD F. G. ORWIN

detailed ultrastructural studies (Orwin, 1976a). This system (see Fig. 1) will be used as a reference guide in this review and is described below. Zone A refers to cells in the mitotic zone, i.e., those below the top of the dermal papilla in nonmedullated fibers or below the widest part of the papilla in medullated fibers. When cells pass above a “critical level” (Auber, 1952) they enter zone B where differentiation begins. This zone is characterized by changes in cell size and shape and the initiation of keratin formation in cortical cells. Zone C begins where Henle’s layer hardens, a stage at which most changes in shape have been completed. This is a region of major protein synthesis for the cortex. Zone D commences where the plasma membrane of the fiber cuticle cells apposed to the inner root sheath (IRS) is associated with a continuous layer of cuticle keratin. Active synthesis of cortex keratin also continues in this zone. The level at which the remainder of the IRS hardens indicates the start of zone E. This is also the region where the stabilization (keratinization) of cortical keratin occurs. The end of the zone is marked by the disappearance of osmiophilia in the cortex. No further changes have been detected in the fiber but two further zones may be listed. Zone F extends from zone E to the end of the region where the degrading IRS and outer root sheath (ORS) cells fragment and slough into the pilary canal. The remainder of the follicle forms zone G and ends at the skin surface. In this zone the fiber is free from the IRS and is coated with the secretions of the sebaceous and sudoriferous (if present) glands prior to its emergence at the skin surface. Within this general framework each cell line undergoes its own specific differentiation. This review covers the cytology and cytochemistry of the dermal papilla, the mitotic zone (the bulb or zone A), and then each cell line in sequence from the central cell line, the medulla, to the outermost cell line, the outer root sheath.

11. Dermal Papilla

The dermal papilla appears to be an essential component of the follicle although its function is little understood. Dermal cells which later form the dermal papilla are present from the earliest stages of follicle development (Hardy and Lyne, 1956). Transplanted papillae can induce hair follicle formation (Oliver, 1970), even after follicle irradiation (Ibrahim and Wright, 1977). The formation of a new dermal papilla is induced by lower follicle tissue after the removal of the existing papilla and prior to regeneration of rat vibrissae (Oliver, 1966, 1967).

THE WOOL FOLLICLE

335

Normal fibers do not form in the absence of dermal papillae although the production of fragile hairs from some depapillated follicles has been reported in the rat (Butcher, 1965). The size and shape of the papillae parallel the size and shape of the fibers formed (Bums and Clarkson, 1949; Auber, 1952; Rudall, 1956a; Henderson, 1965; Straile, 1965), while changes in papilla volume are also paralleled by changes in fiber volume (Straile, 1965). The processes which cause changes in the size of the dermal papilla are not known. Fibroblast mitoses are unlikely to play a significant role as such mitoses in dermal papillae have yet to be reported (Wessells and Roessner, 1965; Moffat, 1968). The mitotic cells found in the dermal papillae of follicles in early anagen are probably endothelial cells (Montagna and Parakkal, 1974; Sholley and Cotran, 1976). Ultrastructural studies have shown that the dermal papilla is composed mainly of fibroblasts (Fig. 2) in the mouse, guinea pig, and human hair follicles (Roth and Helwig, 1964a; Parakkal, 1966; Carlsen, 1974). A basement membrane about 40 nm thick, which is multilayered in the guinea pig, separates the dermal papilla cells from bulb cells (Roth and Helwig, 1964a; Parakkal, 1966). The fibroblasts typically contain extensive rough endoplasmic reticulum systems and well-developed Golgi complexes. Collagen can be observed in the ground substance between the cells. Acid mucopolysaccharides (probably chondroitin sulfate B) in the ground substance are believed to give rise to the metachromasia found when stains such as azure A are used (Braun-Falco, 1958; Sasai, 1976). Glycogen has been reported in some fibroblasts although periodic acidhchiffpositive material remains after digestion with diastase (Parakkal, 1966; Bell, 1969). Another cell type, which has features similar to those of mast cells, has been reported in the dermal papillae of some wool follicles (Orwin, 1970). Both the membrane-bound granules of these cells and the external surfaces of their plasma membranes are stained nonspecifically by silver methenamine (Fig. 3) (Orwin, 1970). Blood vessels found in the dermal papillae of larger follicles (Ryder, 1956; Parakkal, 1966) are of the fenestrated type in the guinea pig (Parakkal, 1966). Focal appositions of plasma membranes have been reported between neighboring papilla cells (Orwin and Thomson, 1973; Carlsen, 1974). These probably correspond to the intermediate junctions (zonulae adherens) found in fibroblasts in other connective tissues (Ross and Greenlees, 1966). This indicates some organization of papilla fibroblasts, a supposition which is further strengthened by the presence of gap junctions between these cells (Orwin et al., 1973a). Alkaline phosphatase, which has been demonstrated histochemically in dermal papillae (Ryder, 1958a; Lyne and Hollis, 1967), has been located cytochemically at or near the plasma membranes of developing hair follicle papillae of rhesus monkeys (Bell, 1969).

336

DONALD F. G . ORWIN

THE WOOL FOLLICLE

337

Little evidence has been produced to indicate the mechanism of interaction of the dermal papilla with the epithelium-derived zone A cells. An attempt to determine whether dermal papilla RNMprotein synthetic activity induced similar activity in mitotic bulb cells of follicles in early anagen indicated that this was unlikely (Silver and Chase, 1977). Light microscope autoradiography of follicles exposed to a labeled RNA precursor, uridine, showed increased uptake in both dermal papillae and mitotic cells but no pattern was observed.

111. Bulb

The cells of the bulb which surrounds the dermal papilla are, in general, mitotically active; they are the source of the cells which differentiate into the various cell lines of the fiber, the IRS, and the ORS. In wool follicles (Fig. 1) the bulb and the dermal papilla are typically angled away from the axis of the fiber (Auber, 1952). The spatial limits of the population of mitotic cells relative to the dermal papilla vary according to the type of fiber produced. In medullated follicles the widest part of the dermal papilla corresponds to the “critical level” (Auber, 1952). Above this level, differentiation of various cell lines is apparent and is associated with loss of mitotic activity (Fraser, 1964; Epstein and Maibach, 1969). In nonmedullated follicles mitoses extend to a level near the top of the dermal papilla (Fraser, 1964; Epstein and Maibach, 1969). In both hair and wool follicles mitotic cells form a cone around the dermal papilla because the cells in the outer regions of the bulb begin to differentiate at levels below that of the top of the dermal papilla. The orientation of the mitotically active cells of the bulb and the movement of labeled cells out of the bulb suggest that the region around the lower dermal papilla gives rise to the ORS and the IRS while that around the upper dermal papilla produces the fiber (Auber, 1952; Epstein and Maibach, 1969; Orwin, 1971). Ultrastructural studies have shown that hair follicle bulb cells adjoining the dermal papilla are columnar while the remainder are more ellipsoidal (Birbeck and Mercer, 1957a; Roth and Helwig, 1964a). In both wool and hair follicles the cells have large nuclei, many ribosomes and mitochondria, and relatively small amounts of endoplasmic reticulum, vacuoles, lysosomes, and Golgi complexes (Birbeck and Mercer, 1957a; Roth and Helwig, 1964a; Forslind and Swanbeck, FIG.2. Dermal papilla. Two fibroblast cells (F) and an endothelial cell of a capillary containing a red blood cell (RBC) are present. An arrow indicates granules of a mast-like cell. X 17,000. FIG.3. Dermal papilla. Mast-like cells are indicated by nonspecific staining of granules and plasma membranes. Silver methenamine. X 11,000.

338

DONALD F. G. ORWIN

1966; Parakkal, 1969a; Chapman and Gemmell, 1971a; Orwin, 1976a). There are intercellular spaces between many cells as well as plasma membrane differentiations such as desmosomes and gap junctions (Birbeck and Mercer, 1957a; Orwin et al., 1973a,b). Despite the lack of obvious differentiation products in the mitotic cell population, cytochemical studies have revealed that there are asymmetrical distributions of some phosphatase enzymes. A histochemical study has shown that alkaline phosphatase activity is distributed on the convex side of the mitotic zone of wool follicles but not on that of hair follicles (Lyne and Hollis, 1967). In contrast, adenosine triphosphatase (ATPase) is found associated with the plasma membranes on the concave side of the wool follicle bulb (Chapman and Gemmell, 1971b). Both of these findings have been taken to indicate that cell differentiation within the mitotic zone is related to an important characteristic of wool, namely, its crimp or waviness. Crimp is associated with segmentation of the cortex into two types, ortho and para. The ATPase activity was shown to be located on the side of the bulb which gives rise to the orthocortex (Chapman and Gemmell , 1971b) . Histochemical and microbiochemical studies have shown that the bulbs of hair follicles are rich in a variety of enzymes. Glycolytic and glycogen-synthesizing enzymes of the Embden-Meyerhoff, tricarboxylic, and pentose cycles are known to be present (Uno et al., 1969; Adachi and Uno, 1969; Sasai et al., 1977). Other enzymes present are P-glucuronidase, aminopeptidase, cytochrome oxidase, and succinic dehydrogenase (Argyris, 1956; Braun-Falco, 1958; Montagna, 1962). As the location of these enzymes according to organelle or cell type has not been established these findings are difficult to evaluate. Little is known of the basis of the continuation of cell division in the bulb although it is known to be influenced by factors such as photoperiodicity and nutrition (Short et al., 1965; Hutchinson, 1965). Current research into the action of natural and synthetic mitosis inhibitors (chemical shearing) on wool follicle bulb cells (Brinsfield et al., 1972; Reis and Chapman, 1974; Reis et al., 1975; Panaretto et al., 1975; Ward and Harris, 1976) might be expected to provide detailed cytological data on the effects of these substances on bulb function.

IV. Medulla The medulla in the fiber consists of the hardened remnants of highly vacuolated cells and occupies the central axis of some keratinized fibers (Fig. 4). These cells change the light-reflective properties of nonpigmented fibers so that medullated fibers appear whiter than nonmedullated fibers. Medullation has been little studied at the cellular level in the wool follicle despite its significance in some textile properties. Although medullation may be present in the wool of any sheep breed, it is the one cell type not necessary for

THE WOOL FOLLICLE

339

normal fiber production. Nevertheless the percentage of medullated fibers in a fleece is highly heritable (0.50-0.70) (Rae, 1956) and can also be influenced in a Mendelian manner by such genes as the pleiotropic incompletely dominant N or recessive n genes (Dry et al., 1940). The degree of medullation in a fleece can be modified by external factors such as shearing, which causes an increase (Rudall, 1935), level of nutrition [higher levels result in greater fiber diameter (Henderson, 1965) and consequently greater medullation], or time of sampling [the greatest fiber diameter and degree of medullation occur in summer (Story and Ross, 1960)l. The medulla occurs in wool follicles which have characteristic dimensions. These follicles are usually larger than nonmedullated follicles and, in particular, they have larger bulbs and dermal papillae. The greater the width of the medulla, the lower down the maximum width of the papilla (Auber, 1952; Henderson, 1965). Despite this, there is an overlap in the diameters of the largest nonmedullated fibers and the smallest medullated fibers, at least in the Herdwick and Romney breeds (Auber, 1952). A. DIFFERENTIATION

The medulla may be multicellular or unicellular in width and continuous or discontinuous. Light microscope observations have shown that medulla cells arise at or around the top of the dermal papilla. In contrast to other follicle cell types, mitosis has not been observed in the medullary cell line even in those cells adjacent to the dermal papilla (Auber, 1952). Although the ultrastructure of medulla formation in wool follicles has not been studied, light microscope observations (Auber, 1952) suggest that it has many similarities to the events observed in hair follicles. The first signs of differentiation are an increase in cytoplasm and ribosomes (Parakkal and Matoltsy, 1964; Parakkal, 1969a). A few desmosomes are present; tonofilaments from these increase during differentiation and appear to be carried over into the hardened state (Roth and Clarke, 1964; Roth and Helwig, 1964b; Hojiro, 1972). Medullary granules about 30 nm in diameter appear in the cytoplasm of cells during the early stages of differentiation (low zone B) (Parakkal and Matoltsy, 1964; Rogers, 1964; Roth and Helwig, 1964b; Parakkal, 1969a; Hojiro, 1972). These granules are amorphous and are not membrane bound. They gradually increase in size until at about the level where the keratinization of the cortex is initiated (start of zone E) they fuse into an amorphous mass, usually around the periphery of the cell. In addition, fibrils possibly associated with the medullary proteins have been reported (Parakkal and Matoltsy, 1964; Rogers, 1964; Roth and Helwig, 196413; Parakkal, 1969a). Associated with medullary granule production is the formation of vesicles and vacuoles, some of which are derived from degenerating mitochondria during the later stages of differentiation. The vacuoles increase in size, sometimes by

340

DONALD F. G . ORWIN

THE WOOL FOLLICLE

34 1

coalescence, until they are incorporated into the hardened medulla (Roth and Helwig, 1964b; Parakkal, 1969a). The presence of glycogen has been noted in the cytoplasm of cells actively producing medullary granules (Roth and Clarke, 1964; Roth and Helwig, 1964b; Parakkal, 1969a; Hojiro, 1972). The intercellular space between apposed medulla cells increases prior to hardening (Roth and Helwig, 1964b; Hojiro, 1972).

B . CYTOCHEMISTRY Medullary granules are composed of an amorphous protein with many morphological and chemical similarities to the trichohyalin in the IRS (Auber, 1952; Rogers, 1964; Rogers and Harding, 1976). Both types of granule have been shown histochemically to be low in cystine (Bannett and Sognnaes, 1962) and citrulline but high in arginine. In the hardened state both types of protein are high in citrulline and low in arginine (Rogers, 1963). One major difference is the ability of medullary granules to harden without undergoing the transformation to a fibrous form shown by trichohyalin (Roth and Helwig, 1964b; Parakkal, 1969a). Some attempts have been made to isolate and characterize medullary granules biochemically. An early method of isolating medullary cells using potassium hydroxide (Matoltsy, 1953) was later shown to extract up to 43% of the medullary protein obtained with a milder method using formic acid (Bradbury and O’Shea, 1969). Nevertheless, both these studies confirmed that medulla cells were low in cystine and high in citrulline, glutamic acid, and glutamine. Tryptic digests and sequence studies, usually of medulla dissected from porcupine quills, have been used to establish that citrulline is covalently bound to proteins by peptide linkages (Rogers, 1962, 1964; Steinert et al., 1969). An enzyme has been isolated which converts arginine in peptides into citrulline (Rogers and Harding, 1976). The presence of e(y-glutamyl) lysine linkages in medullary protein has also been demonstrated (Harding and Rogers, 1971).

FIG.4. Medulla. Scanning electron micrograph shows the vacuolar nature of the medulla (M) of an obliquely cut keratinized Drysdale fiber. The cuticle scale pattern (CS) is present on part of the fiber. At x, a cuticle scale has split away to reveal underlying components. Co, Cortex. An arrow indicates the direction of growth. X 8 0 0 . FIG.5 . Surface structure of keratinized Romney fiber. Scale edges indicate direction of growth (arrow). Cuticle cell margins (small arrows) do not necessarily coincide with scale edges. Corrugations are visible on major cuticle surfaces. X6500. FIG.6 . Scalloped scale pattern of large-diameter keratinized fiber from Drysdale. Large arrow indicates direction of fiber growth and small arrows indicate cuticle cell margins. ~ 6 5 0 0 .

342

DONALD F. G. ORWlN

V. Cortex The cortex constitutes the bulk of most fibers and is responsible for properties such as strength, elasticity, and waviness (crimp). It is formed from cells located around the dermal papilla in the upper region of the bulb in nonmedullated.fibers and from those adjacent to the widest region of the dermal papilla (the mid, not the upper, region) in medullated fibers. The diameter of the cortex is largely determined by the number of cells capable of mitosis and the rate at which they divide. Cell size plays a lesser role (Schinckel, 1961; White and Henderson, 1973). The number of available cells entering the cortex is a relatively low proportion [ca. 15-18% (Short et al., 1965) or 35% (see Ryder and Stephenson, 1968)] of the total population of dividing cells. A. PLASMA MEMBRANE DIFFERENTIATION Differentiation of the cortical cell plasma membranes seems to have the ultimate aim of stabilizing cell-cell contacts. Even at an early stage of fiber development there is evidence of increasing cell adhesion. The intercellular spaces seen in zone A rapidly decrease as the cells move through zone B and the plasma membranes become closely apposed (Birbeck and Mercer, 1957a; Roth and Helwig, 1964a; Forslind and Swanbeck, 1966; Orwin, 1970). Apposed plasma membranes about 7 nm wide are separated by a gap of about 12-15 nm containing surface or intercellular material (Fig. 7). In the lower keratogenous zone (C) of the hair follicle the membranes become wider and the distance between them increases to 20-30 nm (Birbeck and Mercer, 1957a; Rogers, 1964; Parakkal, 1969a) but this has not been confirmed in the wool follicle (Orwin and Thomson, 1972a). In zones C and D there appears to be increasing interaction between the plasma membranes and the developing macrofibrils of keratin. Short fibrils extend at right angles from the macrofibrils to the plasma membranes (Fig. 8) in increasing amounts until by zone E they fill the macrofibril-plasma membrane space. At this stage the electron density of the cytoplasmic leaflet of the membrane is indistinguishable from that of fibrils when conventional electron stains are used (Figs. 8 and 9) (Orwin and Thomson, 1973). The trilaminar structure of the plasma membranes appears to be lost and only the intercellular material (ycomponent) and the electron-lucent region of the membrane are visible (Fig. 9) (Rogers, 1959a). However, this may indicate only that the electron density of the macrofibrils and intercellular material has masked the presence of the cytoplasmic leaflet. The increase in the numbers of membrane particles exposed by freeze-etching between zones B and D (Orwin and Thomson, 1973) probably indicates an

THE WOOL FOLLICLE

343

increase in plasma membrane proteins. Interaction between these particles and the macrofibrillar fibrils has yet to be established but stabilization of membrane proteins (cell-cell contacts) is a possibility. There is other evidence for considerable changes in plasma membrane proteins during differentiation. The periodic acidsilver methenamine technique has shown that surface polysaccharide material, possibly glycoproteins, is most clearly seen between cells in zones A and B. This material decreases in zone C where the cells attain their approximate final shape and is absent or only faintly detectable in zone D (Orwin, 1970). When glycol methacrylate-embedded follicles are stained with phosphotungstic acid at a low pH to show what are probably polysaccharides (Dermer, 1973), the plasma membrane surface material is most intensely stained in zones A and low B; the staining decreases by zone C and is faintly detectable in zone E (Orwin, 1976a,c). The maximum concentration of surface polysaccharides is therefore associated with increasing cell adhesion. Once cell adhesion and cell shape changes are nearing completion the amount of polysaccharide seems to be reduced. Acid phosphatase activity in the plasma membrane surface material of the intercellular space of cortical cells was seen only from zone C upward. Its presence coincides with the region where macrofibrils interact with the plasma membranes (Orwin, 1976a,c). However, the significance of these findings is not known. It is interesting to note that coated vesicles, including those associated with the plasma membranes, reach a maximum in zone B and decline through zone C to a point where they are rarely detected in zone D. Their role in membrane differentiation, if any, has not been determined (Orwin and Thomson, 1972b). Cytological, freeze-etch, and lanthanum-staining techniques have shown that gap junctions are present in the cortex in zones A through D but not in zone E. They cover a maximum of about 7% of the cell surface in zone B and about 4% in other zones (Orwin et al., 1973a). Gap junctions are presumably involved in the intercellular transport of low-molecular-weight molecules. Similar techniques have shown the presence of desmosomes. The percentage of the plasma membrane present as desmosomes increases from about 2% in zone A to 5% in zone B. Dedifferentiation occurs in upper zone B and no desmosomes are detected in zone C. Desmosomes seem to be associated with increasing cell adhesion and the initiation of peripheral macrofibrils (Fig. 7) (Orwin et al., 1973b). Another possible mechanism of stabilizing membrane proteins has arisen from studies of keratinized fibers. When keratin is removed from wool or cortical cells by extraction with performic acid and ammonium hydroxide, an insoluble residue composed in part of cell membranes (cell membrane complex) is left. Aminoacid analyses have shown that this material has a high content of glutamic acid and lysine. It is postulated that e(y-glutamyl) lysine crosslinks are involved in stabilizing cell membrane complex proteins (Peters and Bradbury, 1976).

THE WOOL FOLLICLE

345

B . NONKERATIN CYTOPLASMIC COMPONENTS Little cytochemical work has been done on the differentiation of the major nucleic acid-containing cell components, i.e., the nucleus and ribosomes. Autoradiographic and liquid scintillation techniques have indicated that tritiated thymidine- or cytidine-labeled nuclei lose their label by the top of zone D (Downes et al., 1966a; Sims, 1967; Chapman, 1971). In hair follicles Feulgen staining of nuclear DNA is most intense in zones A and B, but is progressively lost in the lower keratogenous zone (Hardy, 1952; Montagna, 1962). A similar result has been obtained with an electron-cytochemical technique using silver methenamine (Swift, 1977). Nevertheless, both nuclear and ribosomal remnants of unknown composition are retained in keratinized cortical cells (Rogers, 1964; Forslind and Swanbeck, 1966; Fraser et al., 1972; Orwin, 1976~). Histochemical studies, usually of hair follicles, have revealed the presence of several enzymes in the cortex such as succinic dehydrogenase (Argyris, 1956; Braun-Falco, 1958; Montagna, 1962); esterases (Argyris, 1956; Braun-Falco, 1958); cytochrome oxidase, P-glucuronidase, and aminopeptidase (Braun-Falco, 1958); and acid phosphatase (Braun-Falco, 1958; Ryder, 1958a; Rowden, 1967). With the exception of acid phosphatase, the cytological basis of these histochemical findings has not been studied. An electron microscope study of acid phosphatase activity has been used to show that a lysosomal system is present in wool follicles (Orwin, 1976a). Although Golgi complexes, rough endoplasmic reticulum, and vacuoles had been reported in zone A cells previously, this study demonstrated the presence of these

FIG. 7. Longitudinal section of cortex, zone B. At this early stage of differentiation the plasma membrane (PM) has no particular components associated with it. However, the filaments (F) associated with the cytoplasmic plaques of the desmosomes (D) represent the early stages of macrofibril development. X 100,000. FIG. 8. Longitudinal section of cortex, zone D. The well-developed macrofibrils (Mf) run parallel to the plasma membrane (PM). In most regions, the plasma membranes have the normal trilaminar appearance. However, material extends from the macrofibrils to the membranes in some regions (arrowheads). Because of the similar electron density of the membranes, their trilaminar appearance is lost. x 100.000. FIG. 9 . Longitudinal section of cortex and cuticle, zone E. The completed macrofibril (Mf), by continuation of the process described for Fig. 8, apposes the plasma membrane (PM) entirely with the consequent loss of its trilaminar appearance. The cuticle (Cu)comprises the A-layer (A), exocuticle (Exo), and endocuticle (Endo). More electron-dense material is evident where the endocuticle apposes the plasma membranes (arrowheads). In contrast to the amorphous structure of the exocuticle, a fibrillar substructure is visible in the cortical macrofibril. X 125,000. FIG. 10. Cross section of cortex in zone D. Lead deposits indicate sites of acid phosphatase activity in lysosomes (L) and endoplasmic reticulum (ER). The macrofibril (Mf) cross sections show the microfibrilhnatrix structure. Near-hexagonal packing is visible in some regions (arrows). M, Mitochondrion. x 70,000.

346

DONALD F. G. ORWIN

organelles throughout the unkeratinized part of the cortex. In addition, multivesicular bodies, lysosomes, and smooth vesicles were identified in the same zones (A through D). Acid phosphatase activity was found in some Golgi complexes and vacuoles but mainly in lysosomes, endoplasmic reticulum (ER), and small smooth vesicles (Fig. 10). In Golgi complexes, acid phosphatase activity was located on the inner aspect of the stack while thiamine pyrophosphatase was found in the middle of the stack. The numbers of labeled lysosomes, ER profiles, and smooth vesicles rose markedly from zone A to zones C and D. It was suggested that this increased lysosomal activity was associated with the degradation of nonessential cytoplasmic components during cortical differentiation. This, in turn, may be the basis for the decrease in wool fiber diameter of about 25% in zone D (Auber, 1952). Lysosomal-like bodies detected in keratinized cortical cells are probably residual bodies from earlier lysosome activity (Orwin, 1976~). The presence of various atypical substances such as heavy metals in hair and wool could indicate that fibers act as a pathway for the elimination of these substances from the body. The occurrence of residual lysosomes in keratinized fibers suggests that the lysosomal system could form the basis of such a pathway (Orwin, 1976~).Zinc and copper are known to be present in wool and to vary seasonally (Bums et al., 1964; Stevenson and Wickham, 1976). Their location is uncertain although labeled zinc has been detected in the keratogenous zone of hair follicles (Ryder, 1959; Kapur et al., 1974). In contrast to some elements such as lead, zinc and copper are probably required for enzyme activity. Copper-deficiency studies indicate that a copper-containing oxidase is necessary for the conversion of sulfhydryl groups to disulfide bonds during keratinization (Gillespie, 1973). C. KERATINDIFFERENTIATION The early stages of keratin (macrofibril) formation in zone B occur in cells which are increasing in both volume and length (Auber, 1952). Little is known of the mechanism of elongation although the presence of microtubules, usually oriented parallel to the axis of the fiber, has been reported (Orwin and Thomson, 1973). The first sign of keratin formation in both hair and wool follicles is the occurrence of filaments (microfibrils) in the cytoplasm which quickly aggregate to form filamentous bundles of macrofibrils (Figs. 7, 11, and 12). These increase in size and length by further accretion of microfibrils at their outer edges until growth ceases at the top of zone D and keratin stabilization commences. The cortical cells consist mainly of macrofibrils at this stage (Birbeck and Mercer, 1957a; Charles, 1959; Roth and Helwig, 1964b; Forslind and Swanbeck, 1966;

THE WOOL FOLLICLE

347

Parakkal, 1969a; Chapman and Gemmell , 1971a; Forslind et al., 1971; Hojiro, 1972). The main initiation (nucleation) of macrofibrils occurs at an early stage of differentiation, i.e., zone B, as quantitative measurements indicate that the increase in keratin in the keratogenous zone (zones C and D) is attributable to increases in diameter of existing macrofibrils rather than to initiation of new macrofibrils (Forslind and Swanbeck, 1966). Microradiographic and autoradiographic data indicate that both dry mass and sulfur increase to a maximum level near the top of zone D (Forslind et al., 1971). There appear to be at least two sites of macrofibrillar nucleation, desmosomal and cytoplasmic (Figs. 7, 11, and 12) (Chapman and Gemmell, 1971a; Roth and Helwig, 1964b; Orwin et al., 1973b). The desmosome-associated macrofibrils usually come to lie adjacent to and be associated with the plasma membranes (Orwin et al., 1973b; Orwin and Thomson, 1973). Whether they differ from those arising in the cytoplasm is not known. Although the initial orientation of macrofibrils is random, by upper zone B they are oriented puallel to the axis of the fiber. Some may be twisted slightly around their long axis while others lie close to the plasma membranes and nucleus (Figs. 11 and 12) (Birbeck and Mercer, 1957a; Roth and Helwig, 1964b; Forslind and Swanbeck, 1966; Chapman and Gemmell, 1971a). Differential staining behavior, e.g., with methylene blue, led to the realization that the fiber cortex is made up of two types of cells which form the ortho- and paracortex (Fig. 13). In fibers with bilateral segmentation it has been established that the orthocortex is on the outside of any fiber curve while the paracortex is on the inside (Horio and Kondo, 1953; Mercer, 1953). Bilateral segmentation is more common among fibers of small diameter (Fraser and Rogers, 1956). Although there are considerable differences in the form of the macrofibrils in the ortho- and paracortices of keratinized wool fibers (Rogers, 1959a,b) it is only recently that the origin of these differences has been studied. Cross sections of wool follicles at an early stage of differentiation show clearly distinguishable ortho and para cells (Chapman and Gemmell, 1971a). Macrofibrils in the ortho cells are usually circular in cross section and are discrete. Macrofibrils adjacent to the plasma membranes in para cells develop as adjoined masses. These patterns persist into the keratinized fiber (Fig. 13). In longitudinal sections, the macrofibrils of the orthocortex are shorter, thinner, and more numerous in the cytoplasm than paracortical macrofibrils which also seem preferentially aligned along the plasma membranes (Figs. 11 and 12). There is light and electron microscope evidence that the paracortex develops and keratinizes earlier than the orthocortex and that it forms preferentially on the thin side of the follicle (Ryder and Stephenson, 1968; Chapman, 1976). A type of cell in which the macrofibrillar arrangement is intermediate between orthocortex and paracortex has been described and termed mesocortex (Brown and Onions, 1960; Dobb and Sikorski, 1961; Bones

THE WOOL FOLLICLE

349

FIG. 13. Cross section of keratinized Romney wool fiber. Differences in the macrofibrillat structure of orthocortex (0)and paracortex (P) are evident. CM, Cell membrane complex; CR, cytoplasmic remnants; Cu, cuticle. ~ 7 0 0 0 .

and Sikorski, 1967; Whiteley and Kaplin, 1977). In a study of merino sheep Chapman ( 1 976) found that the numbers of intermediate-type cells were greater in sheep fed a poor diet. Once the macrofibrils attain their maximum size (by zone E) the cortical cells undergo consolidation, the macrofibrils becoming more closely apposed as the FIG. 11. Longitudinal section of paracortical cells, zone B. The macrofibrils (arrowheads) are long, relatively few in number, and preferentially located near the plasma membranes (PM). x 12,000. FIG.12. Longitudinal section of orthocortical cells at the same level as in Fig. 11. The macrofibrils (arrowheads) are shorter and more numerous and are found throughout the cytoplasm. PM, Plasma membrane. x 12.000.

350

DONALD F. G. ORWIN

cytoplasm consolidates between them. Few recognizable cytoplasmic components remain by mid-zone E (Fig. 13) (Rogers, 1959a, 1964; Orwin, 1976~).The process of consolidation is little understood but cortical cell dehydration and chemical changes within the macrofibrils are probably involved (Forslind and Swanbeck, 1966; Chapman and Gemmell, 1971a). Keratinization or the stabilization of keratin proteins by conversion of sulfhydryl groups to disulfide bonds is also a feature of zone E although the precise region of zone E where this occurs is disputed (Auber, 1952; Braun-Falco, 1958; Montagna, 1962; Chapman and Gemmell, 1971a). OF MACROFIBRILS D. ULTRASTRUCTURE

Early X-ray diffraction work suggested that a-keratin is composed of crystallites of low sulfur content embedded in a sulfur-rich matrix (Astbury and Dickinson, 1940). Birbeck and Mercer (1957a) showed that macrofibrils in cross section appear as bundles of light-staining circular (sometimes annular) structures about 7 nm in diameter (microfibrils) in a darker staining osmiophilic material (matrix) (Fig. 10). The reaction of osmium with cystine together with the histochemical evidence for sulfhydryl groups in the keratogenous zone of the cortex (Giroud and Bulliard, 1935; Hardy, 1952; Ryder, 1958b) led Birbeck and Mercer (1957a) to conclude that the matrix is rich in sulfur. The specificity of osmium for sulfur is dubious (Sikorski and Simpson, 1959; Litman and Barrnett, 1972). The use of more specific labels has included the reduction or oxidation of the disulfide bonds in the wool fiber and the extraction of high-sulfur proteins prior to osmium staining (Rogers, 1959a); the reaction of untreated wool fibers with ammoniacal silver nitrate (Kassenbeck, 1967); reduction and S-carboxymethylation of wool followed by staining with uranyl acetate (Kassenbeck, 1967); oxidation followed by staining with lead hydroxide as well as with potassium permanganate (Rogers and Filshie, 1962); treatment of hair and wool with silver methenamine (Swift, 1968); and treatment of reduced wool with methylmercuric iodide (Dobb er al., 1972). The last study established that there was virtually stoichiometric uptake of methylmercuric iodide by sulfhydryl groups following tributylphosphine reduction of wool. In all these studies the evidence supports a macrofibrillar structure of microfibrils embedded in a sulfur-rich matrix. Three major groups of proteins can be extracted from wool. The matrix is believed to be the source of high-sulfur proteins (see Crewther, 1976, for a review). The second group, the a-helix-promoting, low-sulfur proteins, is believed to be derived from microfibrils. Evidence to support this at the follicle level has been obtained from the similarity of the amino acid composition of the low-sulfur proteins and that of a filamentous fraction collected from lysed cortical cells of the keratogenous zone of rat vibrissae (Jones, 1975). A third group, the high-glycine proteins, is also believed to be derived from the matrix as the

THE WOOL FOLLICLE

35 1

proportions of this type of protein vary considerably in different types of wool (Gillespie and Darskus, 1971). They are therefore unlikely to be accommodated in more structured regions of the fiber. Details of the types, heterogeneity, and structure of the three groups of fiber proteins are contained in a number of reviews (Fraser et al., 1972; Bradbury, 1973; Crewther, 1976; Sikorski, 1976). Early ultrastructural studies of wool fibers revealed differences in the fine structure of macrofibrils in the ortho- and paracortices (Kassenbeck, 1958; Rogers, 1959a). The whorl pattern of microfibrils with small amounts of matrix between them, observed in macrofibrils of hair follicles (Birbeck and Mercer, 1957a; Forslind and Swanbeck, 1966), is typical of orthocortical macrofibrils; the whorl arrangement of microfibrils is believed to be due to the twisting of peripheral microfibrils around a central core. Macrofibrils in the paracortex show a near-hexagonal m a y of microfibrils with greater amounts of matrix (Rogers, 1959a; Rogers and Filshie, 1963; for reviews, see Fraser et al., 1972; Bradbury, 1973). X-Ray diffraction and chemical data support these observations (Leach et al., 1964; Dobb, 1970). The average proportion of microfibrils in the paracortex is about 33% by volume, and in the orthocortex, 60%. Chapman and Gemmell (1971a) have established that the ultrastructural differences arise during macrofibril formation in the merino wool follicle. The relationship between matrix and microfibrils is not always consistent as high-cystine diets may markedly increase the matrix content of macrofibrils (Rogers, 1964). This is also reflected in the changes in protein type which can be induced by diet or drugs (Frenkel et al., 1974; see Crewther, 1976, for a review). Separation and isolation of ortho- and paracortical cells from wool fibers and subsequent extraction of their proteins has shown a somewhat greater content of high-sulfur proteins in the paracortex, in keeping with its higher cystine content (Kulkarni et al., 1971; Kulkarni and Bradbury, 1974; Kulkarni, 1976; see Bradbury, 1973, for a review). A third type of microfibril/matrix arrangement thought to be specific to mesocortical cells has been recognized (Whiteley and Kaplin, 1977). In their classification the fused paracortical macrofibrils have disordered arrays of microfibrils with only occasional hexagonal arrays; mesocortical macrofibrils have predominantly hexagonal mays of microfibrils and are less extensively fused, while orthocortical macrofibrils consist of discrete whorls of poorly resolved microfibrils . E. KERATIN SYNTHESIS Biochemical studies have shown that groups of proteins similar to those that can be extracted from fibers can also be extracted from the wool follicle (Frater, 1966; Downes et al., 1966b; Fraser, 1969a). This has been supported by immunological studies. After immunohistochemical treatment the keratogenous

352

DONALD F. G. ORWlN

zone of follicles was labeled both by antibodies prepared against S-carboxymethylated low-sulfur protein from hair fibers and by a purified high-sulfur wool fraction (Kemp and Rogers, 1970; Frater, 1976). Although the resolution obtainable by these techniques meant that the cytological site of labeling was not determined, there seems little doubt that the proteins of the fiber are synthesized in the cortical cells and that only minor modifications are required during their passage into the keratinized state. The mode of keratin synthesis has been shown to follow the classical RNA pathway. Polysomes capable of synthesizing high- and low-sulfur proteins may be isolated from both hair and wool follicles (Freedberg, 1970a,b; Clarke and Rogers, 1970; Wilkinson, 1970, 1971; Steinert and Rogers, 1971, 1973). There is some evidence that the polysomes may be attached to the developing macrofibrils (Freedberg, 1970b; Fraser et al., 1972) which raises the possibility that newly synthesized proteins could be added directly to the macrofibrils. The messenger RNA involved in keratin synthesis is likely to be long-lived. Autoradiographic studies at both light and electron microscope levels have indicated that the nuclei of the mid and upper keratogenous zones (zone D) do not take up labeled RNA precursors even though this is the region of maximum keratin synthesis (Sims, 1967; Fraser et al., 1972). In the wool follicle actinomycin D, an inhibitor of DNA-directed RNA production, does not affect polysome profiles for up to 4 hours (Wilkinson, 1970). Injection of actinomycin D into mice prevented keratinization in the early anagen stages of the hair cycle but not when it was injected in mid or late anagen (VI) (Moffat, 1974). Another approach has been to follow the uptake of labeled amino acids by cortical proteins; [35S]cystinehas often been used because of the relatively high proportion of sulfur in keratin. Light microscope autoradiographic studies of both hair and wool follicles have shown that [35S]cystineis incorporated preferentially into zone B of the cortex (where macrofibrils are first detected) and spreads rapidly throughout zones C and D. This process takes about 30 minutes in the mouse and rat but 60 minutes in sheep (Harkness and Bern, 1957; Ryder, 1958b, 1959; Downes et al., 1962). It is much faster than can be accounted for by the passage of labeled cells to higher regions (Downes et al., 1962). Sulfur administered as [35S]sulfate accumulates in different regions of the follicle and much more slowly, indicating that it is cystine itself that is incorporated in the keratogenous zone (Montagna and Hill, 1957; Ryder, 1958b). The uptake of [3H]tyrosineis very similar to that of cystine in rat follicles (Sims, 1964). Electron microscope autoradiographic studies have confirmed that, in comparison with the bulb, the keratogenous zones C and D show greater uptake of [35S]cystineand that the radioactivity increases quantitatively with time (Nakai, 1964). Furthermore, at 3 and 6 hours after injection the label is associated with the macrofibrils. Similar findings have been reported in the mouse follicle although the location of the label in zone B is cytoplasmic rather than macrofibrillar (Forslind, 1971). No account was taken of the differences between ortho- and

THE WOOL FOLLICLE

353

paracortices in previous work; as there is evidence that paracortex contains greater amounts of cystine than orthocortex (Kulkarni et al., 1971) the results may have been biased. Because of this Chapman and Gemmell(l973) studied the uptake of [35S]cystineby both cortices. After 1 hour similar uptakes of label were observed in zone B of both cortices but there was a greater uptake in the paracortex in zones C and D. At 5 hours the uptake of label per unit area in the paracortex in zones C and D was approximately double that in the orthocortex. It was concluded that the results supported previous findings of a higher cystine content in the paracortex and that the two cortices differ in their rates of uptake. As the uptake of cystine may reflect only the synthesis of high-sulfur proteins, an attempt has been made to follow low-sulfur microfibrillar protein synthesis by using [3H]methionine, an amino acid lacking in high-sulfur proteins. Wilson et al. (1971) followed the uptake of the label by electron microscope autoradiography and determinations of the specific activity of high- and low-sulfur proteins. The uptake was rapid and reached a maximum at the 4-hour sampling with the greatest levels in the bulb and diminishing levels in the higher unkeratinized regions of the cortex. Although the label was distributed fairly generally over organelles in the bulb, in zones C and D the label was localized in the macrofibrils 1-8 hours after injection. The specific activity of the low-sulfur proteins was 7-14 times greater than that of the high-sulfur proteins. It was concluded that the low-sulfur proteins are mainly synthesized in the bulb (zone A). It has been suggested that keratin may be synthesized in two stages (Rudall, 1956b), an initial synthesis of low-sulfur microfibrillar proteins followed by the synthesis of high-sulfur matrix proteins. The studies of Wilson et al. (1971) and of Downes et al. (1963) supported this. However, other biochemical data indi-. cated that some high-sulfur proteins may be synthesized concurrently with the low-sulfur proteins although their maximum synthesis probably occurs in zone D (Downes et al., 1966b; Fraser, 1969b). Cytological studies of macrofibril cross sections in ortho- and paracortical cells have shown that differences exist in the early stages of their differentiation (Chapman and Gemmell, 1971a). Only microfibrils are initially observed in the paracortex; matrix then begins to form and eventually it catches up with microfibril production (by zone C) the two then being concurrently synthesized until keratinization occurs. No clear-cut pattern was observed in the orthocortex during the early stages of differentiation and it is only at the later stages of macrofibril formation (zone D) that the microfibril/ matrix pattern is resolved. This evidence suggests that there may well be different sequences of protein synthesis for ortho- and paracortical cells.

VI. Fiber Cuticle The cells of the fiber cuticle form the external layer of the fiber and are therefore responsible for its surface properties. Despite their importance, little is known of their cytochemistry.

354

DONALD F. G. ORWIN

A. CUTICLESHAPEDIFFERENTIATION The cuticle forms a single layer of cells over the periphery of the fiber (Rogers, 1959a,b). Light microscope studies have shown that changes in the shapes of fiber cuticle cells are related to those in the IRS cuticle cells (Auber, 1952). The two cell layers are similar in size and are closely apposed in upper zone A. The first sign of differentiation of the fiber cuticle in zone A is a smoothing out of the plasma membranes and a consequent decrease in the number of intercellular spaces (Birbeck and Mercer, 1975b). The cells are characteristically cuboidal in shape. Peripheral flattening of both cell lines in lower zone B is followed by tilting of the fiber cuticle cells toward the skin surface, causing them to overlap by approximately one-sixth of their length (Bradbury and Leeder, 1970). During tilting the Golgi complex moves to a perinuclear position and the nucleus to the lower end of the cell. A fibrillar material of unknown function (Happey and Johnson, 1962; Roth and Helwig, 1964b) is associated initially with desmosomal cytoplasmic fibers and increases in length and diameter during differentiation (Roth and Helwig, 1964b). Various speculations on the cause of tilting include upward pressure by elongating Huxley ’s cells (Auber, 1952), relatively greater elongation of IRS cuticle cells (Montagna and Parakkal, 1974), zipper-like spread of cell contacts (Birbeck and Mercer, 1957b), and internally directed morphogenetic changes within the cells themselves (Montagna and Parakkal, 1974). It may be relevant that studies of [3H]thymidine-labeled cells show that IRS cells move more quickly out of the bulb than cortical cells (Epstein and Maibach, 1969; Chapman, 1971), but the relationship of this to specific movements of IRS cuticle and fiber cuticle cells is not known. After the fiber cuticle cells have tilted the IRS cuticle cells bulge into them (in upper zone B). The resultant cell shapes are stabilized by zone D, giving the keratinized fiber surface a ridged appearance (scale pattern). The tips of the ridges point toward the skin surface (Figs. 4-6 and 14) and usually occur near the region of fiber cuticle cell overlap. As there may be more than one ridge per cell and the

FIG. 14. Longitudinal section, zone D. The buildup of exocuticle protein (Exo) derived from the amorphous droplets (arrowheads) occurs along the cuticle cell membrane apposing the inner root sheath cuticle (IRSC) but not along that plasma membrane apposing the cortex (Co). The trichohyalin (T) and fibrils (F) of the inner root sheath cuticle predominate in that part of the cell apposing the cuticle, while the nucleus and cytoplasm occupy that part apposing Huxley’s layer (Hu). Sc, Differentiating scale edge. X 15,OOO. FIG. 15. Longitudinal section of Henle’s cell, zone B . Trichohyalin (T) and associated fibrils largely fill the cell. Some of these fibrils are also associated with desmosomes (arrows). Aggregations of fibrils in the elongated companion cell (Comp) show as electron-dense material (arrowheads). Hu, Huxley’s layer; ORS, outer root sheath. x 17,000.

356

DONALD F. G. ORWIN

distance between them is variable, the scale pattern does not reflect the boundaries of the fiber cuticle cells (Figs. 5 and 6) (Kassenbeck, 1958; Dobb er al., 1961; Bradbury and Rogers, 1963). Minor corrugations parallel with the fiber axis may also be found in coarser wools (Fig. 5) (Wolfram. 1972). Scalloped ridges have been described in very coarse fibers (kemps and hair) (Fig. 6) which also have larger fiber cuticle cells (Auber, 1952). The cuticle scale patterns can take several forms (see Bradbury, 1973, for a review). Fine wool fibers show an irregular mosaic pattern, medium and coarse fibers an irregular mosaic pattern, usually waved, while the largest-diameter fibers (kemps and hair) show a regular mosaic in the lower region and an irregular mosaic pattern in the tip region (Wildman, 1956). The shape and degree of cell overlap of fiber cuticle cells, however, have been shown to be similar in both a coarse and a fine wool (Bradbury and Leeder, 1970). It is likely that all cell margins are smooth initially but that a crenate or undulating pattern can be induced by wear. Such margins have been produced experimentally by the use of abrasives (Bums, 1962). Some association between scale patterns and growth rate is shown by the increased distance between scale edges when growth rates increase after thyroxine injections into thyroidectomized sheep (Rougeot, 1965). B. PLASMA MEMBRANE DIFFERENTIATION Differentiation of fiber cuticle membranes, particularly with regard to cell adhesion, must be complex as a cuticle cell apposes at least three different cell types during its development, i.e., cortex, cuticle, and IRS cuticle. In the last case, the plasma membranes of these cells obviously differentiate in a different manner as they separate when the IRS breaks down in zone F. Apart from plasma-membrane-related cell adhesion, projections from both cortical and cuticle cells into neighboring cuticle cells have a role in interlocking the cells (Rogers, 1959a,b). During early differentiation the plasma membranes become increasingly apposed (Birbeck and Mercer, 1957b; Orwin and Thomson, 1972a). About midzone B, the intercellular space has been reported as increasing in width from 12-15 to 20-30 nm (Birbeck and Mercer, 1957b; Rogers, 1964; Parakkal, 1969a). However, measurements of plasma membranes in the wool follicle showed that once apposition had taken place in zone A, no further changes were detected until keratinization occurred (Orwin and Thomson, 1972a). This was true for apposed cortical/fiber cuticle or cuticle/cuticle cells. It is of interest that separation of the cuticle from the IRS cuticle sometimes occurs in zone E in electron microscope preparations prior to keratinization. The plane of separation seems to involve the fiber cuticle plasma membrane and possibly the intercellular material close to it. This may indicate that the fiber cuticle surface is composed of plasma membrane remnants following sloughing of the IRS cuticle in zone F (Orwin and Thomson, 1972a).

357

THE WOOL FOLLICLE

Various junctions have been described in differentiating cuticle plasma membranes. Desmosomes have been found in early differentiating cells (Charles, 1959; Happey and Johnson, 1962; Roth and Clarke, 1964; Orwin et al., 1973b). Both freeze-fracture and morphological studies have shown that desmosomes of apposed cortical/fiber cuticle and fiber cuticle/IRS cuticle cells decrease in number from zone B to zone C and disappear by zone D. The association of the cytoplasmic fibrils of cuticle desmosomes with the fibrillar protein of the cuticle has already been noted. Gap junctions appear to be rare in the plasma membranes of apposed cortical/ fiber cuticle cells, apposed fiber cuticle cells, and fiber cuticle/IRS cuticle cells. Although present in zone B they are absent by zone D, in contrast with their presence in this zone in the cortex (Orwin et al., 1973a). The loss of gap junctions in this region suggests that intercellular communication and transport between the fiber cuticle and the neighboring cell lines is reduced or no longer necessary. c.

CYTOCHEMISTRY OF THE CUTICLE PLASMA

MEMBRANES

The periodic acidhilver methenamine staining technique has shown that polysaccharide material occurs in the intercellular space of cuticle cells (Orwin, 1970). As in the cortex, these membrane-associated polysaccharides appear to undergo modification during differentiation. After the cuticle cells have attained their approximate final shape at the start of zone D similar staining is not observed and polysaccharides may no longer be present. Acid phosphatase activity can be demonstrated in the intercellular spaces of apposed corticaVfiber cuticle cells above zone B but not in fiber cuticle/fiber cuticle or fiber cuticle/IRS cuticle intercellular material. This points to a further specialization in the interaction of fiber cuticle plasma membranes with neighboring cells (Orwin, 1976a). The plasma membranes of the differentiating cuticle cells give rise to the so-called ‘‘cell membrane complex of the keratinized cuticle. This appears, using conventional electron stains, to consist of two electron-lucent layers (p or inert layers) separated by an electron-dense layer (y or intercellular cement) (Fig. 9) (Rogers, 1959a,b). This intercellular material is believed to be proteinaceous because of its ability to take up electron stains. As previously mentioned for cortical membranes, the loss of the trilaminar appearance of the plasma membranes may be misleading because of the similarity in electron density of the adjacent material (Orwin and Thomson, 1973). There are differences between the reactions of cortical and cuticle cells with various salts. The uptake of uranyl and lead salts was found to be greater in apposed fiber cuticle cells than in apposed cortical cells, whereas silver salts showed no uptake in apposed cuticle cells and a low uptake in apposed cortical cells (Orwin and Thomson, 1972a; Nakamura et al., 1976). Formic acid modifies the ultrastructural appearance of the plasma membranes of keratinized ”

358

DONALD F. G . ORWIN

cortical cells but leaves the keratinized cuticle cell membranes apparently unchanged (Peters and Bradbury, 1976). However, more data to support this would be desirable as formic acid extracts a protein constituting about 1% of the fiber and believed to be derived from a cell membrane complex (Bradbury and King, 1967). This may include the cuticle cell membrane complex. Amino acid analysis of formic acid-treated cuticle cell membrane complexes showed the presence of citrulline, whereas those of cortical cells isolated by performic acid and ammonia or urea extraction did not (Bradbury et al., 1971). The high lysine content of both types of membranes led to the hypothesis that they are stabilized by e(y-glutamyl) lysine crosslinks (Peters and Bradbury, 1976). This may explain the apparent inability of enzymes such as trypsin, Pronase, and papaid dithiothreitol to modify the cuticle cell membrane complex (Swift and Bews, 1974; Swift, 1976). The presence of another component, a lipid, can be demonstrated either by following the rate of penetration of phosphotungstic acid along the cuticle cell membrane complex after ethanol extraction of hair fibers (Swift and Holmes, 1965) or by analysis of isolated membranes from wool fibers (Bradbury et al., 1971). D. NONKERATIN CYTOPLASMIC COMPONENTS The presence of nuclei, ribosomes, mitochondria, and Golgi complexes during differentiation of fiber cuticle cells (Fig. 14) and their condensation to form the endocuticle (Fig. 9) has been described (Birbeck and Mercer, 1957b; Parakkal and Matoltsy, 1964; Roth and Helwig, 1964b; Orwin, 1976a). Orwin (1976a) also observed that lysosomes, multivesicular bodies, and vacuoles were present; endoplasmic reticulum was not present to the same extent as in the cortex. Acid phosphatase activity in Golgi complexes, lysosomes, and vesicles increases during differentiation but not to the same extent as in the cortex. It was suggested that the lower degree of lysosomal activity probably reflects a lower degree of autophagy with the result that many more cytoplasmic components are incorporated in the keratinized cuticle cells than in the cortical cells. The cytoplasm of keratinized cuticle cells (endocuticle) remains susceptible to attack by enzymes such as trypsin and Pronase (Birbeck and Mercer, 1957b; Swift and Holmes, 1965; Bradbury and Ley, 1972) but not to keratinolytic agents such as peracetic acidammonia (Rogers, 1959a,b).

E. KERATIN DIFFERENTIATION The formation of cuticle keratin in the hair follicle starts in upper zone B at about the level where the trichohyalin of Henle’s layer becomes fibrous (Birbeck and Mercer, 1957b; Parakkal and Matoltsy, 1964; Roth and Helwig, 1964b;

THE WOOL FOLLICLE

359

Parakkal, 1969a). Changes in cell shape are nearing completion at this stage. Amorphous droplets, about 30 nm in diameter, are synthesized in the cytoplasm and move toward the IRS cuticle side of the fiber cuticle cell (Fig. 14). In the mouse, they may reach a size of 500 nm by deposition of further material at their surface (Parakkal, 1969a). The droplets tend to coalesce into larger discrete lumps leaving a space near the plasma membrane (Birbeck and Mercer, 1957b). The deposition of keratin near the same membrane continues until roughly half to two-thirds of the cell is filled. This results in the nucleus and the remaining cytoplasm being found on the cortical side of the cell (Fig. 14). The fibrous protein mentioned earlier lies on the cortical side of the amorphous cuticle keratin (Happey and Johnson, 1962; Roth and Helwig, 1964b). Cuticle keratin coalesces on keratinization (top of zone D) to form an amorphous layer known as the exocuticle (Fig. 9). Evidence for its amorphous nature has been obtained from X-ray and birefringence studies (see Bradbury, 1973, for a review). A further layer of protein on the surface of the exocuticle called the A-layer (Lagermalm, 1954) also becomes apparent during keratinization (Fig. 9) (Swift, 1977) and the space between the cuticle keratin and the plasma membrane disappears (Birbeck and Mercer, 1957b). The cytoplasm condenses during this stage to form the endocuticle which contains recognizable remnants of the nuclei and other organelles (Fig. 9) (Birbeck and Mercer, 1957b; Orwin, 1976~).

F. CYTOCHEMISTRY OF CUTICLE KERATIN As cuticle keratin is amorphous and shows a high uptake of osmium it has been suggested that it may be similar to the high-sulfur proteins of the cortex (Birbeck and Mercer, 1957b). Further studies of sulfur location using osmium tetroxide and silver nitrate, especially with prior reduction (Rogers, 1959b; Dobb and Sikorski, 1961; Kassenbeck, 1961), sodium plumbite (Sikorski and Simpson, 1958), and silver methenamine (Swift, 1968, 1969) have indicated that the amount of sulfur (cystine) in each layer decreases in the following order: A-layer, exocuticle, endocuticle. Dobb et al. (1972) confirmed these results by using methylmercuric iodide to determine the sulfhydryl groups of reduced fibers. They also noted that an even more cystine-rich layer occurred along the external edge of the A-layer but not on the overlapped parts of the neighboring cuticle cells. On the innermost surface of the cells another cystine-rich layer occurs irregularly against the plasma membrane adjoining the endocuticle (Swift, 1968; Fraser et al., 1972). Studies of fibers treated with chlorine water (Allworden, 1916) have indicated that another component, the epicuticle, may exist on the outermost surface of the fiber (for reviews, see Bradbury, 1973; Swift, 1977). This component exhibits semipermeable properties when underlying cuticle proteins are solubilized by chlorine water. Because these proteins cannot pass through the epicuticle a

360

DONALD F. G . ORWlN

bubble or sac is raised on the surface of the cuticle (Allworden sac). The epicuticle is not normally stained by heavy metals and its structural and biological location has not been rigorously established. It may correspond with the gap (2.5 nm wide) found between the cuticle surface and the metal coating in transverse sections of metal-coated fibers. This gap is believed to represent the lipid remnants of the plasma membrane (Swift and Holmes, 1965). Swift and Holmes (1965) have also observed that Allworden sacs are not formed on fibers that have been extracted with ethanol. Confirmatoq evidence for the presence of cystine in the fiber cuticle has been provided from autoradiographic and histochemical studies of both hair and wool follicles (Ryder, 1958b; Montagna, 1962; Nakai, 1964; Forslind, 1971). Amino acid analyses of cuticle cells or of regions isolated from wool fibers by treatments involving enzymes (Birbeck and Mercer, 1957b; Swift and Holmes, 1965) and formic acid (Bradbury and Ley, 1972) have confirmed that the exocuticle contains more cystine than the endocuticle and is higher in proline (see Bradbury, 1973, for a review). Sequential digestion of keratinized cuticle fragments showed that trypsin preferentially removed the nuclear remnants, Pronase the remainder of the endocuticle, and papain/dithiothreitol the exocuticle followed by the A-layer (Swift, 1976). The cell membrane complex did not appear to be attacked by any of the treatments. Amino acid analyses confirmed previous findings of cystine levels except for surprisingly low levels of cystine in the A-layer. The presence of e(y-glutamyl) lysine crosslinks was also noted in the combined A-layer/cell membrane complex fraction. The possibility of similarities between high-sulfur proteins of the cortex and the cuticle has been strengthened recently. Frater (1976) found that peroxidaselabeled antibodies raised against a high-sulfur fraction from wool cortex would also bind to the cuticle. Similarly, a reasonably pure preparation of cuticle cells released by papain treatment of keratinized hair fibers contained only the y-keratose protein fraction when extracted by peracetic acid. This fraction consists mainly of high-sulfur proteins (Asquith and Parkinson, 1966).

VII. Inner Root Sheath The IRS is composed of three different concentric layers each normally one cell thick (Auber, 1952; Gemmell and Chapman, 1971). Surrounding the fiber cuticle is the IRS cuticle followed by Huxley’s layer and Henle’s layer, the last being adjacent to the outer root sheath. The IRS does not form part of the emergent fiber as it is degraded and sloughed in the upper follicle (Fig. 1; zone F). The three layers in both hair and wool follicles show marked similarities in their differentiation and are characterized by the production of a precursor protein, trichohyalin, which transforms to the filamentous/matrix component of the

THE WOOL FOLLICLE

36 1

hardened cells. Henle’s layer differentiates first and hardens at the start of zone C (Auber, 1952; Gemmell and Chapman, 1971). Huxley ’s layer forms trichohyalin earlier than the IRS cuticle but hardens later according to Gemmell and Chapman (1971); Auber (1952) states the reverse. Hardening occurs in these two cell lines at about the level where fiber keratinization commences (start of zone E). Henle’s and Huxley’s cells become lanceolate in shape while the IRS cuticle has one face which interdigitates with the apposing fiber cuticle surface and the cells are circumferentially wider than longer (Auber, 1952). In wool follicles with deflected bulbs, Henle’s and Huxley’s cells are larger on the outside of the deflection (Fig. 1) (Auber, 1952). While little is known directly of the function of these cell layers, there is evidence that they may be involved as follows. The IRS cuticle appears to be closely associated with the formation of the scale pattern on the surface of the fiber cuticle (Auber, 1952). Huxley’s layer may be more than one cell deep, the variations seemingly being associated with preventing major changes in the diameter of the IRS/fiber complex when marked changes occur in the diameter or shape of the fiber. For instance guard hairs in rabbits or zig-zag hairs in mice have dumbbell-shaped cross sections or constrictions, respectively, yet the IRS/fiber complex remains roughly circular in cross section and of constant diamter because of variations in the numbers of cells in Huxley’s layer (Durward and Rudall, 1956; Priestley and Rudall, 1965; Straile, 1965). In Herdwick wool follicles winter thinning of fibers is associated with Huxley’s layer becoming more than one cell deep on one side of the follicle (Priestley and Rudall, 1965). Huxley’s layer may also have a role in the transport of metabolites as there are gaps between hardened Henle’s cells through which Huxley’s cells make contact with companion cells of the ORS (Happey and Johnson, 1962; Gemmell and Chapman, 1971). Henle’s layer may have a role in determining cell shape in other layers (Auber, 1952). More likely it is involved in the movement of the IRS/fibercomplex toward the skin surface because of its association with companion cells in the ORS (Straile, 1962; Orwin, 1971). It may also influence the passage of metabolites from the ORS to the fiber complex (Orwin et al., 1973a,b,c). A. PLASMA MEMBRANE DIFFERENTIATION An increase in the apposition of the plasma membranes of the IRS has been noted in human hair and wool follicles in upper zone A and zone B (Birbeck and Mercer, 1957c; Orwin and Thomson, 1972a). These membranes later differentiate as a cell membrane complex (Birbeck and Mercer, 1957c) about 10 cell lengths before hardening in the wool follicle (Orwin and Thomson, 1972a). This takes the form of an increase in the intercellular gap from about 13 to 23 nm and the appearance of electron-dense material in the central region of the gap. The

362

DONALD F. G. ORWlN

total width of the complex in hardened cells is about 34 nm (Birbeck and Mercer, 1957c; Orwin and Thomson, 1972a). In addition, a band of electron-dense material about 12 nm wide is laid down one cell prior to hardening on the cytoplasmic side of the plasma membranes. The appearance of the hardened cell membrane complex is similar to that of a desmosome. In fact the cell membrane complex between desmosomes in hardened Henle’s layer is more poorly defined than described, the majority of the cell membrane complex being derived from desmosomal remnants (Orwin and Thomson, 1972a). Increasing numbers of desmosomes have been associated with increasing differentiation and adhesion in the hair (Charles, 1959; Parakkal and Matoltsy, 1964; Roth and Helwig, 1964a) and wool follicles (Orwin et al., 1973b). In the last study the numbers of desmosomes were found to more than double during differentiation, to be larger in Henle’s layer than in the cortex, and to retain both freeze-etch and morphological characteristics in the hardened state (Orwin et al., 1973b). The attachment of trichohyalin-associated filaments to the cytoplasmic plaques of desmosomes has been observed in both hair (Charles, 1959; Parakkal and Matoltsy, 1964; Rogers, 1964; Roth and Helwig, 1964a) and wool follicles (Orwin et al., 1973b). Gap junctions are present in cells of the differentiating IRS of the wool follicle (Orwin et al., 1973a). They increase in number from zones A to B in Henle’s layer but disappear on hardening (Orwin et al., 1973a), although regions of close apposition may still occur (Orwin et al., 1973b). They were found in the apposed membranes of the IRS cuticle/fiber cuticle in zone C but not in zone D. In apposed Henle’s cells and Henle’s/companion cells, alternating arrays of gap junctions and desmosomes were often seen. In view of the association of gap junctions with direct intercellular communication it has been suggested that the hardening of Henle’s cells might limit communication from the ORS to the unhardened IRS/fiber complex. However, areas where Huxley ’s cells are in direct contact with the ORS may still retain this function. Tight-junction zonulae have been observed in freeze-etch and thin-section preparations of wool follicles (Orwin et al., 1973~).They were restricted to the plasma membranes of the cell lines apposing Henle’s cells from zone B to zone D, i.e., even after Henle’s cells had hardened. In some cases, they were associated with gap junctions and desmosomes. A function of Henle’s layer might be to seal off the extracellular space of the ORS from the developing IRS/fiber complex. The extensive system of gap junctions in the differentiating cortex may perhaps be explained in terms of loss of intercellular communication through Henle’s layer. As a result, metabolite transport might be expected to be preferentially oriented along the axis of the developing fiber. Although there is no direct evidence for this, the uptake of labeled amino acids, e.g., [35S]cystinefollows this pattern (Ryder, 1958b, 1959; Downes et al., 1962; Chapman and Gemmell, 1973).

THE WOOL FOLLICLE

363

B. CYTOCHEMISTRY OF THE PLASMA MEMBRANE The presence of a polysaccharide-containing component at the surface of IRS cells throughout differentiation has been demonstrated by the periodic acidhiher methenamine technique (Orwin, 1970). A nonspecifically stained layer about 16 nm wide on the cytoplasmic side of the hardened plasma membrane (Swift, 1969; Orwin, 1970) is believed to be rich in cystine (Swift, 1969). It may correspond with the band of material laid down just prior to hardening in the same location (Orwin and Thomson, 1972a). The polysaccharide component of the intercellular space is retained in hardened cells in contrast to its marked reduction in the cortex (Orwin, 1970, 1976a,c). This intercellular material could also be stained with phosphotungstic acid (Orwin, 1976b). Sites where desmosomes were located showed a double line of silver grains in the intercellular space after silver methenamine staining. This suggests that the central intercellular desmosomal component does not contain polysaccharides (Orwin, 1970). The asymmetric distribution of nucleoside triphosphatase activity found in the plasma membranes of bulb cells of merino wool follicles also occurs in the cells of the IRS (Chapman and Gemmell, 1971b). This asymmetry has been correlated with ortho- and paracortex differentiation. However, no similar asymmetric differentiation has been reported for the IRS as yet. C. NONTRICHOHYALIN CYTOPLASMIC COMFQNENTS IRS cells in both hair and wool follicles contain nuclei, many ribosomes, mitochondria, Golgi complexes, and rough endoplasmic reticulum (Fig. 15) (Parakkal and Matoltsy, 1964; Rogers, 1964; Roth and Helwig, 1964a; Orwin, 1976b). More recently, lysosomes, vesicles, vacuoles, and multivesicular bodies have been observed in wool follicle IRS cells (Orwin, 1976b). Of interest is the presence in Henle’s layer of large dilations of endoplasmic reticulum in upper zone B and single membrane-bound vesicles (about 100-1 80 nm in diameter) with electron-dense contents found near the plasma membranes of the same region (Orwin, 1976b). The role of either component is not clear. However, although most cell organelles can be detected just prior to hardening (Roth and Helwig, 1964a; Orwin, 1976b), the dilated endoplasmic reticulum disappears or becomes markedly reduced in extent. This may indicate that its contents are involved in the process of hardening. It is generally recognized that nuclear remnants are incorporated in the hardened cell (Birbeck and Mercer, 1957c; Parakkal and Matoltsy, 1964; Roth and Helwig, 1964a). Other organelles have been reported as disintegrating and being partially eliminated from the cell before hardening (Parakkal, 1969a). However, this has been challenged by the finding of identifiable cell components such iaS

364

DONALD F. G. ORWIN

lysosomes and mitochondria in hardened cells (Orwin, 1976~).It seems likely that cell components present just prior to hardening become trapped during the transformation. Although the IRS is known to show positive histochemical reactions for phosphorylase only in Henle’s layer, acid phosphatase and P-glucuronidase in the IRS generally, lipid prior to the appearance of trichohyalin, and RNA which decreases during differentiation but is detectable in hardened cells (for reviews, see Braun-Falco, 1958; Montagna, 1962), little is known of the cytochemistry of these cells. The presence of a lysosomal system has been established in the wool follicle IRS since acid phosphatase activity has been found in Golgi complexes, lysosomes, and vesicles (Orwin, 1976b). In contrast to the cortex (Orwin, 1976a) there appears to be little development of the IRS lysosomal system during differentiation, suggesting that extensive autophagy of nontrichohyalin components does not occur. This is supported by the presence of organelles in hardened IRS cells (Orwin, 1976~).As the dilated endoplasmic reticulum of Henle’s cells was not found to contain acid phosphatase activity it is presumably not part of the lysosomal system.

D. TRICHOHYALIN Trichohyalin is usually regarded as one of the earliest differentiation products which distinguish IRS cells from other types. It is normally associated with fibrils about 7-8 nm wide (Fig. 15), but whether the trichohyalin or the fibrils arise first is disputed. In Henle’s layer, which differentiates first, filaments are reported to precede the appearance of the associated amorphous trichohyalin droplets in the follicles of the mouse (Roth and Helwig, 1964a; Parakkal, 1969a; Hojiro, 1972), guinea pig (Parakkal and Matoltsy, 1964), man (Charles, 1959), and sheep (Gemmell and Chapman, 1971). Birbeck and Mercer (1957~)reported that trichohyalin droplets appear first in the human hair follicle. They suggested that the trichohyalin transforms into a filamentous/matrix component about the middle of zone B. This theory has been supported by Rogers (1964) but Parakkal and Matoltsy (1964) believe that the trichohyalin forms an amorphous matrix among the filaments. There is general agreement that the trichohyalin and associated filaments increase in extent and number during differentiation and may coalesce to form larger droplets. On hardening, the trichohyalin disappears and the major part of the cytoplasm becomes filled with material within which tubular filaments can be observed in a matrix of low electron density. The filaments are oriented parallel to the axis of the fiber (Steinert et al., 1971). Similar events have been observed in the differentiation of Huxley ’s cells and IRS cuticle cells. It has been noted that, in the later stages of differentiation of

THE WOOL FOLLICLE

365

the IRS cuticle, the filaments are predominantly located near the plasma membranes apposing the fiber cuticle (Fig. 14) (Gemmell and Chapman, 1971). Histochemical tests have shown that hair follicle trichohyalin is rich in arginine but not citrulline, whereas hardened IRS cells react strongly for citrulline but not arginine (Rogers, 1963). Biochemical studies of hair follicles have since shown that isolated hardened IRS cells and their tryptic digests have amino acid compositions which are very low in cystine (cf. the cortex and fiber cuticle) and high in glutamic acid, aspartic acid, leucine, and lysine (Steinert el al., 1971). These findings confirm the proteinaceous nature of hardened IRS cell contents, i.e., transformed trichohyalin, as indicated by staining and histochemical reactions (Auber, 1952; Braun-Falco, 1958; Rogers, 1963). Further biochemical studies have shown that citrulline is covalently bound by peptide linkages to IRS proteins of hardened cells (Rogers, 1959b, 1962; Steinert et al., 1969); that an enzyme is present in hair roots which converts protein-bound arginine residues into citrulline residues (Rogers and Harding, 1976); and that e(y-glutamy1) lysine crosslinks are present in IRS proteins (Harding and Rogers, 1971). Protein fractions have been isolated from hair roots, i.e., prior to hardening, which may be derived from trichohyalin or medullary granules (Rogers and Harding, 1976). The similarities and differences between trichohyalin and medullary granules have been referred in Section IV.

E. BREAKDOWN O F THE INNER ROOTSHEATH Electron microscope observations in the wool follicle have shown that changes can be seen at the surface of Henle’s cells adjacent to companion cells about 100 pm above where Huxley’s layer hardens (zone E). These take the form of plasma-membrane-bounded protrusions into the ORS cells (Gemmell and Chapman, 1971; Orwin, 1976b). Further toward the skin surface, some of these protrusions appear as lamellated structures with a lamella periodicity of about 55 nm. Both lamellated and nonlamellated protrusions become more frequent and interrelated as degradation of the IRS proceeds. Increased amounts of electrondense intercellular material in continuity with the protrusions are observed in Henle’s and Huxley’s layers during the later stages of IRS degradation. These developments are accompanied by an extracted appearance of the cell contents in Henle’s and Huxley’s layers and later in the IRS cuticle. In zone F, the degraded IRS cells finally compact, fragment, and slough into the pilary canal (Gemmell and Chapman, 1971). A cytochemical study of the IRS breakdown zone (zone F) has failed to demonstrate the presence of the lysosomal enzyme, acid phosphatase, in degrading IRS cells. Furthermore, little evidence of acid phosphatase activity was found in the membrane-bound protrusions of IRS cells into companion cells although it

366

DONALD F. G. ORWIN

did occur occasionally. There was also little cytological evidence that lysosomes or vesicles in companion cells were actively transporting material to Henle ’s cells (Orwin, 1976b). Although these results need confirmation, for instance, with other lysosomal enzymes, as yet there seems to be no strong evidence for the degradation of the IRS by a lysosomal system.

VIII. Outer Root Sheath The outer root sheath forms the nonkeratinizing or nonhardening region of the follicle and is continuous with the epidermis. It has potentialities not associated with fiber growth. For instance, the healing of a hairy skin graft in mouse skin involves the sloughing of the graft’s epidermis and its replacement by ORS cells from the follicles, sometimes with the consequent loss of the fiber-forming function of the follicles (Sanford et al., 1965). StIaile (1962) has distinguished several zones within the ORS according to their probable function. Apart from the bulb, these are the zone of proliferation (zones B through D); zone of migration (zone E); zone of sloughing (zone F); and zone of the upper ORS which is similar to and continuous with the epidermis (zone G). In wool follicles with deflected bulbs the IRS/fiber complex is located eccentrically so that the ORS is thicker on the deflected side of the follicle (Fig. 1). In the relatively infrequent follicles with nondeflected bulbs the IRS/fiber complex is located centrally (Auber, 1952). A. DIFFERENTIATION

Around the bulb (zone A) the ORS is regarded as being one to two cells thick in the human hair (Montagna and Parakkal, 1974), mouse (Roth and Helwig, 1964a), and sheep (Auber, 1952) follicles (Fig. 1). Electron microscope observations have shown quite clearly that two distinct ORS cell layers differentiate at the periphery of the wool follicle bulb (Orwin, 1971). Although the cells are still mitotically active the two layers can be distinguished four to five cells from the base of the dermal papilla. The cell layer next to Henle’s cells adheres more closely to Henle’s layer than to the outer layer of the ORS. At higher levels in the follicle, i.e., zone B and above, its flattened form and location distinguish it from the other ORS cells. Similar cell layers at these upper levels have been recognized in the follicles of man (Pinkus, 1927; Montagna and Parakkal, 1974), sheep (Auber, 1952; Chapman, 1971), mouse and guinea pig (Straile, 1962, 1965), Australian opossum (Gibbs, 1938), and rat (Rogers, 1964). This layer, named the companion cell layer by Orwin (1971) because of its association with

THE WOOL FOLLICLE

367

Henle’s cells rather than with the other ORS cells, is probably involved with the movement of the IRS/fiber complex toward the skin surface (Pinkus, 1927; Straile, 1962, 1965; Rogers, 1964; Orwin, 1971). It has been noted that labeled cells adjacent to Henle’s layer, i.e., presumed companion cells, have been found to move up both wool and human follicles at the same rate as the IRS (Epstein and Maibach, 1969; Chapman, 1971). The role of companion cells in the breakdown of the IRS has been referred to in Section VII. The cells of the peripheral layer of the bulb, the presumptive ORS layer, although initially flattened become more cuboidal by the top of zone A (Fig. 1). They are characterized by marked convulutions of their surfaces suggesting that they do not adhere closely to neighboring cells of the same type or to companion cells. However, they are closely apposed to the basement membrane enclosing the epithelium-derived cells of the follicle and continuous with that of the dermal papilla. In zone B, these cells may form several layers according to the thickness of the overall ORS. This region forms the beginning of the zone of proliferation within which ORS cells (but not companion cells) may undergo further division (Auber, 1952; Straile, 1962, 1965; Epstein and Maibach, 1969; Chapman, 1971). These cells then move up toward the skin surface through the zone of migration (Straile, 1962). In so doing, they move first toward the IRS side of the ORS and become elongated (Straile, 1962; Epstein and Maibach, 1969; Chapman, 1971). In the later stages of differentiation (zone F) they undergo a form of cornification and finally disintegrate into fragments which are sloughed into the pilary canal along with the degraded IRS cells (Straile, 1962; Gemmell and Chapman, 1971). Not all cell layers of the ORS slough, the peripheral cells being cuboidal in shape (Montagna, 1962) and continuous with those of the follicle neck. These cells (zone G) show many characteristics of the epidermis (Montagna, 1962; Straile, 1962; Roth and Helwig, 1964a; Parakkal, 1969a; Gemmell and Chapman, 1971). The cells immediately above zone F (the IRS breakdown zone) have little keratohyalin and show little evidence of hardening or sloughing, while those closer to the follicle orifice flatten, harden, and slough in the same manner as normal epidermis (Gemmell and Chapman, 1971). Langerhans cells have also been identified in the wool follicle neck (Hollis and Lyne, 1972). ORS cells of both hair and wool follicles contain many vacuoles, Golgi complexes, both smooth and rough endoplasmic reticulum, mitochondria, ribosomes, lysosomes, and multivesicular bodies (Rogers, 1964; Roth and Helwig, 1964a; Parakkal, 1969a; Orwin, 1976b). In comparison with other cell lines, ORS cells have a more extensive cytomembrane system. Plasma membrane differentiations include desmosomes (Rogers, 1964; Roth and Helwig, 1964a; Parakkal, 1969a; Orwin, 1970) and gap junctions (Orwin et al., 1973a). An interesting feature of both companion and other ORS cells is the development of

368

DONALD F. G . ORWIN

a fibrillar material, often associated with desmosomes, which reaches its greatest extent in zone E (Rogers, 1964; Roth and Helwig, 1964a; Orwin, 1971). The fibrils form around the periphery of companion cells during early development to the extent that other cytoplasmic components may be excluded (Fig. 15) (Orwin, 1971). At later stages of development (zones B to F) bundles of more electrondense filaments accumulate (Fig. 15) (Rogers, 1964; Orwin, 1971). Their relationship to those that develop earlier is not clear although Rogers (1964) likens them to keratin filaments. The cornification of ORS cells prior to sloughing is reported to involve little keratohyalin and it is therefore different from epidermal keratinization (Gemmell and Chapman, 1971).

B. CYTOCHEMISTRY The ORS of the hair follicle contains several enzymes whose cytological localization is unknown. Alkaline phosphatase increases in activity near the zone of sloughing (zone F). Esterase activity is high in the same zone while the upper regions of the ORS contain amylophosphorylase, cytochrome oxidase, and P-glucuronidase (see Montagna, 1962, for a review). Considerable quantities of glycogen are stored in ORS cells (but not in companion cells) below the level of the sebaceous gland and especially in zones B, C, and D (Ryder, 1958a; Montagna, 1962; Rogers, 1964; Parakkal, 1969a; Orwin, 1971). This glycogen is lost when hair follicles are cultured in the presence of excess vitamin A (Bellows and Hardy, 1977). Biochemical studies of isolated ORS have confirmed active glycogen metabolism (Adachi and Uno, 1969; Uno et al., 1969). Sasai et al. (1977) concluded that the accumulation of glycogen in the ORS of hair follicles is due to a decrease in phosphorylase activity relative to glycogen synthetase. The significance of glycogen storage remains obscure. The presence of polysaccharide material in the intercellular spaces of apposed wool follicle ORS cells has been reported (Orwin, 1970). Contrary to findings in the fiber cuticle and cortex, no changes in reactivity were noted up to zone E. The asymmetric distribution of nucleoside phosphatase activity in the intercellular spaces of apposed cells in the wool follicle bulb also includes presumptive ORS cells (Chapman and Gemmell, 1971b). Although the fibrils of the ORS have been described as keratin-like (Rogers, 1964), no cystine has been detected (Swift, 1969). A poorly developed lysosomal system is present in the lower regions of the ORS (Orwin, 1976b). Near zone F, large lysosomes have been reported in histochemical (Diengdoh, 1964) and cytochemical (Rowden, 1967; Orwin, 1976b) studies of hair and wool follicles. Their presence seemed more consistent with autophagy of the degenerating ORS cells prior to sloughing than involvement in the breakdown of the IRS (Orwin, 1976b).

THE WOOL FOLLICLE

369

IX. Connective Tissue Sheath The hair or wool follicle is surrounded by a basement membrane around which lies the connective-tissue sheath. In hair follicles the basement membrane is slightly periodic acidSchiff-positive; it contains highly sulfated glycosaminoglycans and is about 20 nm thick (Rogers, 1957; Parakkal, 1969b; Butler, 1975). Neuroreceptor (Merkel) cells have been ultrastructurally identified attached to the basement membrane of the ORS of human hair follicles (Marhle and Orfanos, 1974; see Montagna and Parakkal, 1974, for a review of skin innervation). The connective-tissue sheath in the hair follicle consists of two layers of collagen fibers, the inner lying parallel with the fiber axis and the outer at right angles to it. Associated with the connective-tissue sheath are fibroblasts (Rogers, 1957) and blood vessels (see Ryder, 1956, for wool follicle distribution).

X. Concluding Remarks Histological and histochemical studies have provided much knowledge about events occurring during the formation of the wool or hair fiber. The complexity of the follicle is now well established, particularly with respect to the interrelationships between dermal and epithelial components and the different cell lines during differentiation. Cytological and cytochemical studies, although still in their infancy, have already indicated that many more levels of complexity exist. Furthermore, some of the processes for which more mechanistic explanations have been used are being shown to have a cytological basis. This trend may be expected to continue. It is becoming increasingly evident that the full significance of many studies will not be revealed unless they can be related to known and specified cell components. Cytological techniques will play an important part in this area.

ACKNOWLEDGMENTS The excellent assistance of Miss J. L. Woods during the preparation of this review is gratefully acknowledged. The author is indebted to Mrs. J. Onvin, Dr. W. S. Simpson, and Dr. L. F. Story for critically reading the manuscript.

REFERENCES Adachi, K.,and Uno, H. (1969). Adv. B i d . Skin 9, 51 1. Argyris, T. S. (1956). Anat. Rec. 125, 105.

370

DONALD F. G. ORWIN

Asquith, R. S., and Parkinson, D. C. (1966). Text. Res. J . 36, 1064. Astbury, W. T., and Dickinson, S. (1940). Proc. R . SOC. London, Ser. B 129, 307. A u k , L. (1952). Trans. R . SOC. Edinburgh 62, 191. Barmen, R. J., and Sognnaes, R. F. (1962). Publ. Am. Assoc. Adv. Sci. 70, 27. Bell, M. (1969). Adv. Biol. Skin 9, 61. Bellows, C. G., and Hardy, M. H. (1977). Anat. Rec. 187, 257. Birbeck, M. S. C., and Mercer, E. H. (1957a). J. Biophys. Biochem. Cyrol. 3, 203. Birbeck, M. S. C., and Mercer, E. H. (1957b). J . Biophys. Biochem. Cyrol. 3, 215. Birbeck, M. S. C., and Mercer, E. H. (1957~).J . Biophys. Biochem. Cyrol. 3, 223. Bones, R. M., and Sikorski, J. (1967). J . Text. Inst. 58, 521. Bradbury, J. H. (1973). Adv. Protein Chem. 27, 1 1 1 . Bradbury, J. H., and King, N. L. R. (1967). Ausr. J. Chem. 20, 2803. Bradbury, J. H., and Leeder, J. D. (1970). Ausr. J . Biol. Sci. 23, 843. Bradbury, J. H., and Ley, K. F. (1972). Ausr. J. Biol. Sci. 25, 1235. Bradbury, J. H., and O’Shea, J. M. (1969). Ausr. J. Biol. Sci. 22, 1205. Bradbury, J. H., and Rogers, G. E. (1963). Text. Res. J . 33, 452. Bradbury, J. H., Leeder, J. D., and Watt, I. C. (1971). Appl. Polym. Symp. 18, 227. Braun-Falco, 0. (1958). In “The Biology of Hair Growth” (W. Montagna and R. A. Ellis, eds.), p. 65. Academic Press, New York. Brinsfield, T. H., Clark, M. V., Dolnick, E. H., and Reynolds, P. J. (1972). J. Anim. Sci. 34,273. Brown, T. D., and Onions, W. J. (1960). Nature (London) 186, 93. Bums, M. (1962). Nature (London) 195, 512. Bums, M., and Clarkson, H. (1949). J. Agric. Sci. 39, 315. Bums, R. H., Johnston, A., Hamilton, J. W., McColloch, R. J., Duncan, W. E., and Fisk, H. G. (1964). J . Anim. Sci. 23, 5 . Butcher, E. 0. (1965). Anat. Rec. 151, 231. Butler, W. F. (1975). Histochem. J . 7, 67. Carlsen, R. A. (1974). J. Invest. Dermatol. 63, 206. Chapman, R. E. (1971). J. Cell Sci. 9, 791. Chapman, R. E. (1976). Proc. Int. Woof Text. Res. Conf.., 5rh, 1975 Vol. 11, p. 152. Chapman, R. E., and Gemmell, R. T. (1971a). J. Llltrastruct. Res. 36, 342. Chapman, R. E., and Gemmell, R. T. (1971b). J . Invest. Dermatol. 57, 377. Chapman, R. E., and Gemmell, R. T. (1973). J. Cell Sci. 13, 81 1 . Charles, A. (1959). Exp. Cell Res. 18, 138. Chase, H. B., Rauch, H., and Smith, V. W. (1951). Physiol. Zool. 24, 1 . Clarke, R. M., and Rogers, G. E. (1970). J . Invest. Dermatol. 55, 425. Crewther, W. G. (1976). Proc. Int. Wool Text. Res. Conf..,5th. 1975 Vol. I, p. 1 . Dermer, G. B. (1973). J . Llltrastruct. Res. 45, 183. Diengdoh, J . V. (1964). Q. J . Microsc. Sci. “ 3 . 1 105, 73. Dobb, M. G. (1970). J. Text. Inst. 61, 232. Dobb, M. G., and Sikorski, J. (1961). I n “Structure de la laine,” p. 37. Inst. Text. Fr., Paris. Dobb, M. G., Johnston, F. R., Nott, J. A., Oster, L., Sikorski, J., and Simpson, W. S. (1961). J. Text. Inst. 52, T153. Dobb, M. G., Murray, R., and Sikorski, J. (1972). J . Microsc. (Oxford) 96, 285. Downes, A. M., Lyne, A. G., and Clarke, W. H. (1962). Ausr. J . Biol. Sci. 15, 713. Downes, A. M., Sharry, L. F., and Rogers, G. E. (1963). Nuture (London) 199, 1059. Downes, A. M., Chapman, R. E., Till, A. R., and Wilson, P. A. (1966a). Nature (London) 212, 477. Downes, A. M., Ferguson, K. A., Gillespie, J. M., and Harrap, B. S. (1966b). Anat. J . Biol. Sci. 19, 319.

THE WOOL FOLLICLE

371

Dry, F. W. (1926). J. Genet. 16, 287. Dry, F. W., McMahon, P. R., and Sutherland, J. A. (1940). Nature (London) 145, 390. Dunvard, A., and Rudall, K. M. (1956). Proc. Int. Wool Text. Res. Conf., 1955 Vol. F, p. 112. Epstein, W. L., and Maibach, H. I. (1969). Adv. Eiol. Skin 9, 83. Forslind, B. (1971). Acta Derm.-Venereol. 51, 9. Forslind, B., and Swanbeck, G. (1966). Exp. Cell Res. 43, 191. Forslind, B., Lindstrom, B., and Swanbeck, G. (1971). Acta Derm.-Venereol. 51, 81. Fraser, I. E. B. (1964). Aust. J . Biol. Sci. 17, 521. Fraser, I. E. B. (1969a). Ausr. J. Biol. Sci. 22, 213. Fraser, I. E. B. (1969b). Ausr. J. BioL. Sci. 22, 231. Fraser, R. D. B., and Rogers, G. E. (1956). Proc. Int. Wool Text. Res. Conf.., 1955 Vol. F, p. 151. Fraser, R. D. B., MacRae, T. P.,and Rogers, G. E. (1972). “Keratins. Their Composition, Structure and Biosynthesis. ” Thomas, Springfield, Illinois. Frater, R. (1966). Aust. J . Biol. Sci. 19, 699. Frater, R. (1976). Aust. J. Eiol. Sci. 29, 453. Freedberg, I. M. (1970a). J . Invest. Dermatol. 54, 108. Freedberg, I. M. (1970b). Eiochem. Biophys. Acra 224, 219. Frenkel, M. J., Gillespie, I . M., and Reis, P.J. (1974). Ausr. J . B i d . Sci. 27, 31. Gemmell, R. T., and Chapman, R. E. (1971). J. Ultrasrruct. Res. 36, 355. Gibbs, H. F. (1938). Proc. Zool. SOC. London, Ser. B 108, 611. Gillespie, J. M. (1973). Ausr. J . Dermatol. 14, 127. Gillespie, J . M., and Darskus, R. L. (1971). Aust. J. Eiol. Sci. 24, 1189. Giroud, A., and Bulliard, H. (1935). Arch. Anat. Microsc. 31, 271. Happey, F., and Johnson, A. G. (1962). J . Ultrastrucr. Res. 7, 316. Harding, H.W. J., and Rogers, G. E. (1971). Biochemistry 10, 624. Hardy, M. H. (1952). Am. J . Anat. 90, 285. Hardy, M. H., and Lyne, A. G. (1956). Proc. I n t . Wool Text. Res. Conf., 1955 Vol. F, p. 26. Harkness, D. R., and Bern, H. A. (1957). Acta Anat. 31, 35. Henderson, A. E. (1965). In “Biology of the Skin and Hair Growth” (A. G. Lyne and B. F. Short, eds.), p. 447. Angus & Robertson, Sydney, Australia. Hojiro. 0. (1972). J. Electron Microsc. 21, 127. Hollis, D. E., and Lyne, A. G. (1972). J. Invest. Dermatol. 58, 211. Horio, M., and Kondo, T. (1953). Text. Res. J . 23, 373. Hutchison, J. C. D. (1965). I n “Biology of the Skin and Hair Growth” (A. G. Lyne and B. F. Short, eds.), p. 565. Angus 8r Robertson, Sydney, Australia. Ibrahim, L., and Wright, E. A. (1977). Nature (London) 265, 733. Jones, L. N. (1975). Biochim. Eiophys. Acta 412, 91. Kapur, S. P., Bhussry, B. R., Rao, S., and Harmuth-Hoene. (1974). Proc. SOC. Exp. Eiol. Med. 145, 932. Kassenbeck, P. (1958). Bull. Inst. Text. Fr., No. 76, p. 7. Kassenbeck, P. (1961). In “Structure de la laine,” p. 51. Inst. Text. Fr., Paris. Kassenbeck, P. (1967). J. Polym. Sci., Part C 20, 49. Kemp, D. J., and Rogers, G.E. (1970). J . Cell Sci. 7, 273. Kulkami, V. G. (1976). J . Text. Inst. 67, 319. Kulkami, V. G., and Bradbury, J. H. (1974). Aust. J. Eiol. Sci. 27, 383. Kulkami, V. G., Robson, R. M., and Robson, A. (1971). Appl. Polym. Symp. 18, 127. Lagennalm, G. (1954). Text Res. J . 24, 17. Leach, S. J . , Rogers, G. E., and Filshie, B. K. (1964). Arch. Eiochem. Eiophys. 105, 270. Litman, R. B., and Barrnett, R. J. (1972). J. Ultrasrruct. Res. 38, 63. Lyne, A. G., and Hollis, D. E. (1967). J. Invest. Dermatol. 48, 197.

312

DONALD F. G. ORWIN

Mahrle, G., and Orfanos, C. E. (1974). Arch. Derrnatol. Forsch. 251, 19. Matoltsy, A. G. (1953). Exp. Cell Res 5, 98. Mercer, E. H. (1953). Texr. Res. J . 23, 388. Moffat, G . H. (1968). J . Anur. 102, 527. Moffat, G. H. (1974). J . Invest. Dermarol. 63, 199. Montagna, W. (1962). “The Structure and Function of Skin,’’ 2nd ed., p. 174. Academic Press, New York. Montagna, W., and Hill, R. C. (1957). Anar. Rec. 127, 163. Montagna, W., and Parakkal, P. F. (1974). “The Structure and Function of Skin,” 3rd ed., p. 172. Academic Press, New York. Nakai, T. (1964). J. Cell Biol. 21, 63. Nakamura, Y.,Kanoh, T., Kondo, T., and Inagaki, H. (1976). Proc. Inr. Wool T a r . Res. Conf., 5rh. 1975 Vol. 11, p. 23. Oliver, R. F. (1966). J. Embryol. Exp. Morphol. 15, 331. Oliver, R. F. (1967). J. Embryol. Exp. Morphol. 17, 27. Oliver, R. F. (1970). J. Embryol. Exp. Morphol. 23, 219. Onvin, D. F. G. (1970). Ausr. J. Biol. Sci. 23, 623. Orwin, D. F. G. (1971). Ausr. J . Biol. Sci. 24, 989. Onvin, D. F. G. (1976a). J . Ulrrasrrucr. Res. 55, 312. Onvin, D. F. G. (1976b). J. Ulrrasrrucr. Res. 55, 325. Onvin, D. F. G. (1976~).J. Ultrasrrucr. Res. 55, 335. Onvin, D. F. G., and Thomson, R. W. (1972a). J. Cell Sci. 11, 205. Onvin, D. F. G., and Thomson, R. W. (1972b). Ausr. J. Biol. Sci. 25, 573. Onvin, D. F. G., and Thomson, R. W. (1973). J . Ulrrasrrucr. Res. 45, 41. Onvin, D. F. G., Thomson, R. W., and Flower, N. E. (1973a). J . Ulrrasrrucr. Res. 45, 1. Onvin, D. F. G., Thomson, R. W., and Flower, N. E. (1973b). J . Ulrrasrrucr. Res. 45, 15. Onvin, D. F. G., Thomson, R. W., and Flower, N. E. (1973~).J . Ulrrasrrucr. Res. 45, 30. Panaretto, B. A,, Chapman, R. E., Downes, A. M., Reis, P. J., and Wallace, A. L.C. (1975). A m . J . Exp. Agric. Anim. Husb. 15, 193. Parakkal, P. F. (1966). J. Ulrrasrrucr. Res. 14, 133. Parakkal, P. F. (1969a). Adv. Biol. Skin 9, 441. Parakkal, P. F. (1969b). J. Cell Biol. 40, 561. Parakkal, P. F., and Matoltsy, A. G. (1964). J. Invest. Dermarol. 43, 23. Peters, D. E.,and Bradbury, J. H. (1976). Aust. J. Biol. Sci. 29, 43. Pinkus, F. (1927). In “Handbuch der Haut- und Geschlechts-krankheite” (J. Jodassohn, ed.), Vol. I, Part I, p. 116. Springer-Verlag. Berlin and New York. Priestley, G. C., and Rudall, K. M. (1965). I n “Biology of the Skin and Hair Growth” (A. G. Lyne and B. F. Short, eds.), p. 165. Angus & Robertson, Sydney, Australia. Rae, A. L. (1956). A h . Genet. 8, 189. Reis, P. J., and Chapman, R. E. (1974). A m . J. Agric. Res. 25, 931. Reis, P. J., Tunks,D. A., and Chapman, R. E. (1975). Ausr. J . Biol. Sci. 28, 69. Rogers, G. E. (1957). Exp. CellRes. 13, 521. Rogers, G. E. (1959a). J. Ulfrasrrucr. Res. 2, 309. Rogers, G. E. (1959b). Ann. N.Y.Acad. Sci. 83, 378. Rogers, G. E. (1962). Nature (London) 194, 1149. Rogers, G. E . (1963). J. Hisrochem. Cyrochem. 11, 700. Rogers, G. E. (1964). I n “The Epidermis” (W. Montagna and W. C. h b i t z , Jr., eds.), p. 179. Academic Press, New York. Rogers, G. E., and Filshie, B. K. (1962). Electron Microsc. Proc. Inr. Congr., 5th, 1962 Vol. 2, p. 1.

373

THE WOOL FOLLICLE

Rogers, G. E., and Filshie, B. K. (1963). In “Ultrastructure of Protein Fibers” (R. Borasky, ed.), p. 123. Academic Press, New York. Rogers, G. E., and Harding, H.W. J. (1976). Proc. Int. Wool Text. Res. Conf. 5th, 1975 Vol. 11, p. 212. Ross, R., and Greenlees, T. K., Jr. (1966). Science 153, 997. Roth, S. I., and Clarke, W. H.(1964). In “The Epidermis” (W. Montagna and W. C. Lobitz, Jr., eds.), p. 303. Academic Press, New York. Roth, S. I., and Helwig, E. B. (1964a). J. Ultrastruct. Res. 11, 33. Roth, S. I . , and Helwig, E. B. (1964b). J . Ulrrastrucr. Res. 11, 52. Rougeot, J . (1965). I n “Biology of the Skin and Hair Growth” (A. G. Lyne and B. F. Short, eds.), p. 625. Angus & Robertson, Sydney, Australia. Rowden, G. (1967). J . Invest. Dermatol. 49, 181. Rudall, K. M. (1935). J . Text. Inst. 26, T358. Rudall, K . M. (1956a). Proc. Int. Wool Text. Res. Conf., 1955 Vol. F, p. 9. Rudall, K. M. (1956b). Proc. Int. Wool Text. Res. Conf., I955 Vol. F, p. 176. Ryder, M. L. (1956). Proc. Inr. Wool Text. Res. Conf., 1955 Vol. F, p. 63. Ryder, M. L. (1958a). Q. J . Microsc. Sci. [N.S.] 99, 221. Ryder, M. L. (1958b). Proc. R. SOC.Edinburgh 67, 65. Ryder, M. L. (1959). J. Histochem. Cytochem. 7, 133. Ryder, M. L., and Stephenson, S . K. (1968). “Wool Growth.” Academic Press, New York. Sanford, B. H., Chase, H.B., Carroll, S. B., and Arsenault, C. T. (1965). Anar. Rec. 152, 17. Sasai, Y. (1976). Hisrochem. J. 8, 51. Sasai, Y., Inoue, K., and Kameyama, A. (1977). Acta Histochem. Cytochem. 10, 197. Schinckel, P. G. (1961). Aust. J . Biol. Sci. 14, 659. Sholley, M. M . , and Cotran, R. S. (1976). Am. J . Anar. 147, 243. Short, B. F., Wilson, P. A,, and Schinckel, P. G. (1965). In “Biology of the Skin and Hair Growth” (A. G. Lyne and B. F. Short, eds.), p. 409. Angus & Robertson, Sydney, Australia. Sikorski, J. (1976). Proc. Int. Wool Text. Res. Conf., 5th. 1975 Vol. I, 103. Sikorski, J., and Simpson, W. S. (1958). Nature (London) 182, 1235. Sikorski, J . , and Simpson, W. S. (1959). J. R. Microsc. SOC.78, 35. Silver, A. F., and Chase, H. B. (1977). J. Invest. Dermarol. 68, 201. Sims, R. T. (1964). J. Cell Biol. 22, 403. Sims, R. T. (1967). Nature (London) 213, 387. Steinert, P. M., and Rogers, G. E. (1971). Biochim. Biophys. Acra 238, 150. Steinert, P. M., and Rogers, G. E. (1973). Biochim. Biophys. Acra 312, 403. Steinert, P. M., Harding, H. W. J., and Rogers, G. E. (1969). Biochim. Biophys. Acta 175, 1 . Steinert, P. M., Dyer, P. Y., and Rogers, G. E. (1971). J . Invest. Dermatol. 56, 49. Stevenson, J . R., and Wickham, G. A. (1976). N.Z. J . Agric. Res. 19, 217. Story, L. F., and Ross, D. A. (1960). N.Z. J. Agric. Res. 3, 113. Straile, W. E. (1962). J. Exp. 2001.150, 207. Straile, W. E. (1965). In “Biology of the Skin and Hair Growth” (A. G. Lyne and B. F. Short, eds.), p. 35. Angus & Robertson, Sydney, Australia. Swift, J. A. (1968). J. R . Microsc. SOC. 88, 449. Swift, J . A. (1969). Histochemie 19, 88. Swift, J. A. (1976). Proc. Int. Wool Text. Res. Conf., 5th, 1975 Vol. 11, p. 12. Swift, J. A. (1977). In “Chemistry of Natural Protein Fibers” (R. S. Asquith, ed.), p. 81. Wiley, New York. Swift, J. A,, and Bews, B. (1974). J . SOC. Cosmet. Chem. 25, 355. Swift, J . A., and Holmes, A. W. (1965). Text. Res. J . 35, 1014. Uno, H . , Adachi, K., and Montagna, W. (1969). Adv. Biol. Skin 9, 221.

.

314

DONALD F. G. ORWIN

von Allworden, K. (1916). Z. Angew. Chem. 29, 77. Ward, K. A., and Harris, R. L. N. (1976). Aust. J . Biol. Sci. 29, 189. Wessels, N. K . , and Roessner, K. D. (1965). Dev. B i d . 12, 419. White, D. H . , and Henderson, A. E. (1973). N.Z.J . Agn'c. Res. 16, 539. Whiteley, K . J . , and Kaplin, 1. J . (1977). J. Text. I n s / . 68, 384. Wildman, A. B . (1956). Proc. Int. Wool Text. Res. Conf., 1955 Vol. F, p. 156. Wilkinson, B. R . (1970). Ausf. J . Biol. Sci. 23, 127. Wilkinson, B. R . (1971). Biochem. J . 125, 371. Wilson, P. A., Henrikson, R. C., and Downes, A. M. (1971). J . Cell Sci. 8, 489. Wolfram, L. J. (1972). Text. Res. J. 42, 252.

Subject Index

A

C

Actin characteristics, 62-63 in mitotic spindle, 63-65 Amino acid, donation, viral tRNA-like structures and, 16-17 Amphibians hypothalamohypophyseal LH-RH tracts of, 206-207 LH-RH-reactive perikarya of, 200-201 Antibody locally applied, spreading of, 133 production, macrophage-B-cell interactions in, 171-172 Antigen spreading in heterokaryons, 132-133 uptake by macrophages, 152-153

Calcium, in mitotic cell, 74-75 Calcium-adenosine triphosphatase, mitotic in cell cycle, 70-71 characteristics, 71 function, 71-74 Calcium-dependent regulator protein characteristics, 66-68 in mitotic spindle, 68 Carotid body embryology and development, 320-322 histological features blood vessels, 276-277 cell clusters, 274-276 ganglion cells, 279 nerve fibers, 277-278 other cellular elements, 279 type I cell innervation, 289-290 effects of denervation, 300-301 functional interpretation of type I cell-nerve ending relationships, 301-307 noninnervated cells, 307 origin of nerve endings on, 290-292 ultrastructure of junctions between nerve endings and cells, 295-300 ultrastructure of nerve endings on cells, 292-295 ultrastructural changes after stimulation of chemoreceptor changes in nerve endings, 318 changes in type I cells, 315-318 pathology, 318-320 ultrastructure type I cells, 279-289 type I1 cells, 289

B Bacteriorhodopsin, mobility of, 135 B-cell, antibody production, macrophage interaction and, 171-172 Birds hypothalamohypophyseal LH-RH tracts of, 206 LH-RH-reactive perikarya of, 199-200 Blast cells behavior and morphological variations in, 100-105 occurrence of two separate nuclear masses in, 105-107 Bone marrow, grain counts in vesicular nuclei of, 111-114 Brome mosaic virus, aminoacylatable RNA nucleotide sequence of 3’-OHend, 10-1 1

375

376

SUBJECT INDEX

vascular innervation and efferent inhibition, 307-308 barosensory innervation, 312-313 problem of efferent inhibition, 313-315 vasomotor innervation, 308-3 12 Cell myosin in, 65 recovery during action of injurious agents, 225-229 repair after elimination of injurious agent, 229 evidence of protein renativation, 259-263 mechanism, 263-264 resynthesis or reactivation, 257-259 repair of thermal injuries, 230-232 characteristic of thermal injury, 232 conditions affecting repair, 243-246 recovery speed, 240-243 repair of separate functions and cell as a whole, 232-235 repair zone, 235-240 stimulation of cell reparability, 246-252 Cell cycle, Ca*+-ATPase and, 70-71 Cell membrane, see also Plasma membrane lipid fluidity and, 121-130 protein mobility and, 130-132 bacteriorhodopsin, I35 fluorescence photobleaching recovery, 137-142 interpretation of diffusion constants, 143 movement of membrane glycoproteins in an electric field, 142-143 rhodopsin, 133-135 rotational diffusion of proteins, 135- 137 spreading of antigens in heterokaryons, 132- 133 spreading of locally applied antibody, 133 Cell surface, DNA of, 38-49 Coat protein, assembly, viral tRNA-like structures and, 19-20 Cytoplasm, DNA in, 28-38

Q Deoxyribonucleic acid biological implications of cytoplasmic and cell surface DNA, 49 cell surface observations, 38-45 possible source of, 45-49 cytoplasmic

biochemical evidence, 33-38 morphological evidence, 28-33 fluorescence method for, 98-99 Diffusion, rotational, of proteins, 135-137 Diffusion constants, interpretation of, 143 Dynein characteristics, 68-69 in mitotic spindle, 69-70

E Elongation factors, competition for, by viral tRNA-like structures, 17-19 Erythrocytes, young mature, formation of clone cells from, 114-116 Erythropoiesis, in blood of vertebrates with nucleated erythrocytes, 114-1 16

F Fishes hypothalamohypophyseal LH-RH tracts of, 206-207 LH-RH-reactive perikarya of, 200-201 Fluorescence, photobleaching recovery and, 137-142 Fluorescence method, for DNA and RNA, 98-99

0 Genetic analysis, of determinants mediating macrophage-T-cell interaction, 158-163 Glycoproteins, membrane, movement in an electric field, 142-143

H Heat, injurious action, mechanism of, 252-257 Hemopoietic tissue fluorescence method for DNA and RNA, 98-99 imprint, smear, fixation and staining methods, 96-98 tissue culture in virro, 99-100 tritiated thymidine method and, 98 results, 109-114 Heterokaryons, antigen spreading in, 132-133

I Immune response gene, function in macrophage-T-cell interaction, 155-158 Immunocytochemistry, of luteinizing hormonereleasing hormone, 182-188

377

SUBJECT INDEX

K Keratin cytochemistry, 359-360 differentiation, 346-350, 358-359 synthesis, 351 -353

L Leukemic cells, differentiating, fate of peripheral nuclear mass and inner nuclear mass in, 107-109 Lipid, fluidity, cell membrane and, 121-130 Luteinizing hormone-releasing hormone reactive perikarya morphology, 188-194 techniques of study immunoc ytochemistry , 182- 188 preparative histological techniques, 180- 182 topography of reactive perikarya amphibians and fishes, 200-201 birds, 199-200 mammals, 1 94- 199 Luteinizing hormone-releasing hormone tracts extrahypothalamic, 209-2 13 hypothalamohypophyseal, 201 -202 of amphibians and fishes, 206-207 of birds, 206 of mammals, 202-206 preopticoterminal, 207-209 Lymphocytes, interaction with macrophages, 153-154 M Macrophage function in antigen specific T-cell proliferations, 151-152 antigen uptake by macrophages, 152-153 genetic analysis of determinants mediating interaction, 158-163 immune response gene function and, 155158 mode of interaction, 153-154 function in helper T-cell induction, genetic restrictions, 167-171 nature of interaction, 163-167 B-lymphocyte interactions in antibody production, 171- 172 Mammals hypothalamohypophyseal LH-RH tracts of, 202-206 LH-RH-reactive perikarya of, 194-199 Mitotic cell, calcium in, 74-75

Mitotic centers, 79-81 Mitotic spindle actin in, 63-65 Cazt-ATPase of cell cycle and, 70-71 characteristics, 71 function, 7 1-74 calcium-dependent regulator protein in, 68 dynein in, 69-70 isolation procedure, 75-77 models, 78-79 myosin in, 65-66 tubulin in, 59-62 Myosin in cells, 65 in mitotic spindle, 65-66

P Plasma membrane differentiation, wool follicle cortex, 342-344 fiber cuticle, 356-357 inner root sheath, 361-362 Protein mobility in cell membranes, 130-132 bacteriorhodopsin, 135 fluorescence and photobleaching recovery, 137-142 interpretation of diffusion constants, 143 movement of membrane glycoproteins in an electric field, 142-143 rhodopsin, 133-135 rotational diffusion of proteins, 135-137 spreading of antigens in heterokaryons, 132- 133 spreading of locally applied antibody, 133

R Replication, viral tRNA-like structures and, 20-22 Rhodopsin, mobility of, 133-135 Ribonucleic acid animal virus, covalently bound aminoacylatable sequences in, 7-8 fluorescence method for, 98-99 plant virus, covalently bound aminoacylatable sequences in, 2-7 transfer, noncovalent association with viral RNAs, 8-9 Ribosome, competition for, by viral tRNA-like structures, 17

378

SUBJECT INDEX

S Spleen grain counts in vesicular nuclei of, 111-1 14 reassociation of vesicles and nuclear granules in, 109-111

T T-cell, helper, function of macrophage in induction, 163-171 T-cell progiferation, antigen-specific, macrophage function in, 151-163 Thermal injuries, repair of, 230-252 Tobacco mosaic virus, aminoacylatable RNA, nucleotide sequence of 3'-OH end, 12-13 Trichohyalin, of wool follicle, 364-365 Tubulin isolation, 54-55 in mitotic spinde, 59-62 properties, 55-58 Turnip yellow mosaic virus, arninoacylatable 'RNA, nucleotide sequence of 3'-OH end, 9-10 V Viral transfer ribonucleic acid-like structures biological function, 13-16 as amino acid donor, 16-17 in coat protein assembly, 19-20 in competition for elongation factors, 17-19 in competing with host mRNA for ribosomes, 17 in replication, 20-22 nucleotide sequence of 3'-OH ends of aminoacylatable viral RNAs brome mosaic virus, 10-1 1 tobacco mosaic virus, 12-13 turnip yellow mosaic virus, 9-10 occurrence and properties covalently bound aminoacylatable sequences

in animal virus RNAs, 7-8 in plant virus RNAs, 2-7 noncovalent association of tRNAs with viral RNAs, 8-9

W

Wool follicle bulb, 337-338 connective tissue sheath, 369 cortex keratin differentiation, 346-350 keratin synthesis, 351-353 nonkeratic cytoplasmic components, 345346 plasma membrane differentiation, 342-344 ultrastructure of macrofibrils, 350-351 dermal papilla, 334-337 fiber cuticle, 353 cytochemistry of keratin, 359-360 cytochemistry of plasma membrane, 357358 keratin differentiation, 358-359 nonkeratin cytoplasmic components, 358 plasma membrane differentiation, 356-357 shape differentiation, 354-356 general description, 332-334 inner root sheath, 360-361 breakdown of, 365-366 cytochemistry of plasma membrane, 363 nontrichohyalin cytoplasmic components, 363-364 plasma membrane differentiation, 361 -362 trichohyalin, 364-365 medulla, 338-339 cytochemistry, 341 differentiation, 339-341 outer root sheath cytochemistry, 368 differentiation, 366-368

Contents of Previous Volumes Quantitative Histochemistry of PhosphatasesWILLIAM L. DOYLE Some Historical Features in Cell BiologyAlkaline Phosphatase of the Nucleus-M. ARTHUR HUGHES ND C H ~ V R E M OAN T H. FIRKET Nuclear Reproduction-C. LEONARD HUSKINS Gustatory and Olfactory Epithelia-A. F. Enzymic Capacities and Their Relation to Cell BARADI A N D G. H. BOURNE Nutrition in Animak"E0RGE W. KIDDER Growth and Differentiation of Explanted TisThe Application of Freezing and Drying Techsues-P. J. GAILLARD niques in Cytology-L. G . E. BELL Electron Microscopy of Tissue Sections-A. J. Enzymatic Processes in Cell Membrane PenetraDALTON A N D w . WILBRANDT tion-TH. ROSENBERG A Redox Pump for the Biological Performance Bacterial Cytology-K. A. BISSET of Osmotic Work. and Its Relation to the Protoplast Surface Enzymes and Absorption of Kinetics of Free Ion Diffusion across MemSugar-R. BROWN branes-E. J. CONWAY Reproduction of Bacteriophage-A. D. HERA Critical Survey of Current Approaches in SHEY Quantitative Histo- and CytochemistryThe Folding and Unfolding of Protein Molecules CLICK DAVID GOLDas a Basis of Osmotic Work-R. J. Nucleo-cytoplasmic Relationships in the DeACRE velopment of Acetabularia-J. HAMMERLING Nucleo-Cytoplasmic Relations in Amphibian Report of Conference of Tissue Culture Workers Development-4. FRANK-HAUSER J. Held at Cooperstown, New York-D. Structural Agents in Mitosis-M. M. SWANN H ETH ERINGTON Factors Which Control the Staining of Tissue AUTHOR INDEX-SUBJECT INDEX Sections with Acid and Basic Dyes-MARCUS SINGER Volume 3 The Behavior of Spermatozoa in the Neighborhood of Eggs-LORD ROTHSCHILD The Nutrition of Animal CellS-cHARITY The Cytology of Mammalian Epidermis and WAYMOUTH Sebaceous Glands-WILLIAM MONTAGNA Caryometric Studies of Tissue Cu~tures-oTTo The Electron-Microscopic Investigation of TisBUCHER sue Sections-L. H. BRETSCHNEIDER The Properties of Urethan Considered in RelaThe Histochemistry of Esterases-C. GOMORI tion to Its Action on Mitosis-IvoR CORNMAN AUTHOR INDEX-SUBJECT INDEX Composition and Structure of Giant Chromosomes-MAX ALFERT How Many Chromosomes in Mammalian SoVolume 2 matic Cells?-R. A. BEATTY The Significance of Enzyme Studies on Isolated Quantitative Abpects of Nuclear NucleoproCell Nuclei-ALEXANDER L. DOUNCE teins-HEWSON SWIFT The Use of DiffeRntial Centrifugation in the Ascorbic Acid and Its lntracellular Localization. with Special Reference to Plants-J. C H A Y E N Study of Tissue Enzymes-CHR. DE DUVE A N D J . BERTHET Aspects of Bacteria as Cells and as OrganEnzymatic Aspects of Embryonic DifferentiaD. DE iSmS-sTUART M U D D A N D EDWARD tion-TRYGGVE GUSTAFSON LAMATER Azo Dye Methods i n Enzyme HistochemistryIon Secretion in Plants-J. F. SUTCLIFFE A. G. EVERSON PEARSE Multienzyme Sequences in Soluble ExtractsMicroscopic Studies i n Living Mammals with HENRYR. M A H L E R Transparent Chamber Methods-Roy G. WILThe Nature and Specificity of the Feulgen LIAMS Nuclear Reaction-M. A. LESSLER

Volume 1

319

380

CONTENTS OF PREVIOUS VOLUMES

The Mast C e l l - G . ASBOE-HANSEN Elastic Tissue-EDWARDS w. DEMPSEYA N D ALBERTI . LANSING The Composition of the Nerve Cell Studied with New Methods-SvEN-OLoE BRATTGARD AND HOLGERHYDEN AUTHOR INDEX-SUBJECT

INDEX

Volume 4 Cytochemical Micrurgy-M. 1. KOPAC Amoebocytes-L. E. WACGE Problems of Fixation in Cytology, Histology, and Histochemistry-M. WOLMAN Bacterial Cytology-ALFRED MARSHAK Histochemistry of Bacteria-R. VENDRELY Recent Studies on Plant Mitochondria-DAVID P. HACKETT The Structure of Chloroplasts-K. MOHLE-

on Cytokinesis and Amoeboid MovementDOUGLAS MARSLAND lntracellular pH-PETER C. CALDWELL The Activity of Enzymes in Metabolism and Transport in the Red Cell-T. A. J. PRANKERD Uptake and Transfer of Macromolecules by Cells with Special Reference to Growth and Development-A. M. SCHECHTMAN Cell Secretion: A Study of Pancreas and Salivary Glands-L. c. J. J U N Q U E I R A A N D G . c . HlRsCH The Acrosome Reaction-JEAN C. DAN Cytology of Spermatogenesis-VisHwA NATH The Ultrastructure of Cells, as Revealed by the Electron Microscope-FRITloF S . SJOSTRAND AUTHOR INDEX-SUBJECT

INDEX

Volume 6

THALER

Histochemistry of Nucleic Acids-N.

B. KUR-

NICK

Structure and Chemistry of Nucleoli-W. S . VINCENT On Goblet Cells, Especially of the Intestine of Some Mammalian Species-HARALD MOE Localization of Cholinesterases at Neuromuscular Junctions-R. COUTEAUX Evidence for a Redox Pump in the Active Transport of Cations-E. J. CONWAY AUTHOR INDEX-SUBJECT

INDEX

Volume 5 Histochemistry with Labeled Antibody-ALBERT H. COONS The Chemical Composition of the Bacterial Cell Wall<. S. CUMMINS Theories of Enzyme Adaptation in Microorganisms-J. MANDELSTAM The Cytochondria of Cardiac and Skeletal Muscle-JoHN w . HARMON The Mitochondria of the Neuron-WARREN ANDREW The Results of Cytophotometry in the Study of the Deoxyribonucleic Acid (DNA) Content of VENDRELYAND C. VENthe Nucleus-R. DRELY

Protoplasmic Contractility in Relation to Gel Structure: Temperature-Pressure Experiments

The Antigen System ofParumecium 4urelia-G. H. BEALE The Chromosome Cytology of the Ascites Tumors of Rats, with Special Reference to the Concept of the Stemline Cell-.!iAJlRO MAKIN0

The Structure of the Golgi Apparatus-ARTHUR w . POLLISTER A N D PRISCHIA F. POLLISTER An Analysis of the Process of Fertilization and Activation of the Egg-A. MONROY The Role of the Electron Microscope in Virus Research-RoBLEY C. WILLIAMS The Histochemistry of Polysaccharides-ARTHUR J. HALE The Dynamic Cytology of the Thyroid Gland-J. GROSS

Recent Histochemical Results of Studies on Embryos of Some Birds and Mammals-ELio BORGHESE Carbohydrate Metabolism and Embryonic Determination-R. J. O’CONNOR Enzymatic and Metabolic Studies on Isolated Nuclei-G. SIEBERT A N D R. M. s. ShlELLlE Recent Approaches of the Cytochemical Study of Mammalian TissuesaEORGE H. HOGEBOOM. EDWARD L. KUFF, A N D WALTERC. SCHNEIDER The Kinetics of the Penetration of Nonelectrolytes into the Mammalian ErythrocyteFREDABOWYER

38 1

CONTENTS OF PREVIOUS VOLUMES AUTHOR INDEX-SUBJECT

INDEX

CUMULATIVE SUBJECT INDEX (VOLUMES

1-5)

Volume 7 Some Biological Aspects of Experimental Radiology: A Historical Review-F. G.SPEAR The Effect of Carcinogens, Hormones, and Vitamins on Organ Cultures-ILsE LASNITZKI Recent Advances in the Study of the Kinetochore-A. LIM A-DE-FARIA Autoradiographic Studies with S3'-Sulfate-D. D. DZIEWIATKOWSKI The Structure of the Mammalian Spermatozoon-Don W. FAWCETT The Lymphocyte-0. A. TROWELL The Structure and Innervation of Lamellibranch Muscle-J. BOWDEN Hypothalamo-neurohypophysial NeurosecEtionJ. C. SLOPER Cell Contact-PAUL WEISS The Ergastoplasm: Its History, Ultrastructure, and Biochemistry-FRANColSE HAGUENAU Anatomy of Kidney Tubules-JOHANNES RHODIN

Structure and Innervation of the Inner Ear Sensory Epithelia-HANS ENGSTROM AND JAN WERSALL The Isolation of Living Cells from Animal Tissues-L. M. RINALDINI A U T H O R INDEX-SUBJECT

INDEX

Volume 8 The Structure

A Survey of Metabolic Studies on Isolated Mammalian Nuclei-D. B. ROODYN Trace Elements in Cellular Function-BERT L. VALLEE A N D FREDERIC L. HOCH Osmotic Properties of Living Cells-D. A. T. DICK Sodium and Potassium Movements in Nerve, Muscle, and Red Cells-I. M. GLYNN Pinocytosis-H. HOLTER AUTHOR INDEX-SUBJECT

Volume 9 The Influence of Cultural Conditions on Bacterial A N D J. P. DUCytology-J. F. WILKINSON GUID

Organizational Patterns within ChromosomesBERWIND P. KAUFMANN.HELEN GAY. A N D MARGARET R. MCDONALD Enzymic Processes in Cells-JAY BOYDBEST The Adhesion of Cells-LEONARD WElss Physiological and Pathological Changes in Mitochondrial Morphology-CH. ROUILLER The Study of Drug Effects at the Cytological L e v e l - C . B. WILSON Histochemistry of Lipids in Oogenesis-VisHwA NATH Cyto-Embryology of Echinoderms and Amphibia-KUTSUMA DAN The Cytochemistry of Nonenzyme ProteinsRONALD R. COWDEN AUTHOR INDEX-SUBJECT

Of

INDEX

INDEX

CYtOplaSm-cHARLES OBER-

LING

Wall Organization in Plant Cells-R.

D. PRES-

TON

Submicroscopic Morphology of the SynapseEDUARD DE~ ROBERTIS The Cell Surface ofParameriurn-C. F. EHRET A N D E. L. POWERS MIRIAM The Mammalian RetiCUlOCyte-LEAH LOWENSTEIN The Physiology of Chromatophores-MILTON FINGERMAN The Fibrous Components of Connective Tissue with Special Reference to the Elastic FiberA. HALL DAVID Experimental Hererotopic Ossification-J. B. BRIDGES

Volume 10 The Chemistry of Shiff's Reagent-FREDERICK H. KASTEN Spontaneous and Chemically Induced Chromosome Breaks-ARuN K U M A R S H A R MAAN D ARCHANA SHARMA The Ultrastructure of the Nucleus and Nucleocytoplasmic Relations-SAUL WISCHNITZER The Mechanics and Mechanism of CleavageLEWISWOLPERT The Growth of the Liver with Special Reference to Mammals-F. DOLJANSKI Cytology Studies on the Affinity of the Carcinogenic Azo Dyes for Cytoplasmic CompoIl~nlS-YOSHIMA NAGATANI

382

CONTENTS OF PREVIOUS VOLUMES

Epidermal Cells in Culture-A.

GEDEONMA-

TOLTSY AUTHOR INDEX-SUBJECT INDEX CUMULATIVE SUBJECT INDEX (VOLUMES

Sequential Gene Action, Protein Synthesis, and Cellular Differentiation-REED A. FLICKINGER

1-9)

Volume 11 Electron Microscopic Analysis of the Secretion Mechanism-K. KUROSUMI The Fine Structure of Insect Sense OrgansELEANOR H. SLIFER Cytology of the Developing Eye-ALFRED J. COULOMBRE The Photoreceptor Structures-J. J. WOLKEN Use of Inhibiting Agents in Studies on Fertilization Mechanisms--CHARLEsB. METZ The Growth-Duplication Cycle of the Cell-D. M. PRESCOTT Histochemistry of Ossification-RoMuLo L. CABRINI Cinematography, Indispensable Tool for Cytol0 g y - C . M. POMERAT

The composition of the Mitochondria1 Membrane in Relation to Its Structure and Function- ERICG. BALLA N D CLIFFED. JOEL Pathways of Metabolism in Nucleate and Anucleate Erythrocytes-H. A. SCHWEIGER Some Recent Developmentsin the Field of Alkali Cation Transport-W. WILBRANDT Chromosome Aberrations Induced by lonizing Radiations-H. J. EVANS Cytochemistry of Protozoa, with Particular Reference to the Golgi Apparatus and the Mitochondria-VISHWA NATH A N D G. P. DUTTA A N D CHOCell Renewal-FELIX BERTALANFFY SEN LAU AUTHOR INDEX-SUBJECT

INDEX

Volume 14

Inhibitionofcell Division: A Critical and Experimental Analysis-SEYMOUR GELFANT Electron Microscopy of Plant Protoplasm-R. BUVAT Volume 12 Cytophysiology and Cytochemistry of the Organ Sex Chromatin and Human Chromosomes-of Corti: A Cytochemical Theory of HearJOHN L. HAMERTON AND L. K. TITOVA ing-J. A. VINNIKOV Chromosomal Evolution in Cell Populations-T. Connective Tissue and Serum Proteins-R. E. MANCINI C. Hsu The Biology and Chemistry of the Cell Walls of Chromosome Structure with Special Referenceto H. Higher Plants, Algae, and Fungi-D. the Role of Metal lons-D~LE M. STEFFENNORTHCOTE SEN Electron Microscopy of Human White Blood Development of Drug Resistance by StaphyloCells and Their Stem Cells-MARCEL BESSIS cocci in Vim and in Vivo-MARY BARBER A N D JEAN-PAUL THIERY Cytological and Cytochemical Effects of Agents I n Vivo Implantation as a Technique in Skeletal Implicated in Various Pathological CondiBiology-WILLIAM J. L. FELTS tions: The Effect of Viruses and of Cigarette The Nature and Stability of Nerve Myelin-J. B. Smoke on the Cell and Its Nucleic AcidFINEAN CEClLlE LEUCHTENBERGER A N D RUDOLF Fertilizationof Mammalian Eggsin Virro-C. R. LEUCHTENBERGER AUSTIN The Tissue Mast Wall-DouG~~s E. SMITH Physiology of Fertilization in Fish Eggs-ToKI-0 AUTHOR INDEX-SUBJECT INDEX Y AMAMOTO AUTHOR INDEX-SUBJECT

INDEX

AUTHOR INDEX-SUBJECT

INDEX

Volume 13 The Coding Hypothesis-MARTYNAs YCAS Chromosome Reproduction-J. HERBERTTAYLOR

Volume 15 The Nature of Lampbrush Chromosomes-H. G. CALLAN The lntracellular Transfer of Genetic Information-J. L. SIRLIN Mechanisms of Gametic Approach in Plants-

383

CONTENTS OF PREVIOUS VOLUMES LEONARD MACHLISA N D ERIKA RAWITSCHERKUNKEL The Cellular Basis of Morphogenesis and Sea A N D L. Urchin Development-T. GUSTAFSON WOLPERT Plant Tissue Culture in Relation to Development CytOlOgy
The Cells of the Adenohypophysis and Their Functional Significance-MARC HERLANT AUTHOR INDEX-SUBJECT

INDEX

Volume 18

The Cell of Langerhans-A. S. BREATHNACH The Structure of the Mammalian Egg-ROBERT HADEK Cytoplasmic Inclusions in Oogenesis-M. D. L. SRIVASTAVA The Classification and Partial Tabulation of Enzyme Studies on Subcellular Fractions Isolated by Differential Centrifuging-D. B. ROODYN AUTHOR INDEX-SUBJECT INDEX HistochemicalLocalization of Enzyme Activities by Substrate Film Methods: Ribonucleases, Deoxyribonucleases, Proteases, Amylase, and Volume 16 Hyaluronidase-R. DAOUST Ribosomal Functions Related to Protein Synthe- Cytoplasmic DeoxyribonucleicAcid-P. B. GASis-TORE HULTIN H A N A N D J . CHAYEN Physiology and Cytology of Chloroplast Forma- Malignant Transformation of Cells in Vitrotion and "Loss" in Euglena-M. GRENSON KATHERINEK. SANFORD Cell Structures and Their Significance for Ame- Deuterium Isotope Effects in Cytology-E. boid Movement-K. E. WOHLFARTH-BOTT- FLAUMENHAFT, S. BOSE,H. I . CRESPI, A N D J. ERMAN J. KATZ Microbeam and Partial Cell Irradiation-€. L. The Use of Heavy Metal Salts as Electron SMITH Stains-€. RICHARDZOBELA N D MICHAEL Nuclear-Cytoplasmic Interaction with Ionizing BEER Radiation-M. A. LESSLER AUTHOR INDEX-SUBJECT INDEX I n Vivo Studies of Myelinated Nerve FibersCARLCASKEY SPEIDEL Respiratory Tissue: Structure, Histophysiology, Volume 19 Cytodynamics.Part I: Review and Basic Cyto"Metabolic" DNA: A Cytochemical Study-H. morphology-FELIX D. BERTALANFFY ROELS AUTHOR INDEX-SUBJECT INDEX The Significance of the Sex Chromatin-MuRRAY L. BARR Volume 17 Some Functions of the Nucleus-J. M. MITCHISON The Growth of Plant Cell Walls-K. WILSON Synaptic Morphology on the Normal and DegenReproduction and Heredity in Trypanosomes: A erating Nervous System-E. G.GRAYA N D R. Critical Review Dealing Mainly with the AfW. GUILLERY rican Species in the Mammalian Host-P. J. Neurosecretion-W. BARGMANN WALKER The Blood Platelet: Electron Microscopic Stud- Some Aspects of Muscle Regeneration-E. H. BETZ,H. FIRKET, A N D M. REZNIK ies-J. F. DAVID-FERREIRA The Gibberellins as Hormones-P. W. BRIAN The Histochemistry of MucopolysaccharidesPhototaxis in PlantS-WOLFGANG H A U ~ T ROBERTC. CURRAN S. Respiratory Tissue Structure, Histophysiology, Phosphorus Metabolism in Plants-K. ROWAN Cytodynamics. Pan 11. New Approaches and Interpretations-FELIX D. BERTALANFFY A U T H O R INDEX-SUBJECT INDEX

384

CONTENTS OF PREVIOUS VOLUMES

Volume 20 The Chemical Organization of the Plasma Membrane of Animal Cells-A. H. MADDY Subunits of Chloroplast Structure and Quantum Conversion in Photosynthesis-RODERIC B. PARK Control of Chloroplast Structure by Light-LEsTER PACKER A N D PAUL-ANDRBSIEGENTHALER

The Dynamism of Cell Division during Early Cleavage Stages of the Egg-N. FAUTREZA N D J. FAUTREZ FIRLEFYN Lymphopoiesis in the Thymus and Other Tissues: AND Functional Implications-N. B. EVERETT RUTHW. TYLER(CAFFREY) Structure and Organization of the Myoneural J u n c t i o n x . CoERS The Ecdysial Glands of Arthropods-WILLIAM S. HERMAN Cytokinins in Plants-B. I . SAHAISRIVASTAVA

The Role of Potassium and Sodium Ions as Studied in Mammalian Brain-H. HILLMAN AUTHOR INDEX-SUBJECT I N D E X Triggering of Ovulation by Coitus in the RatCUMULATIVE SUBJECT tNDEX (VOLUMES 1-21) CLAUDE ARON,GITTAASCH.A N D JAQUELINE RWs Cytology and Cytophysiology of Non-MelanoVolume 23 phore Pigment Cells-JOSEPH T. BAGNARA Transformationlike Phenomena in Somatic The Fine Structure and Histochemistry of ProstaCells-J. M. OLENOV tic Glands in Relation to Sex HormonesRecent Developments in the Theory of Control DAVIDBRANDES and Regulation of Cellular Processes-ROBCerebellar Enzymology-LuclE ARVY ERT ROSEN AUTHOR IN DEX-SU BJECT I N D E X Contractile Properties of Protein Threads from Sea Urchin Eggs in Relation to Cell DivisionVolume 21 HIKOICHI SAKAI Electron Microscopic Morphology of OogeneHistochemistry of Lysosomes-P. B. GAHAN SiS-ARNE NQRREVANG Physiological Clocks-R. L. BRAHMACHARY Dynamic Aspects of Phospholipids during ProCiliary Movement and Coordination in Ciliatestein Secretion-LOWELL E. HOKIN BELAPARDUCA The Golgi Apparatus: Structure and FunctionElectromyography: Its Structural and Neural H. W. BEAMSA N D R. G. KESSEL Basis-JOHN V . BASMAJIAN The Chromosomal Basis of Sex DeterminationCytochemical Studies with Acridine Orange and KENNETHR. LEWISA N D BERNARD JOHN the Influence of Dye Contaminants in the

AUTHOR INDEX-SUBJECT INDEX H. Staining of Nucleic Acids-FREDERICK KASTEN Experimental Cytology of the Shoot Apical Cells Volume 24 during Vegetative Growth and Flowering-A. Synchronous Cell DifferentiationaEoRGE M. NOUGARBDE A N D I V A N L. CAMERON PADILLA Nature and Origin of Perisynaptic Cells of the Mast Cells in the Nervous System-YNGVE Motor End Plate-T. R. SHANTHAVEERAPPA OLSSON A N D G . H. BOURNE Development Phases in lntermitosis and the AUTHOR INDEX-SUBJECT INDEX Preparation for Mitosis of Mammalian Cells ifr VilrO-BLAGOJE A. NESKOVI~. Volume 22 Antimitotic Substances-Guv DEYSSON The Form and Function of the Sieve Tube: A Current Techniques in Biomedical Electron MiProblem in Reconciliation-P, E. WEATHERcroscopy-SauL WISCHNITZER LEY A N D R. P. C. JOHNSON The Cellular Morphology of Tissue Repair-R. Analysis of Antibody Staining Patterns Obtained M. H. MCMINN with Striated Myofibrils in Fluorescence MiStructural Organization and Embryonic Differencroscopy and Electron Microscopy-FRANK tiatiOn4AJANAN v . SHERBET A N D M. s. A. PEPE LAKSHMI

385

CONTENTS OF PREVIOUS VOLUMES Cytology of Intestinal Epithelial Cdk-PETER G. TONER Liquid Junction Potentials and Their Effects on Potential Measurements in Biology Systems-P. C. CALDWELL AUTHOR INDEX-SUBJECT

INDEX

Volume 25 Cytoplasmic Control over the Nuclear Events of Cell Reproduction-NOEL DE TERRA Coordination of the Rhythm of Beat in Some Ciliary Systems-M. A. SLEIGH The Significance of the Structural and Functional Similarities of Bacteria and MitochondriaSYLVAN NASS The Effects of Steroid Hormones on Macrophage Activity-B. VERNON-ROBERTS The Fine Structure of Malaria Parasites-MARIA A. RUDZINSKA The Growth of Liver Parenchymal Nuclei and Its Endocrine Regulation-RITA CARRIERE Strandedness of Chromosomes-SHELDON WOLFF Isozymes: Classification, Frequency, and Significance-CHARLES R. SHAW The Enzymes of the Embryonic NephronLUCIEARVY Protein Metabolism in Nerve Cells-B. DROZ Freeze-Etching-HANS MOOR AUTHOR INDEX-SU

BJECT INDEX

Volume 26 A New Model for the Living Cell: A Summary of the Theory and Recent Experimental Evidence in Its S U ~ ~ O ~ ~ ~ NI. LLING B E R T The Cell Penphery-LEONAKD WEISS Mitochondria1 DNA: Physicochemical Properties, Replication, and Genetic Function-P. BORSTA N D A. M. KROON Metabolism and Enucleated Cdk-KONRAD KECK Stereological Principles for Morphometry in Electron Microscopic Cytology-EwALD R. WEIBEL Some Possible Roles for lsozymic Substitutions during Cold Hardening in Plants-D. W. A. ROBERTS AUTHOR INDEX-SUBJECT

INDEX

Volume 27 Wound-Healing in Higher Plants-JACQUES LIPETZ Chloroplasts as Symbiotic Organelles-DENNIS L. TAYLOR The Annulate Lamellae-SAUL WISCHN~TZER Gametogenesis and Egg Fertilization in Planarians-4. BENAZZI LENTATI Ultrastructure of the Mammalian Adrenal CorteX-sIMON IDELMAN The Fine Structure of the Mammalian Lymphoreticular SyStem-IAN CARR lmmunoenzyme Technique: Enzymes as Markers for the Localization of Antigens and Antibodies-STRATIS AVRAMEAS AUTHOR INDEX-SUBJECT

INDEX

Volume 28 The Cortical and Subcortical Cytoplasm of Lymntrerr E ~ - C H R I S T I AP. AN RAVEN The Environment and Function of Invertebrate Nerve Cells-J. E. TREHERNE A N D R. B. MORETON Virus Uptake, Cell Wall Regeneration. and Virus Multiplication in Isolated Plant ProtoplastsE. C. COCKING The Meiotic Behavior of the Drosophila 00cyte-ROBERT c . KING The Nucleus: Action of Chemical and Physical Agents-RENB SIMARD The Origin of Bone Cek-MAUREEN OWEN Regeneration and Differentiation of Sieve Tube Elements-WILLIAM P. JACOBS Cells, Solutes, and Growth: Salt Accumulation in Plants Reexamined-F. C. STEWARD A N D R. L. MOTT AUTHOR INDEX-SUBJECT

INDEX

Volume 29 Gram Staining and Its Molecular Mechanism8 . B. BISWAS,P. S . BASU.A N D M K. . PAL The Surface Coats of Animal Cells-A. MARTfNEZ-PALOMO Carbohydrates i n Cell Surfaces-RICHARD J. WINZLER Differential Gene Activation in Isolated Chromosomes-MARKUS LEZZI lntraribosomal Environment of the Nascent Peptide Chain-HIDEKO KAJI

386

CONTENTS OF PREVIOUS VOLUMES

Location and Measurementof Enzymes in Single Cells by Isotopic Methods Part I-E. A. BARNARD

Location and Measurement of Enzymes in Single Cells by Isotopic Methods Part 1 1 4 . C. BUDD Neuronal and Glial Perikarya Preparations: An Appraisal of Present Methods-PATRICIA v. JOHNSTON A N D BETTYI. ROOTS Functional Electron Microscopy of the HypothaKOBAlamic Median Eminence-HlDEsHI YASHI, TOKUZO MATSUI. A N D SUSUMI ISHII Early Development in Callus Cultures- MICHAEL M. YEOMAN AUTHOR INDEX-SUWECT

INDEX

Volume 30

Volume 32 Highly Repetitive Sequencesof DNA in Chromosomes-W. G.FLAMM The Origin of the Wide Species Variation in Nuclear DNA Content-H. REESA N D R. N. JONES Polarized lntracellular Particle Transport: Saltatory Movements and Cytoplasmic Stream1. REBHUN ing- LIONEL The Kinetoplast of the HemOflagellateS-LARRY SIMPSON Transport across the Intestinal Mucosal Cell: AND Hierarchies of Function-D. S. PARSONS C. A. R. BOYD Wound Healing and Regeneration in the Crab Pararelphusa hydrodromous-RITA G. ADIYODI

The Use of Ferritin-Conjugated Antibodies in Electron Microscopy~ouNclLMAN MOR-

High-pressureStudies in Cell Biology-ARTHUR M. ZIMMERMAN GAN Micrurgical Studies with Large Free-Living Metabolic DNA in Ciliated Protozoa, Salivary Amebas-K. W. JEONA N D J. F. DANIELLI Gland Chromosomes, and Mammalian The Practice and Application of Electron MicroCells-S. R. PELC scope Autoradiography-J. JACOB AUTHOR INDEX-SUBJECT INDEX Applicationsof Scanning Electron Microscopy in Biology-K. E. CARR Acid Mucopolysaccharides in Calcified TisSUeS-SHlNJlRO KOBAYASHI AUTHOR INDEX-SUBJECT

INDEX

CUMULATIVE SUBJECT I N D E X (VOLUMES 1-29)

Volume 33

Visualization of RNA Synthesis on ChromoJ R . A N D BARBARA A. somes-0. L. MILLER, HAMKALO Studies on Freeze-Etching of Cell MembranesCell Disjunction (“Mitosis”) in Somatic Cell KURT M ~ H L E T H A L E R G. DIACUMAKOS, Reproduction-ELAINE Recent Developments in Light and Electron A N D PAULINE PECORA SCOTTHOLLAND, Microscope Radioautography4. C. BUDD Neuronal Microtubules. Neurofilaments. and Morphological and Histochemical Aspects of Microfilaments-RAYMOND B. WUERKER Glycoproteins at the Surface of Animal A N D JOELB. KIRKPATRICK Cells- A. RAMBOURG Lymphocyte Interactions in Antibody ReDNA Biosynthesis-H. S. JANSZ,D. V A N DER sponses-J. F. A. P. MILLER MEI. A N D G. M. ZANDVLIET Laser Microbeams for Partial Cell lrradiationCytokinesis in Animal Cells-R. RAPPA~ORT MICHAELW. BERNSA N D CHRISTIAN SALET The Control ofcell Division in Ocular Lens<. Mechanisms of Virus-Induced Cell FusionV. HARDING, J . R. REDDAN, N . J . UNAKAR, GEORGE POSTE AND M. BAGCHI Freeze-Etching of BaCteria-cHARLES C. REMThe Cytokinins-HANS KENDE SEN A N D STANLEY W. WATSON Cytophysiologyof the Teleost Pituitary- MAR- The Cytophysiology of Mammalian Adipose T I N SAGEA N D HOWARDA. BERN CellS-BERNARD G. SLAVIN

Volume 31

AUTHOR INDEX-SUBJECT

INDEX

AUTHOR INDEX-SUBJECT

INDEX

387

CONTENTS OF PREVIOUS VOLUMES

tigophora and Opalinata (Protozoa)-G. P. DUTTA The Submicroscopic Morphology of the Inter- Chloroplasts and Algae as Symbionts in Molp h a Nucleus-SAUL WISCHNITZER IWS-LEONARD MUSCATINEA N D RICHARD The Energy State and Structure of Isolated W. GREENE Chloroplasts: The Oxidative Reactions Involv- The Macrophage-SAlMoN GORDONA N D ZANing the Water-Splitting Step of PhotosyntheVIL A. COHN sis- ROBERTL. HEATH Degeneration and Regeneration of NeurosecreTranspn in Neurospora-ciENE A. SCARtory SyStemS-HORST-DIETER DELLMANN

Volume 34

BOROUGH

Mechanisms of Ion Transport through Plant Cell Membranes-EMANUEL ERSTEIN Cell Motility: Mechanisms in Protoplasmic Streaming and Ameboid Movement-H. KOMNICK, w . STOCKEM, A N D K. E. WOHLE-

FARTH-BOTTERMANN The Gliointerstitial System of Molluscs- GHISLAIN NICAISE Colchicine-Sensitive Microtubules-LYNN MARGULIS AUTHOR INDEX-SUBJECT

INDEX

Volume 37 Units of DNA Replication in Chromosomes of Eukaroytes-J. HERBERT TAYLOR Viruses and Evolution-D. C. REANNEY Electron Microscope Studies on Spenniogenesis in Various Animal SpeCieS-GONPACHIRO YASUZUMl

INDEX

Volume 35 The Structure of Mammalian ChromosomesELTONSTUBBLEFIELD Synthetic Activity of Polytene ChromosomesHANSD. BERENDES Mechanismsof Chromosome Synapsisat Meiotic Prophase-PETER B. MOENS Structural Aspects of Ribosomes-N. NANNINGA

AUTHOR INDEX-SUBJECT

Morphology, Histochemistry, and Biochemistry of Human Oogenesis and OVUlatiOn-sARDUL S. GURAYA Functional Morphology of the Distal LungK A Y EH. KILBURN comparative Studies of the Juxtaglomerular Apparatus-HIROFUMI SOKABE A N D MlZUHO OGAWA The Ultrastructure of the Local Cellular Reaction to Neoplasia-IAN CARRANDJ.C. E.UNDERWOOD

Scanning Electron Microscopy in the Ultrastructural Analysis of the Mammalian Cerebral Ventricular System-D. E. SCOTT,G. P. KOZLOWSKI,AND M. N. SHERIDAH

Comparative Ultrastructure of the Cerebrospinal Fluid-Contacting Neurons-B. VIGH A N D 1. VIGH-TEICHMANN AUTHOR INDEX-SUBJECT INDEX Maturation-Inducing Substance in StarfishesHARUOKANATANI The Limonium Salt Gland: A Biophysical and Structural Study-A. E. HILLA N D B. S.HILL Volume 38 Toxic Oxygen Effects-HAROLD M. SWARTZ Genetic Engineering and Life Synthesis: An AUTHOR INDEX-SUBJECT INDEX Introduction to the Review by R. Widdus and C. Auk-JAMES F. DANlELLl Progress in Research Related to Genetic EngiVolume 36 neering and Life Synthesis-Roy WIDDUS A N D CHARLES R. AULT Molecular Hybridization of DNA and RNA in The Genetics of C-Type RNA Tumor VirusesSitU-WOLFGANG HENNIG J. A. WYKE The Relationship between the Plasmalemma and Three-Dimensional Reconstruction from ProjecPlant Cell Wall-JEAN-CLAUDE ROLAND tions: A Review of Algorithms-RICHARD Recent Advances in the Cytochemistry and UltraA N D GABOR T. HERMAN GORDON structure of Cytoplasmic lnclusions in Mas-

388

CONTENTS OF PREVIOUS VOLUMES

The Cytophysiology of Thyroid Celk-vLADlMIR R. PANTIC The Mechanisms of Neural Tube FormationPERRYKARFUNKEL The Behavior of the XY Pair in MammalsALBERTO J. SOLARI Fine-Structural Aspects of Morphogenesis in Acetabu1aria-G. WERZ Cell Separation by Gradient Centrifugation-R. HARWOOD SUBJECT INDEX

Volume 39 Androgen Receptors in the Nonhistone Protein Fractions of Prostatic Chromatin-TUNG Y U E WANGA N D LEROY M. NYBERG Nucleocytoplasmic Interactions in Development of Amphibian Hybrids-STEPHEN SUBTELNY The Interactions of Lectins with Animal Cell SUrfaceSaARTH L. NICOLSON Structure and Function of Intercellular Junctions-L. ANDREW STAEHELIN Recent Advances in Cytochemistry and Ultrastructure of Cytoplasmic Inclusions in Ciliophora (Protozoa)-C. P. DUTTA Structure and Development of the Renal Glomerulus as Revealed by Scanning Electron Microscopy-FRANC0 SPINELL1 Recent Progress with Laser Microbeams- MICHAEL w . BERNS The Problem of Germ Cell Determinants-H. W. BEAMS AND R. G . KESSEL SUBJECT INDEX

Volume 40 B-Chromosome Systems in Flowering Plants and Animal Species-R. N. JONES The Intracellul'ar Neutral SH-DependentProtease Associated with Inflammatory ReactionsHIDEOHAYASHI The Specificity of Pituitary Cells and Regulation O f Their AC1iVitie.S-VLADIMIR R. PANTIC Fine Structure of the Thyroid Gland-HisAo FUJITA Postnatal Gliogenesis in the Mammalian BrainA. PRIVAT Three-Dimensional Reconstruction from Serial Sections-RANDLE w . W A R E A N D VINCENT LOPREST1 SUBJECT INDEX

Volume 41 The Attachment of the Bacterial Chromosome to AND the Cell Membrane-PAUL J. LEIBOWITZ MOSELIO SCHAECHTER Regulation of the Lactose Operon in Escherichia coli by cAMP-G. CARPENTER A N D B. H. SELLS Regulation of Microtubules in TetrahymenaNORMANE. WILLIAMS Cellular Receptors and Mechanismsof Action of Steroid Hormones-SHUTSUNG LIAO A Cell Culture Approach to the Study of Anterior Pituitary Cells-A. TIXIER-VIDAL,D. GOURDJI, A N D c . TOUGARD lmmunohistochemical Demonstration of Neurophysin in the Hypothalamoneurohypophysial System-W. B. WATKINS The Visual System of the Horseshoe Crab Limulus polyphemus-WOLF H. FAHRENBACH SUBJECT INDEX

Volume 42 Regulators of Cell Division: Endogenous Mitotic Inhibitors of Mammalian Celk-BlSMARCK B. C . BAMLozzio, CARMEN B. Lozzio, ELENA BERGER, A N D STEPHEN v . LAIR Ultrastructure of Mammalian Chromosome A N D WALTER Aberrations-B. R. BRINKLEY N. HITTELMAN ComputerProcessing of Electron Micrographs:A Nonmathematical Account-P. W. HAWKES Cyclic Changes in the Fine Structure of the Epithelial Cells of Human EndometriumMILDREDGORDON The Ultrastructure of the Organ of CortiROBERTS. KIMURA Endocrine Cells of the Gastric Mucosa-ENRIco VASSOLCIA,CARLOCAPELLA, GABIUELE SALLO, A N D ROBERTOBUFFA Membrane Transport of Purine and Pyrimidine Bases and Nucleosides in Animal CellsRICHARD D. BERLINA N D JANET M. OLIVER SUBJECT INDEX

Volume 43 The Evolutionary Origin of the Mitochondrion:A Nonsymbiotic Model-HENRY R. MAHLER A N D RUDOLFA. RAFF Biochemical Studies of Mitochondria1Transcrip-

CONTENTS OF PREVIOUS VOLUMES tion and Translation-C. SACCONE AND E. QUAGLIARIELLO The Evolutionof the Mitotic Spindk-hNNA F. KUBAl

Germ Plasma and the Differentiation of the Germ Cell Line-E. M. EDDY Gene Expression in Cultured Mammalian Cells-RoDY P. COX A N D JAMESc . KING Morphology and Cytology of the Accessory Sex Glands in Invertebrates-K. G . ADlYODl A N D R. G . ADlYODl SUBJECT INDEX

389

Small Lymphocyte and TransitionalCell Populations of the Bone Marrow; Their Role in the Mediation of Immune and Hemopoietic Progenitor Cell FU~C~~O~S-CORNELIUS ROSE The Structure and Properties of the Cell Surface Coat-J. H. LUFT Uptake and Transport Activity of the Median Eminenceof the Hypothalamus-K. M. KNIGGE, S. A. JOSEPH,J. R. SLADEK,M. F. M. MORRIS, D. K. SUNDBERG,M. A. NOTTER, G. E. HOFFMAN,A N D L. HOLZWARTH, O’BRIEN SUBJECT INDEX

Volume 44 The Nucleolar Structure-SlBDAs GHOSH The Function of the Nucleolus in the Expression of Genetic Information: Studies with Hybrid Animal Cells-E. SIDEBOTTOM A N D I . I . DEhK Phylogenetic Diversity of the Proteins Regulating Muscular C o n t r act i o n - W ~ ~ ~ la LEHMAN ~ Cell Size and Nuclear DNA Content in Vertebrates-HENRYK SZARSKl Ultrastructural Localization of DNA in Ultrathin Tissue SeCtiOnS-ALAlN GAUTIER Cytological Basis for Permanent Vaginal Changes in Mice Treated Neonatally with Steroid Hormones-NoBoRu TAKASUGI On the Morphogenesis of the Cell Wall of Staphybcocci-PETER GIESBRECHT, JORG WECKE, A N D BERNHARD RElNlCKE Cyclic AMP and Cell Behavior in Cultured Cells-MARK c . W I L L I N G H A M Recent Advances in the Morphology, Histochemistry, and Biochemistry of Steroid-Synthesizing Cellular Sites in the Nonmammalian Vertebrate OVV-SARDUL S. GURAYA SUBJECT INDEX

Volume 45 Approaches to the Analysis of Fidelity of DNA Repair in Mammalian Cdk-MICHAEL W. LIEBERMAN The Variable Condition of Euchromatin and Heterochromatin-FRIEDRICH BACK Effects of 5-Bromodeoxyuridine on Tumorigenicity, Immunogenicity, Virus Production, Plasminogen Activator, and Melanogenesisof Mouse Melanoma Celk-sELMA SILAGI Mitosis in F u n g i - M ~ ~ v lS~. FULLER

Volume 46 Neurosecretion by Exocytosis-ToM CHRISTIAN NORMANN Genetic and Morphogenetic Factors in Hemoglobin Synthesisduring Higher Vertebrate Development: An Approach to Cell Differentiation Mechanisms-VICTOR NlGoN A N D JACQUELINE GODET Cytophysiology of Corpuscles of Stannius-V. G. KRISHNAMURTHY Ultrastructure of Human Bone Marrow Cell A N D F. Maturation-J. BRETON-GORIUS REYES Evolution and Function of Calcium-Binding Proteins-R. H. KRETSINGER SUBJECT INDEX

Volume 47 Responses of Mammary Cells to Hormones-M. R. BANEWEE Recent Advances in the Morphology, Histochemistry, and Biochemistry of Steroid-Synthesizing Cellular Sites in the Testes of Nonmammalian VertebrateS-sARDUL S.GURAYA Epithelial-Stromal Inkactions in Development Of the Urogenital TTaCt-ERALD R. CUNHA Chemical Nature and Systematization of Substances Regulating Animal Tissue GrowthVICTOR A. KONVSHEV Structure and Function of the Choroid Plexus and Other Sites of Cerebrospinal Fluid Formation- THOMAS H. MILHORAT The Control of Gene Expression in Somatic Cell Hybrids-H. P. BERNHARD PrC.CUrSOr Cells Of Mechanocytes-ALEXANDER J. FRIEDENSTEIN SUBJECT INDEX

390

CONTENTS OF PREVIOUS VOLUMES

Volume 48 Mechanisms of Chromatin Activation and ReA N D VAUGHAN preSSiOn-NORMAN MACLEAN A. HlLDER Origin and Ultrastructure of Cells in Vim-L. A N D PATRICIA D. WILSON M. FRANKS Electrophysiology of the Neurosecretory CellKINJIYAGI A N D S ~ I Z U KIWASAKI O Reparative Processes in Mammalian Wound Healing: The Role of Contractile PhenomA N D DENYSMONena- GIULIOGABBIANI

New Aspects of the Ultrastructure of Frog Rod Outer SegmentS-JURGEN ROSENKRANZ Mechanisms of Morphogenesis in Cell CultUres-J. M. VASlLIEY AND 1. M. GELFAND Cell Polyploidy: Its Relation to Tissue Growth and Functions-W. Y A . BRODSKY A N D I . V. URYVAEVA Action of Testosterone on the Differentiation and Secretory Activity of a Target Organ: The Submaxillary Gland of the MOUS~-MONIQUE CHR~TIEN SUBJECT INDEX

TANDON

Smooth Endoplasmic Reticulum in Rat Hepatocytes during Glycogen Deposition and DepleVolume 51 tion-ROBERT R. CARDELL, JR. Potential and Limitations of Enzyme Cytochem- Circulating Nucleic Acids in Higher Orgaistry: Studies of the lntracellularDigestive Apnisms-MAURICE STROUN, PHlLlPPE ANKER, paratus of Cells in Tissue Culture-M. HIINDA N D PETERB. GAHAN PIERRE MAURICE, GEN Recent Advances in the Morphology, HistoUptake of Foreign Genetic Material by Plant chemistry, and Biochemistry of the DevelopProtoplasts-E. C. COCKING s. GURAYA ing Mammalian OV~IY-SARDUL The Bursa of Fabricius and Immunoglobulin Morphological Modulations in Helical Muscles Synthesis-BRUCE GLICK (Aschelminthes and A n n e l i d a ) - G ~ u ~ i o SUBJECT INDEX LANZAVECCHIA Interrelations of the Proliferation and Differentiation Processes during Cardiac Myogenesis and Volume 49 Regeneration-PAvEL P. RUMYANTSEV Cyclic Nucleotides, Calcium, and Cell Divi- The Kurloff Cell-PETER A. REVELL Sion-LIoNEL I. REBHUN Circadian Rhythms in Unicellular Organisms: An Spontaneous and Induced Sister Chromatid ExEndeavor to Explain the Molecular Mechanchanges as Revealed by the BUdR-Labeling A N D MANism- HANS-GEORG SCHWEIGER Method-HATAo KATO FRED SCHWEIGER Structural, Electrophysiological, Biochemical, SUBJECT INDEX and Pharmacological Properties of Neuroblastoma-Glioma Cell Hybrids in Cell Culture-B. HAMPRECHT Volume 52 Cellular Dynamics in Invertebrate NeurosecreCytophysiology of Thyroid Parafollicular tory SyStemS-ALLAN BERLIND Cytophysiology of the Avian Adrenal MedullaD. CellS-ELADIO A. NUNEZA N D MICHAEL ASOKGHOSH GERSHON Chloride Cells and Chloride Epithelia of Aquatic Cytophysiology of the Amphibian Thyroid Gland Insects-H. KOMNICK through Larval Development and MetamorCytosomes (Yellow Pigment Granules) of MolPhOSiS-ELIANE REGARD luscs as Cell Organelles of Anoxic Energy ProThe Macrophage as a Secretory Cell-Roy C. duction-IMRE ZS-NAGY PAGE, PHILIP DAVlES. A N D A. c . ALLISON SUBJECT INDEX Biogenesis of the Photochemical ApparatusTIMOTHY TREFFRY Extrusive Organelles in Protists-KLAUS HAUSVolume 50 MANN

Cell Surface Enzymes: Effects on Mitotic Activkctins-JAY c. BROWN A N D RICHARDC. HUNT ity and Cell Adhesion-H. BRUCE BOSMANN SUBJECT INDEX

CONTENTS OF PREVIOUS VOLUMES Volume 53 Regular Arrays of Macromolecules on Bacterial Cell Walls: Structure, Chemistry, Assembly, and Function-UwE B. SLEYTR Cellular Adhesiveness and Extracellular SubStrat&FREDElUCK GRINNELL Chemosensory Responses of Swimming Algae AND D. C. and Protozoa-M. LEVANDOWSKY R. HAUSER Morphology, Biochemistry, and Genetics of Plastid Development in Euglena gracilis-V. NEON AND P. HEIZMANN Plant Embryological Investigations and Fluorescence Microscopy: An Assessment of Integration-R. N. KAPILAND S. C. TIWARI The Cytochemical Approach to Hormone Assay-J. CHAYEN SUBJECT INDEX Volume 54

391

The Isolated Mitotic Apparatus and Chromosome Motion-H. SAKAJ Contact Inhibition of Locomotion: A Reappraisal-JOAN E. M. HEAYSMAN Morphological Correlates of Electrical and Other Interactions through Low-Resistance Pathways between Neurons of the Vertebrate Central Nervous System-C. SOTELOA N D H. KORN Biological and Biochemical Effects of Phenylalanine Analogs-D. N. WHEATLEY Recent Advances in the Morphology, Histochemistry, Biochemistry, and Physiology of Interstitial Gland Cells of Mammalian GURAYA O V ~ S A R D SU. L Correlation of Morphometry and Stereology with Biochemical Analysis of Cell Fractions-R. P. BOLENDER Cytophysiology of the Adrenal Zona FasCiCUlabAASTONE G. NUSSDORFER. GIUSEPPINA MAZZOCCHI,AND VIRGILIOMENEGHELLI

SUBJECTINDEX Microtubule Assembly and Nucleation-MARC W. KIRSCHNER The Mammalian Sperm Surface: Studies with Volume 56 Specific Labeling Techniques-JAMES K. Synapses of CephalOpodS-COLLETTE DUCROS KOEHLER The Glutathione status of CellS-NECHAMA s. Scanning Electron Microscope Studies on the Development of the Nervous System in Vivo KOSOWERA N D EDWARD M. KOSOWER and in Virro-K. MELLER Cells and SeneSCenCe-ROBERT ROSEN Immunocytology of Pituitary Cells from Teleost Cytoplasmic Structure and Contractility in AND Fishes-E. FOLL~NIUS, J. DOERR-SCHOTT, Amoeboid Cells-D. LANSINGTAYLOR JOHNS . CONDEELIS AND M. P. DUBOIS Follicular Atresia in the Ovaries of Nonmamma- Methods of Measuring Intracellular CalCiUm-ANTHONY H. CASWELL lian Vertebrates-SRINlvAs K. SAIDAPUR Hypothalamic Neuroanatomy: Steroid Hormone Electron Microscope Autoradiography of Calcified Tissues-ROBERT M. FRANK Binding and Patterns of Axonal RojectionS-DONALD w.PFAFF AND LILY c . A. Some Aspects of Double-Stranded Hairpin Structures in Heterogeneous Nuclear RNACONRAD HIROTONAORA Ancient Locomotion: Rokaryotic Motility SyStemS-LELENG P. TO AND LYNNMAR- Microchemistry of Microdissected Hypothalamic Nuclear Areas-M. PALKOVITS GULIS SUBJECTINDEX An Enzyme Profile of the Nuclear EnvelopeI. B. ZBARSKY SUBJECT INDEX Volume 57 Volume 55 The Corpora Allata of Insects-hRRE CASSIER Kinetic Analysis of Cellular Populations by Chromatin Structure and Gene Transcription: Means of the Quantitative RadioNucleosomes Permit a New Synthesisautography-I .-C . BISCONTE THORUPEDERWN

392

CONTENTS OF PREVIOUS VOLUMES

Cellular Mechanisms of Insect Photoreception-F. G. GRIBAKIN Oocyte Maturation-YosHIo MASUIAND HUGH J. CLARKE The Chromaftin and Chromaffin-likeCells in the Autonomic Nervous SyStem-JACQUES TAXI The Synapses of the Nervous System-A. A. MANINA SUBJECTINDEX

Genetic Control of Meiosis-I.

N. GOLUBOV-

SKAYA

Hypothalamic Neurons in Cell Culture-A. TIXIER-VIDAL AND F. DE VITRY The Subfornical Organ-H. DIETERDELLMANN AND JOHNB. SIMPSON SUBJECT INDEX

Volume 59 The Control of Microtubule Assembly in C. RAFF V~VO-ELIZABETH Functional Aspects of Satellite DNA and Membrane-Coating Granules-A. F. HAYWARD HeterOChrOmatin-BERNARD JOHN AND Innervation of the Gastrointestinal TractGEORGE L. GABOR MIKLOS GIORGIO GABELLA Determination of Subcellular Elemental Concen- Effects of Irradiation on Germ Cells and Emtration through Ultrahigh Resolution Electron bryonic Development in Teleosts-Nosuo Microprobe Analysis-THOMAS E. HUTCHIN- EGAMIAND KEN-ICHIIJIRI SON Recent Advances in the Morphology, Cytochemistry, and Function of Balbiani's The Chromaffin Granule and Possible Mechanisms of Exocytosis-HARVEY B. POLLARD, Vitelline Body in Animal Oocytes-SARDuL J. PAZOLES,CARLE. CREUTZ, S. GURAYA CHRISTOPHER AND ORENZINDER Cultivationof Isolated Protoplastsand HybridizaThe Golgi Apparatus, the Plasma Membrane, and tion of Somatic Plant Cells-RAIsA G. AND Functional Integration-W. G. WHALEY BUTENKO MARIANNE DAUWALDER SUBJECT INDEX

Volume 58