Current Topics in Developmental Biology, Volume 46

Current Topics in Developmental Biology Volume 46

Series Editors Roger A. Pedersen

and

Reproductive Genetics Division Department of Obstetrics, Gynecology, and Reproductive Sciences University of California San Francisco, California 94143

Gerald P. Schatten Departments of Obstetrics-Gynecology and Cell and Developmental Biology Oregon Regional Primate Research Center Oregon Health Sciences University Beaverton, Oregon 97006-3499

Editorial Board Peter Gruss Max-Planck-Instituteof Biophysical Chemistry Gottingen, Germany

Philip lngham University of Sheffield, United Kingdom

Mary Lou King University of Miami, Florida

Story C. Landis National Institutes of Health/ National Institute of Neurological Disorders and Stroke Bethesda, Maryland

David R. McClay Duke University, Durham, North Carolina

Yos hi taka Nagahama National Institute for Basic Biology, Okazaki, Japan

Susan Strome Indiana University, Bloomington, Indiana

Virginia Walbot Stanford University, Palo Alto, California

Founding Editors A. A. Moscona Alberto Monroy

Current Topics in Developmental Biology Volume 46 Edited by

Roger A. Pedersen Reproductive Genetics Division Department of Obstetrics, Gynecology, and Reproductive Sciences University of California San Francisco, California

Gerald P. Schatten Departments of Obstetrics- Gynecology and Cell and Developmental Biology Oregon Regional Primate Research Center Oregon Health Sciences University Bea verton, Oregon

Academic Press San Diego London

Boston

New York

Sydney Tokyo Toronto

Cover photograph: Fluorescencemicrograph of male and female pronuclei in a mouse zygote. lkro condensed sperm nuclei and the second polar body are also visible. For more details see Chapter 5 “Sperm Nuclear Activation during Fertilization” by Shirley J. Wright.

This book is printed on acid-free paper. @ Copyright 0 1999 by ACADEMIC PRESS All Rights Reserved. No part of this publication may be reproduted or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-1999 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0070-2 153/99 $30.00 Explicit permission from Academic Press is not required to reproduce a maximum of two figures or tables from an Academic Press chapter in another scientific or research publication provided that the material has not been credited to another source and that full credit to the Academic Press chapter is given.

Academic Press A Harcourt Science and Technology Company 525 B Street, Suite 1900, SanDiego, California 92101-4495, USA http://www.apnet.com

Academic Press 24-28 Oval Road, London NWl 7DX, UK

http://www.hbuk.co.uk!ap/ International Standard Book Number: 0-12-153146-5 PRINTED IN THE UNITED STATES OF AMERICA 99 0 0 0 1 02 03 0 4 E B 9 8 7 6 5

4

3 2

1

Contents

Contributors Preface xi

ix

1 Maternal Cytoplasmic Factors for Generation of Unique Cleavage Patterns in Animal Embryos Hiroki Nishida, lunji Morokuma, and Takahito Nishikata

I. Introduction

2

11. General Tendencies of Cleavage Plane Positioning

III. Micromere Formation in Sea Urchin Embryos IV. Spiral Cleavage in Gastropod Embryos 8 V. Unequal Cleavage in Ascidian Embryos 11 VI. Unequal Cleavage in Annelid Embryos 17 VII. par Mutants in Nematode Embryos 20 V I E Concluding Remarks 31 References 32

3 5

2 Multiple Endo-l,4-f3-~-glucanase(Cellulase) Genes in Arabidopsis Elena del Campillo

I. Introduction

40

11. Cellulase Genes in General 40 111. How Does Cellulase Relate to Cellulose and the Plant Cell Wall? More Questions

42 Than Answers Cellulase Genes in Plants 43 Molecular Characterization of EGase Genes in Arabidopsis 45 Expression of Three Distinct EGase Genes in Arabidopsis Tissues 54 EGase and Cell Growth EGase Mutants in Arabidopsis 55 IX.Conclusions 57 References 58

IV. V. VI. VII. VIII.

52

V

vi

Contents

3 The Anterior Margin of the Mammalian Gastrula: Comparative and Phylogenetic Aspects of Its Role in Axis Formation and Head Induction Christoph Viebahn

I. Introduction 64 11. Morphology 65 76 111. Changes at the Anterior Margin during Development 81 IV. The Anterior Margin in Different Vertebrate Classes 87 V. Gene Expression Related to the Anterior Margin 89 VI. A View on Phylogenetic Implications VII. Conclusions and Outlook 93 References 95

4 The Other Side of the Embryo: An Appreciation of the Non-D Quadrants in leech Embryos David A. Weisblat, FranGoise Z. Huang, Deborah E. Isaksen, Nai-Jia L. Liu, and Paul Chang

I. Introduction and Overview of Leech Development 108 11. Macromere Behavior during Cleavage 111. Syncytial Yolk Cell Formation 113 IV. Regulation of Macromere Fusion 119 V. Epiboly 121 VI. Conclusions 126 130 References

105

5 Sperm Nuclear Activation during Fertilization Shirley J. Wright

I. 11. 111. IV. V. VI. VII.

Introduction 134 The Sperm Nucleus 134 142 Egg Stage at Time of Fertilization Transformation of the Sperm Nucleus into a Male Pronucleus 143 Asynchronous Behavior of the Paternal and Maternal Chromatin 161 Technological Advances to Combat Human Infertility 164 Conclusions and Future Directions 166 References 167

vi i

Contents

6 Fibroblast Growth Factor Signaling Regulates Growth and Morphogenesis at Multiple Steps during Brain Development Flora M . Vaccarino, Michael L. Schwartz, Rossana Raballo, Julianne Rhee, and Richard Lyn-Cook

I. Overview 180 11. Developmental Control Genes and Morphogenesis within the CNS 111. The Fibroblast Growth Factor Family 183 IV.FGFs Regulate Patterning of the Neuroepithelium 185 V. Effect of FGFs on Cell Fate and Cortical Lineages 192 VI. Conclusions and Future Prospects 194 References 195

Index 201 Contents of Previous Volumes

209

18 I

This Page Intentionally Left Blank

Contributors

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

Paul Chang (lOS), Department of Molecular and Cell Biology, University of California, Berkeley, California 94720 Elena del Campillo (39), Department of Cell Biology and Molecular Genetics, University of Maryland at College Park, College Park, Maryland 20742 Franqoise Z. Huang (103, Department of Molecular and Cell Biology, University of California, Berkeley, California 94720 Deborah E. Isaksen (105), Department of Molecular and Cell Biology, University of California, Berkeley, California 94720 Nai-Jia L. Liu (105), Department of Molecular and Cell Biology, University of California, Berkeley, California 94720 Richard Lyn-Cook (179), Child Study Center, Yale University School of Medicine, New Haven, Connecticut 06520 Junji Morokuma (l),Department of Life Science, Tokyo Institute of Technology, Nagatsuta, Midori-Ku, Yokohama 226-8501, Japan Hiroki Nishida (I), Department of Life Science, Tokyo Institute of Technology, Nagatsuta, Midori-Ku, Yokohama 226-8501, Japan Takahito Nishikata (l), Department of Biology, Faculty of Science, Konan University, Kobe 658-8501, Japan Rossana Raballo (179), Child Study Center, Yale University School of Medicine, New Haven, Connecticut 06520 Julianne Rhee (179), Child Study Center, Yale University School of Medicine, New Haven, Connecticut 06520 Michael L. Schwartz (179), Section of Neurobiology, Yale University School of Medicine, New Haven, Connecticut 06520

ix

X

Contributors

Flora M. Vaccarino (179), Child Study Center and Section of Neurobiology, Yale University School of Medicine, New Haven, Connecticut 06520 Christoph Viebahn (63), Institute of Anatomy, Rheinische Friedrich-WillhelmsUniversitat, 53 115 Bonn, Germany David A. Weisblat (103, Department of Molecular and Cell Biology, University of California, Berkeley, California 94720 Shirley J. Wright (133), Department of Biology, University of Dayton, Dayton, Ohio 45469

Preface

This volume of Current Topics in Developmental Biology will be important to all developmental biologists because it considers the remarkable breadth and depth of significant discoveries in a variety of experimental systems. One of the most significant discoveries has been the molecular basis of axis determination and the breaking of asymmetry during early development. In Chapter 1, Hiroki Nishida, Junji Morokuma, and Takahito Nishikata discuss maternal cytoplasmic factors for the generation of unique cleavage patterns in animal embryos, and they address the classic question of how a differentiated and polarized offspring results from the seemingly homogeneous symmetrical egg. Gastrulation in mammals, which seemed an intractable problem years ago, is advanced by Christoph Viebahn in Chapter 3 on the anterior margin of the mammalian gastrula and its role in axis formation and head induction. David A. Weisblat, FranCoise Z. Huang, Deborah E. Isaksen, Nai-Jia L. Liu, and Paul Chang consider this remarkable progress in their chapter on the appreciation of non-D quadrants in leech embryos. Plant development biologists will be particularly interested in Chapter 2, by Elena del Campillo, on multiple endo- 1,4-P-~-glucanasegenes in Arubidopsis, because it adds to the extraordinary understanding of the development of this model. Shirley J. Wright, in her paper on sperm nuclear activation during fertilization, details our knowledge on the manner in which the sperm nucleus in transformed into a decondensed active genomic partner. Perhaps it is fitting that Volume 46, which begins with a consideration of the egg, ends with an exciting article on brain development. In Chapter 6, Flora M. Vaccarino, Michael L. Schwartz, Rossana Raballo, Julianne Rhee, and Richard Lyn-Cook consider how fibroblast growth factor signaling regulates growth and morphogenesis at multiple steps during brain development. Together with the other volumes in this series, this volume provides a comprehensive survey of major issues at the forefront of modem molecular mechanisms of developmental biology. These chapters should be valuable to researchers in the fields of plant and animal development, as well as to students and other professionals who want an introduction to current topics in neurobiology; cellular, molecular, and genetic approaches to developmental biology; and plant biology. This volume in particular will be essential reading for anyone interested in plant de-

xi

xii

Preface

velopment and plant biology, morphogenesis and embryo formation, gene regulation of development, development in invertebrates, growth factors, signal transduction, and the molecular basis of mammalian embryogenesis. This volume has benefited from the ongoing cooperation of a team of participants who are jointly responsible for the content and quality of its material. The authors deserve full credit for their success in covering their subjects in depth, yet with clarity, and for challenging the reader to think about these topics in new ways. We thank the members of the Editorial Board for their suggestions of topics and authors, and Liana Hartanto and Michelle Emme for their exemplary administrative and editorial support. We are grateful for the unwavering support of Craig Panner and Hilary Rowe in the editorial office at Academic Press in San Diego. We are also grateful to the scientists who prepared chapters for this volume and to their funding agencies for supporting their research. Gerald P. Schatten Roger A. Pedersen

Maternal Cytoplasmic Factors for Generation of Unique Cleavage Patterns in Animal Embryos

'

Hiroki Nishida, *, Junji Morokuma, * and Takahito Nishikataj *Department of Life Science Tokyo Institute of Technology Nagatsuta, Midori-Ku,Yokohama 226-850 1, Japan +Departmentof Biology Faculty of Science Konan University Kobe 658-8501, Japan

I. Introduction A. Role of an Invariant Cleavage Pattern B. Intrinsic and Extrinsic Cues 11. General Tendencies of Cleavage Plane Positioning A. Furrow Formation, Mitotic Apparatus, and Centrosome B. Orthogonal Pattern C. Cell Divisions in Single-Layered Epithelium 111. Micromere Formation in Sea Urchin Embryos A. Role of Maternal Factors in Unequal Cleavage B. Involvement of the Vegetal Cortex in Micromere Formation C. Maternally Localized Substance and the First Cleavage IV. Spiral Cleavage in Gastropod Embryos A. Spiral Cleavage B. Maternal Factor Determining Cleavage Handedness C. Mechanisms of Spiral Cleavage D. Cleavage Clock V. Unequal Cleavage in Ascidian Embryos A. Centrosome- Attracting Body B. Formation and Ultrastructure of the Centrosome-Attracting Body C. Role of Posterior Egg Cytoplasm VI. Unequal Cleavage in Annelid Embryos A. Unequal Division at the First Cleavage B. Unequal Divisions at the Second Cleavage and Micromere Formation VII. par Mutants in Nematode Embryos 'Author to whom correspondence should be addressed. Current Topics in Developmental Biology, Vol. 46 Copyright B 1999 by Academic Press. All rights of reproduction in any fonn reserved 0070-2153/99 $30.00

1

2

Hiroki Nishida et al.

A. Asymmetric Cleavage in Cuenorhubdiris elegans B. The par Genes VIII. Concluding Remarks References

1. Introduction A. Role of an Invariant Cleavage Pattern

Stereotyped cleavage patterns are observed in embryos of various kinds of animals. The orientation of cleavages determines the overall organization of the embryo. Invariant cleavage patterns also play important roles in embryonic cell fate determination in two ways. First, they ensure the proper partitioning of localized developmental determinants to specific daughter cells (Freeman, 1979). In the nematode Caenorhabditis elegans, specific components of the ooplasm, P granules, are segregated to germ-line precursor cells through invariant cleavages (Strome and Wood, 1982; Hird et al., 1996). In ascidian embryos, various kinds of maternal determinants, i.e., muscle, endoderm, and epidermis determinants, and determinants for gastrulation movements and axis specification, are localized in the egg cytoplasm (Satoh, 1994; Nishida, 1997). These factors are partitioned into specific blastomeres during the cleavage stage and determine their developmental fate. During neurogenesis in the Drosophila embryo, the Numb and Prosper0 proteins play crucial roles in cell fate determination. These proteins are localized in a neuroblast, and asymmetric cell division partitions these proteins into one of the daughter cells, a ganglion mother cell (Doe, 1996; Jan and Jan, 1998).Thus, specific positioning of the division plane can contribute to the precise partitioning of cytoplasmic determinants. It has been revealed that there is coupling of determinant localization and positioning of the cleavage plane. These two processes may occur independently of each other. Of course, certain treatments can affect one of the two processes, but studies have indicated the presence of an underlying mechanism or machinery that controls both localization and cell division. This mechanism couples the two processes to ensure precise partitioning of determinants into daughter cells, because the cell division plane must be perpendicular to the axis of localization to segregate determinants into one of the two daughter cells, resulting in asymmetric cell division. In par mutants of C. elegans, localization of P granules and various cell fate determinants, as well as the cleavage pattern, become abnormal. The par mutants will be described in detail in Section VII of this review. Similarly, in inscuteable mutants of Drosophila, localization of the Numb and Prosper0 proteins and positioning of the division plane are improperly executed (Kraut et al., 1996; Doe, 1996; Jan and Jan, 1998). In ascidians, ooplasmic seg-

1. Maternal Factors and Embryonic Cleavage Patterns

3

regation occurs before cleavage starts. The movement of the egg cytoplasm brings muscle determinants and factors that control cleavage patterns together at the posterior pole (Nishida, 1992, 1994; Nishikata et al., 1999) (see Section V), thus ensuring the invariant and precise partitioning of muscle determinants into posterior muscle lineage cells. The second role of invariant cleavage patterns in embryonic cell fate specification is to ensure proper spatial arrangements of interacting cells. For example, the notochord of an ascidian larva is induced at the 32-cell stage through inductive cell interaction with the neighboring cells (Nakatani and Nishida, 1994). In C. elegans, a number of cell interactions occur during the cleavage stage (Schnabel, 1991; Schnabel and Priess, 1997). Precise positioning of interacting cells is essential for these cell interaction processes.

B. Intrinsic and Extrinsic Cues

The position of the cleavage plane is specified by intrinsic cues, and also by extrinsic ones such as cell interactions. In this review, we focus mainly on intrinsic cues, especially maternal cytoplasmic factors in early embryos. Current advances related to this issue will be reviewed. The results of various early studies have been well documented and reviewed by Freeman (1979,1983). The position specification of the cleavage plane by cell interactions has been analyzed in a few species. Blastomere isolation and recombination experiments have revealed that, in C. elegans, orientation of cleavage of the EMS blastomere of four-cell embryos depends on the attachment site of the neighboring P, blastomere (Goldstein, 1995). Similarly, formation of an asymmetric spindle in the CD blastomere at the two-cell stage of the oligochaete Tubifex is dependent on contact with the AB blastomere (Takahashi and Shimizu, 1997) (see Section VI). Meshcheryakov [1976, 1978a,b; reviewed in Freeman (1983)] proposed a role of blastomere shape in specifying spindle orientation in the early cleavages of gastropod embryos (see Section IV). In this case, a change in cell shape through contact with neighboring cells may provide a mechanical force for orienting the spindle. Another case, where the spindle is oriented parallel to the embryo surface, will be discussed in the next section.

11. General Tendencies of Cleavage Plane Positioning A. Furrow Formation, Mitotic Apparatus, and Centrosome

The placement of the cleavage plane is an essential element in determining cleavage patterns (Freeman, 1983). The mechanism by which the cell division plane is positioned is one of the most intriguing problems in cell and developmental

4

Hiroki Nishida et al.

biology. The plane of cell division is determined by the orientation and position of the mitotic apparatus. Cytokinesis occurs perpendicular to the axis of the mitotic spindle at the position of the metaphase plane, and the asters of the spindle dictate where the cleavage furrow will form (Rappaport, 1986; Strome, 1993; White and Strome, 1996).Thus, it is important to investigate how the position of the mitotic apparatus is specified in order to understand the mechanism of division plane positioning. A mitotic spindle forms between the two poles, nucleated by centrosomes or microtubule organizing centers. Therefore, the pattern of migration of the centrosome eventually directs the cleavage plane, as will be seen in most cases reviewed in this article. B. Orthogonal Pattern

Generally, there is a tendency for the cleavage furrow to form orthogonally to the previous cleavage plane (Hertwig, 1885). Figure 1 shows such an example in an ascidian embryo. In embryos that show radial cleavage patterns, and even in spiralian embryos, this rule is applicable to most of the cleavages (see Section

31

\ 38

I

63

33

a9.

54

b9.

Fig. 1 Diagrammatic representation of the arrangement of cells in the animal hemisphere and the orientation of their cell division in the ascidian embryo. The right half of the figure shows the cell numbers of the ninth generation after the eighth cleavage. The left half shows the orientation of three successive cleavages. The bidirectional arrows indicate the orientation of the seventh cleavage. The wavy lines indicate the plane of the eighth division. The dashed lines indicate the plane of the ninth division. From Nishida (1986).

5 IV). A succession of orthogonally oriented cleavage planes can be explained by the movement of centrosomes. After duplication of the centrosome, the two daughter centrosomes migrate 90" away from their original position at the previous cleavage and away from one another to opposite sides of the nucleus, where they serve as the poles of the mitotic apparatus. Then, in the next cell cycle the same sequence of events takes place (Strome, 1993; White and Strome, 1996). Various deviations from this orthogonal pattern exist. Most investigations on the mechanisms of cleavage plane positioning have dealt with cases that do not conform to this pattern, as will be discussed later. 1. Maternal Factors and Embryonic Cleavage Patterns

C. Cell Divisions in Single-Layered Epithelium

Another type of tendency, although not a general one, is observed in single-layered epithelium in late-cleavage-stage and blastula embryos of a wide variety of animals. The division plane forms perpendicular to the embryo surface, so that both daughter cells remain facing the surface, and maintain their position in the single-layered epithelium. This has been investigated in the cleavages of shrimp embryos (Wang et al., 1997). When blastomeres are isolated from the embryos, the orientation of the spindle changes so that it becomes parallel to the surface of the partial embryo. In this pattern, the mechanism that orients the spindle is not yet known, although the positional relationship with neighboring cells is apparently important.

111. Micromere Formation in Sea Urchin Embryos A. Role of Maternal Factors in Unequal Cleavage

In many echinoderm species, including starfish and sea cucumber, the cleavage pattern is radial and equal. Although the cleavage pattern of sea urchins is equal in most blastomeres, the fourth and fifth cleavages in the vegetalmost blastomeres produce cells of distinct size, i.e., micromere and macromere, and small and large micromeres, respectively. These two rounds of unequal cleavage always produce smaller blastomeres at the vegetal pole. The cellular mechanism of this unequal cleavage has been studied mainly during the fourth cleavage of the vegetal blastomeres, which produces micromeres at the vegetal pole. The cell lineage assignment of the micromeres is skeletogenic cells (large micromere) and coelomic sac constituents (small micromere) (Okazaki, 1975; Cameron and Davidson, 1991). In the sea urchin egg, the site of micromere formation is specified prior to fertilization (Horstadius, 1937,1939). When unfertilized eggs of Arbacia and Paracentrotus are cut into animal and vegetal halves and then inseminated, the vegetal halves form micromeres, gastrulate, generate a skeleton, and develop into

6

Hiroki Nishida e t al.

pluteus larvae. By contrast, the animal halves cleave equally and do not gastrulate. Vegetal quartets of eight-cell embryos also form micromeres and give rise to relatively normal-looking pluteus larvae, whereas animal quartets do not. These results suggest that the maternal factors responsible for micromere formation and cell fate specification are localized in the vegetal half of the unfertilized egg and partitioned into the vegetal half of the early embryo. B. Involvement of the Vegetal Cortex in Micromere Formation

Prior to the fourth cleavage that generates micromeres, interphase nuclei of vegetal blastomeres of the eight-cell embryo shift toward the vegetal cortex (Dan, 1979). This nuclear migration is dependent on intact microtubules but not actin filaments (Lutz and InouC, 1982). After nuclear migration, the nuclear envelope breaks down and an asymmetric mitotic apparatus is formed (Dan and Nakajima, 1956). One spindle pole is very close to the vegetal cortex, and its centrosome is located only 3-4 pm from the cell surface (Fig. 2A). This spindle pole seems firmly attached to the cortex, because when the mitotic apparatus is isolated from the vegetal blastomere, a piece of plasma membrane of the vegetal pole is usually isolated together with the aster (Fig. 2B) (Holy and Schatten, 1991).As a result of this attachment, the aster is flattened and truncated, instead of being the typical radiate shape. This asymmetric mitotic apparatus is very

Fig. 2 Asymmetrically positioned mitotic apparatus during the fourth cleavage of the sea urchin embryo. (A) Extracted Strongylocentrotus purpuratus blastomeres at late metaphaselearly anaphase. The symmetric mitotic apparatus of the animal blastomere (upper cell) is centrally located, whereas the asymmetric mitotic apparatus of the vegetal blastomere (lower cell) is apposed to the vegetal pole (arrowhead). Bar, 25 pm. (B) Metaphase spindle isolated from the vegetal blastomere. The micromere aster (mi) is flattened whereas the macromere aster (ma) is radiate. Bar, 10 pm. (C) Transmission electron microscopy of two vegetal blastomeres of the eight-cell stage of the Hemicentmtus pulcherrimus embryo. The eccentrically shaped asters of the asymmetric mitotic apparatus are anchored in the vesicle-free area (g) at the vegetal pole. Bar, 10 pm. A and B from Holy and Schatten (1991). with the permission of the Company of Biologists, Limited; C from Dan er al. (1983).

1. Maternal Factors and Embryonic Cleavage Patterns

7

similar to that of the first cleavage of the mollusc, Spisula solidissirnu (Dan and Ito, 1984). Migration and anchorage of one pole of the mitotic apparatus to the cortex are also observed as common phenomena in polar body formation (e.g., Lutz et al., 1988). The vegetal cortex that the nucleus approaches has a feature different from other parts of the cortex (Dan et al., 1983). Observation of blastomeres of Hemicentrotus using transmission electron microscope has revealed that the plasma membranes facing the surface of the embryo are lined with a row of vesicles (Uemura and Endo, 1976; Tanaka, 1979). The asymmetric mitotic apparatus in the micromere-forming blastomere is attached to a unique area of vegetal cortex where vesicles are absent, the so-called vesicle-free area (Fig. 2C) (Dan et aZ., 1983). This cortical difference is first recognizable at the four-cell stage. Morgan (1894) also found that the red pigment granules in the egg of Arbacia at the fourcell stage migrate away from the vegetal pole; consequently, the micromere is less pigmented. When the cortical specification is disturbed by treatment with the detergent sodium dodecyl sulfate (SDS), during the four-cell stage, the unequal division at the fourth cleavage is changed to an equal division (Tanaka, 1976). Under this condition, the mitotic apparatus is positioned symmetrically and oriented orthogonally to that in the previous cleavage. These findings suggest that unknown maternal factors localized to the vegetal half of the unfertilized egg promote cortical changes at the vegetal pole during the early cleavage stages, which might attract the nucleus to the vegetal pole.

C. Maternally localized Substance and the First Cleavage

Vlahou et al. (1996) obtained intriguing results suggesting that maternal mRNA of the steroid receptor SpCOUP-TF is localized in oocytes, eggs, and embryos of the sea urchin, Strongylocentrotus purpuratus. SpCOUP-TF nucleic acid encodes a sea urchin homolog of vertebrate COUP-TFs and the Drosophila seven up subfamily of transcription factors, which are members of the orphan steroid hormone receptors. The maternal mRNA of SpCOUP-TF is localized asymmetrically along an axis orthogonal to the animal-vegetal axis in the oocyte, unfertilized egg, and embryo. During the cleavage stage, these transcripts are localized with their highest concentration at a 90" angle to the first cleavage plane. Therefore, the first cleavage plane bisects the egg into two halves, one with SpCOUP-TF mRNA and the other without. The mRNA is partitioned into only one of the two blastomeres in most cases. Using SpCOUP-TF as a probe, COUPTF maternal mRNA was proved to be localized also in Lytechinus pictus and Lytechhinus variegutus eggs and embryos. In Lytechinus, SpCOUP-TF transcripts exhibit localization at 45" relative to the first cleavage plane in most cases. Although there are species differences in the relationship between mRNA localization and the first cleavage plane, the relationship is almost invariable with-

8

Hiroki Nishida et al.

in a species. Thus, the position of the first cleavage plane is predictable even in the unfertilized egg. It is not known whether COUP-TF messages are involved in the positioning of the cleavage plane, but localization of the mRNA is likely to indicate the presence of some kind of egg organization. Therefore, the orientation of the first cleavage plane may depend on the cytoplasmic organization of the unfertilized egg. On the other hand, Schatten (1981) has reported a close relationship between the pathway taken by the male pronucleus and the sperm aster, and the orientation of the spindle for the first cleavage. The first cleavage spindle forms perpendicular to this path. Further analysis is required to clarify the relationship between egg organization, the pronucleus pathway, and the first cleavage plane.

IV. Spiral Cleavage in Gastropod Embryos A. Spiral Cleavage

The spiralian group includes the class Gastropoda, the subphylum Diasoma (e.g., Mytilus, Dentaliurn), the class Turbellaria (e.g., Dugesia), the phylum Nemertea (e.g., Cerebratulus), and the phylum Annelida (cf. Section VI). Many wellknown examples of spiral cleavage are found in gastropods (reviewed by Collier, 1997).In typical cases of spiral cleavage, the first and second cleavages run almost along the animal-vegetal axis, dividing the egg into four quadrants (A, B, C, and D blastomeres) (Fig. 3). At the third cleavage, a quartet of micromeres

C

D

C

a!

Fig. 3 Diagram of dextral (A-D) and sinistral (a-d) spiral cleavage patterns; animal views. (A, a) The 2-cell stage. The embryos represent the rotation of the nuclei (arrows),which corresponds with the handedness of the spiral cleavage. (B,b) The 4-cell stage; (C, c) the 8-cell stage; (D, d) the 16-cell stage. The direction of the cleavage (arrows) changes at the right angle in an alternate manner at the 8and 16-cell stages. From Wilson (1925).

1. Maternal Factors and Embryonic Cleavage Patterns

9 is formed at the animal pole, but this is displaced to the right or to the left of its sister macromere. Similarly, during successive cleavages, blastomeres are given off along the animal-vegetal axis in quartets. These quartets are always displaced obliquely toward one side or the other of the basal sister blastomeres in an alternate manner at each cleavage in a regular order (Fig. 3). The first quartet (micromeres) at the eight-cell stage is typically rotated in a clockwise direction as observed looking down on the animal pole, and the second quartet is rotated counterclockwise. This cleavage pattern reflects the fact that the mitotic apparatus is oriented obliquely with reference to the animal-vegetal axis and takes an orthogonal orientation to the previous one at each cleavage. Of course, there are several modifications of this basic cleavage pattern among various species.

B. Maternal Factor Determining Cleavage Handedness

The typical spiral cleavage pattern is dextral (clockwise at the third cleavage), although some species, Physa, for example, have a sinistral (counterclockwise) spiral cleavage pattern. Moreover, in Lymnaea, there are both dextral and sinistral individuals. This handedness of the cleavage has a strict correlation with the handedness of the spiral of the adult shell. The shell shape emerges through the left or right location of the D macromere, the orientation of the mesodermal bands, and twisting of the visceral mass during development. Therefore, in this case, the handedness of the cleavage pattern is important for determining the entire body pattern of the adult. In Lymnaea peregra, the handedness of both the cleavage and the shell is determined by the maternal inheritance of a single locus, dextrality being dominant to sinistrality (Sturtevant, 1923; Freeman and Lundelius, 1982). This pattern of inheritance suggests that the products of this locus are produced during oogenesis. Cytoplasmic transfer experiments have shown that injection of cytoplasm from dextral eggs into sinistral eggs changes the cleavage pattern of sinistral eggs to dextral. By contrast, injection of cytoplasm from sinistral eggs into dextral eggs has no effect (Freeman and Lundelius, 1982). These results coincide well with the fact that dextrality is dominant in genetics. This strongly suggests that the maternal dextral gene produces a cytoplasmic factor that influences the cleavage pattern. Cleavage occurs in a dextral manner when this dextral gene product is present, but in a sinistral manner in its absence. The cytoplasm from dextral eggs can change the sinistral cleavage pattern only when cytoplasmic transfer is done before formation of the second polar body. This result suggests that some events determining the handedness of the cleavage pattern occur around the time of second polar body formation and before the beginning of cleavage. This event may change the cortex or the cytoplasmic organization of the egg and specify the future orientation of the spindle. The molecular nature of this dextral gene product has yet to be determined.

10

Hiroki Nishida et al.

C. Mechanisms of Spiral Cleavage

The handedness of micromere formation in the gastropod embryo is also related to earlier embryonic events. Conklin (1897) pointed out that rotation of the nuclei, asters, and protoplasmic areas occurs after the first cleavage in the Crepidula egg (Fig. 3A). The direction of the rotation is dextral, in accordance with the handedness of the spiral cleavage. Meshcheryakov and Beloussov (1975) observed movement of the surface of blastomeres marked with carbon particles during cytokinesis of the first to fourth cleavages in Lymnaea (dextral cleavage species) and Physa (sinistral cleavage species) embryos. They designated this movement “asymmetrical rotation” and revealed the close relationship between the direction of surface movement and the eventual handedness of the spiral cleavage pattern. These early events are likely under the maternal control discussed above. Meshcheryakov (1976, 1978a,b; reviewed in Freeman, 1983) pointed out the role of extrinsic factors in generating the spiral cleavage pattern. These were intercellular contact between blastomeres, and blastomere shape at the time of spindle formation, which is affected by the packing arrangement of the blastomeres. At the two-cell stage, the Lymnaea embryo, by controlling the Ca2+ concentration of the culture solution, the size of the contact zone between the two blastomeres, and consequent blastomere shape, can be changed (Meshcheryakov, 1978a). The angle between the spindle and the contact zone is closely related to the size of the contact zone and blastomere shape. The greater the size of the contact zone between the two blastomeres, the more the spindles become oriented parallel to the contact zone. In other experiment, the blastomeres are isolated prior to mitotic apparatus formation at the four-cell stage. The isolated blastomere cleaves equally and meridionally instead of forming micromeres (Meshcheryakov, 1976). When the blastomere is similarly isolated at metaphase, the cleavage is not affected by the isolation. When eight-cell-stage embryos are treated with trypsin in order to disturb the cell contacts and change the blastomere shape, the mitotic apparatus of the macromere forms parallel to whichever edge where the macromere has broader contact with one of the micromeres by chance (Meshcheryakov, 1978b). These observations highlight the role of the blastomere packing arrangement and shape in the determination of spindle orientation. Thus, both maternal intrinsic factors and extrinsic factors are involved in generation of the spiral cleavage pattern. Maternal factors determine the handedness of the spiral at the beginning of the process, and cell contact and blastomere shape influence the execution of spiral cleavages during the cleavage stage. This is contrary to the argument that the orthogonal pattern observed in spiral cleavages can be simply explained by centrosome movements around the nucleus after their duplication.

1. Maternal Factors and Embryonic Cleavage Patterns

11

D. Cleavage Clock

One of the mechanisms implicated in the positioning of the cleavage plane is referred to as the “cleavage clock.” There have been several studies of the cleavage clock in sea urchins (reviewed in Horstadius, 1973) and ctenophores (Freeman, 1976). In this article, we summarize some experiments that have been performed with gastropod embryos. Cleavages of some gastropod embryos can be reversibly inhibited by treatment with an appropriate concentration of cytochalasin B, which disturbs actin cytoskeletal filaments, and the cleavage resumes after the inhibitor has been washed out. Even when the second cleavage is inhibited reversibly in Crepidula (ConMin, 1912) and Zlyanassa (reviewed in Freeman, 1983), the micromeres are produced in a clockwise direction with the same timing as that in the control embryo. This results in a four-celled embryo with two macromeres and two micromeres. When the third cleavage of Zlyanassa is skipped, the next division occurs in a counterclockwise manner. This is the same cleavage direction shown by untreated control embryos when they cleave to form the second quartet. These results imply that spindle orientation and the timing of micromere formation are not dependent on the previous cleavage history, and that they are determined by a “clock” that counts the timing for developmental events independently of cleavage. The cleavage clock is responsible for the timing and the spindle orientation for micromere formation in molluscan species. A similar clock mechanism has also been suggested for micromere formation in the sea urchin embryo (Dan and Ikeda, 1971; Horstadius, 1973). Many cell cycle-controlling genes, some of which directly or indirectly control the organization of the cytoskeletal filaments, have been identified. It would be worthwhile studying the possibility that some of the products of these genes are involved in the cleavage clock.

V. Unequal Cleavage in Ascidian Embryos A. Centrosome-Attracting Body

The cleavage pattern of ascidian embryos is unique and invariant (Conklin, 1905; Satoh, 1979; Nishida, 1986) and progresses in a bilaterally symmetric manner. The cleavage pattern in the animal hemisphere (Fig. 1) and that in the anteriorvegetal region are simple and almost radially symmetrical. In contrast, the posterior-vegetal region cleaves in a complicated manner, due mostly to three rounds of unequal cleavage that occur at the posterior pole, as described below. The first three cleavages produce blastomeres that are almost equal in size. After the eightcell stage, three successive unequal cleavages occur in the cells of the posteriormost blastomere pair at each stage, always producing smaller cells posteriorly.

12

Hiroki Nishida et al.

All the other cleavages are equal in terms of blastomere size. The smallest blastomere pair at the posterior pole of the 64-cell embryo will cease dividing and give rise to two cells in the endodermal strand of the tadpole larva (Nishida, 1987).This cleavage pattern suggests that the posterior pole may attract the spindle every time, with resulting production of smaller cells at this pole. Hibino et al. (1998) investigated the process of unequal cleavage in embryos of the ascidian Halocynthia roretzi. Figure 4A shows an embryo immunostained with antitubulin antibody at the 16-cell stage. Posterior is to the right. Larger (B5.1) and smaller (B5.2) vegetal blastomeres generated by the previous unequal cleavage can be seen. The posterior (B5.2) blastomeres will cleave unequally again in the next cleavage. In all blastomeres except for the B5.2 blastomeres, astral microtubules are organized symmetrically on both sides of the nuclei. In the B5.2 cells, microtubules from the posterior centrosome formed an unusual bundle and were focused on the cortex of the posterior pole, whereas those from the anterior centrosome showed an ordinary radial distribution. By extracting ascidian eggs with buffer containing Triton X-100, a unique structure, designated the centrosome-attracting body (CAB), was found in the posterior cortex of the posteriormost blastomere pair (Fig. 4B, arrowheads). The CAB is observable as a structure with high refraction under differential interference contrast (Nomarski) optics. The CAB exists only in blastomeres that cleave unequally. The position of the CAB coincides well with the focal point of the microtubule bundle (Fig. 4A, arrows). Unequal cleavage in the ascidian is preceded by migration of the nucleus, led by one centrosome in the posterior direction (Fig. 5). After the fourth cleavage, the nuclear envelope appears close to the previous cleavage plane (Fig. 5A, B). A single centrosome is present in the central region, from which astral microtubules have radially emerged. The nucleus then migrates to the center to meet with the centrosome, and the centrosome duplicates (Fig. 5C, D). Microtubule arrays extending from the posterior centrosome gather around the posterior cortex, and the posterior ends then focus on the CAB. A thick microtubule bundle is formed between the centrosome and the CAB, and seems to connect the two structures. Then, in accordance with shortening of the microtubule bundle, the interphase nucleus with the centrosome shifts posteriorly and approaches the CAB (Fig. 5E, F). Consequently, an asymmetrically located mitotic apparatus is formed (Fig. 5G, H), one pole remaining anchored to the CAB. Then, unequal division takes place, producing a smaller daughter cell that inherits the CAB. Similar events are observed in all unequally cleaving blastomeres from the 8- to the 64-cell stage, but never in the other equally cleaving blastomeres during the cleavage stages. These observations suggest that the CAB plays an important role in producing unequal cleavages. Translocation of the nucleus involves shortening of the microtubule bundle. The microtubule bundle between the centrosome and the CAB appears to be essential for nuclear migration, because the microtubule inhibitor, nocodazole, suppresses

1. Maternal Factors and Embryonic Cleavage Patterns

13

Fig. 4 Microtubule array and the centrosome-attracting body in the 16-cell-stage ascidian embryo. (A) An embryo stained with antitubulin antibody: vegetal view. Anterior is to the left and posterior is to the right. A bundle of microtubules is present between one of the centrosomes and the position of the centrosome-attracting body (arrows) in the posterionnost blastomere (B5.2). (B) An embryo extracted with extraction buffer and observed using Nomarski optics. Apair of centrosome-attracting bodies is present at the posterior pole of the embryo (arrowheads); interphase. Nuclei are visible. Bar, 100 pm. (C) After removal of the posterior-vegetal cytoplasm from the eggs, the cleavage pattern becomes radialized. Unequal cleavage was not observed and the centrosome-attracting body disappeared. (D) After transplantation of the posterior-vegetal cytoplasm to the anterior-vegetal position of another egg, unequal cleavages are observed on both the anterior and posterior sides. Centrosome-attracting bodies (arrowheads) have formed ectopically in the anterior region of the embryo. A and B from Hibino et al. (1998);C and D from Nishkata et al. (1999).

nuclear movement toward the CAB (Nishikata et al., 1999). By contrast, disruptions of microfilaments do not affect the movement. The CAB can be stained immunohistochemically with a monoclonal antibody against bovine brain kinesin, suggestingthat a microtubule motor protein, a kinesin or kinesinlike molecule, may be associated with the CAB and involved in attraction of the centrosome.

B. Formation and Ultrastructure of the Centrosome-AttractingBody

The CAB is first recognizable as precursors, which appear as dozens of small dots in the posterior cortex of the 2-cell-stage embryo. These particles gradual-

14

Hiroki Nishida et al.

Fig. 5 The CAB and microtubule array in the B5.2 blastomeres of the 16-cell stage embryo. Posterior is to the right. Embryos were extracted at 10 min (A, B), 30 min (C, D), 40 min (E, F), and 65 min (G, H) after the appearance of the previous fourth cleavage furrow. Next cleavage started at 75 min. (A, C, E, and G ) Extracted embryos were observed using Nomarski optics. The CABs, which are situated at the posterior pole, are indicated by arrowheads. Arrows indicate nuclei and the spindle. (B, D, F, and H) Embryos were stained with antitubulin antibody. CABs were also faintly observable in the tubulin-stained embryos. Bar, 50 p n . From Hibino et al. (1998).

ly assemble and form a slender cluster by the late 4-cell stage. During the 8-cell stage, the particles fuse together to form the CAB, which has a uniform appearance. The CAB is present continuously throughout the 8- to 64-cell stages, and throughout each cell cycle, and is observable until the gastrula stage (Hibino et al., 1998).

1. Maternal Factors and Embryonic Cleavage Patterns

15

Drug treatments have indicated that microtubule organization is unnecessary for formation of the CAB after completion of ooplasmic segregation that occurs prior to the first cleavage (Nishikata et al., 1999). Accumulation of the kinesin epitope in the CAB is also resistant to microtubule inhibitor. By contrast, maintenance of CAB integrity requires microfilaments. Filamentous actin lines the plasma membrane. The actin network may therefore provide a scaffold for maintaining the integrity of the CAB. Figure 6A shows the entire CAB region at the 16-cell stage. It exits just beneath the plasma membrane. The CAB has a clear boundary, but there is no membrane between it and the surrounding cytoplasm. Ultrastructurally, the CAB consists mainly of electron-dense cytoplasm (Fig. 6B). Many granular and

Fig. 6 Electron micrographs of the CAB. (A) Whole view of a CAB in a B5.2 cell of a 16-cell embryo. Posterior is down. The CAB appears as a relatively electron-dense cytoplasmic region. Bar, 5 pm.(B) Higher magnification. Electron-dense cytoplasm, granular structures, and vesicular structures (arrows) can be seen. Bar, 1 wm.

16

Hiroki Nishida et al.

some vesicular structures are also observed in the CAB (Iseto and Nishida, 1999). C. Role of Posterior Egg Cytoplasm

The ooplasm of ascidian eggs undergoes dramatic movements between fertilization and the beginning of the first cleavage, and this process is known as ooplasmic segregation (Conklin, 1905; Jeffery and Meier, 1983; Sawada, 1988; Jeffery and Bates, 1989; Sardet et al., 1989). Movement of the ooplasm occurs in two phases. In Halocynthia eggs, the vegetal pole egg cytoplasm after the first phase of ooplasmic segregation is translocated to the posterior-vegetal region during the second phase. These cytoplasm are required for generation of the posterior cleavage pattern. When these parts of eggs are removed from the egg, the cleavage pattern becomes radialized along the animal-vegetal axis, and no unequal cleavage occurs in the embryo (Fig. 4C) (Nishida, 1994,1996). Transplantation of the posterior-vegetal cytoplasm (PVC) after the second phase of ooplasmic segregation to the anterior-vegetal position of another egg causes duplication of the posterior cleavage pattern, resulting in successive unequal cleavages in the anterior and posterior sides (Fig. 4D).Therefore, ooplasmic factors localized in the posteriorvegetal cytoplasm direct the posterior pattern of cleavage by generating three rounds of unequal cleavage at the posterior pole of the embryo. Nishikata et al. (1999) showed that, when the PVC is removed, the embryo does not form the CAB, and no thick microtubule bundle is observed (Fig. 4C). Moreover, nuclear translocation is prevented. There is a good correlation between loss of the CAB and abolishment of unequal cleavage. Brief treatment of fertilized eggs with the detergent sodium dodecyl sulfate gives results similar to those when the PVC is removed. By contrast, when the PVC is transplanted to the anterior region of another intact egg, ectopic CABS are formed at the anterior pole of the embryo (Fig. 4D). Thick microtubule bundles are formed on both the anterior and posterior sides, and unequal cleavages take place on both sides. In these experiments, there is strict coincidence between CAB formation, appearance of the microtubule bundle, nuclear translocation, and the occurrence of unequal cleavage (Fig. 7). As mentioned before, the CAB formation starts after the two-cell stage. Therefore, maternal factors in the PVC are likely to mediate unequal cleavage through formation of the CAB. Various maternal transcripts have been shown to be localized in the posteriorvegetal ooplasm (Yoshida e l al., 1996; Satou and Satoh, 1997; Sasakura et al., 1998a,b). Interestingly, most of them seem to be colocalized with the CAB during the cleavage stage. Thus, there is a possibility that these mRNAs contribute to formation of the CAB. Another possibility is that the CAB plays an additional role in the localization and maintenance of these mRNAs. If so, then the CAB may have another role in the segregation of maternal information. The CAB couples both the localization of determinants and the orientation of cleavages to seg-

1. Maternal Factors and Embryonic Cleavage Patterns

17

Fig. 7 Experiments involving removal and transplantation of the posterior-vegetal cytoplasm in ascidian eggs. In the left part of this figure, experimental designs depict the lateral view of the eggs. Anterior is to the left. In the right part of the figure, the results of experiments are represented schematically. (A) Untreated control (CAB, centrosome-attractingbody). (B) Removal of the posterior-vegetal cytoplasm from the fertilized egg. (C) Transplantation of the posterior-vegetal cytoplasm to the anterior position of another intact egg.

regate these mRNAs into one daughter cell, because the cell division planes are always perpendicular to the axis of localization. A stereotyped cleavage pattern will play a role in the precise partitioning of maternal cytoplasmic determinants, and, in ascidians, the generation of a specific cleavage pattern itself is controlled by localized maternal cytoplasmic factors. Ascidian embryos offer a novel experimental system for analyzing the mechanisms of unequal cleavage.

VI. Unequal Cleavage in Annelid Embryos The cleavage pattern and early development of annelids have been well documented in several species of Polychaeta (Nereis, Cuetopterus, Subellaria) (e.g., Reverberi, 1971), Oligochaeta (Tubifex) (e.g., Shimizu, 1982), and Hirudinea (Helobdellu)(e.g., Fernandez and Olea, 1982). A review of the cumulative work

18

Hiroki Nishida et al.

done by Shimizu (1998) revealed some of the cellular mechanisms underlying unequal cleavages in the Tubifex embryo. The early development of Tubifex consists of a stereotyped sequence of cell divisions (Penners, 1922; Shimizu, 1982). The first cleavage is unequal and meridional, and gives rise to a smaller AB blastomere and a larger CD blastomere. At the second cleavage, the CD blastomere divides into a smaller C blastomere and a larger D blastomere, while the AB blastomere divides into blastomeres A and B of various sizes. The resulting four blastomeres are called quadrants. From the third cleavage onward, quadrants A-D repeat unequal divisions, producing micromeres at the animal side (which corresponds to the future dorsal side of the embryo) and macromeres at the vegetal side (corresponding to the future ventral side). These unequal divisions occur in an oblique fashion, shift orthogonally at each cleavage, and show a typical clockwise spiral cleavage pattern.

A. Unequal Division at the First Cleavage

Mature Tubifex eggs are oviposited at the metaphase of the first meiosis and extrude first and second polar bodies before the first cleavage (Shimizu, 1982). Fertilization occurs near the vegetal pole (Hirao, 1968). The zygotic nucleus is located at the center of the egg, and the metaphase mitotic apparatus of the first cleavage is formed centrally (Ishii and Shimizu, 1995). This mitotic apparatus is asymmetric in shape and has unique features (Ishii and Shimizu, 1995; Shimizu, 1995). The spindle is organized bipolarly, and metaphase chromosomes are located at its midpoint; however, it is not amphiastral but monoastral (Fig. 8A-C). The one pole of the mitotic apparatus that has an aster is stained with anti-y-tubulin antibody whereas the other is not, indicating that the astral pole has a centrosome but the anastral pole might be organized through a microtubule organizing center other than a centrosome (Shimizu, 1996).The centrosome participating in the mitotic apparatus assembly is not paternal but maternal in origin. The single centrosome of the meiotic spindle is utilized without duplication. During cytokinesis, a larger CD blastomere is formed on the astral side and a smaller AB blastomere is formed on the anastral side. Thus, unequal cleavage is attributable not to the asymmetric position of the spindle but to asymmetric formation of the asters. To examine how such an asymmetrical mitotic apparatus is produced, several experiments were carried out. In one experiment, eggs were compressed or elongated along the egg axis to examine whether the relative positions between the cortex and mitotic apparatus contribute to the generation of asymmetry. In spite of these disturbances, asymmetric mitotic apparatuses were always formed, and unequal cleavages occurred (Ishii and Shimizu, 1995). In a second study, by suppressing formation of the polar body, eggs were manipulated to inherit two

1. Maternal Factors and Embryonic Cleavage Patterns

19

Fig. 8 Antitubulin immunostaining of the mitotic apparatus during the first (A-C) and second (D) cleavages of the Tubifex embryo. Animal views. Arrows in A-C point to the anastral spindle pole. (A) Early anaphase. (B) Cleavage furrows (arrowheads) begin to form. (C) End of cytokinesis of the first cleavage. Smaller AB blastomere (AB) and larger CD blastomere (CD) are produced. (D) Asymmetric mitotic apparatus in the CD blastomere during the second cleavage. Anterior side of one of the spindle poles is apposed to the previous cleavage plane (arrowhead). Arrow indicates the AB nucleus. Bar, 100 pm. A-C from Ishii and Shimizu (1995); D from Takahashi and Shimizu (1997).

maternal centrosomes during first cleavage. These two centrosomes did not duplicate and located at both poles of the first meiotic spindle. They form an amphiastral mitotic apparatus at the first cleavage. The mitotic apparatus was located at the center of the egg and resulted in an equal division (Ishii and Shimizu, 1997a). These two experiments indicate that, unlike other animals discussed in this article, the egg cortex of Tubifex does not affect the generation of asymmetry during mitotic apparatus organization. In a third experiment, to examine whether centrosome duplication is under the control of the egg cytoplasm, eggs in meiosis I1 and eggs in mitosis I were fused so that they shared a

20

Hiroki Nishida et al.

common cytoplasm (Ishii and Shimizu, 1997b).After completion of their meiosis/ mitosis, one centrosome associated with the mitosis I spindle was duplicated, whereas the other centrosome associated with the meiosis 11 spindle was not. This suggested that the meiosis I1 centrosome of the oocyte is intrinsically distinct from the mitotic centrosome, and that inhibitory factors for centrosome duplication are associated with the meiotic centrosome. The first-cleavage mitotic apparatus orients perpendicular to the animal/vegeta1 axis. The shape of the Tubifex egg, which is an oblate spheroid, may contribute to the orientation of the mitotic apparatus (Ishii and Shimizu, 1995). In addition, the pole plasm, which is localized in a bipolar fashion at both the animal and the vegetal poles, may also have a role in orienting the mitotic apparatus by interacting with astral microtubules (Shimizu, 1988, 1989). B. Unequal Divisions at the Second Cleavage and Micromere Formation

At the second cleavage, an asymmetrical amphiastral mitotic apparatus is formed in the CD blastomere (Shimizu, 1993).The anterior side (the side of the AB blastomere in the embryo) of the CD spindle is apposed to the anterior cortex near the midbody of the previous first cleavage, and the anterior aster is flattened and truncated (Fig. 8D).As the CD blastomere isolated from the embryo forms a symmetrical mitotic apparatus in the center of the blastomere and divides equally, attraction of the spindle toward the anterior cortex is dependent on cell contact with the AB blastomere (Takahashi and Shimizu, 1997). The mitotic apparatus in the D quadrant of the four-cell embryo is located at the animal side of the cell, orients parallel to the animal-vegetal axis, and associates with the animal cortex by one of its spindle poles (Shimizu, 1988, 1989). Even in the isolated D quadrant, this asymmetrical positioning and micromere formation occur. Therefore, the positioning mechanism in the D quadrant is independent of cell contact. Thus, the unique cleavage pattern of the Tubifex embryo, consisting of several rounds of unequal cleavages, is controlled by distinct mechanisms, including both intrinsic and extrinsic cues.

VII. par Mutants in Nematode Embryos To accomplish asymmetric cleavage, polarized distribution of cytoplasmic components and proper orientation of the mitotic spindle are critical. In the eggs of the nematode Caenorhabditis elegans, the definition and maintenance of initial polarity and the control of mitotic spindle orientation are considered to be determined by specific maternal cytoplasmic factors. So far, six maternal par (partitioning-defective) genes, par-1 through par-6, have been isolated. In this tion, we summarize current knowledge of how maternal par genes maintain the

sec-

21

1. Maternal Factors and Embryonic Cleavage Patterns

initial polarity and control the orientation of the mitotic spindles in asymmetric cleavage during the first and second cleavages of the C. elegans embryo. A. Asymmetric Cleavage in Caenorhabdifis eregans

The early cell lineage of the C. elegans embryo is illustrated in Fig. 9, showing that early cleavage forms six “founder” cells (Sulston et al., 1983). Five of the early cleavages are asymmetric and produce daughter cells differing in cell size, cell fate, successive mitotic spindle orientation, and cell cycle. Cleavage of the germ-line lineage (Po-P3) is always asymmetric, and the smaller blastomere continually remains in the germ-line fate (P1-P4). The first and second asymmetric cleavages in the C. elegans embryo occur as a result of a sequence of specific events (Fig. 10). The ellipsoid C. elegans oocyte has a distally positioned pronucleus at one end, but no other developmental asymmetries can be observed before fertilization (Strome, 1986). Goldstein and Hird (1996) have shown that the initial asymmetry is cued by entry of the sperm. The ellipsoid’s end, close to the sperm entry point, becomes the future posterior pole (Fig. 10A). This suggests that the C. elegans oocyte has no intrinsic polarity or, if any, it can be overruled easily. Although the underlying mechanism is still to be explored, sperm entry establishes the initial anterior-posterior (A-P)

Anterior

Postehor

~

b2 Hypodermls Neurons Muscle Others

@

9d,

@JW

@

C

MS Muscle Neurons Glands Somatic Gonads Others

lntestlns

eJL@ p4

D

Muscle Hypodermis Neurons

Germllne Muscle

Fig. 9 Early cell lineage of the Cuenorhabditis elegans embryo. AB,MS, E, P,, D, and C are designated “founder” cells, with cell fates as shown. AB continues to divide with an orthogonal cleavage pattern. In Po, P,, P,, P,, and EMS, the centrosome-nuclear complex rotates to carry out asymmetric cleavage. Modified from White and Strome (1996); copyright 1996 Cell Press.

22

Hiroki Nishida et al. polnl of ap nn entry cytopkrmlcflow mlpratlonot cornpomnla

K%%!'i:

postertor)

P gnnulaa

cenhueom oocyte pronuclotis

rpenn pronucleus

(with centroeorn)

Fig. 10 Schematic sequence of the first and second cleavage of the Caenorhabditis elegans embryo. To aid understanding, G, H and I are depicted as if the egg shell has been removed.

polarity along the long axis of the fertilized egg, Po.After fertilization, the oocyte pronucleus migrates to unite with the sperm pronucleus at the posterior end (Fig. 10A). Simultaneously, a dynamic cytoplasmic rearrangement takes place: anterior-to-posterior internal flow, and posterior-to-anterior flow at the cortex of the posterior end (Fig. IOA) (Hird and White, 1993). Coincidentally, the cortical actin accumulates at the anterior (Strome, 1986). Although any direct relationship with the cytoplasmic flow is unclear, germ-line lineage-specific particles known as P granules also accumulate at the posterior pole (Strome and Wood, 1982; Hird et aL, 1996). In addition to providing the initial A-P polarity, the sperm also supplies the centrosome, which the oocyte lacks (Albertson, 1984). After duplication of the sperm centrosome and the meeting of both pronuclei, the daughter centrosomes migrate 90" around the nucleus so that they lie on the opposite sides (Hyman, 1989), and the nucleus of the zygote migrates to the center of Po (Fig. 10B). Later, the centrosome-nucleus complex rotates 90" so that the axis between the two centrosomes becomes aligned parallel to the A-P axis (Fig. 1OC).Then the mitotic spindle forms and migrates slightly in a posterior direction (Fig. lOD). As a result, an unequal first cleavage occurs, creating a larger anterior blastomere (AB) and a smaller posterior one (PI) carrying the P granules (Fig. 10E). Next, at the two-cell stage, in both the AB and P, blastomeres, the centrosomes duplicate and migrate 90" around each nucleus, as if the second cleavage is about to occur orthogonally to the first cleavage (Fig. 10F). In the AB blastomere, cleavage continues with this orientation to produce two symmetric blastomeres. The AB lineage cells continue to cleave in this orthogonal manner, and their

1. Maternal Factors and Embryonic Cleavage Patterns

23

cleavages occur synchronously. On the other hand, the centrosome-nucleus complex of the smaller P, blastomere rotates 90" to become oriented parallel to the A-Paxis (Fig. 10G).This rotation can be disrupted by a microtubule inhibitor (Hyman and White, 1987). When P, undergoes centrosome rotation, short and straight microtubules are seen running from a centrosome to the area of contact between AB and P,. When the microtubules attached to the putative cortical attachment site are disrupted by a microlaser beam, the rotation is inhibited (Hyman, 1989;White and Strome, 1996).These results indicate that the centrosomenucleus complex rotates by capture and shortening of the microtubules from one centrosome using a specialized cortical attachment site. This model is very similar to that proposed for unequal cleavage in ascidians mentioned above. It has been reported that actin and actin-capping protein transiently accumulate in the region of the putative attachment site in P,, although the role is not clear (Waddle et al., 1994). After spindle formation in P,, the spindle migrates slightly in a posterior direction. Then P, divides into a larger EMS and a smaller P, (Figs. 10H, I), and the posterior P, inherits the P granules. Later in the division of germline lineage P, and P,, similar unequal cleavage occurs to produce larger and smaller blastomeres: the latter contain the P granules and are the germ-line lineage cells. Blastomere isolation experiments have shown that asymmetric cleavage in the germ-line lineage occurs cell-autonomously (Schierenberg, 1988). The P granules, unequally inherited in the P, -P, germ-line lineage, are considered to contain ribonucleoprotein that plays an important role in germ-line development (Gruidl et al., 1996).Although not described here, other well-known cell fate determinants, including GLP- 1, SKN- 1, MEX-1, PAL-1, and PIE- 1, which are equally dispersed in oocytes, also show a cell lineage-specific distribution throughout early development (Bowerman et al., 1993; Schnabel et al., 1996; Crittenden et al., 1997;Tenenhaus et al., 1998;reviewed in Rose and Kemphues, 1998a; Bowerman, 1998; Schnabel and Priess, 1997). These are critical for determination of cell fate. 6. The par Genes

The par genes were first identified through analysis of isolated maternal-effect lethal mutants with disordered early embryonic asymmetries (Kemphues et al., 1988). So far, six par genes, par-I to par-6, have been characterized (Guo and Kemphues, 1996a; Watts et al., 1996). The common phenotypes ofpar embryos (embryos from homozygous par mutant mothers) include equal first cleavage (except for the par-4 embryo), synchronous second cleavage, and altered spindle orientation at second cleavage. Also, a defect in P granule localization is observed. Thus, various kinds of process are affected by par mutations. Probably, these spatial regulations utilize a common machinery. Moreover, altered distribution of cell fate determinants (GLP-1, SKN-1, PAL-1, PIE-1, etc.) is observed

24

Hiroki Nishida et al.

in par embryos in a mutation-specific manner (reviewed in Rose and Kemphues, 1998a;Bowerman, 1998; Schnabel and Priess, 1997). In par embryos, except for par-4, equal-sized AB and P, blastomeres are formed at the first cleavage. Therefore, these genes are required for asymmetric positioning of the first mitotic apparatus. All of the par embryos show altered spindle orientation at second cleavage, although the abnormality of spindle orientation differs among the mutations. The phenotypes of the par embryos are summarized in Fig. 11, together with the known characteristics of the par gene products (reviewed in Rose and Kemphues, 1998a). It is interesting that each par phenotype becomes distinct only at the second cleavage. Figure 12 shows the distribution of par gene products

Fig. 11 Phenotypes of Caenorhabditis elegans mutant embryos. Bidirectional arrows indicate variable spindle orientation. See Fig. 10 for explanation of other symbols. Wild-type (WT),par mutants, and pkc-3 and nmy-2 mutants are shown. The known characteristics of each of the par gene products are indicated.

1. Maternal Factors and Embryonic Cleavage Patterns

25 (PAR-1, PAR-2, and PAR-3 proteins) and the orientation of the mitotic spindle during the one-cell and two-cell stages of wild-type and par embryos. The par-1 gene is considered to have an important role in establishing asymmetric distribution of cell fate determinants such as P granules (Kirby et al., 1990), SKN-1 (Bowerman et al., 1993), and PIE-1 (Tenenhaus et al., 1998), and also in unequal first cleavage (Kemphues et al., 1988). The par-1 embryos exhibit variable mitotic spindle orientation at the second cleavage (Fig. 11). The Ser/Thr kinase domain of the par-1 gene product, PAR- 1 protein, is required for establishing asymmetry during early cleavage (Guo and Kemphues, 1995). The accumulation of PAR-1 in the posterior cortex suggests that the protein acts by modifying the cytoskeleton in this region. However, it is also known that PAR-1 need not be localized cortically in order to mediate the localization of P granules and SKN-1 (Boyd etal., 1996; Bowerman et al., 1997).The activities of thepar2, par-3, par-5, and par-6 genes are required for the posterior distribution of PAR-1 (Fig. 12) (Boyd et al., 1996; Guo and Kemphues, 1996a; Watts et al., 1996), and par-1 has no known ability to control other par genes (EtemadMoghadam et al., 1995; Boyd et al., 1996). Taken together, these findings suggest that par-1 is “downstream” of the par-2, par-3, par-5, and par-6 genes. Analysis of the PAR-1 mammalian homologs, mPAR-1 (Bohm et al., 1997) and MARK (Drewes et al., 1997), which are members of the novel microtubule-associated protein (MAP) kinase family, suggests that PAR-1 of C. elegans also regulates microtubule dynamics and establishes cellular asymmetry in the embryo (reviewed in Nelson and Grindstaff, 1997). It is understood that thepar-2 gene is responsible for the positioning of the mitotic spindle during the first and second cleavages, as well as for establishing asymmetry (Kemphues et al., 1988; Kirby et al., 1990; Levitan et al., 1994; Cheng et al., 1995; Boyd et al., 1996; Watts et al., 1996). The typical par-2 embryo phenotype is similar to that of the par-1 embryo, except that both mitotic spindles at the two-cell stage do not rotate at all, and as a result the second cleavage occurs transverse to the initial A-P polarity (Fig. 11). Activity of par-2 is required for the function of thepar-1, par-3, andpar-4 genes. Thepar-2 gene product, PAR-2 protein, accumulates at the posterior cortex, similarly to PAR-1 (Fig. 12). The posterior distribution of PAR-2 depends on the activities of the par-3, par-5, and par-6 genes, but not on par-1 or par-4. The par-2 embryo conversely lacks anterior localization of PAR-3. Activity ofpar-2 is also required for proper localization of PAR- 1 to the cortex. The proposed role of PAR-2 is to exclude PAR-3 from the posterior periphery and also to direct PAR-1 localization to the cortex (Fig. 13). The par-3 gene also plays a critical role in defining spindle orientation during the first and second cleavages. The specific phenotype of thepar-3 embryo is that the mitotic spindles at the two-cell stage both rotate, unlike the par-2 embryo phenotype, so that they align parallel to the A-P axis (Fig. 11) (Kemphues et al., 1988; Cheng et al., 1995). The product of the par-3 gene, PAR-3, is localized at

Fig. 12 Distribution of PAR-I, PAR-2, PAR-3, NMY-2, and PKC-3 proteins in specific mutant backgrounds. Shaded regions indicate the type of distribution of defined proteins: solid border, cortical localization; hatched border, variable distribution; filled, dispersed distribution. See Fig. 10 for explanation of other symbols (WT, wild type).

Fig. 13 Model of PAR protein localization. For details, see text for references; also see Fig. 10 for explanation of symbols.

U h)

28

Hiroki Nishida et al.

the anterior periphery, reciprocal to PAR- 1 and PAR-2, and is considered to restrict PAR-1 and PAR-2 to the posterior region. The proper distribution of PAR-3 is controlled by par-2, par-5, and par-6 activities (Etemad-Moghadam et al., 1995; Watts et al., 1996; Guo and Kemphues, 1996a). Therefore, PAR-2 and PAR-3 are suggested to repel each other, resulting in mutually exclusive distribution in the fertilized egg (Fig. 13). the par-2par-3 double mutants show apar3-like phenotype whereby both the AB and P, spindles rotate. Therefore, neither par-2 norpar-3 is required for spindle rotation (Kirby et al., 1990; Cheng et al., 1995). Spindle rotation during the two-cell stage is prevented by PAR-3, whereas PAR-2 appears to prevent PAR-3 from functioning, thereby permitting some other process to rotate the spindle. Thepar-4 embryo has most of thepar-l phenotype, except that the first cleavage is unequal (Fig. 11) (Kemphues et al., 1988; Kirby et al., 1990; Morton et al., 1992; Boyd et al., 1996).Although little is known at present, PAR-4 protein is unique among the PAR products in being distributed uniformly in the cortex throughout early embryogenesis (Bowerman, 1998). PAR-4 is suggested to interact with PAR-2 (Morton et al., 1992), although neither PAR-I, PAR-2, nor PAR-3 requirespar-4 function for distribution. Thus, it is assumed thatpar-4 acts independently or “downstream” of the others. At present, not much is known about the par-5 gene. The mutant phenotype is similar to that of the par-2 embryo; a symmetric first cleavage, and no rotation of the mitotic spindle at the two-cell stage (Fig. 11). It is known thatpar-5 activity is required for proper localization of PAR-1, PAR-2, and PAR-3 (Guo and Kemphues, 1996a), and it is thought to be required by the otherpar genes as well (Rose and Kemphues, 1998a). Therefore,par-5 acts “upstream” of the others. The par5 gene encodes a member of the “14-3-3” protein family (Rose and Kemphues, 1998a), which modulate a wide variety of cellular processes and are ubiquitous and highly conserved throughout the plant and animal kingdoms (Wang and Shakes, 1996). PAR-3 also contains a 14-3-3 binding consensus (Rose and Kemphues, 1998a), which indicates direct interaction between PAR-5 and PAR-3. In order to clarify the function of the par-5 gene, analysis of PAR-5 distribution in otherpar embryos is a potentially very fruitful issue of future research. The newest of the group, the par-6 embryo, has apar-3-like phenotype, showing rotation of both mitotic spindles at the two-cell stage (Figs. 11 and 12) (Watts et al., 1996). in par-6 embryos, PAR-1, PAR-2, and PAR-3 exhibit altered localization (Watts et al., 1996). Both PAR-3 and PAR-6 contain PDZ domains (Etemad-Moghadam et al., 1995; Rose and Kemphues, 1998a), which are considered to form protein-protein interacting domains (reviewed in Ponting et al., 1997).Also, PAR-6 and PAR-3 are considered to be mutually required for each other’s localization at the periphery (Watts et al., 1996; Rose and Kemphues, 1998a). The probable role of PAR-6 is localizing or maintaining PAR-3 at the cell periphery by binding to PAR-3, thus signaling or maintaining the initial embryonic A-P polarity during early embryogenesis.

1. Maternal Factors and Embryonic Cleavage Patterns

29 Further understanding has been obtained by identifying factors that interact with the PAR proteins. Through the screening of such factors that bind to PAR-1, a nonmuscle cytoplasmic myosin, NMY-2, has been identified (Guo and Kemphues, 1996b). Together with the Ser/Thr kinase domain, PAR-1 contains a nonmuscle myosin binding domain at the C terminus. It has also been shown that NMY-2 and PAR-1 can interact both in vitro and in vivo. Use of the RNA interference method (reviewed in Montgomery and Fire, 1998) has revealed that NMY-2-depleted embryos show altered localization of PAR-1, PAR-2, and PAR-3, a symmetric first cleavage, and no spindle rotation in both AB and P, (Figs. 11 and 12), similar topar-2 andpar-5 embryos. This suggests that NMY2 is required for proper localization of the PAR proteins and also that actomyosinbased motility is involved in polarizing the embryo. An atypical protein kinase C homolog of C. elegans, PKC-3, has been cloned (Wu et al., 1998; Tabuse et al., 1998).Analysis of PKC-3 has shown that it binds to PAR-3 in vitro. Also, PKC-3 colocalizes with PAR-3 at the anterior cortex of the one-cell embryo. The PKC-3-depleted phenotypes show similarity to those of par-3 and par-6 embryos (Fig. 11). For proper localization of PKC-3, the activities of par-2, par-5, andpar-6 genes are required (Fig. 12). Furthermore, the localizations of PKC-3 and PAR-3 are mutually dependent on each other (Tabuse et al., 1998). From these findings, it can be hypothesized that PAR-3, PKC-3, and PAR-6 form a functional complex in the cortex of the embryo, to interact and modify the cytoskeleton for promotion of asymmetric cleavages (see Rose and Kemphues, 1998a). On the basis of these results, a model of how PAR-1, PAR-2, and PAR-3 are localized and function during early development is proposed (Fig. 13) (Guo and Kemphues, 1996a; Kemphues and Strome, 1997). PAR-3 locates with a gradient at the periphery of Po, in response to the initial polarity cued by sperm entry. PAR-2 mutually interacts with PAR3 to restrict PAR-3 to the anterior and PAR-2 to the posterior. Then, PAR-1 becomes localized to the posterior periphery by PAR-2 and PAR-3. These processes require PAR-6, nonmuscle myosin, and probably atypical protein kinase C. Then, by an unknown mechanism, the spindle migrates posteriorly and the distribution of other cellular components becomes polarized. As a result, the first cleavage becomes asymmetric, and the two blastomeres, AB and P,, acquire different cell sizes and fates. At the two-cell stage, PAR-3 is localized throughout the entire cortex of AB and to the anterior periphery of PI. PAR-1 and PAR-2 reciprocally occupy the posterior cortex of P,, and rotation of the spindle occurs in P,. PAR-3 probably inhibits spindle rotation by strongly promoting an association between the cortex and the microtubules of the aster. In AB, PAR-3 equally distributed in the cortex mediates tight and broad association of the centrosome-nuclear complex with the inner cellular surface, probably the cytoskeleton, making the complex unable to move or rotate. In contrast, because PAR-3 is restricted to the anterior periphery in PI, association with the posterior pole will not occur. Either one of the two centrosomes

30

Hiroki Nishida et al.

in PI is pulled to the anterior periphery, in order to make the centrosome-nuclear complex rotate. PAR- 1 is restricted to the posterior periphery in P, by PAR-2 and PAR-3, and hence the cytoplasmic components are localized asymmetrically. Again by an unknown mechanism, the spindle migrates posteriorly, and unequal cleavage ensues in P,, to yield daughter cells with different sizes and fates (EMS, P2). Another plausible possibility is that cortical PAR-3 is not directly involved in stabilization of the spindle at the two-cell stage, and that PAR-2 and PAR-3 might function to polarize the distribution of other unknown factors, resulting in only PI acquiring the machinery necessary for spindle capture and rotation (Bowerman, 1998). From the information available so far, it is possible to construct a model of the manner in which the PAR proteins interact during early asymmetric cleavage (Fig. 14) (Kemphues and Strome, 1997; Rose and Kemphues, 1998a). PAR-5 is required for asymmetric localization of all the par gene products (except for PAR-4), and therefore is considered to be the most “upstream” of these genes. PAR-3, PAR-6, and PKC-3 form a functional complex and mutually interact with PAR-2. PAR-1 is the most “downstream” of all the genes (except for PAR-4) and requires all the upstream genes. The role of PAR-4 in the pathway is still unclear. This model is based mostly on the distribution of each protein in each mutant. However, all of the par embryos show dissimilar phenotypes in terms of spindle orientation at the two-cell stage and in the distribution of cell fate determinants such as SKN- 1 and PAL- 1. Therefore, the par genes may compose a network rather than a linear pathway (Bowerman, 1998). How is the initial sperm-cued A-Ppolarity signaled to thepar genes? How do the par genes distribute such numerous cellular components? How do the par genes control spindle orientation? How do the PAR proteins interact with both microfilaments and microtubules? How do the PAR proteins interact with each

PAR-5

--

PAR4 7

PAR1

Fig. 14 Schematic model of PAR protein interaction. Proteins considered “downstream”in the pathway are placed to the right of those considered “upstream.”Boxed region indicates formation of a functional complex. Interaction of PKC-3 is also shown.

1. Maternal Factors and Embryonic Cleavage Patterns

31 other to establish their respective distributions? Here are some additional clues that may help to answer these questions. Schierenberg (1988) has shown that the posterior cytoplasm directs asymmetric cleavage. Cortical actin accumulates at the anterior pole after fertilization (Strome, 1986; Waddle et al., 1994). Treatment of embryos with cytoskeleton inhibitors, such as cytochalasin D and nocodazole, indicates that the rotation of the mitotic spindle depends on intact cytoplasmic actin and microtubules (Hyman, 1989; Hyman and White, 1987; reviewed in White and Strome, 1996). Another gene, ler-99, is also known to be involved in early cleavage. The let-99 gene is a newly discovered maternal gene that is required for proper spindle orientation after A-P polarity has been established, because PAR-1, PAR-2, and PAR-3 all show a normal distribution in let99 mutants (Rose and Kemphues, 1998b). The let-99 gene is postulated to play a role in mediating or regulating connections between astral microtubules and the cortical cytoskeleton. Another maternal-effect gene, mes-1, is required for rotation of centrosomes specifically in P, and P, (Strome et al., 1995), because mes-1 embryos show normal division and P granule segregation in Po and P,. At present, there is no known sequence similarity between the polarity-establishing par genes in C. elegans and known Drosophila genes involved in asymmetric cell division (reviewed in Lu et al., 1998).

VIII. Concluding Remarks In this article, we have dealt mainly with maternal control of the generation of unique cleavage patterns. In ascidian embryos, experimental transplantation of egg cytoplasm has revealed the presence of localized factors in the egg cytoplasm. In gastropods and nematodes, maternal-effect variants and mutants show genetic evidence for the maternal contribution in controlling the pattern of cleavage. In nematodes, the products of such genes, the PAR proteins, show clear localization in the egg cortex. These studies indicate the importance of egg organization and localized factors for generation of unique cleavage patterns. Future studies of par genes will help to clarify the underlying machinery responsible for coupling the localization of developmental determinants and the positioning of the cleavage plane. To create divergence of the cleavage pattern from the “default” equal and orthogonal pattern, various mechanisms for positioning the mitotic apparatus are used in different animal systems. In sea urchin, ascidian, and nematode embryos, movements of centrosomes during interphase play key roles in the orientation and asymmetric positioning of the mitotic apparatus. The movements of the centrosome involve the cortical site, the centrosome, and the microtubules between them. In the nematode embryo, and in the second and third cleavages of oligochaete eggs, the mitotic spindle migrates during the mitotic phase. Microfilaments are required for the migration in the latter case. And in the first cleav-

32

Hiroki Nishida et al.

age of oligochaete, a difference in the size of the asters causes unequal cleavage. In most cases, the cell cortex provides essential spatial cues. The vegetal cortex of sea urchin embryos, the CAB of ascidian embryos, and cortical PAR proteins in nematode embryos capture microtubules from the centrosome and mitotic apparatus. In the sea urchin, ascidian, gastropod, and annelid, cytological observations have enabled great progress. By contrast, in C. elegans and Drosophila, genetic and molecular studies have been the trend. To understand fully the mechanisms involved in positioning of the cell division plane, both types of approaches are expected to be combined in future research.

Acknowledgments Our work reviewed in this article was supported by the “Research for the Future” Program from the Japanese Society for the Promotion of Science (96L00404). We thank Tohru Iseto for reading our manuscript.

References Albertson, D. (1984). Formation of the first cleavage spindle in nematode embryos. Dev. Biol. 101, 61-72. Bohn, H., Brinkmann, V., Drab, M., Henske, A., and Kurzchalia, T.V. (1997). Mammalian homologues of C. elegans PAR-1 are asymmetrically localized in epithelial cells and may influence their polarity. Curt Biol. 7, R603-R606. Bowerman, B. (1998). Maternal control of pattern formation in early Caenorhabdiris elegans embryos. Cum Top. Dev. Biol. 39,73-117. Bowerman, B., Draper, B. W., Mello, C. C., and Priess, J. R. (1993). The maternal gene skn-1 encodes a protein that is distributed unequally in early C. elegans embryos. Cell 74,443-452. Bowerman, B., Ingram, M. K., and Hunter, C. P. (1997). The maternal par genes and the segregation of cell fate specification activities in early Caenorhabdiriselegans embryos. Development 124, 38 15-3826. Boyd, L., Guo, S., Levitan, D., Stinchcomb,D. T.,and Kemphues, K. J. (1996). PAR-2 is asymmetrically distributed and promotes association of P granules and PAR-1 with the cortex in C. elegans embryos. Development 122,3075-3084. Cameron, R. A., and Davidson, E. R. (1991). Cell type specification during sea urchin development. Trends Genet. 7,212-218. Cheng, N. N., Kirby, C. M., and Kemphues, K. J. (1995). Control of cleavage spindle orientation in Caenorhabditis elegans: The role of the genes par-2 and par-3. Genetics 139,549-559. Collier, J. R. (1997). Gastropods, the snails. In “Embryology: Constructing the Organism” (S. F. Gilbert and A. M. Raunio, eds.), pp. 189-217. Sinauer, Sunderland. Conklin, E. G. (1897). The embryology of Crepidula. J. Morph. 13, 1-226. Conklin, E. G. (1905). The organization and cell lineage of the ascidian egg. J. Acad. Narl. Sci. (Philadelphia) 13, 1-1 19. Conklin, E. G. (1912). Experimental studies on nuclear and cell division in the eggs of Crepidula. J. Acad. Natl. Sci. (Philadelphia, Set 2 ) 15,503-591. Crittenden, S . L., Rudel, D., Binder, J., Evans, T. C., and Kimble, J. (1997). Genes required for GLP-1 asymmetry in the early Caenorhabditis elegans embryo. Dev. Biol. 181,36-46.

1. Maternal Factors and Embryonic Cleavage Patterns

33

Dan, K. (1979). Studies on unequal cleavage in sea urchins. I. Migration of the nuclei to the vegetal pole. Dev. Growth Diffe,: 21,527-535. Dan, K., and Ikeda, M. (1971). On the system controlling the time of micromere formation in sea urchin embryos. Dev. Growth Diffe,: 13,285-301. Dan, K., and Ito, S. (1984). Studies of unequal cleavage in molluscs: I. Nuclear behavior and anchorage of a spindle pole to cortex as revealed by isolation technique. Dev. Growth D@,: 26,249262. Dan, K., and Nakajima, T. (1956). On the morphology of the mitotic apparatus isolated from echinoderm eggs. Embryologia 3,187-200. Dan, K., Endo, S., and Uemura, I. (1983). Studies on unequal cleavage in sea urchins. 11. Surface differentiation and direction of nuclear migration. Dev. Growth Diffe,: 25,227-237. Doe, C. Q. (1996). Spindle orientation and asymmetric localization in Dmsophila: Both inscuteable? Cell 86,695-697. Drewes, G., Ebneth, A., F’reuss, U., Mandelkow, E.-M., and Mandelkow, E. (1997). MARK, a novel family of protein kinases that phosphorylate microtubule-associatedproteins and trigger microtubule disruption. Cell 89,297-308. Etemad-Moghadam, B., Guo, S., and Kemphues, K. J. (1995). Asymmetrically distributed PAR-3 protein contributes to cell polarity and spindle alignment in early C. elegans embryos. Cell 83, 743-752. Fernandez, J., and Olea, N. (1982). Embryonic development of glossiphoniid leeches. In “Developmental Biology of Freshwater Invertebrates” (F. W. Harrison er al., eds.), pp. 317-361. Alan R. Liss, New York. Freeman, G. (1976). The role of cleavage in the localization of developmentalpotential in the ctenophore Mnemiopsis leidyi. Dev. Biol. 49, 143-177. Freeman, G. (1979). The multiple roles which cell division can play in the localization of developmental potential. In “Determinants of Spatial Organization”(S. Subtelny and I. Konigsberg, eds.), pp. 53-76. Academic Press, San Diego Freeman, G. (1983). The role of egg organization in the generation of cleavage patterns. In “Time, Space, and Pattern in Embryonic Development” (W. R. Jeffery and R. A. Raff, eds.), pp. 171196. Alan R. Liss, New York. Freeman, G., and Lundelius, J. W. (1982). The developmental genetics of dextrality and sinistrality in the gastropod Lymnaea pregra. Roux’sArch. Dev. Biol. 191,69-83. Goldstein, B. (1995). Cell contacts orient some cell division axis in the Caenorhabditis elegans embryo. J. Cell Biol. 129, 1071-1080. Goldstein, B., and Hird, S. N. (1996). Specification of the anteroposterioraxis in Caenorhabditis elegans. Development 122,1467-1474. Gruidl, M. E., Smith, P.A., Kuznicki, K. A., McCrone, J. S., Kirchner, J., Roussell, D. L., Strome, S., and Bennett, K. L. (1996). Multiple potential germ-line helicases are components of the germline-specific P granules of Caenorhabditis elegans. Proc. Natl. Acad. Sci. U.S.A. 93, 1383713842. Guo, S., and Kemphues, K. J. (1995). par-1, a gene required for establishing polarity in C. elegans embryos, encodes a putative Ser/Thr kinase that is asymmetrically distributed. Cell 81,611-620. Guo, S . , and Kemphues, K. J. (1996a). Molecular genetics of asymmetric cleavage in the early Caenorhabditis elegans embryo. Cur,: Opin. Genet. Dev. 6,408-4 15. Guo, S., and Kemphues, K. J. (1996b). A non-muscle myosin required for embryonic polarity in Caenorhabditis elegans. Nature (London)382,455-458. Hertwig, 0. (1885). Welchen einflufi uber die Schwerkraft auf die teilung der zellen. Jena 2. Narurwiss. 18, 175. Hibino, T., Nishikata, T., and Nishida, H. (1998). Centrosome-attractingbody: A novel structure closely related to unequal cleavages in the ascidian embryo. Dev. Growth Diffe,: 40,85-95. Hirao, Y. (1968). Cytological study of fertilization in Tubifex egg. Zool. Mag. (Tokyo) 77, 340346.

34

Hiroki Nishida et al.

Hird, S. N., and White, J. G. (1993). Cortical and cytoplasmic flow polarity in early embryonic cells of Caenorhabditis elegans. J. Cell Biol. 121, 1343- 1355. Hird, S. N., Paulsen, J. E., and Strome, S. (1996). Segregation of germ granules in living Caenorhabditis elegans embryos: Cell-type-specific mechanisms for cytoplasmic localization. Development 122,1303-1312. Holy, J., and Schatten, G. (1991). Differential behavior of centrosomes in unequally dividing blastomeres during fourth cleavage of sea urchin embryos. J. Cell Sci. 98,423-431. Horstadius, S. (1937). Investigations as to the localization of the micromere-, the skeleton-, and the entoderm-forming material in the unfertilized egg of Arbacia puncrulafa.Biol. Bull. 73,295316. Horstadius, S. (1939). The mechanics of sea urchin development studied by operative methods. B i d . Rev. 14,132-179. Horstadius, S . (1973). “Experimental Embryology of Echinoderms.” Clarendon, Oxford University Press, London. Hyman, A. A. (1989). Centrosome movement in the early divisions of Caenorhabditis elegans: A cortical site determining centrosome position. J. Cell Biol. 109, 1185-1 195. Hyman, A. A., and White, J. G. (1987). Determination of cell division axes in the early embryogenesis of Caenorhabditis elegans. J. Cell Biol. 105,2123-2135. Iseto, T., and Nishida, H. (1999). Ultrastructural studies on the centrosome-attracting body (CAB): The electron dense matrix and its role in unequal cleavages in ascidian embryos. Dev. Growth Dijfer., in press. Ishii, R., and Shimizu, T. (1995). Unequal first cleavage in the Tubifex egg: Involvement of a monoastral mitotic apparatus. Dev. Growth Difser. 37,687-701. Ishii, R., and Shimizu, T. (1997a).Equalization of unequal first cleavage in the Tubifex egg by introduction of an additional centrosome: Implications for the absence of cortical mechanisms for mitotic spindle asymmetry. Dev. Biol. 189,49-56. Ishii, R., and Shimizu, T. (1997b). Reproductive capability of maternal centrosomes in the Tubifex egg. 2001.Sci. (Tokyo) 14,961-968. Jan, Y. N., and Jan, L. Y. (1998). Asymmetric cell division. Nature (London) 392,775-778. Jeffery, W. R., and Bates, W. R. (1989). Ooplasmic segregation in the ascidian Styela. In “The Molecular Biology of Fertilization” (H. Schatten and G. Schatten, eds.), pp. 341-367. Academic Press, San Diego. Jeffery, W. R., and Meier, S. (1983). A yellow crescent cytoskeletal domain in ascidian eggs and its role in early development. Dev. €201.96, 125-143. Kemphues, K. J., and Strome, S. (1997). Fertilization and establishment of polarity in the embryo. In “C. elegans. 11” (D. L. Riddle, T.Blumenthal, B. J. Meyer, and J. R.Priess, eds.), pp. 335-360. Cold Spring Harbor Laboratory, New York. Kemphues, K. J., Priess, J. R.,Morton, D. G., and Cheng, N. (1988). Identification of genes required for cytoplasmic localization in early C. elegans embryos. Cell 52, 3 1 1 -320. Kirby, C., Kusch, M., and Kemphues, K. (1990). Mutations in the par genes of Caenorhabditis elegans affect cytoplasmic reorganization during the first cell cycle. Dev. Biol. 142,203-215. Kraut, R., Chia, W., Jan, L. Y.,Jan, Y. N., and Knoblich, J. A. (1996). Role of inscureable in orienting a symmetric cell division in Drosophila. Nature (London) 383,50-55. Levitan, D. J., Boyd, L., Mello, C. C., Kemphues, K. J., and Stinchcomb, D. T. (1994).par-2, a gene required for blastomere asymmetry in Caenorhabditis elegans, encodes zinc-finger and ATPbinding motifs. Proc. Natl. Acad. Sci. U.S.A. 91,6108-61 12. Lu, B., Jan, L. Y., and Jan, Y. N. (1998). Asymmetric cell division: Lessons from flies and worms. Curr. Opin. Genet. Dev. 8,392-399. Lutz, D. A., and InouB, S. (1982). Colcemid but not cytochalasin inhibits asymmetric nuclear positioning prior to unequal cell division. Biol. Bull. Mar: Biol.Lab. (Woods Hole) 163,373314.

1. Maternal Factors and Embryonic Cleavage Patterns

35

Lutz, D. A., Hamaguchi, Y., and InouB, S . (1988). Micromanipulation studies of the asymmetric positioning of the maturation spindle in Chaetopterus sp. oocytes: I. Anchorage of the spindle to the cortex and migration of a displaced spindle. Cell Motil. Cytoskel. 11,83-96. Meshcheryakov, V. (1976). Asymmetrical cytotomy and spindle orientation in early isolated blastomeres of gastropod mollusks. Ontogenez 7,558-565. Meshcheryakov, V. (1978a). Orientation of cleavage spindles in pulmonate molluscs. I. Role of blastomere shape in orientation of second cleavage spindles. Ontogenez 9,558-566. Meshcheryakov, V. (l978b). Orientation of cleavage spindles in pulmonate molluscs. 11. Role of structure of intercellular contacts in orientation of the third and fourth cleavage spindles. Ontogenez 9,567-575. Meshcheryakov, V. N., and Belussov, L. V. (1975). Asymmetrical rotations of blastomeres in early cleavage of gastropoda. Wilhelm Rouxk Arch. Enhu Org. 177, 193-203. Montgomery, M. K., and Fire, A. (1998). Double-stranded RNA as a mediator in sequence-specific genetic silencing and co-suppression. Trends Genet. 14,255-258. Morgan, T. H. (1894). Experimental studies on echinoderm eggs. Anut. Anzeig. 9,81. Morton, D. G.,Roos, J. M., and Kemphues, K. J. (1992). par-4, a gene required for cytoplasmic localization and determination of specific cell types in Cuenorhabditis elegans embryogenesis. Genetics 130,77 1-790. Nakatani, Y., and Nishida, H. (1994). Induction of notochord during ascidian embryogenesis. Dev. Biol. 166,289-299. Nelson, W. J., and Grindstaff, K. K. (1997). Cell polarity: Par for the polar course. Cur,: Biol. 7 , R562-R564. Nishida, H. (1986). Cell division pattern during gastrulation of the ascidian, Halocynthia roretzi. Dev. Growth Diff: 28, 191-201. Nishida, H. (1987). Cell lineage analysis in ascidian embryos by intracellular injection of a tracer enzyme. 111. Up to the tissue-restricted stage. Dev. Biol. 121,526-541. Nishida, H. (1992). Regionality of egg cytoplasm that promotes muscle differentiation in embryo of the ascidian, Halocynthia roretzi. Development 116,521-529. Nishida, H. (1994). Localization of determinants for the formation of the anterior-posterior axis in the eggs of the ascidian Halocynthia roretzi. Development 120,3093-3 104. Nishida, H. (1996). Vegetal egg cytoplasm promotes gastrulation and is responsible for specification of vegetal blastomeres in embryos of the ascidian Halocynthia roretzi. Development 122, 1271-1 279. Nishida, H. (1997). Cell fate specification by localized cytoplasmic determinants and cell interactions in ascidian embryos. fnt. Rev. Cytol. 176,245 -306. Nishikata, T., Hibino, T., and Nishida, H. (1999). The centrosome-attracting body, microtubule system, and posterior egg cytoplasm are involved in positioning of cleavage planes in the ascidian embryo. Dev. Biol.,in press. Okazaki, K. (1975). Normal development to metamorphosis. In “The Sea Urchin Embryo: Biochemistry and Morphogenesis” (G. Chihak, ed.), pp, 177-232. Springer-Verlag. Berlin and New York. Penners, A. (1922). Die furchung von Tubifex rivulorum Lam. Zool. Jb. Abt. Anat. Ontog. 43,323367. Ponting, C. P., Phillips, C., Davies, K. E., and Blake, D. J. (1997). PDZ domains: Targeting signaling molecules to sub-membranous sites. BioEssays 19,469-479. Rappaport, R. (1986). Establishment of the mechanism of cytokinesis in animal cells. fnt. Rev. Cytol. 105,245-281. Reverberi, G. (1971). Annelids. In “Experimental Embryology of Marine and Fresh-water Invertebrates” (G. Reverberi, ed.), pp. 126-163. North-Holland, Amsterdam. Rose, L. S . , and Kemphues, K. J. (1998a). Early patterning of the C. elegans embryo. Annu. Rev. Genet. 32,521-545. Rose, L. S., and Kemphues, K. J. (1998b). The let-99 gene is required for proper spindle orientation during cleavage of the C. elegans embryo. Development 125, 1337-1346.

36

Hiroki Nishida et al.

Sardet, C., Speksnijder,J., Inoue, S., and Jaffe, L. (1989). Fertilization and ooplasmic movements in the ascidian egg. Development 105,237-249. Sasakura,Y., Ogasawara, M., and Makabe, K. W. (1998a). HrWnt-5:A maternally expressed ascidian Wnt gene with posterior localization in early embryos. Inr. J. Dev. Biol. 42,573-580. Sasakura, Y., Ogasawara, M., and Makabe, K. W. (1998b). Maternally localized RNA encoding a serinelthreonineprotein kinase in the ascidian, Halocynthia roretzi. Mech. Dev. 76,161 - 163. Satoh, N. (1979). Visualization with scanning electron microscopy of cleavage pattern of the ascidian eggs. Bull. Ma,: Biol. Sm.Asamushi Tohoku Univ. 16,169-178. Satoh, N. (1994). “Developmental Biology of Ascidians.” Cambridge Univ. Press, London and New York. Satou, Y., and Satoh, N. (1997).posterior end mark 2 (pem-2).pem-4, pem-5, andpem-6 Maternal genes with localized mRNA in the ascidian embryo. Dev. Biol. 192,467-481. Sawada, T. (1988). The mechanism of ooplasmic segregation in the ascidian eggs. Zool. Sci. Tokyo 5, 667-675. Schatten,G. (1981). Sperm incorporation,the pronuclear migrations, and their relation to the establishment of the first embryonic axcis: Time-lapse microscopy of the movements during fertilization of the sea urchin Lytechinus variegatus. Dev. Biol. 86,426-437. Schierenberg, E. (1988). Localization and segregation of lineage-specific cleavage potential in embryos of Caenorhabditis elegans. Roux’s Arch. Dev. Biol. 197,282-293. Schnabel, R. (1991). Cellular interactions involved in the determinationof the early C. elegans embryo. Mech. Dev. 34,85-100. Schnabel, R., and Priess, J. R. (1997). Specification of cell fates in the early embryo. In “C. elegans. W (D. L. Riddle, T.Blumenthal, B. J. Meyer, and J. R. Priess, eds.), pp. 361-382. Cold Spring Harbor Laboratory, New York. Schnabel, R., Weigner, C., Hutter, H., Feichtinger, R., and Schnabel, H. (1996). men-1 and the general partitioning of cell fate in the early C. elegans embryo. Mech. Dev. 54, 133-147. Shimizu,T. (1982). Development in the freshwater oligochaete Tubifex. In “Developmental Biology of Freshwater Invertebrates” (F. W. Harrison et al., eds.), pp. 283-316. Alan R. Liss, New York. Shimizu, T. (1988). Localization of actin networks during early development of Tubifex embryos. Dev. Biol. 125,321-331. Shimizu,T. (1989). Asymmetric segregation and polarized redistribution of pole plasm during early cleavages in the Tubifex embryo: Role of actin networks and mitotic apparatus. Dev. Growrh Diffez 31,283-297. Shimizu, T. (1993). Cleavage asynchrony in the Tubifex embryo: Involvement of cytoplasmic and nucleus-associated factors. Dev. Biol. 157, 191-204. Shimizu, T. (1995). Role of the cytoskeleton in the generation of spatial patterns in Tubifex eggs. In “Current Topics in Developmental Biology” (D. Capco, ed.), Vol. 31, pp. 197-234. Academic Press, New York. Shimizu, T. (1996). Behaviour of centrosomes in early Tubifex embryos: Asymmetric segregation and mitotic cycle-dependent duplication.Row’s Arch. Dev. Biol. 205,290-299. Shimizu, T. (1998). Unequal cleavage in the early Tubifex embryo. Dev. Growth Di@z 40,257-266. Strome, S. (1986). Fluorescence visualization of the distribution of microfilaments in gonads and early embryos of the nematode Caenorhabditis elegans. J. Cell Biol. 103,2241-2252. Strome, S. (1993). Determination of cleavage planes. Cell 72,3-6. Strome, S . , and Wood, W. B. (1982). Immunofluorescence visualization of germ-line-specific cytoplasmic granules in embryos, larvae, and adults of Caenorhabditis elegans. Proc. Narl. Acad. Sci. U.S.A. 79, 1558-1562. Strome, S., Martin, P.,Schierenberg, E., and Paulsen, J. (1995). Transformationof the germ line into muscle in mes-1 mutant embryos of Caenorhabditis elegans. Development 121,2961 -2972. Sturtevant, A. H. (1923). Inheritance of direction of coiling in Lymnaea. Science 58,269-270.

1. Maternal Factors and Embryonic Cleavage Patterns

37

Sulston, J., Schierenberg,E., White, J., and Thomson, N. (1983). The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev. Biol. 100,67- 119. Tabuse, Y., Izumi, Y., Piano, F., Kemphues, K. J., Miwa, J., and Ohno, S. (1998). Atypical protein kinase C cooperates with PAR-3 to establish embryonic polarity in Caenorhabditis elegans. Development 125,3607-3614. Takahashi, H., and Shimizu, T. (1997). Role of intercellular contacts in generating an asymmetric mitotic apparatus in the Tubifex embryo. Dev. Growth Diffex 39,351-362. Tanaka, Y. (1976). Effects of the surfactants on the cleavage and further development of the sea urchin embryos: I. The inhibition of micromere formation at the fourth cleavage. Dev. Growth Diffex 18,113-122. Tanaka, Y. (1979). Effects of the surfactants on the cleavage and further development of the sea urchin embryos: II. Disturbance in the arrangement of cortical vesicles and change in cortical appearance. Dev. Growth DiffeK 21,331-342. Tenenhaus, C., and Schubert, C., and Seydoux, G. (1998). Genetic requirements for PIE-I localization and inhibition of gene expression in the embryonic germ lineage of Caenorhabditis elegans. Dev. Biol. 200,212-224. Uemura, I., and Endo, Y. (1976). Electron microscopic observations in the extragranular zone of the embryo of the sea urchin, Hemicentrotus pulcherrimus. Dev. Growth Diffex 18,399-406. Vlahou, A,, Gonzalez-Rimbau,M., and Flytzanis, C. N. (1996). Maternal mRNA encoding the orphan steroid receptor SpCOUP-TF is localized in sea urchin eggs. Development 122,521-526. Waddle, J. A,, Cooper, J. A,, and Waterston, R. H. (1994). Transient localized accumulation of actin in Caenorhabditis elegans blastomeres with oriented asymmetric divisions. Development 120, 2317-2328. Wang, W., and Shakes, D. C. (1996). Molecular evolution of the 14-3-3 protein family. J. Mol. Evol. 43,384-398. Wang, S . W., Griffin, F.J., and Clark, W. H., Jr. (1997). Cell-cell association directed mitotic spindle orientation in the early development of the marine shrimp Sicyonia ingentis. Development 124, 773-780. Watts, J. L., Etemad-Moghadam, B., Guo, S., Boyd, L., Draper, B. W., Mello, C. C., Priess, J. R., and Kemphues, K. J. (1996). par-6, a gene involved in the establishmentof asymmetry in early C. elegans embryos, mediates the asymmetric localization of PAR-3. Development 122,3 133-3 140. White, J., and Strome, S. (1996). Cleavage plane specification in C. elegans: How to divide the spoils. Cell 84, 195-198. Wilson, E. B. (1925). “The Cell in Development and Heredity.” Macmillan, New York. Wu, S.-L.,Staudinger, J., Olson, E. N., and Rubin, C. S. (1998). Structure, expression, and properties of an atypical protein kinase C (PKC3) from Caenorhabditis elegans. J. Biol. Chem. 273, 11301143. Yoshida, S., Marikawa, Y., and Satoh, N. (1996). posterior end murk, a novel maternal gene encoding a localized factor in the ascidian embryo. Development 122,2005-2012.

This Page Intentionally Left Blank

3 Multiple Endo-l,4-p-~-glucanase(Cellulase) Genes in A rabidopsis Elena del Campillo Department of Cell Biology and Molecular Genetics University of Maryland at College Park College Park, Maryland 20742

I. Introduction 11. Cellulase Genes in General 111. How Does Cellulase Relate to Cellulose and the Plant Cell Wall? More Questions Than Answers IV. Cellulase Genes in Plants V. Molecular Characterization of EGase Genes in Arabidopsis VI. Expression of Three Distinct EGase Genes in Arabidopsis Tissues VII. EGase and Cell Growth VIII. EGase Mutants in Arubidopsis IX. Conclusions References

The plant cell wall is modified in coordination with almost all plant developmental processes. Modifications in the cell wall are thought to be mediated by cell wall hydrolases, including those encoded by a large family of genes specifying endo- 1,443D-glucanases (EC 3.2.1.4), which participate in the breakdown of p- 1,4 glucosidic linkages. The enzymes expected to modify cellulose, commonly referred to as cellulases, are encoded by members of this gene family. In Arubidopsis the endo- 1,4-p-~-glucanase(EGase) gene family is extensive (more than 12 members) and encompasses structurally different classes of genes encoding proteins with contrasting enzyme functions. Within the family there are enzymes located at the plasma membrane that are presumed to act at the innermost layers of the cell wall, and enzymes that are secreted and are presumed to act at any stratum within the cell wall, including the outermost layer. Both structural gene groups are members of the glycosyl hydrolase gene Family 9. Evidence suggests that EGases anchored in the plasma membrane play a role in cell wall biosynthetic processes, presumably by editing cellulose synthesis or during the assembly of the cellulose-hemicellulose network. Those EGases that are extracellular play specific roles in cell wall catabolic processes and their activity ranges from partial and localized to massive and catastrophic. This range in activity is linked to processes such as cell growth and cell death, respectively. For all Arubidopsis EGases nothing is known about their true in vivo substrate, mode of action, or to what extent they can act on cellulose or other p-1,4 glucans. The study of the EGase gene family is in its infancy, and because of the possible agronomic implications this group of genes deserves continued attention. o 1999 Academic press. Currpnr Topics in Develupmemzl Biologv. V d . 46 Copynght 0 1999 by Academic Prehs. All rights of reproduction in any form reserved 0070-2153/99 $30.00

39

40

Elena del Campillo

1. Introduction All plant cell walls are composed of the same basic structural polymers-cellulose, hemicelluloses, pectins, and proteins-and yet within a plant, cell walls differ among different cell types and even among cell faces. During plant development, all cell walls are modified in coordination with cell changes in plant size and shape. Also, complex plant processes such as growth, cell separation, or cell death are linked to changes in the cell wall (Sexton el al., 1989; Fischer and Bennett, 1991). The main structural elements of all plant cell walls are cellulose microfibrils, and yet it is unknown if they sustain alterations as changes in the cell wall take place. Changes in the cell wall can range from partial and localized (causing a realignment of cellulose microfibrils?) to massive and catastrophic (causing cellulose breakdown?). These changes are thought to be mediated by cell wall hydrolases, including those encoded by a large family of endo-l,4-pD-glucanase (EC 3.2.1.4)genes, often referred to as cellulases. Little is known about the role of cellulase genes in plant development and to what extent they can mediate cell wall changes. Much more is known about the activity of wall enzymes associated with chemical modifications in hemicelluloses, pectins, and proteins than is known about enzymes affecting the main structural elements of the plant cell wall, the cellulose microfibrils. In fact, we do not know if the plant genes referred to as cellulases can cause hydrolysis of cellulose or other p-1,4 glucans of the plant cell wall, or to what extent they may do so. The intention of this review is to present a current view about the endo-1p-pD-glucanase (cellulase) genes in plants, to clarify their nomenclature, and to summarize the occurrence of these genes in Arubidopsis as known to date. The types of gene sequences that have been referred to as cellulases and how many classes of genes are comprised within this gene family will be described. The deduced properties of each gene, in terms of primary structure, molecular characteristics, and phylogenetic relationships of encoded proteins, will be analyzed. Finally, the functions of the few genes that have been studied and the efforts from various laboratories to isolate a collection of T-DNA insertional mutants to understand ultimately the diversity of functions associated with this gene family will be described. For prokaryotic and fungal forms of the enzymes, readers are referred to excellent reviews on these topics in Gilkes et ul., 1991; Gilbert and Hazelwood, 1993; and BCguin and Aubert, 1994.

It. Cellulase Genes in General Cellulases are, according to the International Union of Biochemistry-Molecular Biology (IUB-MB) Enzyme Nomenclature, enzymes that can degrade cellulose. They are 0-glycosyl hydrolases that cleave the p-1 ,4-glucosidic bonds between

2. Arabidopsis Endo- 1,4-P-D-glUCanaSeGenes

41

two glucose moieties. This definition, based on substrate specificity, encompasses several types of enzymes, such as endoglucanases, cellobiohydrolases, and exoglucanases. The endoglucanases comprise a large group of proteins with different specificities and modes of action. Based on amino acid sequence similarities and hydrophobic cluster analysis, the endoglucanases (EC 3.2.1.4) have been classified into 12 families, each one characterized by a protein motif or signature (Henrissat, 1991; Henrissat and Bairoch, 1993, 1996; Davies and Henrissat, 1995). Each of the 12 families is specified by a numerical denomination as follows: Family 5, Family 6, Family 7, Family 8, Family 9, Family 12, Family 44, Family 45, Family 48, Family 5 1, Family 60, and Family 61. The only family that is relevant to the discussion in this review is Family 9. The catalytic core of bacterial cellulases from Family 9 (formerly referred to as cellulase E family) (Henrissat and Bairoch, 1993) consists of about 300-400 amino acid residues and contains two conserved regions and residues important for catalytic activity (BCguin and Aubert, 1994). The first region, referred to as catalytic signature 1, contains an active-site histidine and the second region, referred to as catalytic signature 2, contains two catalytically important residues: an aspartate (D) and a glutamate (E) (Tomme et al., 1992). Hydrolysis of the glycosidic bond takes place via acid catalysis and requires both a proton donor and a nucleophile/base. The catalytic nucleophile/base is the Asp and the proton donor is the Glu (Baird et al., 1990). The consensus patterns of the catalytic signatures of Family 9 are as follows (Davies and Henrissat, 1995): Catalytic region 1: Catalytic region 2:

[STV]-x-[LIVMFYI-[STV]-x(2)-G-x-[NKRIx(~)-[PLIVM]-H-X-R [FYWI-x-D-x(4)-[FYW]-x(3)-E-x-[STA]-x(3)-NWAI

The substrate specificity of cellulases within a family is variable and most bacterial genomes contain cellulase genes from different families. Some cellulases can degrade native cellulose and are referred to as C1, and some hydrolyze only soluble substituted cellulose derivatives and are referred to as Cx. The Cl-cellulases, capable of complete hydrolysis of crystalline cellulose, consist of a catalytic domain joined to a cellulose-binding domain (CBD) by a short linker sequence rich in proline and/or hydroxy amino acids. The CBD confers the ability to attack amorphous regions of the substrate and consists of about 105 amino acid residues (Meinke et al., 1991). The CBD domain is found either at the N-terminal or at the C-terminal extremity of these enzymes and contains two conserved cysteines, one at each extremity of the domain linked by a disulfide bond. There are also four conserved tryptophan residues that may be involved in the interaction of the CBD with polysaccharides. The consensus pattern for the CBD is as follows:

xCxxxxWxxxxxNxxxWxxxxxxxWxxxxxxxxWNxxxxxxGxxxxxxxxxxCx

42

Elena del Campillo

111. How Does Cellulase Relate to Cellulose and the Plant Cell Wall? More Questions Than Answers The basic structural features of cellulose relative to the action of cellulase relate to the physical complexity of the cellulose molecule, which can exist in crystalline and noncrystalline forms. Cellulose microfibrils are composed of several dozen parallel and linear chains of p- 1,4-glucan. The adjacent glucan chains are tightly cross-linked by intermolecular hydrogen bonds. About two-thirds of each microfibril may be considered crystalline. The remaining one-third is less ordered, has no regular hydrogen bonds between adjacent chains, and is considered paracrystalline. Part of the paracrystalline cellulose is localized at the surface of the microfibrils (Hayashi et al., 1994). Thus, many things are known about the physical structure of cellulose microfibrils in plants, and yet there are many unanswered questions about the action of cellulase on the cellulose substrate. Is paracrystalline cellulose the main cellulase substrate? Are paracrystalline regions of cellulose more conspicuous in places where cellulose microfibrils show a high degree of curvature, such as cell comers or pith regions? The major structural variations in cellulose of different tissues are in the degree of polymerization and in the degree of crystallinity of the glucan chain (Delmer and Amor, 1995). What determines the length of cellulose microfibrils is unknown. Are cellulase genes involved in editing the length of glucan chains? The microfibrils are laid down perpendicular to the axis of growth and they wrap transversally or in a shallow helix around the axis of growth. Microfibrils of cellulose show a high degree of curvature and appear kinked in some areas of bridging or interactions with other molecules (McCann and Roberts, 1994; McCann et al., 1994). However, we do not know the contribution of external bracing by xyloglucan as opposed to internal modification of the microfibril (by cellulase?) to microfibril bending. Xyloglucans appear to enhance the amount of cellulose in a noncrystalline state by interfering with the tendency of cellulose to self-associate into a crystalline state (Hayashi et al., 1994). Moreover, xyloglucans also appear to activate some forms of endo-l,4-P-glucanase (Maclachlan and Brady, 1992). The plant cell wall has two opposed faces displaying compositional differences and bordering different environments: the inner face fronts the plasma membrane and the outer face fronts the middle lamella. The difference between these two layers becomes apparent during elongation because it brings about a reorientation of cellulose microfibrils as well as deposition of new strata of microfibrils primarily in the inner layers of the wall (McCann and Roberts, 1994). Thus, in the inner layers of the wall, the orientation of microfibrils is perpendicular to the axis of growth. In the outer layers the reorientation of microfibrils is less pronounced and the microfibrils are stretched apart. However, because the wall appears to retain a near uniform thickness during growth, it is likely that microfibrils of the outer layers may merge to a certain extent, allowing the spacing

2. Arabidopsis Endo- 1,4-(3-~-glucanase Genes

43 between microfibrils to remain constant (Carpita and Gibeaut, 1993; McCann and Roberts, 1994). How do changes of cellulose come about in this complex environment? Does a change in orientation of cellulose microfibrils require a participation of cellulase or any other chemical modification? Does curvature of cellulose microfibrils require an editing of nascent cellulose molecules as they are released outside of the plasma membrane?

IV. Cellulase Genes in Plants To date, approximately 30 plant genes in the GenBank DNA sequence database have cellulase as a key word. The genes are found in a broad range of species, including tomato, orange, bean, peach, pepper, Arabidopsis, and trees. Of those, only very few have been biochemically characterized and the evidence indicates that they do not work in vitro either on crystalline cellulose or on cellobiose (Brummell et al., 1994). They cleave internal P-glucosidic bonds at random in soluble cellulose derivatives such as carboxymethyl cellulose (CMcellulose), but hydrolysis of xyloglucans is either absent (Durbin and Lewis, 1988; O’Donoghue and Huber, 1992) or very slow (Ohmiya et al., 1995). Little is known about their true in vivo substrate, mode of action, and to what extent they can act on cellulose or other P- 1,4 glucans of the plant cell wall. Very few studies in plants have dealt with the action of cellulase on the insoluble cellulose matrix. The traditional enzyme assays are inadequate to monitor cellulose hydrolysis because even conditions of extensive hydrolysis do not produce appreciable quantities of aqueous-soluble products due to the strong hydrogen bonding of cellulose chains (O’Donoghue et al., 1994). The uncertainty about the substrate specificity of enzymes of this gene family has led to ambiguities about their nomenclature. For example, in a current literature survey the products of the 30 plant genes that have cellulase as a key word were referred to as end0-P- 1,4-glucanase (EGase), P- 1,4-glucanhydrolase (cellulase), glycosyl hydrolase Family 9, and P-glucanase. However, all 30 genes are characterized by specifying the presence of the same active-site signatures peculiar to the glycosy1 hydrolase Family 9 described above and several conserved sequence elements unique to this family (see below). The enzymes are all approximately 500-600 amino acids long, lack a cellulose-binding domain signature, and contain signature 1 and signature 2, which are predicted to be important for catalysis. Based on these common structural characteristic all 30 plant genes are defined as members of a single family, the glycosyl hydrolase gene Family 9 (Henrissat and Bairoch, 1993). This endo-P-1 ,Cglucanase family is structurally distinct from the endoglucanases that hydrolyze xyloglucans, i.e., xyloglucan endotransglycosylases (XETs) (Xu et al., 1996), or hydrolyze cereal P-Dglucans containing 1,3 and 1,4 glycosidic linkages (Woodward and Fincher, 1982; Hoj et al., 1989).

44

Elena del Campillo

The term “cellulase” was originally used to refer to products of this gene family because, decades ago, the first members isolated were detected by monitoring endohydrolytic enzyme activity on a cellulose matrix (CM-cellulose) and were purified from protein extracts by affinity chromatography on a cellulose column (Koehler et d., 1981; Bennett and Christoffersen, 1986). Among the first enzymes of this family characterized as pure proteins were bean abscission cellulase, also referred to as 9.5 cellulase (Durbin and Lewis, 1988), fruit avocado Cx-cellulase (Awad and Lewis, 1980), and two cellulases from auxin-treated pea epicotyls (Byme et al., 1975). The term “cellulase” was also used to describe an enzyme isolated from bean petioles that was found to be associated with the plasma membrane (Koehler et al., 1976). This bean enzyme was distinct from bean abscission cellulase, and was active in most bean tissues (Lewis and Koehler, 1979) but activity decreased prior to petiole abscission (del Campillo et al., 1988). Genes for bean abscission cellulase and fruit avocado Cx-cellulase were the first members cloned, sequenced, and shown to share extensive sequence similarity to the catalytic core of endo- 1,4-p-glucanase genes from cellulolytic bacteria (Cass et al., 1990; Tucker and Milligan, 1991). For example, the avocado fruit Cx-cellulase gene shows extensive homology with the gene for Avicelase I, a Clostridium cellulase that can degrade crystalline cellulose (Jauris et aZ., 1990).A later study by O’Donoghue et ul. (1994) demonstrated that the enzyme encoded by the avocado fruit Cx-cellulase gene can modify in vivo the noncrystalline regions of cellulose microfibrils. It has become apparent that within a plant genome there are multiple members of the glycosyl hydrolase gene Family 9. For example, as known to date, 7 members have been identified in tomato (Maclachlan and Brady, 1992; Catala et aZ., 1997) and 12 in Arubidopsis (see below) and by sequence comparison they have been grouped into at least three structurally different classes (see below). Should all the enzymes of this gene family be referred to as cellulases?Why would a plant require different classes of cellulase genes? Do they represent functional differences among cellulases such that specificity could vary among those that modify cellulose from different locations within the cell wall and to different extents? In fact, for all these genes nothing is known about their products regarding the true in vivo substrate, mode of action, and to what extent they can act on cellulose or other p-1,4 glucans. Moreover, it is completely unknown whether structural differences among divergent classes signify changes in specificity or mode of action. The only information that can be used with certainty to define this gene family in plants is based on the primary structure and not on the function of their products. Thus, a gene family in plants can be defined with all the gene sequences that encode the active-site signatures peculiar to the glycosyl hydrolase Family 9, plus the six amino acid strings that are found in highly conserved positions in all plant enzymes, i.e., Q-[KRI-S-G-[KRI-L-P, L-x-G-G-Y-Y-D-A-G-D, D-H-xCW-[EMQVI-R-P-E-D-M, E [TMV] AAA[FLM]A-A-A-S[ILM] [VA] F, P-NP-N, and G-A-x-V-x-G-P. How should the proteins of this family be named? Cer-

2. Arubidopsis Endo- 1,CP-~-glucanaseGenes

45

tainly the term “cellulase” should be reserved for those endo-@-1,4-glucanases that can modify the embedded cellulose of the plant cell wall. In the past 5 years, a good compromise was reached by adopting the term “endo-1,4-@-~-glucanase” (EGase) to refer to the proteins, and by adopting the abbreviation “cel” as a prefix for the gene name. Still some authors prefer to use the term “cellulase” and some prefer to use both names, i.e., endo-1,4-@-glucanase(cellulase) (Osborne and Henderson, 1998). The latter has the advantage of recognizing important studies carried out in the past wherein cellulase referred to proteins encoded by members of this gene family (Fan and Maclachlan, 1966, 1967; Verma et ul., 1975; Lewis and Varner, 1970). For convenience, EGase will be used herein.

V. Molecular Characterization of EGase Genes in Arabidopsis Currently, with just 30% of the Arubidupsis genome sequenced, the BLAST algorithm finds 12 different but related EGase genes. This suggests that the EGase gene family in Arubidupsis is composed of more than 12 members. Some EGase genes were found in BAC clones that are completely sequenced, and others were reported separately by individual researchers. Table I identifies 10 of the 12 EGase genes by the name of the BAC clone in which the gene resides and 2 genes by the name of the cDNA reported in GenBank. The latter do not yet match any sequenced BAC clone. The table presents some of the properties of the encoded enzymes as deduced from the primary structure of the mature proteins. The comparison shows that among the 12 enzymes there are important differences in p l and degree of N-ASN glycosylation. Table I also shows that among the 12 genes there are differences in gene organization, such that some members contain three introns and others members can have four, five, and up to six introns. Among the products of the 12 Arubidupsis EGase genes shown in Table I, at least two structurally distinct groups can be identified. One group comprises soluble secreted proteins with predictable N-terminal signal peptides and the other comprises proteins without a cleavable signal peptide, which are predicted to be 5 p e I1 (Ncyt Cexo) membrane proteins. The Q p e I1 membrane proteins are usually plasma membrane proteins. The secreted mature proteins have few or no Nglycosylation sites, and are of two types: (1) highly basic, pZ > 9.0, or (2) highly acidic, pZ < 5.5. The EGases with an uncleavable signal peptide are predicted to have toward the N-terminal end a cytoplasmic hydrophilic tail of around 80 amino acids followed by a short hydrophobic transmembrane domain that would anchor the protein to a membrane. The predicted membrane-anchored EGases are basic proteins, which are highly glycosylated. Both structural groups are members of the cellulase Family 9 (Henrissat and Bairoch, 1993). It is noticeable, however, that in the group Cel T21L14.7, F411.37, F411.36, F16B22.5, and F16B22.6, only signature 1 is fully conserved,

Table I Current List of Full-Length EGase Genes Found in the Arubidopsis Genome and Their Protein Propertiesa

Locus CelFl9G 10.16 Cell CelT2H3.5 CellT26J12.2 Ce12 CelF16B22.6 CelF16B22.5 CelF411.36 CelT21L14.7 CelF4I 1.37 CelT1705 CelF5114.14

Exon

Intron

PI

MWb

9.36 9.19 9.05 8.11 7.20 5.46 5.16 8.55 5.42 5.90 8.96 9.31

50.5* 51.2* 55.2* 5 1.2* 52.0* 50.5*

50.1* 50.4* 55.9* 53.4 69.1 69.8

ASNglycosylation 1

1 1 -

1

2 ~

8 10

Signature 1

Signature 2

-

+ + + + + + + + + + +

-

+ +

Signal peptide Cleavahle (1 to 27) Cleavahle (1 to 30) Cleavable (1 to 19) Cleavable (1 to 3 1) Cleavable (1 to 3 1) Cleavable (1 to 30) Cleavable (1 to 28) Cleavable (1 to 28) Cleavable (1 to 24) No signal peptide? No signal peptide No signal peptide

Additional motif

-

Cell attachment Cell attachment Cell attachment Cell attachment Membrane anchoring Membrane anchoring

aGenBank accession numbers for the EGase genes: CelF19GIO.16, AF000657; Cell, X98544; CeIT2H3.5, AF075597; CellT26J12.2, AC002311; Ce12, AF034573; CelF16B22.6, AC003672; CelF16B22.5,AC003672; CelF411.36, AC004521; CeIF411.37,AC004521; CeIT21L14.7, AC003033; CelTI705, B27804; CelF5114.14, AC001229. '*, Molecular weight after cleavage of the signal peptide.

2. Arabidopsis Endo- 1,4-P-~-glucanaseGenes

47

whereas signature 2 has all the amino acids of the consensus pattern but one substitution that is not conserved and hence is not recognized as a signature. In CelF19G10.16 only signature 2 is fully conserved and signature 1 has all the amino acids of the consensus pattern but one substitution that is not conserved. The biochemical significance of changing residues in the catalytic signature is unknown. It is also interesting that those EGases that have only signature 1 fully conserved also have the Arg-Gly-Asp (RGD) motif, which has been identified as a cell attachment signature. This tripeptide is found in fibronectin and is crucial for the interaction with its integrin cell surface receptor (Ruoslahti and Pierschbacher, 1986; d’Souza et ul., 1991). It is also found in a number of proteins that have been shown to play a role in cell adhesion. The alignment of the amino acids specified by all Arubidopsis EGase genes (Fig. 1) shows that the encoded enzymes exhibit approximately 20% positional identity in amino acid sequence and at least five distinct ungapped segments (blocks) that are highly conserved. The distribution of these five blocks in the primary structure is shown schematically below using the letters A through E to identify each block, where each hyphen represents 10 amino acids and every letter represents 12 amino acids. ---------AAA-B-CCCC--------------------DDDDD-EEEEBlocks D and E include the two amino acid signatures located at the C-terminal end thought to be involved in the enzymatic hydrolysis of P-glucosidic bonds. These signatures contain an active-site histidine and two conserved carboxylic acids: an aspartate (D) and a glutamate (E) (Tomme et al., 1992). Block A includes the string L-x-G-G-Y-Y-D-A-G-D, which is also found and conserved in products of cellulase genes from microbes; however, in bacteria the first Y is replaced by a conservative substitution (W). The presence and proximity of the two conserved carboxylic residues (D) in this string suggest that they could be involved in metal binding. Block C includes the string D-H-x-CW[EMQVI-R-P-E-D-M and Block E includes the string P-N-P-N. Both strings are highly conserved among all the plant EGase gene products but not in cellulase gene products from microbes. The hydrophobic string = E[TMV]AAA[FLM]AA-A-S [ILM][VA] F is also highly conserved in both plant and microbe gene products. These conserved amino acid blocks have been very useful to design specific primers, which were then used to identify EGase genes in a variety of plant species (Lashbrook and Bennett, 1992). The alignment in Fig. 1 also highlights two polypeptides, CelT1705 and CelF5114, which are around 100 amino acids longer than the other ten. These longer proteins show a unique stretch of 80 amino acids at the N-terminal domain and a unique stretch of 30 amino acids rich in proline residues at the C-terminus. A phylogenetic tree generated from the alignment of the deduced amino acid sequences of all proteins encoded by Arubidopsis EGase genes (Fig. 2) reveals

49

2. Arabidopsis Endo- 1,4+-~-glucanase Genes

-

7

AraCelT21L14.7

2 s

‘k

AraCelF411.37 AraCelF411.36 AraCelF16B22.5 61

AraCelFl6B22.6

+AraCelTl705 Fig. 2 Phylogeny of the Arabidopsis EGase family. The tree was constructed using parsimony as the optimality criterion and depicts the predicted relationship between the members of the Arabidopsis EGase family. The protein alignment was determined by multiple sequence alignment with hierarchical clustering (Corpet, 1988) for the region defined by Asp-28 to Phe-485 of Arabidopsis Cell. Using PAUP 3.1.1, a heuristic search was performed using simple stepwise addition, TBR branch swapping, MULPARS ON, and a stepmatrix for amino acid substitution from PROTPARS (Felsenstein, 1991). The outgroup was defined as the Arabidopsis CelT1705 and Arabidopsis CelF5114.14 proteins, and the branch length is presented above the branches. The length of the minimal tree is 1772. The origin of each homolog and the amino acid region used for the comparison are represented as AraCell (28-485), AraCelT26J12.2 (28-485), AraCelT2H3.5 (53-509), AraCel2 (43-498). AraCelF19G10.16 (24-479). AraCelT21L14.7 (39-502), AraCelF411.37 (33-490), AraCelF411.36 (33-490). AraCelF16.B22.5 (33-490), AraCelF16.B22.6 (30-491), AraCelT1705 (108-586), and AraCelF5114.14 (116-590).

the existence of three distinct classes. The nonsecreted CelT1705 and CelF5114.14 polypeptides constitute the most divergent class. A second class includes all secreted acidic EGases and those members that have only signature 1 fully conserved, i.e., CelT21L14.7, CelF411.37, CelF411.36, CelF16B22.5, and CelF16B22.6.The third class includes all secreted EGases with p1higher than 9.0, and comprises at least three subgroups within the class: (1) Cell and CelT26J12.2,

Fig. 1 Multiple-alignmentchart for the Arabidopsis EGase gene family of proteins. Sequence analysis and alignment of 12 full-length deduced amino acid sequences for EGase members found in BAC genomic clones and genes registered by individual researchers. Identical residues are represented by black shading and similar residues by grey shading.

50

Elena del Campillo

(2) Ce12 and CelT2H3.5, and (3) CelF19G10.16.The position of CelF19G10.16 in the third class is marginal but the probability of the five genes forming a class is 98%. The confidenceof the grouping of CelF411.36, CelF411.37,CelF16B22.5, and CelF16B22.6 in the second class and the subgroups of Cell and CelT26J12.2 and Ce12 and CelT2H3.5 in the first class is very strong and is confirmed by bootstrapping.The functional significance of organization into classes and subgroups is unknown. Some information could be obtained if the comparison is extended to include EGases from other plant species in which expression patterns as well as biochemical studies have been conducted. A phylogenetic tree generated from alignment of the deduced amino acid sequences encoded by all plant EGase genes, including those from Arubidopsis, is shown in Fig. 3. The tree reveals the existence of at least four distinct classes. Interestingly, one class is encoded solely by five Arubidopsis EGase genes, suggesting that the comparable sequences in other plants have not been found yet. TornCell BeanBAC1

II 1

AraCelT26J12.2 PrunCel2 PrunCell

~

55 TomCelA

TomCel5 PoplarCel CitrusAcdCel AvoCell TomCel7 PeaCel CitrusABCel AraCelFl9GlO. 16 AraCelT21 L14.7 AraCelF4I1.37 AraCelF411.36 AraCelF16822.5 AraCelFl6822.6

Fig. 3 Phylogeny of the plant EGase family. The tree was constructed as described for Fig. 2 for the region defined by Asp-28 to Phe-485 ofArubidopsis Cell. The outgroup was defined as the Arubidopsis CelT1705, Arubidopsis CelF5114.14, and tomato Ce13 proteins and the branch length is presented above the branches. It is one of four trees; the length of the minimal tree is 3395. The origin of each homolog and the amino acid region used for the comparison are represented as TomCel1 (26-478), PeppercaCell (25-477), ElderJETl (33-485), BeanBACl (40-492), TomCel2 (27-484), PeppercaCe13 (24-481), AraCell (28-485), AraCelT26J12.2 (28-485), PrunCel2 (52-510). PrunCel1 (46504). TomCel4 (46-503), AraCelT2H3.5 (53-509), AraCel2 (43-498), TomCel5 (34-493), PoplarCel (30-489), CitrusAcdCel (39-499). AvoCell (28-486). TomCel7 (19-473, PeaCel (27-482), CitrusABCel (27-483), AraCelF19G10.16 (24-479). AraCelT21L14.7 (39-502), AraCelF411.37 (33-490), AraCelF411.36 (33-490), AraCelF16.B22.5 (33-490). AraCelF16.B22.6(30-491), TomCe13 (108-586). AraCelT1705 (108-586), andAraCelF5114.14(116-590).

2. Arabidopsis Endo- 1,4-f%D-ghCanaSe Genes

51

Another class, in contrast, is encoded by genes from several plant species, but no comparable members from A rabidopsis have been found. All EGases from this class have been associated with the development of ethylene-induced abscission. The fourth class is the largest and includes at least four subgroups. One group includes CelF19G10.16, designated a P-glucanase in GenBank, and two IAA-stimulated P- 1,4-endoglucanases, one from pea (EGL1) (Wu et al., 1996) and the other from tomato (Ce17) (Catala et al., 1997). This would suggest that CelF19G10.16 is also an IAA-stimulated P- 1,4-endoglucanase from Arabidopsis. Another group includes TomCel4 with AraCel2 and CelT2H3.5. TomCel4 is a developmentally regulated P- 1,bendoglucanase found in the pistil of tomato flowers (Milligan and Gasser, 1995). The most divergent class comprises the membrane-anchored endo-P-l,4-glucanase from tomato (TomCel3) and the two polypeptides encoded by Arabidopsis genes (AraCelT 1705 and AraCelF5114.14), which, as shown in Table I, have no predictable signal peptide and are likely membrane associated as well. The tomato Ce13 protein was shown to be associated with both golgi and plasma membrane (Brummell et al., 1997a,b) and is very closely related to the Arabidopsis gene referred to as AraCelT1705 here. The function of tomato Ce13 is unknown but it is abundant in vegetative tissues undergoing rapid growth. Thus, it has been suggested that this cellulase could be part of the enzymatic complex related to synthesis of cellulose, as demonstrated in Agrobacterium tumefaciens (Mathysse et al., 1995). The genes Cell and CelT26J12.2 are closely related and show close homology to tomato Ce12 and Capsicum annuum Ce13. Both genes increase during tomato and pepper fruit ripening, respectively. The EGase genes appear to be distributed throughout the Arabidopsis genome. Several EGase genes have been mapped to the upper arm of chromosome I, others map close the centromere of chromosome IV. The lower arm of chromosome I1 has the highest densities of EGase genes found so far and, interestingly, all of them are characterized by encoding only the catalytic signature 1. This includes EGase gene T21L14.14, found at around 62.5 cM, very close to COPI, and a string of four genes that are very similar, are tandemly arranged, and appear to have arisen by gene duplication. This includes the genes F16B22.5, F16B22.6, F411.37, and F411.36. The gene TI705 corresponds to the mRNA for cellulase OR16pep, and is found in chromosome V, 1 cM from the ngaZ29 marker. The differences among Arabidopsis EGases with respect to their gene organization, their primary structure, and the indications that members of this family have different places of action suggest that different members are likely to be involved in distinct functions. In fact, EGases have been correlated in processes whereby EGase activity is limited to discrete modifications, such as those implicated in growth (Fan and Maclachlan, 1966, 1967; Shani et aE., 1997; Inouhe and Nevins, 1991), and in processes whereby EGase activity is extensive, such as those implicated in xylem differentiation (Sheldrake, 1970; Murmanis, 1978), in fruit ripening (Fischer and Bennett, 1991), in abscission (Sexton et al., 1989),

52

Elena del Campillo

and flower reproductive organs (del Campillo and Lewis, 1992). Those that are extracellular and nonglycosylated may reach the outermost layer of the cell wall for their action, whereas those that are positioned in the plasma membrane may be constrained to act only at the innermost layer of the cell wall. Some EGases may modify xyloglucans and some may modify cellulose from different locations within the cell wall and to a different extent. Furthermore, it is possible that the type and number of EGases involved in a process determine the extent of cellulose breakdown and determine also the nature of the intermediate products generated, which in turn may serve as intertissue signals (Lorences et al., 1990).

VI. Expression of Three Distinct EGase Genes in Arabidopsis Tissues EGases have been implicated in both cell growth and cell death processes. Such contrasting functions for the same enzyme activity very likely correspond to EGase genes with a unique biochemical specialization. Such unique specialization is anticipated to show a spatial and temporal regulation of gene expression in accordance with function. Figure 4 shows differential expression of three Arubidopsis EGase genes, CelTl705, CelnH3.5, and CelT21L14.7, as a function of tissue type, developmental stage, and light growth conditions. Note that Arabidopsis hypocotyl growth in dark or light entails primarily longitudinal or radial growth and does not significantly entail cortical or epidermal cell division (Gendreau et ul., 1997). Thus, at the end of the third week, seedlings exposed to continuous light show long roots, green cotyledons, and a short hypocotyl, whereas seedlings exposed to continuous darkness show short roots, yellow, small cotyledons, and a long, etiolated hypocotyl. Growth continues through the second week in both light and dark but in the third week growth is continued only in the light and is stalled in the dark. Figure 4 shows that when growing in continuous light, the CeZTl705 message is present in all tissues and it is most abundant in 1-week-old roots and 3-weekold shoots and stems. More interestingly, expression of this gene in hypocotyls grown in the dark increases 10-fold from the first to the second week, just when hypocotyls are undergoing elongation, but when elongation ceases (at 3 weeks) there is a concomitant halt in the accumulation of this message. Expression in roots was lower but also increased steadily throughout the 3 weeks. In contrast, the CelT2H3.5message is less abundant than the CelTI 7 0 5 message and it is expressed primarily in flowers, flower stems, and in roots of 1-week-old lightgrown seedlings. This gene shows 65% sequence identity to tomato Ce15, a gene that correlates with in pluntu flower abscission (del Campillo and Bennett, 1996). The CelT21Ll4.7 message localizes exclusively in roots of light-grown seedlings and its function is completely unknown. Differential expression of these three genes would suggest that each one per-

53

2. Arabidopsis Endo- 1,4+-~-glucanase Genes DARK Shoots

LIGHT Roots

Shoots

Roots

Sh St Fw Sil

1w 2w 3w 1w 2w 3w 1w 2w 3w 1w 2w 3w 3w 3w 3w 3w ---------------

Fig. 4 Differential expression of members of three distinct classes of EGase genes in Arabidopsis. Northern blot analysis showing ethidium bromide staining of total RNA (bottom) isolated from roots and shoots (Sh) collected after 1,2, and 3 weeks of growth, including flower stems (St), flowers (Fw), and siliques (Sil) of Arabidopsisplants grown in soil for 6 weeks under continuous light. Samples were electrophoresed in an agarose gel and blotted onto a nylon membrane. The blot was hybridized sequentially, first with a 550-bp CelTl7M-specific cDNA probe extending through the 5' untranslated region and the N-terminal membrane-anchoring domain. After removal of the first probe, the blot was probed with a 1200-bp CelT2H3.5 cDNAprobe ranging from half of the gene to the 3' untranslated region. After removal of the second probe, the blot was probed with a 1200-bp CelnlL147cDNAprobe also extending from half the gene to the 3' untranslated region are shown.

forms a specialized biochemical function. That these three EGase genes belong to three distinct phylogenetic groups also suggests that functional diversity among EGase groups is linked to their structural divergence.

54

Elena del Campillo

VII. EGase and Cell Growth The biochemistry of cell wall changes during growth and expansion is complex and is not clearly understood. It is generally accepted that these processes involve cuts and rearrangements of cell wall polymers. Cellulase was among the first enzyme activities correlated with auxin-induced cell elongation. Studies initiated by Maclachlan over 20 years ago concluded that cellulases induced by auxin function in the cleavage of xyloglucans (Fan and Maclachlan, 1966,1967; Hayashi et al., 1984). However, a recent study by Tominaga et al. demonstrated that the apoplast of auxin-treated stems at the early stage of elongation accumulates cello-oligosaccharides in addition to xyloglucan solubilization products. These authors concluded that the cello-oligosaccharides were products of cellulose degradation (Tominaga et al., 1998). The role of cellulase in elongation was questioned after the discovery of two cell wall proteins: expansin, which can interfere with hydrogen bonding between cellulose microfibrils (McQueen-Mason et al., 1992; McQueen-Mason and Cosgrove, 1995), and XET, which can break and join xyloglucan polymers (Fry et al., 1992; Cosgrove and Duratchko, 1994). Cosgrove and Duratchko (1994) found no correlation between wall extension and wall autolysis by endogenous cell wall hydrolases, in an in vitro assay for long-term wall extension (creep). However, they could not rule out the possibility that endoglycanases could have a role in cell wall creep if the products produced by these enzymes remain attached to the wall, as appears to be the case for plant Cx-cellulases (O’Donoghue et al., 1994). Furthermore, they found that addition of fungal cellulases to native walls enhances wall extension, which suggests that cellulases and pectinases may act synergistically with expansin to enhance wall extension. It is unknown if, and to what extent, cellulose microfibrils are modified during cell elongation or radial expansion. Several studies have shown that during growth the cellulose microfibrils from the innermost layers of the wall undergo extensive reorientation, which suggests that this could be mediated by enzymes located in the plasma membrane (Carpita and Gibeaut, 1993; McCann and Roberts, 1994). One possible mechanism for reorientation is that cellulases acting at the innermost layers of the wall modify the length of cellulose microfibrils. However, much less is known about the genes involved in elongation and how their expression relates to growth in living tissues. The link between elongation and EGase expression was strengthened with the cloning of an endo-P-1,4-glucanase gene (EGLl)from pea. The expression of this gene is most abundant in elongating epicotyls of etiolated seedlings and increases 10fold when epicotyl segments are incubated with a synthetic auxin (Wu et al., 1996). Subsequently, characterization of the Arabidopsis Cell gene as encoding an elongation-specific P-1,4-glucanase was reported (Shani et al., 1997). This gene is highly expressed in elongating zones of flowering stems of normal plants, whereas expression in the corresponding zones of dwarf flowering

2. Arabidopsis Endo- 1,4-P-D-glucanaseGenes

55

stems is significantly lower. Furthermore, transgenic tobacco plants transformed with the putative Cell promoter region fused to the GUS reporter gene show significant GUS staining both in the shoot and root of elongating zones (Shani et al., 1997). These results further substantiate the link between Cell expression and plant cell elongation. Questions of whether EGases are involved in radial expansion as a form of growth remain unexamined. It is known that radial expansion increases in response to exogenously applied ethylene (Abeles et al., 1992). Thus, an EGase involved in radial expansion is likely to be an ethylene-dependent activity that performs discrete, or “fine tuning,” modifications on the cell wall.

VIII. ECase Mutants in Arabidopsis A search for EGase mutants in Arabidopsis is imperative to understand the diversity of EGase genes in this plant. One strategy is to analyze T-DNA insertional (knockout) mutants, by polymerase chain reaction (PCR)-based reverse genetics (from the gene to phenotype). The basic approach for their isolation consists of first identifying large mutant pools, then subpools, and ultimately a single plant, wherein it is possible to amplify by PCR the junctions between the T-DNA insert and one EGase gene (McKinney et al., 1995;Azpiroz-Leehan and Feldmann, 1997). Once a pool is identified, a new round of PCR reactions is performed for each of the lines comprising the pool. After each round, the PCR products are analyzed by size and by Southern blot hybridization using EGase-specific probes (Mckinney et al., 1995). Using this method, several knockout EGase mutant lines have been obtained from the collection of T-DNA insertional mutants of Dr. Michael Sussman, in Madison, Wisconsin (E. del Campillo and S. Patterson, unpublished data). The initial search identified mutant pools in which the T-DNA landed in members of both structural classes of EGase genes, those that encode secreted proteins and those that encode membrane-associated proteins. At the final steps of the search, three independent mutant lines for the gene CelF9G10.16 and two allelic mutants of the CelTl705 gene were selected. For all these lines, the analysis of PCR products suggests that the T-DNA has landed either in the promoter or the N-terminal region. A mutant in the promoter region usually suppresses gene expression and yields a null function. However, tissue specificity or gene regulation could also be altered. The segregating population of one line (Cs 6474-2B-11) in which the T-DNA affected gene CelTl705 shows that approximately 20-25% of the germinating seedlings have a dwarf phenotype. The digital capture of video microscope images of this mutant is shown in Fig. 5. The mutant grows slowly, has a short stem compared to the wild type, and the cotyledons are puffy, light green, and look like antlers. The characterization of this mutant as well as the three CelF9G10.16 mutants is in progress.

56

Elena del Campillo

A report described an Arabidopsis EGase mutant referred to as KORRIGAN (Nicol and Hofte, 1998). KOR is a recessive mutation identified by short hypocotyls in dark-grown plants but a nearly wild-type adult phenotype. The sequence of the KOR gene indicates that it encodes an endo-l,4-P-~-glucanase (Nicol and Hofte, 1998; Nicol et aL, 1998). The KOR sequence predicts an integral protein with a short amino terminus in the cytosol and an external catalytic domain. KOR is very closely related to the tomato Ce13 membrane-anchored EGase (Brummell et al., 1997b). Thus KOR corresponds to the gene referred to as Ara CelTl705 in this review. Another Arabidopsis EGase mutant (T-DNA insertional), designated DEC (defective cytokinesis), has been identified. In this mutant the T-DNA landed in the promoter region of the DEC gene. The DEC gene also corresponds to the putative gene for membrane-anchored EGase, referred to as Ara CelTl705 in this review. The DEC mutation caused typical cytokinesis defects, leading to the transformation of adult organs into calli shortly after germination. This strongly suggests that DEC primarily functions in cytokinesis and that proper physical separation of dividing daughter cells is essential for cell differentiation (N.-H. Chua, personal communication). Overall, the results from studies of KOR and DEC suggest that Ara CelTI 7 0 5 is involved in both elongation and cell division and may also be involved in morphogenesis. It is not clear how the cleavage of a P-1,4 glucan would participate in elongation, facilitate the proper cell separation during cytokinesis, or even affect cell shape. A possible explanation is that in vivo CelT1705 may be part of a multisubunit complex involved in editing steps during cell wall biosynthesis. Specifically, CelT1705 may be involved in editing cellulose synthesis and the extent of cellulose trimming may determine the nature of intermediate products, which in turn may serve as developmental intertissue signals (Lorences et al., 1990). Nicol et al. (1998) have proposed that KOR plays a central role in the assembly of the cellulose-hemicellulose network in the expanding cell wall. There are at least two Arabidopsis genes encoding proteins with a transmembrane domain. What is the function of the gene CelF5114, the other gene encoding a putative membrane-anchored protein? Is there any redundancy of the functions of CelFSI14 and CelTl705?

>

Fig. 5 Digital capture of video microscope images of an EGase T-DNA insertional mutant line. Seeds of the Arabidopsismutant line CS 6474-28- 11 were surfacesterilizedby successiveimmersion in 95% ethanol for 10 min, 50% sodium hypochlorite, 0.05%Triton-X100for 5 min, and followed by five rinses with sterile water. Seeds were layered onto 90-mmplastic petri dishes containing half strength basic salt nutrients,pH 5.7,l mllliter of Gambor’svitamins and 0.8% phytoagar.The plates were wrapped with parafilm and aluminum foil and stored at 4°C for 24 hr. Plates were then transferred to a growth chamber and grown in continuous light at 22-24°C. The mutant grows slowly and has a short stem compared to the wild type (A). Within the population of small seedlings, an unusual cotyledon morphology (B-D) is observed in many but not in all seedlings.The cotyledonsare puffy, light green, look like antlers, and sometimes develop an unusual white outgrowth at the tip of one cotyledon (C).

M

This Page Intentionally Left Blank

2. Arabidopsis Endo- 1,4-P-~-glucanase Genes

57

IX. Conclusions Arubidopsis endo- 1,4+-~-glucanases (EGases) comprise a diverse family of enzymes that participate in the breakdown of p-1,4 glucosidic linkages. A variety of polymeric molecules contain this linkage in the plant cell wall, and cellulose is one among them. Thus, the enzymes expected to modify cellulose, commonly referred to as cellulases, are members encoded by a large EGase gene family. The Arubidopsis EGase gene family is extensive (more than 12 members) and encompasses structurally different classes of genes encoding contrasting enzyme functions ranging from cell plate formation and cell elongation to cell-cell separation. Within the EGase family there are members located at the plasma membrane and presumed to act at the innermost layers of the cell wall, and members that are extracellular and presumed to act at any stratum within the cell wall, including the outermost layer. Evidence from mutant analysis suggests that one Arubidopsis EGase anchored in the plasma membrane, Ara CelT 1705, may play a specific role in cell wall biosynthetic processes. Specifically, CelT1705 may be involved in editing cellulose synthesis. The mechanism by which this might occur is unknown, but analysis of the regulation governing this enzyme, as well as the characterization of the substrates and conditions required by Ara CelT1705 will elucidate these issues. EGase genes that are involved in cell wall biosynthetic processes are expected to be up-regulated during growth and downregulated during processes of cell wall disassembly. EGases that are extracellular are likely to function primarily from the outermost layers of the cell wall and play specific roles in cell wall catabolic processes. Both structural groups encompass several genes and it is unknown whether this multiplicity is a reflection of gene redundancy or a reflection of different biochemical specializations within members of a class. Multiplicity of gene members and contrasting functions are features of other plant cell wall genes and proteins. For example, within the extensive expansin gene family (22 members) there are distinct members expressed during elongation and others that are expressed during degradation processes (Rose et ul., 1997). Similarly, members of the XET gene family have been correlated in contrasting processes such as growth, fruit ripening, and airspace formation (Antosiewicz et ul., 1997). The EGase family is the only gene family encoding cell wall hydrolases that have representative members anchored to the plasma membrane. In the future, analysis of EGase knockout mutants at physiological and biochemical levels will be one important approach to define the function of the extensive EGase gene family. Knockout mutants provide strong evidence for gene function because they may have a modified phenotype consistent with the hypothesized function of the gene and show how such genes are regulated and coordinated throughout the life of the plant. If disruption of a particular gene does not have an effect on the expected phenotype because of redundancy, then crosses between individual knockouts may have the expected effect. This approach is

58

Elena del Campillo

currently being pursued in a collaborative effort between the University of Maryland and the University of Wisconsin at Madison. For all Arabidopsis EGase gene products nothing is known about their true in vivo substrate, mode of action, and to what extent they can act on cellulose or other p- 1,Cglucans. Moreover, it is completely unknown whether structural differences among divergent classes signify changes in specificity or mode of action. Thus, it is also imperative to address the question of substrate specificity. Efforts should be made to express, purify, and analyze the activity and substrate specificity of recombinant endo-p- 1,4-glucanases in yeast and bacteria. Studies should address which members modify cellulose versus other plant cell wall p1,4 glucans. Furthermore, site-directed mutagenesis should be used to elucidate the functional significance of each conserved amino acid motif found in the postulated active site as well as outside the catalytic site. Cellulose is the most abundant polymer on earth and all cellulolytic microbes and fungi that use cellulose as a carbon source produce a complex enzymatic system composed of a variety of cellulases with different specificities and modes of action. Plants need to modify cellulose during changes in cell shape and size occurring at different points in their life cycle. For those changes, plants require EGase genes that choreograph either the editing of new cellulose synthesis or the breaking down of the amorphous cellulose already present in the cell wall. In addition, plants also need to modify hemicellulose-containingp- 1,4-glucans as part of many developmental processes associated with cell wall changes. For these processes plants need EGase genes with specialized biochemical functions. The study of this gene family is in its infancy and the agronomic importance of the genes and their enzymes should encourage continued study.

References Abeles, F., Morgan, P., and Saltveit, J. (1992). In “Ethylene in Plant Biology, pp. 147-156. Academic Press, San Diego. Antosiewicz, D., Purugganan, M., Polisensky, D., and Braam, J. (1997). Cellular localization of Arubidopsis xyloglucan endotransglysosilase-relatedproteins during development and after wind stimulation. Plant Physiol. 115, 1319-1328. Awad, M., and Lewis, L. (1980). Avocado cellulase: Extraction and purification. J. Food Sci. 45, 1625- 1628. Azpiroz-Leehan, R., and Feldrnann, K.(1997). T-DNA insertion mutagenesis in Arubidopsis: Going back and forth. Trends Genet. 13,4. Baird, S., Hefford, M., Johnson, D., Sung, W., Yaguchi, M., and Seligy, V. (1990). The Glu residue in the conservedAsn-Glu-Pro sequence of two highly divergent endo-P-l,4-glucanases is essential for enzymatic activity. Biochem. Biophys. Res. Commun. 169, 1035- 1039. BCguin, P., and Aubert, J.-P. (1994). The biological degradation of cellulose. FEMS Microbiol. Rev. 13,25-58. Bennett, A., and Christoffersen,R. E. (1986). Synthesis and processing of cellulase from ripening avocado fruit. Plant Physiol. 81,830-835. Brummell, D., Lashbrook, C. C., and Bennett, A. B. (1994). Plant endo-P-1,4-glucanases:Structure,

2. Arabidopsis Endo- 1,CP-~-glucanaseGenes

59

properties and physiologicalfunction. In “Enzymatic Conversion of Biomass for Fuels Production” (M. Himmel, J. Baker, and R. Overend, eds.), Vol. 566, pp. 100-129. American Chemical Society,Washington. Brummell, D., Bird, C., Schuch, W., and Bennett, A. (1997a). An endo-l,4-p-glucanase expressed at high levels in rapidly expanding tissue. Plant Mol. Biol. 33,87-95. Brummell, D., Catala, C., Lashbrook, C., and Bennett, A. (1997b). A membrane-anchoredE-type endo- 1,4-P-glucanaseis localized on golgi and plasma membranes of higher plants. Proc. Narl. Acad. Sci. U.S.A. 94,4794-4799. Byme, H., Christou, N., Verma, D., and Maclachlan, G. (1975). Purification and characterization of two cellulases from auxin-treated pea epycotyls. J. Biol. Chem. 250,1012-1018. Carpita, C., and Gibeaut, D. (1993). Structural models of primary cell walls in flowering plants: Consistency of molecular structure with the physical properties of the walls during growth. Plant J. 3, 1-32. Cass, L., Kirven, K., and Christoffersen, R. (1990). Isolation and characterization of a cellulase gene family member expressed during avocado fruit ripening. Mol. Gen. Genet. 223,76-86. Catala, C., Rose, J. K. C., and Bennett, A. B. (1997). Auxin-regulation and spatial localization of an endo-l,4-p-glucanase and a xyloglucan endotransglycosylasein expanding tomato hypocotyls. Plant J. 12,417-426. Corpet, F. (1988). Multiple sequence alignment with hierarchical clustering.Nucleic Acids Res. 16, 10881- 10890. Cosgrove, D., and Duratchko, D. (1994). Autolysis and extension of isolated walls from growing cucumber hypocotyls. J. Exp. Bot. 45,1711-1719. Davies, G., and Henrissat, B. (1995). Structures and mechanisms of glycosyl hydrolases. Structure 3, 853-859. del Campillo, E., and Bennett, A. (1996). Pedicel break strength and cellulase gene expression during tomato flower abscission. Plant Physiol. 111,813-820. del Campillo, E., and Lewis, L. (1992). Occurrence of 9.5 cellulase and other hydrolases in flower reproductive organs undergoing major cell wall disruption. Plant Physiol. 99, 1015-1020. del Campillo, E., Durbin, M., and Lewis, L. (1988). Changes in two forms of membrane-associated cellulase during ethylene-inducedabscission. Plant Physiol. 88,904-909. Delmer, P. D., and Amor, Y. (1995). Cellulose biosynthesis. Plant Cell 7,987-1000. d’Souza, S., Ginsberg, M. H., and Plow, E. F. (1991). Arginyl-glycyl-asparticacid (RGD): A cell adhesion motif. Trends Biochem. Sci. 16,246-250. Durbin, M., and Lewis, L. (1988). Cellulases in Phaseolus vulgaris. Methods Enzymol. 160,342351. Fan, D., and Maclachlan, G. (1966). Control of cellulase activity by indolacetic acid. Can. J. Bot. 44, 1025-1034. Fan, D., and Maclachlan, G. (1967). Massive synthesis of ribonucleic acid and cellulases in the pea epicotyl in response to indolacetic acid with or without concurrent cell division. Plant Physiol. 42, 1114-1122. Felsenstein, J. (1991). PHYLIP Manual C Phylogeny Inference Package, version 3.4. University of Washington (distributed by author), Seattle, 69-75. Fischer, R., and Bennett, A. (1991). Role of cell wall hydrolases in fruit ripening. Annu. Rev. PZant Physiol. Plant Mol. Biol. 42,675-703. Fry, S., Smith, R., Renwick, K., Martin, D., Hodge, S., and Matthews, K. (1992). Xyloglucan endotransglycosilase, a new wall-loosening enzyme activity from plants. Biochem. J. 282,821 -828. Gendreau, E., Traas, J., Desnos, T., Grandjean, O., Caboche, M., and Hoffte, H. (1997). Cellular basis of hypocotyl growth in Arabidopsis thaliana. Plant Physiol. 114,295-305. Gilbert, H., and Hazlewood, G. P.(1993). Bacterial cellulases and xylanases. J. Gen. Microbiol. 139, 187-194. Gilkes, N., Henrissat, B., Kilburn, D. R. C., Miller, J., and Warren, R. (1991). Domains in microbial p1.4-glycanases: Sequence conservation function and enzyme families. Microbiol. Rev. 55,303-3 15.

60

Elena del Campillo

Hayashi, T., Wong, W.-S., and Maclachlan, G. (1984). Pea xyloglucan and cellulose. 11. Hydrolysis by pea endo-~-1,4-glucanase.Plant Physiol. 35,419-424. Hayashi, T., Takeda, T., Ogawa, K., and Mitsuishi, Y. (1994). Effects of the degree of polymerization on the binding of xyloglucansto cellulose. Plant Cell Physiol. 35,893-899. Henrissat, B. (1991). A classification of glycosyl hydrolases based on amino-acid sequence similarities. Biochem. J. 280,309-316. Henrissat, B., and Bairoch, A. (1993). New families in the classificationof glycosyl hydrolases based on amino acid sequence similarities.Biochem. J. 293,781-788. Henrissat, B., and Bairoch, A. (1996). Updating the sequence-based classificationof glycosyl hydrolases. Biochem. J. 316,695-696. Hoj, P., Hartman, D., Monice, N., Doan, D., and Fincher, G. (1989). Plant Mol. Biol. 41,339-367. Inouhe, M., and Nevins, D. (1991). Inhibition of auxin-induced cell elongation of maize coleoptiles by antibodies specific for cell wall glycanases. Plant Physiol. 96,426-431. Jauris, S., Rucknagel, K., Schwarz, W., Kratzsch, P., Bronnenmeier, K., and Staudenbauer, W. (1990). Sequence analysis of the Clostridiun stercorarium CelZ gene encoding a thermoactive cellulase (Avicelase I): Identification of catalytic and cellulose-bindingdomains. Mol. Gen. Genet. 223,258-267. Koehler, D., Leonard, R., Vandenvoude, W., Linkins, A., and Lewis, L. (1976). Association of a latent cellulase activity with plasma membrane from kidney bean abscission zones. Plant Physiol. 58,324-330. Koehler, D., Lewis, L., Shannon, L., and Durbin, M. (1981). Purification of abscission zone cellulase. Phytochemistry 20,409-412. Lashbrook, C., and Bennett, A. (1992). Functional analysis of Cx-cellulase (endo-P-1,4-glucanase) gene expression in transgenic tomato fruit. In “Cellular and Molecular Aspects of the Plant Hormone Ethylene” (J. Pech, A. Latche, and C. Balague, eds.), pp. 123-128. Agen, France. Lewis, L., and Koehler, D. (1979). Cellulase in the kidney bean seedling. Planta 146, 1-5. Lewis, L., and Varner, J. (1970). Synthesis of cellulase during abscission of Phaseolus vulgaris leaf explants. Plant Physiol. 46, 194-199. Lorences, E. P.,McDougall, G. J., and Fry, S. C. (1990). Xyloglucan and cello-oligosaccharides.Antagonist of the growth promoting effect of H+.Physiol. Planta,: 80, 109-1 13. Maclachlan, G., and Brady, C. (1992). Mutiple forms of 1,4-P-glucanasein ripening tomato fruits include a xyloglucanase activatable by xyloglucan oligosaccharides.A m . J. Plant Physiol. 19, 137-146. Mathysse, A., White, S., and Lightfoot, R. (1995). Genes required for cellulose synthesis in Agrobacterium tumefaciens. J. Bacteriol. 117, 1069-1075. McCann, M., and Roberts, K. (1994). Changes in cell wall architecture during cell elongation. J. Exp. Bor. 45,1683-1691. McCann, M., Wells, B., and Roberts, K. (1994). Direct visualization of cross-links in the primary plant cell wall. J. Cell Sci. 96,323-334. Mckinney, E., Nazeem, A., Traut, A., Feldmann, K., Belostotsky, D., McDowell, J., and Meaguer, R. (1995). Sequence-basedidentification of T-DNA insertion mutations in Arabidopsis: Actin mutants act2-1 and acr4-I. Plant J. 8,6213-622. McQueen-Mason, S., and Cosgrove, D. J. (1995). Expansin mode of action on cell walls. Plant Physiol. 107,87-100. McQueen-Mason, S., Durachko, D., and Cosgrove, D. (1992). Endogenous proteins that induce cell wall expansion in plants. Plant Cell 4,1425-1433. Meinke, A., Gilkes, N., Kilbum, D., Miller, R. J., and Warren, R. (1991). Protein Seq. Data Anal. 4, 349-353. Milligan, S., and Gasser, C. (1995). Nature and regulation of pistil-expressed genes in tomato. Plant. Mol. Biol. 28,69 1-7 11. Murmanis, L. (1978). Breakdown of end walls in differentiatingvessels of secondary xylem in Quercus rubra L. Ann. Bot. 42,679-682.

2. Arabidopsis Endo- 1,CP-~-glucanaseGenes

61

Nicol, F., and Hofte, H. (1998). Plant cell expansion: Scaling the wall. Curr @in. Plant Biol. 1,1217. Nicol, F., His, I., Jauneau, A., Vernhettes, S., Canut, H., Hofte, H. (1998). A plasma membranebound putative endo-l,4-P-~-glucanaseis required for normal wall assembly and cell elongation in Arabidopsis. EMBO J. 19,5563-5576. O’Donoghue,E., and Huber, D. (1992). Modification of matrix polysaccharides during avocado (Persea americanu) fruit ripening: An assessment of the role of Cx-cellulase. Physiol. Plant 86, 33-42. O’Donoghue,E., Huber, D., Timpa, J., Erdos, G., and Brecht, J. (1994). Influence of avocado (Persea americana) Cx-cellulase on the structural features of avocado cellulose. Planta 194, 573-584. Ohmiya, Y., Takeda, T., Nakamura, S., and Hayashi, T. (1995). Purification and properties of a wallbound endo-1,4-P-glucanasefrom suspension cultured poplar cells. Plant Cell Physiol. 34,607614. Osborne, D., and Henderson, J. (1998). Ethylene as the initiator of intertissue signalling and gene expression cascades in ripening and abscission of oil palm fruit. “Biology and Biotechnology of the Plant Hormone Ethylene.” 11. Santorini, Greece. Rose, J., Lee, H., and Bennett, A. (1997). Expression of a divergent expansin gene is fruit-specific and ripening-regulated.Proc. Natl. Acad. Sci. U.S.A. 94,5955-5960. Ruoslahti, E., and Pierschbacher, M. D. (1986). Arg-Gly-Asp: A versatile cell recognition signal. Cell44,517-518. Sexton, R., Tucker, M., del Campillo, E., and Lewis, L. (1989). The cell biology of bean leafabscission. In “Proceedings of the NATO Advanced Research Workshop on Cell Separation in Plants” (D. Osborne and M. Jackson, eds.), pp. 69-78. Springer-Verlag, Berlin and New York. Shani, Z., Dekel, M., Tsabary, G., and Shoseyov,0. (1997). Cloning and characterizationof elongation specific endo-l,4-beta-glucanase (cell) from Arabidopsis thaliana. Plant Mol. Biol. 34,837842. Sheldrake,A. (1970). Cellulase and cell differentiationin Acerpseudoplatanus. Planta 95, 167-178. Tominaga, R., Samejima, M., Sakai, F.,Hayashi, T. (1999). Occurrence of cello-oligosaccharidesin the apoplast of auxin-treated pea stems. Plant Physiol. 119,249-254. Tomme, P., van Beeumen, I., and Claeyssens, M. (1992). Modification of catalytically important residues in endoglucanase D from Clostridium thermocellum. Biochem. J. 285,3 19-324. Tucker, M., and Milligan, S. (1991). Sequence analysis and comparison of avocado fruit and bean abscission cellulases. Plant Physiol. 95,928-933. Verma, D., Maclachlan, G., Byrne, H., and Ewings, D. (1975). Regulation and in vitro translation of messenger ribonucleic acid for cellulase from auxin-treated pea epicotyls. J. BioZ. Chem. 250, 1019- 1026. Woodward, J., and Fincher, G. (1982). Purification and chemical properties of two 1,3; 1,4-P-glucan endohydrolases from germinating barley. Eur J. Biochem. 121,663-669. Wu, S.-C., Blumer, J., Darvill, A., and Albersheim, P. (1996). Characterizationof an endo-b-1,Cglucanase gene induced by auxin in elongating pea epicotyls. Plant Physiol. 110, 163-170. Xu, W., Campbell, P., Vargheese, A., and Braam, 3. (1996). The Arabidopsis XET-related gene family: Environmental and hormonal regulation of expression. Plant J. 9(6), 879-889.

This Page Intentionally Left Blank

3 The Anterior Margin of the Mammalian Gastrula: Comparative and Phylogenetic Aspects of Its Role in Axis Formation and Head Induction Christoph Viebahn Institute of Anatomy Rheinische Friedrich- Wilhelms-Universitat 531 15 Bonn, Germany

I. Introduction 11. Morphology

A. The Blastocyst B. Implantation C. Amnion Formation D. Gastrula Size E. The Embryonic Disc F. The Anterior Margin of the Gastrula G. Epiblast or Hypoblast Differentiation? H. Earlier Descriptions of Anterior Differentiation 111. Changes at the Anterior Margin during Development A. Changes Leading to Anterior Marginal Crescent Formation B. Changes Following Anterior Marginal Crescent Formation IV. The Anterior Margin in Different Vertebrate Classes A. Birds B. Amphibia C. Bony Fish V. Gene Expression Related to the Anterior Margin A. Early Gastrulation Stages B. Late Gastrulation Stages VI. A View on Phylogenetic Implications A. Homology of the Lower Layer B. Yolk Accumulation in the Vegetal Hemisphere C. Generation of Extraembryonic Membranes VII. Conclusions and Outlook A. Organizing Gastrulation B. Gastrulation Staging C. Evolution References

Current Topics in Developmental Biology. Vol. 46 Copyright B 1999 by Academic Press. All rights ofreproduction in any form resewed W70-2153/99 $30.00

63

64

Christoph Viebahn Recent findings on morphology and gene expression in several mammalian embryos suggest that there is a new landmark and possibly a center with organizer activity in the anterior margin of the embryo at the onset of gastrulation. This review compiles morphological variations and similarities found among mammals during gastrulation stages and, at the same time, stresses the common aspects, at the morphological and the molecular level, of setting up the body plan with regard to axis formation and head induction. Both morphological and functional aspects are then used to draw comparisons with equivalent developmental stages in lower vertebrate species, such as birds, amphibia, and bony fish. Finally, a suggestion is made as to how gastrulation may have evolved in the vertebrate phylum. B 1999 Academic press.

1. Introduction The background of this review, gastrulation, is a cornerstone in the development of all higher animal species. The word “gastrula” is the diminutive of the Greek word for stomach (4yola~+p),and was coined for the stage in amphibian development signaled by the appearance of cells that form the inner surface of the adult stomach and intestines (Haeckel, 1874). The concept that the primary role of gastrulation is to construct a major portion of the milieu interieur of the embryo has not changed over the past 100 years (Keibel, 1917; Lubosch, 1931; Peter, 1941; Wolpert, 1992; De Robertis et al., 1994), though in modem times, the term “gastrula” has become more encompassing, referring to the appearance of the future craniocaudal axis of the fetus and the establishment of the three germ layers, ectoderm, mesoderm, and endoderm, from the epiblast. The future craniocaudal axis of the fetus, called the primitive streak, has until recently been the only criterion known to herald the onset of gastrulation. Its appearance in the posterior half of the round embryonic disc simultaneously fixes two body coordinates, the anteroposterior (future craniocaudal) axis, and the left-right axis. Functionally, the primitive streak is the amniote equivalent of the blastoporus in amphibia, which, during elongation, serves as the gateway through which epiblast cells enter and emerge on the inner surface of the epiblast as mesoderm and endoderm. The remainder of the epiblast is, at the same time, converted into the ectoderm. The anteriormost region of the streak condenses into a structurereferred to as the primitive node, or simply “node,” first described in the rabbit and guinea pig (Hensen, 1876). The node has the impressive capacity of being able to induce or “organize,” under experimental conditions, the complete set of definitive axial organs, among which are the central nervous system and the future vertebral column (notochord) (Spemann and Mangold, 1924). Some molecular components of this ability have been described in several vertebrates (Cho et al., 1991;Blum et al., 1992; Izpisba-Belmonte etal., 1993), and it has become clear that the organizing activity of the node is a conserved feature of all vertebrates (De Robertis et al., 1994). Thus, the appearance of the node may, in addition to formation of the craniocaudal axis and three primary germ layers, be regarded as a major hallmark of gastrulation.

3. Anterior Margin of the Mammalian Gastrula

65 Yet another potentially major feature of gastrulation has been described in mammals, first by morphology in the rabbit, where it was referred to as the “anterior marginal crescent” (Viebahn et al., 1995), and then by gene expression in the mouse (Rosenquist and Martin, 1995; Hermesz et al., 1996;Thomas and Beddington, 1996). Comparison of morphological data from a variety of mammalian embryos, including the mouse, indicates that this early anterior differentiation is a general phenomenon in mammals, as will be shown further below. Together, these observations add a new dimension to the current view of gastrulation, which is that the appearance of the primitive streak is no longer the starting point of gastrulation in mammals. Rather, the anterior marginal crescent may be its first morphological manifestation in this vertebrate class. Closely associated with early anterior midline structures during further development are the development of the heart and the liver (Davis, 1927; Fishman and Chien, 1997; Landry et al., 1997) and several proteins that are candidates to be considered components of a head organizer, distinct from a trunk organizer (Bouwmeester and Leyns, 1997; Glinka et al., 1998). In view of all these findings it seems warranted that the present review focuses on the anterior margin, attempting to link up existing descriptions of the early mammalian embryonic disc with contemporary molecular data. Thereby, striking morphological differences will become apparent on examination of the early gastrula stages of different mammalian species, and, because of the long tradition of morphological embryology, a historical overview of the subject matter will be created alongside. However, the survey is intended mainly to open up an embryological perspective ranging from the earliest signs of the body axes to the emerging head organizer, and it is meant to generate the theoretical basis for elucidating the factors that initiate the laying down of our body plan. We may find ourselves considering that, in addition to axis formation, the second aspect of gastrulation (i.e., the specification of the germ layers) also begins with a differentiation at the anterior margin. As an intriguing counterpoint, structures of the mammalian brain may be specified within the early anterior margin, making the brain the first definite organ to be specified during embryogenesis, whereas the margin may end up as something quite unspectacular, such as the root of our umbilical cord. On the evolutionary level, this “precocious” anterior differentiation of the mammalian embryo prompts a fresh look at phylogenetic interrelationships regarding gastrulation within the vertebrate, particularly the chordate, phylum.

II. Morphology A. The Blastocyst

At the transition from the blastocyst to the gastrula stage the mammalian embryo can be divided morphologically into two cellular parts that have principally different developmental fates: the trophoblast, which will give rise to extraembry-

66

Chnstoph Viebahn

onic tissues such as the peripheral parts of the placenta and fetal membranes, and the embryoblast or embryonic disc, from which all other fetal tissues are derived. The trophoblast has a polar part covering the embryoblast and a mural part that surrounds the blastocyst cavity. In most species it is either the polar or the mural trophoblast that will engage in the attachment to the maternal tissues (see below). The embryoblast consists at this stage of two epithelial cell layers, which, again, have different developmental fates. The upper layer is the epiblast (alternative terms are embryonic ectoderm, primitive ectoderm, or primary ectoderm) and is the sole source for the germ layers (ectoderm, mesoderm, endoderm) and hence for all tissues of the embryo proper (Gardner and Rossant, 1979).The lower layer, facing the blastocyst cavity, is the hypoblast (embryonic endoderm, primitive endoderm, primary endoderm, or visceral endoderm) and will differentiate into the (extraembryonic) yolk sac epithelium (see below) and thus transform the blastocyst cavity into the yolk sac cavity. To this end, the hypoblast will be “pushed aside” by the (definite) endoderm that ingresses from the epiblast at the anterior end of the primitive streak, i.e., from the primitive node, and intercalates into this lower layer (Beddington, 1981; Poelmann, 1981; Lawson et al., 1986; Kadokawa et al., 1987). In most mammals the yolk sac epithelium (or extraembryonic endoderm) is a simple uniform sheet of cells initially covering the inside of the mural trophoblast. Later it is separated from the trophoblast by proliferating extraembryonic mesoderm and the enlarging extraembryonic coelomic cavity and it will form a yolk sac of varying sizes. In the rodent embryo, however, the germ layer inversion (egg cylinder; see Section II,B) has specific consequences for the shape of the yolk sac. Here, the yolk sac epithelium is divided into two parts: (1) a ring of visceral extraembryonic endoderm, which covers the ectoplacental cone (a derivative of the polar trophoblast) close to the cup-shaped embryonic disc, and (2) the parietal extraembryonic endoderm, which covers the rest of the trophoblast (the mural trophoblast). During the early degeneration of the mural trophoblast in rodents it acquires a substantial basement membrane (Reichert’s membrane) (cf. Salamat et al., 1995) and forms the inner lining of the implantation chamber (Kaufman, 1992). The terms “visceral” and “parietal endoderm” were introduced by Sobotta (1903), apparently because of the morphological similarity between longitudinal sections of the rodent egg cylinder and schematic cross-sections of the adult trunk. However, in the latter, these terms qualify parts of the peritoneal epithelium, which derives from the (mesodermal) coelomic epithelium and should not be confused with the yolk sac epithelium. Although the relative arrangement and the developmental fate of the blastocyst components are identical among mammals, there are four aspects with marked interspecies variations that superficially confound the comparison of mammalian gastrula morphology: implantation, amnion formation, gastrula size, and the shape of the embryonic disc.

3. Anterior Margin of the Mammalian Gastrula

67

B. Implantation Most mammals, including the lower primates, use the mural trophoblast for initial attachment to the maternal tissue and implant relatively late, i.e., during or after gastrulation has started. Concomitantly, the polar trophoblast is severely attenuated to a thin layer (Rauber’s layer), degenerates (Williams and Biggers, 1990), and brings the embryonic disc into direct contact with the uterus lumen or the endometrium. In rodents and higher primates, including humans, on the other hand, implantation occurs well before gastrulation and it is the polar trophoblast [itself in intimate contact with the embryoblast and therefore also called “Trager” (Selenka, 1883)] that initiates implantation and develops into a fully fledged placenta. Consequently, the mural trophoblast degenerates, though with different speeds in different species, and opens the door for other fetal tissues (e.g., the yolk sac epithelium in rodents) to attach to the maternal decidua. In marsupials, such as the oppossum and the kangaroo, a fully functional yolk sac placenta can form in this way. In the rodent embryo, the differentiating polar trophoblast takes on a conical shape (ectoplacental cone) as it lengthens without widening. At the same time, the embryoblast transforms not into a flat disc shape (as in all other mammalian orders) but into a cup shape (egg cylinder) in which the germ layers are inverted [ectoderm inside, endoderm outside: entypy of Selenka (1883)l. The reason for this rodent-specific morphogenetic behavior is still unclear, but it may simply be the result of a fairly constant width of the differentiating polar trophoblast combined with a surge of proliferation in the embryoblast. This forces the embryonic disc to bulge into the blastocyst cavity, to form a kind of “embryonic lordosis” (Snow, 1977). However, common to all species in which the polar trophoblast forms the placenta is the fact that the embryonic disc is covered up by the developing placenta and develops well into gastrulation in a rather secluded position [see Assheton (1899), Hubrecht (1909), Mossman (1937), Hamilton and Mossman (1972), and Badwaik et al. (1997) for comparisons of implantation].

C. Amnion Formation

Closely correlated with the mode and time of implantation is the mode of amnion formation. When the amnion was first “invented” by evolution for a semiaquatic embryonic development of the sauropsids and mammals (hence “amniotes”), it seems that the only way an amniotic cavity could be formed was by means of amniotic folds: in the lower amniotes (reptiles, birds, and lower mammals), two or four opposing folds are formed by dorsal outgrowths of the superficial layer near or on the junction between trophoblast and epiblast. These folds converge over the epiblast and fuse to form between them and the epiblast

68

Christoph Viebahn

the amniotic cavity. Following the ventral folding of the whole embryo, the amnion will eventually invest the embryo completely. For mammals with the mural (abembryonic) and late type of implantation (rabbit, dog, sheep, pig, cow, etc.) it seems quite plausible to adopt the method of a “folding amnion” because the embryonic disc develops on the free surface of the conceptus. In contrast, higher primates developed the mechanism of forming an amnion by splitting, i.e., by opening the space between the embryoblast and the overlying polar trophoblast at the blastocyst stage; in these species the polar trophoblast prepares early for implantation, does not degenerate (see Section II,B), and seems to be “in the way” of amniotic folds. Intermediate forms exist, though, such as the hedgehog, which implants relatively late and still has a splitting amnion (Starck, 1975); in some bats (Badwaik et al., 1997) a preliminary amniotic cavity (called proamniotic cavity) forms early by splitting, soon degenerates (together with the polar trophoblast because of abembryonic implantation), and a definite amniotic cavity forms by folding subsequently. Rodents, too, reveal a mixture as they form a proamniotic cavity under the polar trophoblast by splitting. But covered by the rapidly growing ectoplacental cone, the wall of the proamniotic cavity cannot disintegrate and enlarges to make room for amniotic folds to form an amniotic cavity the “traditional way.” In fact, the enlarging proamniotic cavity of the rodents may be an alternative or additional factor (to simple embryoblast proliferation) leading to the rodent-specific bending of the embryonic disc into its typical cup shape, the egg cylinder (see Section 11,B). Interestingly, a difference in height between the two (anterior and posterior) amniotic fold primordia in prestreak stages of the mouse led Sobotta (191 1) to the conclusion that the longitudinal axis may be specified before primitive streak formation. For reviews on amnion formation see Hubrecht (1909), Hill (1932), and Starck (1975).

D. Gastrula Size The size of the embryonic disc varies widely among different mammalian species and correlates loosely with the type of implantation and amnion formation; in species with early implantation and a splitting amnion the embryonic disc tends to be small (rodents, apes, and humans). However, embryo size does not correlate with the size of the adult individual: lemurs, for example, have an embryonic disc about five times the diapeter of the human disc at a comparable stage (cf. Hill and Florian, 1963; O’Rahilly and Miiller, 1987). However, size differences are an as yet unexplained biological phenomenon in the adult as well as during development (cf. Raff, 1996), although they may play an important role in the survival of species in an evolutionary context (Gould, 1997).

3. Anterior Margin of the Mammalian Gastrula

69

E. The Embryonic Disc

The three interspecies differences mentioned so far, namely, the time and mode of implantation, the mode of amnion formation, and the size of the embryonic disc, make for difficult comparisons of early gastrula stages among different mammalian species, at least at first sight. Direct inspection of the whole embryonic disc is impossible in higher primate species such as Galago (Hendrickx, 1971), the hominoid monkeys, and humans (O’Rahilly and Muller, 1987), in which a small embryonic disc is combined with early implantation and the splitting mode of amnion formation. In these species, the disc has to be dissected free of these extraembryonic tissues in order to obtain a simple en face view of the complete disc. Even more problematic is the situation in monotremes, some primates (e.g., the macaque monkey), and rodents, in which the developing embryonic disc is bent into the cup shape of the embryocyst (before gastrulation) (Hill, 1932) or the egg cylinder (during gastrulation; see Section II,B), respectively. Here, the embryonic disc cannot be viewed in toto from any one side of the embryo at any time without superimposition of its parts. Thus, in these species one has to resort to reconstructing serial histological sections and, further, to constructing two-dimensional (“geographical”) projection maps in order to get a view of the overall shape and differentiation in the epiblast of the early gastrula (cf. Lawson et al., 1991).

F. The Anterior Margin of the Gastrula

As summarized above, most mammalian gastrulae have a flat embryonic disc open to simple inspection after removal from the uterus (Fig. lA), and among these the rabbit lends itself most conveniently to investigation. Using a translucent embedding medium and a combined fixation with glutaraldehyde and 0smium (Viebahn et al., 1995), it is possible to produce en face views with high contrast, which accentuates morphological differences between embryonic and extraembryonic tissues and within regions of the embryo (Fig. 1C). Thus, it is easy to see that immediately prior to primitive streak formation the embryonic disc (of the late blastocyst) is more or less round. It is not radially symmetrical, however, because the circumference of its anterior third or half exhibits a sharp but smooth contour separating dense embryonic (dark in osmium-fixed preparations) and loose extraembryonic (light) tissues; in the remainder of the circumference the difference in density between these tissue compartments is less pronounced and the contour is more irregular, almost ragged in places (Fig. 1C). The embryonic disc appears to be most dense near the anterior margin in a bandlike area whose posterior limit is more difficult to define in some cases (Fig. 1C) than in others (Viebahn et al., 1995). However, together with the sharp and smooth

70

Christoph Viebahn

Fig. 1 Morphology of anterior differentiation in the early mammalian gastrula as seen in surface views (A-C) or sagittal sections (D-H) of rabbit (A-F), Sorex (G), and cat (H). (A) Darkfield photograph of a complete blastocyst of the rabbit (fixed in 4% paraformaldehyde) at 6.0 days postconception with a prestreak embryonic disc (center) on its surface, i.e., the embryonic disc is integrated into the blastocyst wall. Arrowheads mark the lateral limits of the anterior marginal crescent (AMC). (B) Darkfield photograph at the same magnification as in A of an early-streak embryonic disc on the surface of a blastocyst. The blastocyst had been recovered from the uterus at 6.0 days postconception and cultured in v i m for 12 hr. The arrow marks the elongating primitive streak. (C) Surface view of a prestreak embryonic disc of the rabbit fixed with glutaraldehyde and osmium tetroxide and embedded in polyester resin. Note differences in staining intensity between anterior margin (top) and posterior margin (bottom). (D) Semithin (1 pn) median sagittal section of the embryonic disc shown in c . (E) High

3. Anterior Margin of the Mammalian Gastrula

71

anterior contour of the disc this bandlike area can be described as the anterior marginal crescent (AMC) (Viebahn el al., 1995). This happens to be an almost exact translation of “Vorderer Randbogen,” a term originally introduced by Kolliker (1882) for the same anterior structure in later (primitive streak) stages of the rabbit (cf. Fig. 1B). The differences in tissue density in these en face views correspond to differences in cellular height and cell numbers within the two cell layers present at this stage (epiblast and hypoblast), as seen in sagittal 1-pm sections (Fig. 1D). The sharp and smooth anterior margin is caused by a marked step between the cuboidal (embryonic) epiblast and the squamous (extraembryonic) trophoblast cells (Fig. lD, E). In addition, epiblast and hypoblast cells are more numerous anteriorly than posteriorly, with a gradual decline between these extremes along the anteroposterior axis (Fig. 1D; cf. Fig. lE, F). This gradual decline explains the indistinct posterior border of the anterior marginal crescent. Similar histological characteristics in the epiblast and hypoblast of the anterior margin are also found in older embryos, which have a primitive streak (Viebahn et al., 1995) and are thus instrumental for the claim that the anterior marginal crescent is, indeed, an anterior structure. Sagittal sections of other lower mammals showing similar histological features are available, to date, from Sorex (Fig. 1E) (Hubrecht, 1890), the bat Vespertilio rnurinus (van Beneden, 1911), and cat (Fig. 1F) (Hill and Tribe, 1924). Conclusive evidence for the precocious differentiation of the anterior margin is, in fact, difficult to obtain in higher primates due to the secluded position of the embryonic disc in the blastocyst (and the uterine mucosa) and the consequent need for serial reconstructions, not to mention the practical problems and the ethical dilemma connected with investigating suitable stages of monkey and human development. However, in several existing sections from rhesus monkey (Fig. 2A) taken from Heuser and Streeter (1941) or human (Fig. 2B) embryos there are signs of an early anterior differentiation that are reminiscent of the situations observed in rabbit, sorex, and cat (cf. Fig. lD, G, and H) (see also Hertig et al., 1956; Luckett, 1978; Enders et al., 1986; O’Rahilly and Miiller, 1987). More difficult to interpret is the histology of the early gastrula, the egg cylinder of rodents, due to the marked lordosis (backward bending) into which the early embryonic disc is forced in these species; in mice and rats, there seems to