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

<

er magnification of the anterior margin present in the section in D. The asterisk marks the embryonic/ extraembryonic border (between epiblast and trophoblast and hypoblast and yolk sac epithelium).Arrowheads mark some persisting cells of the polar trophoblast (Rauber’s layer) lying on top of the epiblast. (F) Higher magnification of the posterior margin present in the section in D; labeling as in E. ( G ) Reproduction of a figure from Hubrecht (1890) showing the drawing of a median sagittal paraffin section of a prestreak Sorex embryo. (H) Reproduction of a figure from Hill and Tribe (1924) showing a photograph of a median sagittal paraffin section of a prestreak cat embryo (reproduced with the permission of the Company of Biologists Ltd.). The scale bar (bottom), for A through H, is 260,260, 170, 50, 1 1 , 11, 17, and 17 pm, respectively.

72

Christoph Viebahn

Fig. 2 Morphologyof anterior differentiation in the prestreak mammalian gastrula as seen in sagittal sections of a 12/13-daypostconception rhesus monkey [A, reproduction of a figure from Heuser and Streeter (1941)l and a 10-day postconception human embryo (B, with kind permission from Dr. Noe, Camegie Collection of Human Embryos, Department of Pathology, Walter Reed Hospital, Washington D.C.). The epiblast and the hypoblast of both the rhesus monkey (A) and the human embryo (B) are higher in the left two-thirds than in the right one-third. Left is most likely anterior. The amniotic cavity (a) is between the epiblast (below) and the amnion epithelium (above),which on the left is seen to be continuous with the epiblast. Y.S., Yolk sac epithelium; y, yolk sac lumen. The scale bar is 60 pm for A and B.

3. Anterior Margin of the Mammalian Gastrula

73 be a stretch of hypoblast anteriorly that can be distinguished from the adjacent hypoblast (posteriorly) and the yolk sac epithelium (visceral endoderm, anteriorly) by increased cellular height and the absence of a yolky cytoplasm apically (Fig. 3A, B) (Bonnevie, 1950; see also Kaufman, 1992; Huber, 1915) or by increased cellular height only (Varlet et al., 1997). Similarly, the epiblast may appear higher on one side of the egg cylinder in whole-mount views of unfixed mouse embryos such as presented by Downs and Davies (1993). However, a histological reinvestigation of 6.5-day postconception mouse embryos using paraf-

Fig. 3 Anterior differentiation in the mouse egg cylinder. (A, C) Prestreak (possibly early streak) stage. (B, C) Primitive streak stage. A and B are reproductions of figures from Bonnevie (1950); B is rotated by 90" in order to bring the anterior end of the embryo to the left. At both stages the egg cylinder shows anterior differentiation in the hypoblast: higher cell density and increased cellular height when compared with the remaining endoderm (most likely extraembryonic, visceral endoderm). Note the presence of apical vacuoles in the yolk sac epithelium at the primitive streak stage. (C and D) Schematic representation of prestreak and primitive streak stage egg cylinders indicating the body axes and the planes of sectioning in A and B, respectively. Original labeling: I.L., inner layer; P.A., proamniotic cavity; ENT, endoderm; PR.N, primitive node (of Bonnevie, 1950); EP.AMN, epamnion (proamniotic cavity); ALL, allantois. Additional labeling: arrowheads mark the lateral extent of the anterior differentiation (high columnar epithelium) in the endoderm (hypoblast); aye, anterior visceral endoderm (hypoblast); m, mesoderm; pa, proamniotic cavity; eb, epiblast; ve, visceral endoderm (hypoblast) or yolk sac epithelium. The scale bar is 60 p,m for A and B.

74

Christoph Viebahn

fin as well as plastic embedding (Viebahn and Downs, 1999) revealed that in embryos both before and after primitive streak formation only the hypoblast has features that are typical of the anterior marginal crescent seen in the rabbit.

G. Epiblast or Hypoblast Differentiation?

Comparing the histological data available from various mammalian species to date it would seem that the differentiation in the hypoblast is the most prominent feature of the early anterior margin. However, in the rabbit the anterior epiblast is differentiated, too, when compared with the remainder of the epiblast (Fig. 1E) (Viebahn et al., 1995), even at earlier stages of development (Liebke and Viebahn, 1999); a similar anterior thickening of the epiblast seems to be present also in higher primates (cf. Hertig et al., 1956; Heuser and Streeter, 1941), but in these species the embryonic disc is comparably small and may not show fine histological differences after paraffin embedding (which was used by these authors.) In the rodent embryo it still needs to be determined whether differences in height at the marginal stretch of epiblast described by Sobotta (191 1 ) and Selenka (1884) in sagittal histological sections correlate with the differences in cellular height between different parts of the epiblast, seen also in modern whole-mount preparations (Downs & Davies, 1993). However, a stretch of high-columnar epiblast correlates with the anterior limiting furrow on the outside of the egg cylinder (Downs and Davies, 1993; Selenka, 1884). Perhaps molecular methods such as in situ hybridization of anterior- and germ layer-specific mRNA (Thomas et al., 1998) combined with labeling the anterior marginal furrow of freshly isolated embryos (Gardner et d., 1992) will show whether the anterior marginal furrow may be used as a landmark for the anteroposterior axis in the living prestreak mouse embryo. On the other hand, it seems possible that forcing the embryonic disc into the shape of the egg cylinder may lead to an irregular but physiological planar compression of the epiblast epithelium; in many cases, this may level out differences in cellular height for mechanical reasons. Optimal fixation and plastic embedding methods have to be applied to more mammalian species with a flat embryonic disc in order to find out whether epiblast differentiation forms an integral part of the early anterior margin.

H. Earlier Descriptions of Anterior Differentiation

Because the AMC is a morphologically defined structure, it may not come as a surprise that anterior differentiation in mammalian pre-primitive-streak embryos has been described, albeit tentatively in some cases, previously. Investigating the

3. Anterior Margin of the Mammalian Gastrula

75

rabbit embryo, Kolliker (1882) found a sharp delineation of the embryonic disc and an increased contrast at the anterior margin of early primitive-streak embryos in surface views and thus coined the term “Vorderer Randbogen” (anterior marginal crescent). However, despite the masterly production of serial sections of these delicate embryos, his fixation and microscopical methods were apparently still inadequate to the extent that Kolliker was unsure about the histology of this structure and about its existence in pre-primitive-streak embryos. Also investigating the rabbit, Assheton (1895) and Rabl (1915) illustrated an extensive series of surface views of pre-primitive-streak embryos with a definite AMC but the emphasis of their work was on later stages, and the histology of the anterior margin in these early stages was not clarified beyond doubt. The first publication of an anterior differentiation preceding the formation of the primitive streak came, in fact, from Hubrecht (1890) in his histological analysis of the shrew (Sorex vulgaris). He showed evidence of a patch of thickened hypoblast (Fig. IG), which is present before formation of the primitive streak (then called “gastrula ridge”). According to Brachet, a pupil of the well-known Belgian embryologist van Beneden, there was also a manuscript by van Beneden from 1889 describing in detail a similar anterior differentiation of the lower layer in the bat Vespertilio murinus and in the rabbit; this was published by Brachet in 1912 under van Beneden’s name posthumously (van Beneden, 1912), with some additions by Brachet, which are difficult to identify retrospectively (Rabl 1915). In between those reports, Bonnet (1901) described a similar anterior differentiation in the dog embryo and called it “Erganzungsplatte” (“a plate supplementary to the notochord”). However, around the turn of the century feelings ran high on different subjects such as the question as to the origin of the mesoderm and a general definition of gastrulation (Rabl, 1915; Keibel, 1917). Indeed, Hubrecht (1890) assumed, and maintained for some time (Hubrecht, 1909), that his patch of hypoblast was an anterior contribution to the notochord (and mesoderm) and called it the “protochordal plate.” Yet, he could not prevent “many of the best modem embryologists from following Kolliker in his negation of the participation of the endoderm towards the production of mesoblast” [Hubrecht (1909, p. 32); see also van der Stricht (1923) and Hill and Tribe (1924) for discussion]. So, only a few years later the early appearance of this anterior differentiation had already been forgotten and (only marginally!) more modem embryologists were surprised to find that their description of an early anterior differentiation in the embryonic disc of the cat (Fig. 1G) was not an entirely new suggestion (Hill and Tribe, 1924, p. 596). Thereafter, early anterior differentiation in the mammalian embryo was confined, again, to the embryological backwaters (this time perhaps due to the overwhelming interest in Spemann’s organizer concept) (Spemann and Mangold, 1924; Spemann, 1938) and in many of the recent papers and textbook chapters on the early mammalian embryo a possible early anterior differentiation receives only a vague mention, if any.

76

Christoph Viebahn

111. Changes at the Anterior Margin during Development Like other early axial organs such as the primitive streak and the primitive node, the AMC is a transient structure. Therefore, the following sections will summarize the events leading up to the formation of the AMC and those that follow its full differentiation at the beginning of gastrulation. A. Changes leading to Anterior Marginal Crescent Formation

Histological studies of implanted mouse egg cylinders suggest that the anteroposterior axis may be visible initially due to a tilt of the inner cell mass (ICM) within the blastocyst (Smith, 1985), the lower part of the tilted ICM pointing toward the posterior pole of the embryo. Comparison with a tilt of the ectoplacental cone in gastrula stages led to the idea that the anteroposterior axis may be preserved up to primitive streak stages and the ICM tilt may, hence, be the first sign of anteroposterior axis polarity. Using horseradish peroxidase, Gardner et al. (1992) labeled small patches of the hypoblast overlying the posterior aspect of the tilt in explanted egg cylinders at early gastrulation stages. Subsequent histological analysis showed the label consistently on the anteroposterior axis (as defined by ingressing mesoderm cells in the primitive streak) but the labeled cells were just as often found overlying the primitive streak as overlying the wall of the egg cylinder opposite the primitive streak, i.e., overlying the anterior margin. This suggested that the initial ICM tilt may indeed indicate the plane of the anteroposterior axis of the conceptus but that the polarity of the axis within this plane is defined later through the development of the primitive streak. Alternatively, only the tilting plane of the early ICM but not the direction of the tilt may be preserved until the gastrulation stages. Perhaps the marking of the ICM tilt in earlier blastocysts and a subsequent histological search for the anterior differentiation in the hypoblast (Bonnevie, 1950) (Fig. 3) of the labeled blastocysts will decide between these alternatives. A direct link between the anteroposterior axis of the late embryo and the very first steps of embryonic development or even oogenesis is a common feature in lower vertebrates (Xenopus)(Gerhart et al., 1989),urochordates (Nishida, 1994), and invertebrates [C. elegans (Goldstein and Hird, 1996) and Drosophila (St. Johnston et al., 1989; Munn and Steward, 1995; Goldstein and Freeman, 1997)l. Such a link was rendered plausible for the first time in a higher vertebrate (the mouse) by Gardner (1997). Using a combination of morphological analysis and experimental micromanipulation Gardner (1997) showed that a polar body, which regularly persists until the blastocyst stage, remains attached, by a thin intercellular bridge (or “tether”), to the blastocyst somewhere in its plane of bilateral symmetry (Smith, 1985). On obtaining polarity through the development of the AMC or of the primitive streak (Gardner et al., 1992), this axis of bilateral

3. Anterior Margin of the Mammalian Gastrula

77

symmetry will subsequently turn into the anteroposterior axis of the fetus. Because the original place of polar body attachment is closely related to the animal pole of the mouse zygote [as in lower chordates, i.e., Amphioxus (Conklin, 1932)] and because the polar body was not found to move during blastulation (Gardner, 1997), these results suggest that the anteroposterior axis of the embryo is aligned with the animal-vegetal axis of the zygote. It remains to be seen whether the axial organization of the zygote, which is a prerequisite for this model, is born out by asymmetrical distribution patterns of gene products even in the oocyte, as has been shown in Drosophilu (St. Johnston, 1995). If such a cytoplasmic axial organization were to be found in the human oocyte, though, it should be anticipated that intracytoplasmic sperm injection (ICSI), which is now frequently used to circumvent male infertility, could lead to an increased rate of congenital axis malformations, such as monochorial and “Siamese” twinning. However, the AMC may help to close the gap between earlier signs (if any) of axis formation and the definite body axes of the fetus, which become apparent at the beginning of gastrulation.

B. Changes Following Anterior Marginal Crescent Formation

1. General Topography The next significant step after establishing the AMC is the formation of the primitive streak. From that stage onward, the morphogenetic movements of gastrulation, or “Gestaltungsbewegungen” of Vogt (1923, make sure that the overall morphology of the embryonic disc is very similar among all amniotes: for example, heart and liver primordia, with the neural plate among the earliest definitive organ anlagen of the body, are established near the anterior extremity of the embryo (Fig. 4, see color plate). As a consequence of the arrival of these two voluminous organ anlagen, the original differentiation of the AMC is no longer visible and cell labeling studies will have to demonstrate the exact fate of the cells involved in the early anterior differentiation, especially for its hypoblast part (see Section III,B,4). Posterior to the heart anlage follow, in the midline, the pharyngeal membrane, the prechordal plate, and the notochordal process, which underlies the neural plate (cf. Fig. 4).At the time when these organ anlagen form, the anterior margin bends ventrally and then posteriorly, with the anterior end of the neural plate seeming to serve as a hinge. As a consequence, a new anterior extremity of the embryo emerges. Being formed by the junction between epidermal ectoderm and neural plate, and supported from underneath by the prechordal plate (see below), this new anterior extremity can now be called “cranial” in a literal sense because the elements contained in it will contribute most of the tissues of the adult head. Ventrally, the foregut pocket is deepened by this bending process, which is really the result of increased proliferation of the neural and the

78

Christoph Viebahn

prechordal plates and concomitant decreased proliferation of the anterior marginal tissues. By combining these movements, previously anterior regions such as heart and liver anlagen come to lie in a plane ventral and caudal to the pharyngeal membrane and find, at this stage already, the topographical positions allocated to them in the adult body. The principal topographical relationships of these organ anlagen and of their origins are schematically represented in Fig. 4 in sagittal sections from two stages of gastrulation; in order to facilitate the comparison between stages, the more advanced stage is represented as a flat disc by artificially flattening out the anterior foregut pocket. 2. Ectoderm When considering the differentiation of the three germ layers in the anterior region, the general fate of the outer layer, the epiblast, can be described quite satisfactorily by comparing the morphology of different stages (Fig. 4): although certain to be influenced by the germ layers lying underneath, during this phase of development the anterior epiblast is morphologically self-contained with no cells known to leave or enter the epithelial compartment. Anteriorly, the epiblast gives rise, thus, to the epidermal ectoderm (or surface ectoderm) and, posteriorly adjacent, to the neural plate. In the epidermal ectoderm are embedded, from anterior to posterior and at subsequent developmental stages, the pharyngeal membrane, Rathke’s pouch epithelium (the primordium of the anterior lobe of the pituitary gland), and, lateral to the midline, the olfactory placodes. The latter lie closest to the neural plate and are likely to be split off from the neural epithelium similar to neural crest cells on both sides of the neural plate further caudally (Verwoerd and van Oostrom, 1979; Smith et al., 1994). The part of the epiblast anterior to the pharyngeal membrane, i.e., the part that bends ventrally, will contribute to the epidermal ectoderm covering of the trunk in the heart and liver region.

3. Mesoderm The mesodermal germ layer adds a new layer to the anterior margin. Originating from the primitive streak, it arrives by way of migration along both sides of the notochordal process between the ectoderm and endoderm. This part of the mesoderm occupies an inverted U-shaped area anterior and lateral to the pharyngeal membrane and the notochordal process (Fig. 4B). In it develops (1) the transverse septum, which surrounds the endodermal liver diverticulum and which will contribute to the diaphragm, and (2) the cardiogenic plate, which varies in size between a small median patch of mesodermal cells in rodents and a paired primordium separated at first by a median patch of seemingly uncommitted, possibly vasculogenic mesodermal cells in most other mammals. This cardiogenic potency was used as an indicator in fate mapping studies that elucidated the origin of this part of the mesoderm: it develops through anterolateral migration of cells

3. Anterior Margin of the Mammalian Gastrula

79

that had ingressed through the middle third of the primitive streak (Rosenquist, 1970; Garcia-Martinez et al., 1993; Garcia-Martinez and Schoenwolf, 1993; Inagaki et al., 1993; Tam et al., 1997). This might explain why rodents start off with a single median cardiogenic plate: the cup shape of the embryo permits the cells from the middle third of the primitive streak to reach the midline anterior to the notochord along circumferential trajectories much faster than in species with a flat embryonic disc. Whether the human embryo has a single median or two bilateral heart anlagen is still a matter of debate (de Vries, 1981; Kuhn and Liebhem, 1987). Although the cardiogenic mesoderm lies immediately adjacent to the prechordal plate, it has never been described as containing labeled cells in fate mapping studies of the mouse embryo in which the primitive node was labeled (Beddington, 1994; Sulik et d.,1994). In bony fish, on the other hand, cardiac progenitor cells have been described among mesoderm cells immediately adjacent to the prechordal plate (Serbedzija et al., 1998). 4. Endoderm

Describing developmental fates in the lower layer of the anterior margin is complicated by the fact that the existing lower layer, the hypoblast, is split medially and is pushed aside (starting posteriorly) by cells that leave the epiblast and ingress through the anterior part of the primitive node. By way of ingression a cellular sheet is formed, consisting of (1) endoderm (also called “definitive endoderm” to distinguish it from primitive or visceral endoderm), (2) the prechordal plate, and (3) the notochordal process. The latter two migrate in line and anteriorly, thus establishing a rodlike formation within the median plane, whereas the definitive endodermal cells are thought to form a halo straddling and accompanying this rod in the shape of an inverted U. Unfortunately, definitive endoderm and hypoblast are difficult to distinguish because there are no clear-cut morphological differences between the two populations (Poelmann, 1981). However, some genes, such as the homeobox gene Hex (Thomas et al., 1998), seem to be specifically expressed in areas presumed to be definitive endoderm and may help to clarify the extent of migration in this population (see below). In particular, specific expression of the secreted inhibitory cerberus protein in both the liver primordium and in the cells ingressing earlier from the primitive node equivalent in the Xenopus embryo (the dorsal blastopore lip) suggests that the definitive endoderm cells (ingressing from the primitive node in amniotes) migrate as far as the liver anlage (Bouwmeester ef d., 1996). By the same token, the endoderm cells lying “posteriorly” to the liver (see Fig. 4B), i.e., endoderm covering the heart anlage and endoderm forming the inner lining of the pharyngeal membrane, are very likely derived from the cells ingressing early and through the anterior part of the primitive node. Less clear is the extent to which the cells ingressing through the primitive node contribute to endodermal derivatives “anterior” to the liver diverticulum (Fig. 4).

80

Christoph Viebahn

In principle, the whole of the fetal endoderm (the future gastrointestinal tube) and its glandular derivatives such as liver and pancreas are derived from the definitive endoderm, whereas the yolk sac epithelium is derived from the hypoblast. After organogenesis is complete, the junction between these two sorts of endoderm (embryonic and extraembryonic) lies at the level of the umbilical ring in the omphaloenteric duct, which is the connection between the (regressing) yolk sac and the small intestine. Therefore, it is reasonable to assume that definitive endoderm emigrating from the primitive node anteriorly toward the midline will not only contribute to the liver diverticulum but also to the ventral wall of the small intestine, i.e., the midgut anterior (later becoming caudal) to the liver (Fig. 4B). Alternatively, there may be the possibility that the hypoblast of the AMC is not completely displaced toward the extraembryonic areas and also contributes to the midgut epithelium.

5. The Prechordal Plate Anterior to the tip of the early notochordal process advancing anteriorly, there is a small batch of cells that also ingresses through the primitive node, but which migrates as a loosely connected group of cells, some of them inserted into the endoderm, others migrating between ectoderm and endoderm. This has been demonstrated in chicks (Pera and Kessel, 1997), mice (Sulik et al., 1994), and rabbits (C. Viebahn, unpublished). On reaching the vicinity of the anterior margin, the central cells of this group come to form an epithelium-like sheet, the prechordal plate of Oppel (1890), which forms an insertion in anterior endoderm. The peripheral cells stay connected to the prechordal plate initially and, remaining in the space between ectoderm and endoderm, could be called prechordal mesoderm [the “prosencephalic mesenchyme” of Seifert et al., (1993)l. These prechordal mesoderm cells may intervene between the two seemingly independent and topographically separated bilateral heart anlagen of most species (see above). In species with a median heart anlage (i.e., in the mouse and rat), however, there may be only very little prechordal mesoderm reaching the anterior margin in time to separate the cardiogenic mesoderm wings. The prechordal mesoderm (and the prechordal plate) are subsequently separated from the cardiogenic plate in the median plane by a patch of close apposition between ectoderm and definite endoderm: the pharyngeal membrane (Fig. 4B). Eventually, the prechordal mesoderm and the cells of the prechordal plate proliferate and take part in the formation of the head mesenchyme, including the premandibular cavities [a kind of coelomic cavity of the head (Oppel, 1890; Parker, 1917; Adelmann, 1926)] and some of the external muscles of the eye (Couly et al., 1992).In addition, cells originating from the prechordal plate contribute to the formation of the definitive endoderm immediately posterior to the pharyngeal membrane [preoral gut or Seessel’s pocket (Seessel, 1877)] and to the endodermal lining of the oral cavity (Parker, 1917); because of this contribution to

3. Anterior Margin of the Mammalian Gastrula

81

mesoderm and endoderm, the prechordal plate is commonly referred to as the mesendoderm. Interestingly, the prechordal plate seems also to play a pivotal role in the patterning of the overlying forebrain (see Section V,B), whereas the prechordal mesoderm is as yet less well defined in molecular or functional terms. The morphologically obvious connection between head mesenchyme and the prechordal plate may have been the reason why Hubrecht (1909) suggested, in line with the assumed mesoderm formation by his protochordal plate (see Section II,H and Fig. lG), that the protochordal plate was the origin of the prechordal plate and therefore deserved its prefix “proto.” Subsequently, some authors simply called the first anterior thickening of the hypoblast in prestreak embryonic discs the “prechordal plate” (Hill and Tribe, 1924; cf. Fig. 1H). This may be the reason for some of the confusion surrounding the origin, fate, and significance of the prechordal plate and may have led to the neglect of the early anterior differentiation in most modern embryology text books.

IV. The Anterior Margin in Different Vertebrate Classes A. Birds

From the primitive streak stages onward, through organogenesis and beyond, the chick embryo behaves very much like a mammalian embryo with regard to both morphology and morphogenetic movements. Morphological and functional results obtained in the chick are thus an essential part of many textbooks on the human embryo. Less obvious, however, is the fact that during the developmental phase immediately preceding the primitive streak stage, there are principal differences in these two vertebrate classes; these concern the gross morphology of the embryonic disc, the formation and function of the hypoblast, and the initial accumulation of the primordial germ cells. 1. Koller’s Sickle Just before the primitive streak is formed in the chick (stages XI-XII EGK) (Eyal-Giladi and Kochav, 1976) and during the time when the primitive streak is already visible (stage 2 HH) (Hamburger and Hamilton, 1993), there is a crescent-shaped thickening in the embryonic disc, intriguingly similar to the mammalian AMC; however, but this thickening lies within the posterior margin and not within the anterior margin as in mammals. The thickening was initially suggested by Rauber (1876) to be an inconsistent feature probably involving both layers present at the posterior margin at this time of development (epiblast and hypoblast or endophyll); subsequently, Koller (1882) described it more concisely as a regular differentiation that, however, can be found in the lower layer (hy-

82

Christoph Viebahn

poblast or endophyll) only, and since then the thickening has carried the label “Koller’s sickle.” In contemporary experimental attempts at elucidating the morphogenetic potencies of the posterior margin, the definition of Koller’s sickle was limited to a stretch of the upper layer (epiblast) within the posterior margin (cf. Stem, 1990; Eyal-Giladi et al., 1992). Whatever its definition, Koller’s sickle may be similar in shape to the anterior marginal crescent of mammals (albeit at the opposite margin of the embryonic disc) and a feature in common may be the active role during differentiation of the anteroposterior axes (Stem, 1990; EyalGiladi et al., 1992; Callebaut and van Nueten, 1994; Khaner, 1998), but a striking difference is the involvement of Koller’s sickle in the generation of the hypoblast (Stem, 1990; see also below). 2. Hypoblast

In the chick, the hypoblast layer forms relatively late, i.e., from a time point onward when the embryonic disc measures several hundreds of cells in diameter (stage X EGK) until the time when formation of the third germ layer, the mesoderm, has started (stage 2 HH). Initially, it forms by polyingression from the epiblast, i.e., by ingression of epiblast cells through the basement membrane of the epiblast at multiple locations in the anterior part of the embryonic disc (Harrisson et al., 1991); later, it is supplemented by anterior migration of lower layer cells underlying the posterior margin (Stem, 1990). Whether the upper layer of the posterior margin contributes to the hypoblast at this stage of development is still controversial (cf. Stem, 1990 and Eyal-Giladi et al., 1992); in fact, several other models have been put forward in the past (see Eyal-Giladi et al., 1992). Confounding the issue is, perhaps, the fact that hypoblast formation in the chick is difficult to separate both in time and space from formation of the mesoderm: at the stage with a full-sized Koller’s sickle and ongoing hypoblast formation (stage XI11 EGK), mesoderm is starting to be formed near the posterior margin, and also uses ingression from the epiblast as its mode of formation. In the case of the mesoderm, though, epiblast ingression carries on to epitheliomesenchymal transformation (Hay, 1995), whereas, in the case of the hypoblast, ingression from the epiblast leads to the formation of another simple epithelium. In the mammalian embryo, in contrast to the chick, no thickening such as within Koller’s sickle exists posteriorly; the posterior area of the early mammalian embryonic disc may even be particularly thin at the position where the primitive streak will form (Fig. 1C). Moreover, the hypoblast forms relatively early (see Section 11,A). Shortly after the inner cell mass, which in most species is just a few dozen of cells, has separated from the trophoblast during the blastocyst stage, hypoblast and epiblast can be distinguished within this inner cell mass (Nadijcka and Hillman, 1974; Enders et al., 1978; Gardner, 1986; Salamat et al., 1995). Once formed, the hypoblast starts to spread extraembryonically, forming an in-

a3 ner lining of the expanding blastocyst cavity, the yolk sac epithelium. In the mammalian embryo, the hypoblast is thus a fully established lower layer well before the ingression of mesoderm from the epiblast starts. Compared to the overwhelming amount of experimental data obtained in the chick embryo, there are only relatively few published photographs of median sagittal sections of pre-primitive-streak stage embryos of the chick showing both the anterior and the posterior margins of the embryonic disc (Wetzel, 1929; Peter, 1938b; Spratt, 1946; Vakaet, 1962; Spratt and Haas, 1965; Harrisson er al., 1991). In none of these sections can a regional differentiation be ascertained in the anterior parts of the epiblast or hypoblast at stages prior to the presence of the primitive streak, i.e., before stage 2 HH, although Waddington (1932) reported on a round patch of coherent hypoblast cells sometimes to be seen anterior to the center of prestreak embryonic disc of the chick. But only in later stages (4 and 5 HH) may there be a consistent area of specialized hypoblast cells anterior to the primitive node and the notochordal process [the "Entodennhof" of Wetzel (1929)l. However, in view of the early report by Schauinsland (1903) about an anterior differentiation in the lower layer of prestreak embryos of the sparrow and the starling, an investigation specifically addressing the question of early anterior margin differentiation, using semithin section technology and in situ hybridization of specifically expressed genes (see below), may be warranted in these and other bird species; interestingly, Schauinsland (1903) stated that an anterior differentiation could not be found in the six other bird species also included in this investigation, nor in the chick. In fact, the bird species with a suggestion of an AMC-like formation have a small and compact embryonic disc at the beginning of gastrulation, whereas the seven other species mentioned are larger at the stage in question. It is thus just possible that there is, indeed, an anterior differentiation similar to the mammalian AMC in birds: it may consist of only a limited number of hypoblast and epiblast cells that are attenuated in species with a large disc and are, therefore, morphologically not visible. Unfortunately, Schauinsland (1903) did not clarify the relationship between this early differentiation in the lower layer and the definitive endoderm or the prechordal plate, and it remains to be seen whether this small population of cells could have the same functional potential as the anterior hypoblast in mammals. Functionally, the hypoblast was shown by several authors (Azar and Eyal-Giladi, 1981; Mitrani et al., 1990; Waddington, 1933a) to have axis-inducing activity [see also, however, Khaner (1995,1998)l possibly through the chicken analog of the Xenopus growth factor Vgl (Seleiro et al., 1996). By analogy, and following recent experimental data (Thomas and Beddington, 1996, and Rhinn et al., 1998), the mammalian hypoblast may also have an axis-inducing ability. But, given the anteroposterior differentiation of the mammalian hypoblast, this influence on the epiblast would probably emanate anteriorly and not posteriorly as in the chick (Stern, 1990; Eyal-Giladi et al., 1992; Callebaut and van Nueten, 1994; Khaner, 1998). 3. Anterior Margin of the Mammalian Gastrula

a4

Christoph Viebahn

3. Primordial Germ Cells Another case of a diametrical difference between mammals and birds concerns the distribution of the primordial germ cells (PGCs) immediately following their separation from the somatic cell line. In the chick, PGCs originate (centrally) in the epiblast at primitive-streak stages (Muniesa and Dominguez, 1990;Ginsburg, 1994) and can then be found anteriorly within the lower layer (hypoblast), in the “germinal crescent” (Swift, 1914; Simon, 1960). In mammals, PGCs also originate in the epiblast, albeit peripherally (Tam and Zhou, 1996), but they are subsequently found posteriorly in the primitive streak. It is tempting to speculate, therefore, that specific activities (such as axis induction by Vgl) (Seleiro et al., 1996) residing in or near Koller’s sickle preclude a posterior accumulation of PGCs in the chick and that, conversely, PGCs in mammals have to accumulate posteriorly because of axis-inducing activities residing in the anterior margin.

4. Axial Inversion? Taken together, these differences in the avian and mammalian pre-streak-stage embryos suggest that there might be a transient back-to-front inversion along the anteroposterioraxis between these two vertebrate classes. This inversion is, however, corrected early with the appearance of the primitive streak. The recent suggestion and confirmation of an upside-down inversion in the dorsoventral axis of arthropods and vertebrates, initially proposed by Geoffroy Saint-Hilaire (1822), showed that axis inversions may, indeed, occur during evolution (Arendt and Nubler-Jung, 1994; De Robertis and Sasai, 1996).

B. Amphibia

At the stage when the internal germ layers are about to be formed, i.e., at the stage equivalent to the early embryonic disc stage in birds and mammals, the amphibian blastula is a more or less hollow ball of cells and, in most forms, will in its entirety contribute to the formation of the embryo (holoblastic development).Although there is a distinction between a (yolk-containing) vegetal and a (nonyolky) animal hemisphere, which may be reminiscent of the embryonic and extraembryonic subdivision of amniotes, there are no extraembryonic structures that are shed during later development (as in the meroblastic development of amniotes). Consequently, there is no obvious morphological equivalent to the margin of the embryonic disc of amniotes and we do not immediately know where to look for an early anterior marginal differentiation. However, the question of analogy may be approached from a dynamic angle, by comparing homologous developments during gastrulation in the mammalian and amphibian embryos.

3. Anterior Margin of the Mammalian Gastrula

85

Following the application, in the wake of Darwin’s work, of phylogenetic principles to embryonic development (Haeckel, 1874; Rauber, 1876; Kolliker, 1879), van Beneden was probably the first to realize that the blastopore is the amphibian homolog of the primitive streak in amniotes (Kollmann, 1886). Early embryologists, therefore, called the primitive streak the gastrula ridge (Hubrecht, 1890). In addition, the inner lining of the amphibian vegetal hemisphere was soon considered to correspond to the avian and mammalian hypoblast (Hubrecht, 1890; Waddington, 1956; Eyal-Giladi, 1997).Analogous development is particularly evident with regard to the notochordal process, which originates in the dorsal blastopore lip in the amphibian and in the primitive node in amniotes, and this was the basis for Waddington’s organizer transplantation experiments in amniotes involving chick and duck (Waddington, 1933b), as well as chick and rabbit (Waddington, 1934).As stated in Section I, there is a wealth of molecular data now corroborating this analogy (De Robertis et al. 1994). However, preceding the appearance of the dorsal blastopore lip and the involution of the notochordal process, there is no morphological landmark so far described that could represent a precocious (“preblastoporal”)anterior differentiationin the amphibian embryo. On the contrary, the only area with a specific activity, in fact with the ability to induce the organizer, is described in the late blastula as the so-called Nieuwkoop center (Gerhart et al., 1989), and this lies in the dorsal part of the vegetal hemisphere. This area may be regarded as lying posteriorly and acquiring an anterior position (equivalent to the AMC) only when involution of the notochordal process is almost complete (Bauer et al., 1994). Drawing conclusions as to the analogy of the hypoblastic part of the mammalian AMC with the amphibian Nieuwkoop center has, thus, to await expression analysis of genes closely related to the Nieuwkoop center, such as V g l (Thomsen and Melton, 1993; Seleiro et al., 1996), siamois (Lemaire et al., 1995), and eFGF genes (Pownall et al., 1996), in the mammalian embryo. However, there is an apparent accumulation of primordial germ cells anteriorly in the amphibian blastula (Cleine, 1986) just as in the chick (see Section IV,A,3), and this suggests that they may also be an early transient anteroposterior inversion between amphibia and mammals in this respect.

C. Bony Fish Superficially similar to the development of birds and mammals is the gastrulation process in teleost fish. In the zebrafish, the best described teleost (Kimmel et al., 1995), the blastoderm develops as a cap of cells on the top of a big yolk cell as in the chick and, consequently, it has a distinct margin; but there are, as in amphibia, no extraembryonic tissues because the margin will subsequently grow around, and thus internalize, the yolk cell in a process described as “epi-

86

Christoph Viebahn

boly” (Trinkaus, 1984). A broad thickening, called “embryonic shield,” at the posterior margin of the blastoderm expresses organizer-specificgenes (Abdelilah et al., 1994; De Robertis et al., 1994), but the whole margin of the blastoderm (the germ ring) is engaged in the formation of the internal germ layers: they originate here by way of involution initially as a single layer, frequently called a “hypoblast” [not to be confused with the avian hypoblast; see Kimmel et al., (19991, proceed to doubling back on to the subsurface of the blastoderm, and are thus sandwiched between the blastoderm and the yolk syncytial layer (YSL) underneath (Driever, 1995; Kimmel et al., 1995).As in the comparison between mammalian and amphibian embryos, it is mainly the developmental dynamics that give clues as to possible analogies. The germ ring clearly has the function of the amphibian blastopore lips or the amniote primitive streak (and is not an embryonic-extraembryonic border). In addition, induction capacity for mesoderm (and organizer) formation resides in the YSL, which covers the yolk mass underneath the blastoderm (Oppenheimer, 1936; Mizuno et al., 1996). This and the expression of p-catenin as a component of the axis inducing Wnt pathway (Schneider et al., 1996), as well as the position of the YSL, suggest that the teleost YSL may be homologous to the endoderm of the amphibian vegetal hemisphere and the amniote hypoblast (Eyal-Giladi, 1997). Preceding gastrulation, however, there is indeed a morphological differentiation in that the anteroposterior axis can be identified by a stretch of low-columnar epithelium (within an otherwise high-columnar blastoderm epithelium (at the posterior margin (Schmitz and Campos-Ortega, 1994).As in the mammalian embryo, the thickening in the area of the organizer is formed within the thinnest part of the blastodisc. Another interesting parallel between the zebrafish and mammals concerns the anterior specification of the neural plate: just as the anterior margin of the neural plate is defined through induction by the underlying anterior hypoblast before the appearance of the primitive streak (Thomas and Beddington, 1996), there is a precocious morphological differentiation of the anterior neurectoderm margin in zebrafish before the arrival of the notochordal process at that level of the blastoderm (Houart et al., 1998).Apart from marking this margin, these so-called row-l cells then go on to pattern the forebrain anlage. It will be interesting to find out where the stimulus for this anterior differentiation originates: from posterior or anterior blastoderm regions via planar induction, or from below and anterior, i.e., from the yolk syncytial layer? With regard to the origin and initial settlement of primordial germ cells (PGCs), the zebrafish takes an intermediate position between birds and mammals: as determined by the expression of the vasa gene, PGCs reside at the periphery of the blastoderm during gastrulation. Some clumps of PGCs lie anteriorly and others posterolaterally (Yoon et aZ., 1997) before transiently settling in the YSL (Gevers et al., 1993). There are, thus, no clear indications for an anteroposterior inversion in the teleost fish when compared with the mammalian embryo.

3. Anterior Margin of the Mammalian Gastrula

87

V. Gene Expression Related to the Anterior Margin A. Early Gastrulation Stages

Some transcription factor genes (such as Otx2 and Oct3/4) are expressed globally within the early embryonic disc of the inner cell mass, and growth factor receptors seem to be similarly involved in the sustained differentiation of the whole embryoblast or in development of the inner cell mass (Arman et al., 1998; Gu et al., 1998). However, several genes are now known to be expressed initially within or near the anterior margin [in addition to the early antigenic marker VE-1 (Rosenquist and Martin, 1995)] and, like the globally expressed genes, they fall into two categories: homeobox-containing genes coding for transcription factors and genes involved in the pathway of TGF-P-like signaling molecules. Four of the transcription factor genes are expressed in the hypoblast only at the anterior margin: Hex (Thomas et al., 1998), Liml, gsc, and HNF3P (Belo et al., 1997). Two others (Otx2 and Rpx/Hesx-Z) are sequentially expressed in different layers, first in the hypoblast and subsequently in the overlying neurectoderm (Acampora et al., 1995; Hermesz et al., 1996; Thomas and Beddington, 1996)). Within the ectoderm, Rpx/Hesx-1 marks the anterior half of the emerging neurectoderm. A little later it is confined to the anterior extremity of the anterior neural folds [like the homeobox gene Six-3 (Oliver et al, 1995)] and when organogenesis is almost complete, Rpx/Hesx-1 is specifically expressed in Rathke’s pouch (hence one of the names: Rathke’s pouch homeobox gene) (Hermesz et al., 1996). There are several homologs of this gene known in other vertebrate species (e.g., zebrafish, Xenopus, chick, human) and because one characteristic of these genes is the expression in the early anterior neural folds, these genes, including Rpx/ Hesx-I, were recently subsumed into the family of ANF genes (mterior neural fold) (Kazanskaya et al., 1997). In the mouse, the spatiotemporal sequence of Rpx/ Hesx-I expression may be regulated by the transcription factor because after ablation of the early endoderm, Rpx/Hesx-1 expression is diminished in the ectoderm and head development is significantly impaired (Thomas and Beddington, 1996). As is the case with Rpx/Hesx-I, the expression of Otx2 in the anterior hypoblast is required for neurectoderm induction, and later within the neurectoderm for forebrain and hindbrain patterning (Rhinn et al., 1998), whereas Lim-1 has a more general but still dominating function for head induction (Shawlot and Behringer, 1995). The expression pattern of the Hex gene, on the other hand, reveals another aspect of development at the anterior margin: just before its expression is seen at the anterior margin Hex is expressed in the hypoblast at the (distal) tip of the egg cylinder (Thomas et al., 1998). Through the labeling of these early Hex-expressing cells with a lineage tracer (Dil) they were found to migrate toward the anterior margin (Thomas et al., 1998). Translated to the normal flat embryonic disc of other mammals, these lineage tracing studies suggest that there is a general anterior movement within the hypoblast originating in the

88

ChristophViebahn

center of the embryonic disc, and this fits the general morphology of late prestreak stages. The result of this migration may very well be a cellular congestion at the anterior hypoblast, which produces the morphological characteristics seen in the hypoblast of the AMC (cf. Fig. 1D-F). Now, factors can be sought which support this hitherto unsuspected migration pattern within the hypoblast and thereby initiate anteroposterior axis formation in the mammalian embryo. Apart from the homeobox genes mentioned above, the gene coding for a glycoprotein that is related to the TGF-P antagonist cerberus of the frog (Bouwmeester et al., 1996), and therefore called cer-1, cerr-1, or mCer-I (Belo et al., 1997; Biben et al., 1998b; Shawlot et al., 1998), is specifically expressed in the early anterior hypoblast of the mouse as well. This is particularly intriguing in connection with the expression of the TGF-P-like molecule nodal in the epiblast [as well as in the hypoblast (Varlet et al., 1997)] and the requirement for nodal for mesoderm formation (Zhou et al., 1993; Conlon et al., 1994). Inhibiting the effects of nodal at the anterior margin, cer-Zlcerr-IlmCer-1 would ascertain that mesoderm is formed further posteriorly only, i.e., in the area of the primitive streak. Indeed, such a limited-inhibitionmodel of mesoderm formation would fit ultrastructural changes leading to epithelio-mesenchymal transformation, some of which occurs in the whole of the epiblast (perhaps as a sign of a general growth factor stimulation), whereas other changes (perhaps driven by nodal) are generally prevented except in the area of the primitive streak. Only in the primitive streak do epiblast cells then proceed to overt mesoderm ingression (Viebahn, 1995). This model also fits experiments that have demonstrated localized suppression of mesoderm formation by the activin immediate-early response gene Mix.1 in Xenopus (Lemaire et al., 1998). The lack of mesoderm formation in smad2 knockout mice (Nomura and Li, 1998; Waldrip et al., 1998; Weinstein et al., 1998) further supports the role of TGF-P-like growth factors in this process, and downstream targets of this pathway are beginning to be elucidated (Tang et al., 1998). At the same time, there are differences in gene expression in the anterior and posterior parts of the early primitive streak: gsc is expressed in the anterior half and gsx in the posterior half in the chick, and in the rabbit, the mesoderm-specific transcription factor brachyury is switched on in the posterior half of the primitive streak and simultaneouslyin the presumptive primitive node, while the anterior half of the primitive streak is still negative (Mitchell et al., 1999). These differences suggest that the primitive streak is patterned early, i.e., while it is being formed, and is the result of a complex induction mechanism. They also emphasize the dual role of the brachyury transcription factor: it is active in the development of the notochord (from the primitive node) and the tail (from the posterior mesoderm), at least in protochordates (the closest vertebrate predecessor) such as amphioxus and in all vertebrates (Chapman et al., 1996a,b; Holland et al., 1995). At a more general level of axis formation, a member of the Wnt gene family has been shown to play a role in the mouse in establishment of the anteropostenor axis (Popper1 et al., 1997), and several other genes, such as HNF3P, Shh,

3. Anterior Margin of the Mammalian Gastrula

89

and Activin-receptor ZZb, are apparently involved in left-right differentiation in the chick during midgastrulation (after appearance of the primitive node) (Levin et al., 1997; Pagin-Westphal and Tabin, 1998). But the primary source or signal for their specific expression patterns in the mammalian embryo is still elusive.

B. late Gastrulation Stages Due to the laying down of several specific organ anlagen at the anterior margin (cf. Fig. 4A, B), gene expression patterns near the anterior margin are becoming complex by the late gastrula stage. Principally, they are involved in (1) patterning the forebrain anlage by planar signals within the neural plate or by vertical signaling from the prechordal plate (Oliver et al., 1995; Dale et al., 1997; Pera and Kessel, 1997; Cutiss and Heilig, 1998; Pevny et al., 1998; Rubenstein et al., 1998), (2) induction of cardiogenic differentiation in the mesoderm by the definitive endoderm (Gannon and Bader, 1995; Zhang and Bradley, 1996; Arai et al., 1997; Narita et al., 1997; Biben et al., 1998a), and (3) differentiation of internal organs (liver, lung, and thyroid) in the anterior definitive endoderm through interaction with the mesoderm underlying the anterior epidermal ectoderm, such as the septum transversum (Bouwmeester and Leyns, 1997; Dunwoodie et al., 1998; Pera and Kessel, 1998; Thomas et al., 1998). With regard to the first group of activities (forebrain induction) there is a principal amphibianavian difference: neural induction in Xenopus is the result of inhibition of bone morphogenetic protein (BMP)-type growth factors (expressed in the ectoderm) by the anti-BMP glycoprotein chordin, which is secreted by the organizer. In the chick, by contrast the BMP-chordin antagonism is used not during neurogenesis but for induction of the primitive streak (Kessel and Pera, 1998). At present, the relationship between all these activities and the possible functions of the AMC is not known, although many genes (such as hex, Liml, HNF3p, gsc, and cer-Z) that are expressed early in the anterior hypoblast are also expressed in the cells generated in the primitive node, be it the notochordal process (Liml, HNF3p, gsc) or in the (definitive)endoderm (hex, cer-Z) (Belo et al., 1997). These expression patterns are thus contiguous but nevertheless in tissues with entirely different developmental fates, in many cases. It is hoped that they will not lead to a confusion similar to the one that surrounded the description of the prechordal plate (see Section III,B,5).

VI. A View on Phylogenetic Implications The functional aspects of the anterior gastrula margin seem to have been remodeled with the appearance of mammals during evolution, but there is a new challenge in finding out how the margin might have evolved initially and how this development might fit the long history of ideas on the evolution of gastrulation

90

Christoph Viebahn

(Haeckel, 1874; Kollmann, 1886; Pasteels, 1936; Peter, 1938a; Wolpert, 1992; De Robertis et al., 1994). The following interspecies comparison will, thus, use morphological and functional criteria, as in the case of the homology of the primitive streak and the blastopore lips, although much about the possible functions of the early anterior margin still remains to be proved. Consequently, the evolutionary steps deduced from the homologies are tentative in character and are put forward only in the absence of a current alternative evolutionary model.

A. Homology of the lower layer

The ability to induce mesoderm formation in the ectoderrn or epiblast is a feature shared by the teleost yolk syncytial layer, the amphibian Nieuwkoop center in the endoderm of the vegetal half, the avian hypoblast, and also, possibly, the mammalian hypoblast (cf. Section V). These four structures can, therefore, be regarded as being functionally homologous in addition to earlier suggestions of their morphological homology (Hubrecht, 1890).Mesoderm is similarly induced by cellular interaction in the earliest common ancestors of the vertebrates: the protochordates such as Amphioxus (a cephalochordate) and Halocynthia roretzi (an urochordate). In H . roretzi (family Ascidiaceae) embryos this mesoderm-inducing capacity resides to a greater extent in the vegetal blastomeres, which are fated to become the invaginating endoderrn of the blastoporus at the early gastrula stage (Jeffery, 1992; Nakatani and Nishida, 1994). Given these parallels, variability in two evolutionary events (discussed in Sections VI,B and V1,C) near Fig. 4 Components of the anterior margin during gastrulation. Schematic sagittal sections passing through landmarks near the anterior margin at an early (A) and a late (B) stage of mammalian gastrulation. (C and D) Surface views of equivalent stages in the rabbit; orange lines indicate the position of the sagittal sections shown in A and B, respectively. At the late stage (B, D) the anterior parts of the embryo start folding ventrally to various degrees in different species. using the junction between epidermal and neural ectoderm as a pivot (arrow). As a result, the foregut pocket is formed. For the sake of clarity and in order to facilitate the comparison between stages and between species, however, this curvature was omitted and so the shape of the foregut pocket cannot be seen here. am, Amniotic epithelium; amc, anterior marginal crescent; cc, coelomic cavity (extraembryonic);cp, cardiogenic plate mesoderm; eb, epiblast; ee, epidermal ectoderm; fg, ventral wall of foregut pocket (definitive endoderm); hb, hypoblast; hd, position of future hepatic diverticulum primordium; mg, ventral wall of future midgut (definitive endoderm); n, primitve (Hensen’s) node; ‘n’, presumptive primitive node; nc, notochordal process; np, neural plate; pm, pharyngeal membrane (ectodermal and endodermal part); pp, praechordal plate; ps, primitive streak; rp, Rathke’s pouch primordium; st, mesoderm of septum transversum: tb, trophoblast; td, position of future thyroid diverticulum; ur, position of future umbilical ring: xm, extraembryonic mesoderm; ys, yolk sac epithelium. Asterisk marks the anterior junction between embryonic and extraembryonic tissues. Colors indicate the cell fates according to lineage tracing studies (Lawson erul., 199I ; Quinlan erul., 1995) and gene expression data (see text, Sections II1.B and V,C): blue, derivatives of the epiblast; yellow, derivatives of the hypoblast; green, derivatives of epiblast cells after ingression through the primitive node; red, derivatives of the epiblast ingressed through the primitive streak.

Components of the anterior margin during gastrulation

Fig. 4

3. Anterior Margin of the Mammalian Gastrula

91

the vegetal pole of the blastula could have led to the evolution of different vertebrate gastrula forms from an ancestral protochordate (Fig. 5, see color plate).

B. Yolk Accumulation in the Vegetal Hemisphere

As a first step toward recent gastrula forms, the protochordate blastula wall in the region of the blastopore may have bulged into the blastocoel by increasing its yolk content and thereby pushing the capacity of the gastrulation center (Jeffery, 1992) into the interior of the embryo. Through some variability in this bulging process two principal alternatives may have evolved (Fig. 5). 1. Combined with an overwhelming accumulation of yolk, the whole of the blastoporal lining of the blastocoel could have become enlarged; in this way, a fish gastrula would have evolved with its characteristic big yolk cell vegetally and a cap-shaped blastodenn at the animal pole. The margins of this animal cap are homologous with the amphibian blastopore lips and create the impression of a widely open blastopore, covered by the yolk cell in the manner of an amphibian yolk plug, well before the invagination of the definitive germ layers begins at these margins. Due to its size, the yolk cell compresses the blastocoel to the narrow slit seen between the yolk syncytial layer (YSL) and the blastoderm. An asymmetrical enlargement of the blastocoel lining would have brought the gastrulation center into a dorsal (posterior) position within the YSL, a position in agreement with the mesodem induction results of Oppenheimer (1936) and Mizuno et al. (1 996). 2. The combination of a moderate increase in the yolk content and an asymmetrical enlargement of the blastoporal lining of the blastocoel could have led, alternatively, to the evolution of an amphibian gastrula: the blastocoel cavity is only moderately reduced in relative size and the vegetal half is, again, enlarged anteriorly more than posteriorly, with the result that the gastrulation center, now called the “Nieuwkoop center,” again lies posteriorly near the dorsal blastoporal lip.

The (evolutionary) inward movement during this first step should not be confused with the invagination of the definitive endoderm (and mesoderm) seen in invertebrates or with the vertebrate germ layer involution that starts at the blastoFig. 5 Homologies and possible phylogenetic pathways between different vertebrates at the blastulagastrula transition. Red, pink, brown, and orange denote phylogenetically active parts in the phylogenetically more primitive form and the derivatives of these parts in the evolved form. Blue discs denote the presumed equivalents of the amphibian Nieuwkoop center (N) in the protochordate (G, the ascidian gastrulation center of Jeffery (1992)), in the chick (K, Koller’s sickle), and in the mammalian embryo (A, anterior hypoblast). Green discs denote the equivalent of Spemann’s organizer (the dorsal mesoderm). c, Chorion; tb, trophoblast; ysl, yolk syncytial layer. Asterisks denote the junctions hetween embryonic and extraembryonic regions.

92

Christoph Viebahn

pore lips in amphibia or at the margin of the blastoderm in fish. Instead, the structures that derive from the inner lining of the protochordate blastopore wall are hypoblast equivalents such as the YSL of the zebrafish or the yolky endoderm at the floor of the blastocoel in the frog embryo. These new cellular layers deal with the increased amount of stored nutrient (yolk) and make it available to the embryo. These layers are, therefore, not to be compared with the germ layers, which may be suggested by the alternative names (primitive endoderm, primary endoderm) of the hypoblast, their mammalian homolog. C. Generation of Extraembryonic Membranes

In a second evolutionary step toward recent gastrula forms, truly extraembryonic tissues such as the avian chorion or the mammalian trophoblast are generated in the ancient amphibian blastula, again along two alternative pathways. 1. The outer cellular layer of the vegetal half may break open and retract to pave the way for a further massive increase of yolk material, as seen in the avian embryo. This yolk mass is only partly covered by the former outer layer of the vegetal hemisphere that has transformed into the chorion; the chorion has acquired the novel (long-term) function of gas exchange with the environment and can thus be considered a truly extraembryonic tissue. At the same, time the (endodermal) floor of the blastocoel bulges further toward the animal hemisphere and compresses the blastocoel to a slitlike cavity between the epiblast and hypoblast, the subblastodermic cavity (Hubrecht, 1890; Bonnet, 1918; Eyal-Giladi, 1997). During this bulging process, however, the anterior part of the vegetal hemisphere may have lengthened more than its posterior part; this would displace the Nieuwkoop center further posteriorly and would thus create the region with axis-inducing capacity in the chick, i.e., Koller’s sickle. 2. The alternative pathway, leading to a mammalian gastrula, is based on the generation of a new cavity (the blastocyst cavity) in the central part of the uncleaved yolk mass (“Furchungshohle”) (Peter, 1922), with the latter not necessarily increasing in relative size. In this case, the outer (covering) layer of the vegetal hemisphere transforms into the trophoblast: it acquires a capacity that goes beyond simple gas exchange by making intimate contact apically with the maternal tissues [the “cell biological paradox” of Denker (1993)l. This mode of vegetal transformation would also be accompanied by bulging of the endoderma1 floor of the blastocoel, but this time the posterior part may lengthen more than the anterior part; this would push the Nieuwkoop center anteriorly and create the anterior bulge of the mammalian embryo, the hypoblast part of the AMC with its pattern-forming capacity (see Section V,A). In order to facilitate comparisons between the avian and the mammalian gastrula, the morphology of a lower mammal, such as hedgehog or bat, is used here (Fig. 5) because in these species the trophoblast consists (during early gastrula-

3. Anterior Margin of the Mammalian Gastrula

93 tion) of a mural part only (after degeneration of the polar part), i.e., the margin of the embryonic disc is continuous laterally with the trophoblast, as in the avian embryo. During early (ontogenetic) development, of course, the trophoblast surrounds completely the inner cell mass, but this could be explained by an (evolutionary) overgrowth from laterally as a reaction to the immersion in, or prolonged contact with, the uterine epithelium.As discussed in Section II,B, lower mammals tend to use the mural trophoblast for placenta formation whereas higher mammals use the polar trophoblast (the part covering the inner cell mass), but this cannot be regarded as a principal difference within the order of the mammals. This second asymmetrical enlargement of the hypoblast analog of the ancient amphibian embryo would thus allow for the back-to-front inversion to occur, which is seen in the mammalian and the avian early gastrula with regard to distribution of primordial germ cells. The model has the additional feature (which may be appealing for higher mammals) that if an axial differentiation (and signaling center) appeared earlier and anteriorly (instead of posteriorly), this may give anterior structures (such as the developing brain) in mammals a “head start” over its avian counterparts. On the other hand, it will be intriguing to find out how the molecular pathways used by Koller’s sickle for induction of the primitive streak in the chick (Mitrani et al., 1990)can be matched with those used by the anterior hypoblast for patterning the overlying epiblast (Thomas and Beddington, 1996)and for the subsequent induction of the primitive streak in the mammal (Varlet et al., 1997).The use of the same growth factor system (e.g., BMP4 and chordin) in a heterochronical fashion (neural induction in mammals vs. mesoderm induction in birds; see Section V,B) (Kessel and Pera, 1998) suggests that the evolution of developmental processes may take unexpected turns. Notwithstanding the lower layer’s inductive role for axis differentiation, the timing for completion of the lower layer may still be related to its primary function, communication between embryo and its initial nutrient, the yolk. Depending on the form in which yolk is present and used during development, the lower layer is completed either before gastrulation has commenced (as in amphibia and mammals), or afterward (as in birds).

VII. Conclusions and Outlook A. Organizing Gastrulation

The ready availability of chicken embryos and the relative ease of in v i m cultivation and experimental manipulation of primitive-steak and early organogenesis stages has led to clarification of a wealth of developmental processes, not only during gastrulation. Given the morphological similarities between the avian and mammalian embryo during the later stages of gastrulation, it seems very likely that most of these processes should also act in the mammalian embryo. The striking morphological differences during the stages leading up to gastrulation, how-

94

Christoph Viebahn

ever, make new approaches for the mammalian embryo necessary. In this context, the anterior margin in pregastrulation stages of the rabbit serves both as a convenient landmark and as a piece of tissue for classical induction studies, which, in the rodent embryo, are technically challenging and difficult to reproduce. The cells that will form the primitive streak, and the primitive node for that matter, can now be “pinned down” and experimentally manipulated in vitro before they openly engage in their assignment, i.e., in the forming of the primitive streak and node (Viebahn and HrabC de Angelis, 1995). In addition, once the primitive node is established, the anterior margin is still a prominent feature of the embryo and would be in a prime place for conducting the control of the molecular left-right differentiation in the vicinity of the primitive node, which is well described in the chick (Levin et al., 1995) but still eludes us in the mammalian embryo. Combining the experimental approach with recent molecular advances in the mouse (Beddington and Robertson, 1998) we may now open a new chapter of gastrulation research in mammals. It is hoped that the factors organizing gastrulation in mammals will soon become as accessible to experimental study as they have been for years in lower vertebrates and birds. 6. Gastrulation Staging

According to the definitions of gastrulation given in Section I it would be logical to say that the anterior margin marks the beginning of gastrulation: it is actively engaged in the differentiation of the epiblast into the neural part of the definitive ectoderm and thereby specifies one of the three germ layers characteristic of the gastrula stage. In addition, it is the first morphological landmark of the anteroposterior axis and can for that reason also be regarded as an integral part of gastrulation. It might well turn out to be a useful structure on which a uniform staging system of the mammalian gastrulation could be based. However, the ultimate fate of the cells making up the AMC still needs to be investigated before it can be integrated into the well-known set of morphogenetic movements of gastrulation. C. Evolution

The early differentiation of the anterior margin of the mammalian gastrula invites us to take a fresh look at the evolution of gastrulation. A general principle of gastrulation in chordates seems to be that a cellular layer lying underneath or outside the cell populations giving rise to the body plays a critical role in initiating gastrulation and inducing axis differentiation. This might be the reason why, in amniotes, the hypoblast (as the lower layer) tends to display essentially the same configuration whereas other extraembryonic structures such as amnion and the trophoblast show such a variety of forms, especially in mammals. As the most

3. Anterior Margin of the Mammalian Gastrula

95

constant feature of vertebrate gastrulation, this lower layer may thus help to find homologies for the regulative processes acting during gastrulation in the animal kingdom, which, in turn,may point out the main avenues in the evolution of gastrulation.

Acknowledgments The skillful technical assistance of W. Langmann and the diligent help with the photographic work by L. Droese are gratefully acknowledged. Special thanks go to Klaus Sander and Karen Downs, who have commented on earlier versions of the manuscript. This work is supported by the Deutsche Forscbungsgemeinschaft (Vi 151/1-2 and Vi l5l/3-1) and by the BONFOR research commission of the Medical Faculty, University of Bonn.

References Abdelilah, S., Solnica-Krezel, L., Stainier, D. Y. R., and Driever, W. (1994). Implications for dorsoventral axis determination from the zebrafish mutation janus. Nature (London)370,468-471. Acampora, D., Mazan, S., Lallemand, Y., Avantaggiato, V., Maury, M., Simeone, A,, and Brulet, P. (1995).Forebrain and midbrain regions are deleted in Otx2-’- mutants due to a defective anterior neuroectoderm specification during gastrulation. Development 121,3279-3290. Adelmann, H. B. (1926). The development of the premandibular head cavities and the relations of the anterior end to the notochord in the chick and robin. J. Morphol. Physiol. 42,371-439. Arai, A., Yamamoto, K.,and Toyama, J., (1997). Murine cardiac progenitor cells require visceral embryonic endoderm and primitive streak for terminal differentiation. Dev. Dyn. 210,344353. Arendt, D., and Niibler-Jung, K. (1994). Inversion of dorsoventral axis? Nature (London) 371,26. Arman, E., Haffner-Krausz, R., Chen, Y., Heath, J. K., and Lonai, P. (1998). Targeted disruption of fibroblast growth factor (FGF) receptor 2 suggests a role for FGF signaling in pregastrulation mammalian development. Proc. Nutl. Acad. Sci. U.S.A. 95,5082-5087. Assheton, R. (1895).The primitive streak of the rabbit; the causes which may determine its shape, and the part of the embryo formed by its activity. Q. J. Microsc. Sci. 37, 191-221. Assheton, R. (1 899). The segmentation on the ovum of the sheep, with observations on the hypotheses of a hypoblastic origin for the trophoblast. Q. J. Microsc. Sci. 41,205-262. Azar, Y., and Eyal-Giladi, H. (1981). Interaction of epiblast and hypoblast in the formation of the primitive streak and the embryonic axis in chick, as revealed by hypoblast rotation experiments. J. Embryol. Exp. Morphol. 61, 133-141. Badwaik, N. K., Rasweiler, J. J., 4th, and Oliveira, S. F. (1997).Formation of reticulated endoderm, Reichert’s membrane, and amniogenesis in blastocysts of captive-bred, short-tailed fruit bats, Carollia perspicillata. Anat. Rec. 247,85- 101. Bauer, D. V., Huang, S., and Moody, S. A. (1994). The cleavage stage origin of Spemann’s organizer: Analysis of the movements of blastomere clones before and during gastrulation in Xenopus. Development 120, 1179-1189. Beddington, R. S. P. (1981). An autoradiographic analysis of the potency of embryonic ectoderm in the 8th day postimplantation mouse embryo. J. Embryol. Exp. Morphol. 64,87-104. Beddington, R. S. P. (1994). Induction of a second neural axis by the mouse node. Development 120, 61 3-620. Beddington, R. S. P., and Robertson, E. J. (1998).Anterior patterning in mouse. Trends Genet. 14, 217-284.

96

Christoph Viebahn

Belo, J. A., Bouwmeester, T., Leyns, L., Kertesz, N., Gallo, M., Follettie, M., and De Robertis, E. M. (1997). Cerberus-likeis a secreted factor with neutralizing activity expressed in the anterior primitive endoderm of the mouse gastrula. Mech. Dev. 68,45-57. Biben, C., Hatzistavrou, T., and Harvey, R. P. (1998a). Expression of NK-2 class homeobox gene Nkr2-6 in foregut endoderm and heart. Mech. Dev. 73,125-127. Biben, C., Stanley, E., Fabri, L., Kotecha, S., Rhinn, M., Drinkwater, C., Lah,M., Wang, C. C., Nash, A., Hilton, D., Ang, S. L., Mohun, T., and Harvey, R. P. (1998b). Murine cerberus homologue mCer- 1: A candidate anterior patterning molecule. Dev. Biol., 194, 135- 151. Blum, M., Gaunt, S. J., Cho, K. W. Y.,Blumberg, B., Steinbeisser, H., Bittner, D., and De Robertis, E. M. (1992). Gastrulation in the mouse? The role of the homeobox gene goosecoid. Cell 69, 1097- 1106. Bonnet, R. (1901). Beitrage zur Embryologie des Hundes. Erste Fortsetzung.Anat. Hefte 1(Abt. 51), 16. Bonnet, R., (1918). “Lehrbuch der Entwicklungsgeschichte.” Paul Parey, Berlin. Bonnevie, K. (1950). New facts on mesoderm formation and proamnion derivatives in the normal mouse embryo. J. Morphol. 86,495-546. Bouwmeester, T., and Leyns, L. (1997). Vertebrate head induction by anterior primitive endoderm. B~OESSUYS 19,855-863. Bouwmeester, T., Kim, S. H., Sasai, Y.,Lu, B., and De Robertis, E. M. (1996). Cerberus is a head-inducing secreted factor expressed in the anterior endoderm of Spemann’s organizer. Nature (Lmdon) 382,595-601. Callebaut, M., and van Nueten, E. (1994). Rauber’s (Koller’s) sickle: The early gastrulation organizer of the avian blastoderm. Eu,: J. Morphol. 32,35-48. Chapman, D. L., Agulnik, I., Hancock, S., Silver, L. M., and Papaioannou, V. E. (1996a). Tbx6, a mouse T-box gene implicated in paraxial mesoderm formation at gastrulation Dev. Biol. 180, 534-542. Chapman, D. L., Garvey, N., Hancock, S., Alexiou, M., Agulnik, S. I., Gibsonbrown, J. J., Cebrathomas, J., Bollag, R. J., Silver, L. M., and Papaioannou,V. E. (1996b). Expression of the T-box family genes, Tbxl-Tbx5, during early mouse development. Dev. Dyn. 206,379-390. Cho, K. W. Y., Blumberg, B., Steinbeisser, H., and De Robertis, E. M. (1991). Molecular nature of Spemann’s organizer-The role of the Xenopus homeobox gene goosecoid. Cell 67,1111-1120. Cleine, J. H. (1986). Replacement of posterior by anterior endoderm reduces sterility in embryos from inverted eggs of Xenopus laevis. J. Embryol. Exp. Morphol. 94,83-93. Conklin, E. G. (1932). The embryology of amphioxus. J. Morphol. 54,69-151. Conlon, F. L., Lyons, K. M., Takaesu, N., Barth, K. S., Kispert, A., Herrmann, B., and Robertson, E. J. (1994). A primary requirement for nodal in the formation and maintenance of the primitive streak in the mouse. Development 120, 1919-1928. Cody, G. F., Coltey, P. M., and Le Douarin, N. M. (1992). The developmentalfate of the cephalic mesoderm in quail-chick chimeras. Development 114,l- 15. Curtiss, J., and Heilig, J. S. (1998). DeLIMiting development. BioEssays 20,58-69. Dale, J. K., Vesque, C., Lints, T. J., Sampath, T. K., Furley, A., Dodd, J., and Placzek, M. (1997). Cooperation of BMP7 and SHH in the induction of forebrain ventral midline cells by prechordal mesoderm. Cell 90,257-269. Davis, C. L. (1927). Development of the human heart from its first appearance to the stage found in embryos of twenty paired somites. Catnegie Inst. (Washingron) Publ, 380, Contrib. Embryol. 19, 245-284. Denker, H. W. (1993). Implantation-A cell biological paradox. J, Exp. 2001. 266,541-558. De Robertis, E. M., Fainsod, A., Gont, L. K., and Steinbeisser, H. (1994). The evolution of vertebrate gastrulation.Development (Suppl. 1994), 117-124. De Robertis, E. M., and Sasai, Y.(1996). A common plan for dorsoventrd patterning in bilateria. Nature (London) 380,37-40. de Vries, P. A. (1981). Evolution of precardiac and splanchnic mesoderm in relationship to the in-

3. Anterior Margin of the Mammalian Gastrula

97

fundibulum and truncus. In “Perspectives in Cardiovascular Research; Mechanisms of Cardiac Morphogenesis and Teratogenesis” (T. Pexieder, ed.), Vol. 5, pp. 31-48. Raven Press, New York. Downs, K. M., and Davies, T. (1993). Staging of gastrulating mouse embryos by morphological landmarks in the dissecting microscope. Development 118,1255- 1266. Driever, W. (1995). Axis formation in zebrafish. Cur,: Opin. Genet. Dev. 5,610-618. Dunwoodie, S . L., Rodriguez, T. A., and Beddington, R. S. P. (1998). MsgI and MrgI, founding members of a gene family, show distinct patterns of gene expression during mouse embryogenesis. Mech. Dev. 72,27-40. Enders, A. C., Given, R. L., and Schlafke, S. (1978). Differentiationand migration of endoderm in the rat and mouse at implantation.Anat. Rec. 190,65-78. Enders, A. C., Schlafke, S., and Hendrickx,A. G. (1986). Differentiationof the embryonic disc, arnnion, and yolk sac in the rhesus monkey. Am. J. Anat. 177,161-186. Eyal-Giladi, H. (1997). Establishmentof the axis in chordates: Facts and speculations.Development 124,2285-2296. Eyal-Giladi, H., and Kocbv, S. (1976). From cleavage to primitive streak formation:A complementary normal table and a new look at the first stages of the development of the chick. Dev. B i d . 49, 321-337. Eyal-Giladi, H., Debby, A., and Harel, N. (1992). The posterior section of the chick’s area pellucida and its involvement in hypoblast and primitive streak formation. Development 116,8 19-830. Fishman, M. C., and Chien, K. R. (1997). Fashioning the vertebrate heart: Earliest embryonic decisions. Development 124,2099-2117. Gannon, M., and Bader, D. (1995). Initiation of cardiac differentiation occurs in the absence of antenor endoderm. Development 121,2439-2450. Garcia-Martinez,V., and Schoenwolf, G. C. (1993). Primitive-streak origin of the cardiovascular system in avian embryos. Dev. Biol. 159,706-719. Garcia-Martinez,V., Alvarez, I. S., and Schoenwolf, G. C. (1993). Locations of the ectodermal and nonectodermal subdivisions of the epiblast at stages 3 and 4 of avian gastrulation neurulation. J. EXP.2001. 267,43 1-446. Gardner, R. L. (1986). Cell mingling during mammalian embryogenesis.J. Cell Sci. (Suppl.) 4,337356. Gardner, R. L. (1997). The early blastocyst is bilaterally symmetrical and its axis of symmetry is aligned with the animal-vegetal axis of the zygote in the mouse. Development 124,289-301. Gardner, R. L., and Rossant, J. (1979). Investigation of the fate of 4.5 day post-coitum mouse inner cell mass cells by blastocyst injection. J. Embryol. Exp. Morphol. 52,141- 152. Gardner, R. L., Meredith, M. R., and Altman, D. G. (1992). Is the anterior-posterior axis of the fetus specified before implantation in the mouse? J. Exp. Zool. 264,437-443. Geoffroy Saint-Hilaire,E. (1822). Considerations gbnerales sur la vertkbre. M6m. Mus. Hist. Nat. 9, 89-119. Gerhart, J., Danilchik, M., Doniach, T., Roberts, S . , Browning, B., and Stewart, R. (1989). Cortical rotation of the Xenopus egg: Consequences for the anteroposteriorpattern of embryonic dorsal development. Development (1989 Suppl.), 37-51. Gevers, P.,Dulos, J., Schipper, H., and Timmermans, L. P. M. (1993). Origin of primordial germ cells, as characterized by the presence of nuage, in embryos of the teleost fish Barbus conchonius. Eu,: J. Molphol. 30,195-204. Ginsburg, M. (1994). Primordial germ cell formation in birds. In “Germline Development” (J. Marsh, ed.), pp. 52-61. Wiley, London. Glinka, A., Wu, W., Delius, H., Monaghan, A. P., Blumenstock, C., and Niebrs, C. (1998). Dickkopf1 is a member of a new family of secreted proteins and functions in head induction. Nature (London) 391,357-362. Goldstein, B., and Freeman, G. (1997). Axis specification in animal development. BioEssays 19, 105-116.

98

Christoph Viebahn

Goldstein, B., and Hird, S. N. (1996). Specification of the anteroposterior axis in Caenorhabdiris elegans. Development 122,1467-1474. Gould, S. J. (1997). ‘‘Full House: The Spread of Excellence from Plat0 to Darwin.” Random House, New York. Gu, Z., Nomura, M., Simpson, B. B., Lei, H., Feijen, A,, van den Eijnden-van Raaij, J., Donahoe, P.K., and Li, E. (1998). The type I activin receptor ActRIB is required for egg cylinder organization and gastrulation in the mouse. Genes Dev. 12,844-857. Haeckel, E. (1874). Gastreatheorie. Jen. Zeitschr: Nururwiss. 8. Hamburger, V., and Hamilton, H. L. (1993). A series of normal stages in the development of the chick embryo. Dev. Dynam. 195,23 1-272. Hamilton, W. J., and Mossman, H. W. (1972). “Hamilton, Boyd and Mossman’s Human Embryology. Prenatal Development of Form and Function.” Macmillan, New York. Hanisson, F., Callebaut, M., and Vakaet, L. (1991). Features of polyingression and primitive streak ingression through the basal lamina in the chicken blastoderm. Anat. Rec. 229,369-383. Hay, E. D. (1995). An overview of epithelio-mesenchymal transformation. Acra Anat. 154,8-20. Hendrickx, A. G. (1971). “Embryology of the Baboon.” Univ. Chicago Press, Chicago. Hensen, V. (1876). Beobachtungen uber die Befruchtung und Entwicklung des Kaninchens und Meerschweinchens. Z. Anat. Enfwickl. Gesch. 1,213-273; 353-432. Hermesz, E., Mackem, S., and Mahon, K. A. (1996). Rpx: A novel anterior-restricted homeobox gene progressively activated in the prechordal plate, anterior neural plate and Rathke’s pouch of the mouse embryo. Developmenr 122,41-52. Hertig, A. T., Rock, J., and Adams, E. C. (1956). A description of 34 human ova within the first 17 days of development. Am. J. Anat. 98,435-493. Heuser, C. H., and Streeter, G. L. (1941). Development of the macaque embryo. Contrib. Embryol. Carnegie Insf. 29, 15-55. Hill, J. P. (1932). The developmental history of the primates. Phil. Trans. R. SOC.Lond. Sec B 221, 45-165. Hill, J. P., and Florian, J. (1963). The development of the primitive streak, head-process and annular zone in Tursius, with comparative notes on Loris. In “Bihliotheca Primatologica” (H. Hofer, A. H. Schultz, and D. Starck, eds.), Vol. 2, pp. 1-90. Karger, Basel. Hill, J. P., and Tribe, M. (1924). The early development of the cat (Felis domestica). Q. J. Microsc. Sci. 68,513-602. Holland, P. W. H., Koschorz, B., Holland, L. Z., and H e m a n n , B. G. (1995). Conservation of brachyury (t) genes in amphioxus and vertebrates: Developmental and evolutionary implications. Developmenr 121,4283-4291. Houart, C., Westerfield, M., and Wilson, S. W. (1998). A small population of anterior cells patterns the forebrain during zebrafish gastrulation. Nature (London) 391,788-792. Huber, G. C. (1915). The development of the albino rat, Mus norvegicus albinus. I. From the pronuclear stage to the stage of mesoderm anlage; end of the first to the end of the ninth day. J. Morphol. 26,247-358. Hubrecht, A. A. W. (1890). Studies in mammalian embryology. 11. The development of the germinal layers of Sorex vulguris. Q. J. Microsc. Sci. 31,499-562. Hubrecht, A. A. W. (1909). Early ontogenetic phenomena in mammals and their bearing on our interpretation of the phylogeny of the vertebrates. Q. J. Microsc. Sci. 53, 1- 181. Inagaki, T., Garcia-Martinez, V., and Schoenwolf, G. C. (1993). Regulative ability of the prospective cardiogenic and vasculogenic areas of the primitive streak during avian gastrulation. Dev. Dynam. 197,57-68. Izpislia-Belmonte, J. C., De Robertis, E. M., Storey, K. G., and Stem, C. D. (1993). The homeobox gene goosecoid and the origin of the organizer cells in the early chick blastoderm. Cell 74, 1-20. Jeffery, W. R. (1992). A gastrulation center in the ascidian egg. Developmenr (Suppl. 1992). 53-63. Kadokawa, Y., Kato, Y.,and Eguchi, G. (1987). Cell lineage analysis of the primitive and visceral endoderm of mouse embryos cultured in vitro. Cell D i g 21,69-76.

3. Anterior Margin of the Mammalian Gastrula

99

Kaufman, M. H. (1992). “The Atlas of Mouse Development.” Academic Press, London. Kazanskaya, 0. V., Severtzova, E. A., Barth, K. A., Ermakova, G. V., Lukyanov, S. A,, Benyumov, A. O., Pannese, M., Boncinelli, E., Wilson, S. W., and Zaraisky, A. G. (1997). An$ A novel class of vertebrate homeobox genes expressed at the anterior end of the main embryonic axis. Gene 200,25-34. Keibel, F. (1917). Zu Carl Rabl’s “Edouard van Beneden und der gegenwiktige Stand der wichtigsten von ihm behandelten Probleme.” Arch. Mikrosk. Anat. 89, 1-15. Kessel, M., and Pera, E. (1998). Unexpected requirements for neural induction in the avian embryo. Trends Genet. 14, 169-171. Khaner, 0.(1995). The rotated hypoblast of the chicken embryo does not initiate an ectopic axis in the epiblast. Proc. Nad Acad. Sci. U.S.A. 92, 10733-10737. Khaner, 0. (1988). The ability to initiate an axis in the avian blastula is concentrated mainly at a posterior site. Dev. B i d . 194,257-266. Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B., and Schilling, T. F. (1995). Stages of embryonic development of the zebrafish. Dev. Dynam. 203,253-310. Koller, C. (1882).Untersuchungen iiber die Blatterbildung im Hiihnerkeim. Arch. Mikmsk. Anat. 20, 174-21 1. Kolliker, A. (1879). “Entwicklungsgeschichte des Menschen und der hoheren Thiere.” Wilhelm Engelmann, Leipzig. Kolliker, R. A. (1882). Die Entwicklung der Keimblatter des Kaninchens. In “Festschrift zur 3. Skikularfeier der Alma Julia Maximiliana (Wiirzburg)” pp. 1-51. Wilhelm Vogel, Leipzig. Kollmann, J. (ed.) (1886). “Gastrulasitzung” der 59. Versammlung deutscher Naturforscher und k z t e zu Berlin. Anar. Anz. 1,281-294. Kuhn, H. J., and Liebherr, G. (1987). The early development of the heart of Tupaia belangeri, with reference to other mammals. Anat. Embryol. 176,53-63. Landry, C., Clotman, F., Hioki, T., Oda, H., Picard, J. J., Lemaigre, F. P.,and Rousseau, G. G. (1997). HNF-6 is expressed in endoderm derivatives and nervous system of the mouse embryo and participates in the cross-regulatory network of liver-enriched transcription factors. Dev. Biol. 192,247-257. Lawson, K. A,, Meneses, J. J., and Pedersen, R. A. (1986). Cell fate and cell lineage in the endoderm of the presomite mouse embryo, studied with an intracellular trace. Dev. Biol. 115,325-339. Lawson, K. A., Meneses, J. J., and Pedersen, R. A. (1991). Clonal analysis of epiblast fate during germ layer formation in the mouse embryo. Development 113,891-91 1. Lemaire, P., Garrett, N., and Gurdon, J. B. (1995). Expression cloning of siamois, a Xenopus homeobox gene expressed in dorsal vegetal cells of blastulae and able to induce a complete secondary axis. Cell 81,85-94. Lemaire, P., Darras, S., and Caillol, D., and Kodjabachian, L. (1998). A role for the vegetally expressed Xenopus gene Mix. 1 in endoderm formation and in the restriction of mesoderm to the marginal zone. Development 125,237 1-2380. Levin, M., Johnson, R. L., Stem, C. D., Kuehn, M., and Tabin, C. (1995). A molecular pathway determining left-right asymmetry in chick embryogenesis. Cell 82,803-814. Levin, M., Pagan, S., Roberts, D. J., Cooke, J., Kuehn, M. R., and Tabin, C. J. (1997). Lefthght patterning signals and the independent regulation of different aspects of situs in the chick embryo. Dev. Biol. 189,57-67. Liebke, M., and Viebahn, C. (1999). In preparation. Lubosch, W. (1931). Gastrula und Gastrulation im Lichte der neuesten Forschungen. Gegenbaurs Morphol. Jahrb. 66,583-607. Luckett, W. P. (1978). Origin and differentiation of the yolk sac and extraembryonic mesoderm in presomite human and rhesus monkey embryos. Am. J. Anat. 152,59-98. Mitchell, S. A., Viebahn, C., and Blum, M. (1999). In preparation. Mitrani, E., Ziv, T., Thomsen, G., Shimoni, Y.,Melton, D. A., and Bril, A. (1990). Activin can induce the formation of axial structures and is expressed in the hypoblast of the chick. Cell 63,495-501.

100

Christoph Viebahn

Mizuno, T., Yamaha, E., Wakahara, M., Kuroiwa, A,, and Takeda, H. (1996). Mesoderm induction in zebrafish. Nature (London)383, 131-132. Mossman, H. W. (1937). Comparativemorphogenesisof the fetal membranes and accessory uterine structures. Contrib. Embryol. Carnegie Inst. 26, 129-246. Muniesa, P., and Dominguez, L. (1990). A morphological study of primordial germ cells at pregastrular stages in the chick embryo. Cell Difi Dev. 31, 105-117. Munn, K., and Steward, R. (1995). The anterior-posterior and dorsal-ventral axes have a common origin in Drosophila melanogustex BioEssays 17,920-922. Nadijcka, M., and Hillman, N. (1974). Ultrastructural studies of the mouse blastocyst substages. J. Embryol. Exp. Morphol. 32,675-695. Nakatani, Y., and Nishida, H. (1994). Induction of notochord during ascidian embryogenesis. Dev. Biol. 166,289-299. Narita, N., Bielinska, M., and Wilson, D. B. (1997). Wild-type endoderm abrogates the ventral developmental defects associated with GATA-4 deficiency in the mouse. Dev. Biol. 189,270-274. Nishida, H. (1994). Localization of determinants for formation of the anterior-posterior axis in eggs of the ascidian Halocynthia roretzi. Development 120,3093-3 104. Nomura, M., and Li, E. (1998). Smad.2 role in mesoderm formation, left-right patterning and craniofacial development.Nature (London)393,786-790. Oliver, G., Mailhos, A., Wehr, R., Copeland, N. G., Jenkins, N. A,, and Gruss, P. (1995). Six3, a murine homologue of the sine oculis gene, demarcates the most anterior border of the developing neural plate and is expressed during eye development.Development 121,4045-4055. Oppel, A. (1890). Uber Vorderkopfsomiten und die Kopfhohle von Anguis fragilis. Arch. Mikrosk. Anat. 36,603-627. Oppenheimer, J. M. (1936). The development of isolated blastoderms of Fundulus heterclirus. J. EXP.z00l. 72,274-269. O’Rahilly, R., and Miiller, F. (1987). Developmental stages in human embryos. Carnegie Inst. (Washington)Publ. 637. Pagh-Westphal, S. M., and Tabin, C. J. (1998). The transfer of left-right positional information during chick embryogenesis.Cell 93,25-35. Parker, K. M. (1917). The development of the hypophysis cerebri, pre-oral gut, and related structures in the marsupalia. J. Anat. 51, 13-249. Pasteels, J. (1936). Etudes sur la gastrulation des vert6br6s m6roblastiques. In. Oiseaux. IV. Conclusions g6n6rales. Arch. Biol. (Lidge) 48,381-488. Pera, E. M., and Kessel, M. (1997). Patterning of the chick forebrain anlage by the prechordal plate. Development 124,4153-4162. Pera, E. M., and Kessel, M. (1998). Demarcation of ventral territories by the homeobox gene NKx2.1 during early chick development. Dev. Genes Evol. 208,168-171. Peter, K. (1922). h e r den Begriff “Homologie” und seine Anwendung in der Embryologie.Biol. Zentralbl. 42,308-327. Peter, K. (1938a). Gastrulation und Homologie.Anat. Anz. 86,94-122. Peter, K. (1938b). Untersuchungen iiber die Entwicklung des Dotterendoderms.I. Die Entwicklung des Entoderms beim Htihnchen. 2.Mikrosk. Anat. Forsch. 43,362-415. Peter, K. (1941). Die Genese des Entoderms bei den Wirbeltieren. Ergeb. Anat. Enmick. Gesch. 33, 285-369. Pevny, L. H., Sockanathan,S., Placzek, M., and Lovell-Badge, R. (1998). A role for SOX1 in neural determination. Development 125, 1967- 1978. Poelmann, R. E. (1981). The head process and the formation of the definitive endoderm in the mouse embryo. Anat. Embryol. 162,41-49. Popperl, H., Schmidt, C., Wilson, V., Hume, C. R., Dodd, J., Krumlauf, R., and Beddington, R. S. P. (1997). Misexpression of Cwnt8C in the mouse induces an ectopic embryonic axis and causes a truncation of the anterior neuroectoderm. Development 124,2997-3005.

3. Anterior Margin of the Mammalian Gastrula

101

Pownall, M. E., Tucker, A. S., Slack, J. M. W., and Isaacs, H. V. (1996). eFGF, Xcad3 and Hox genes form a molecular pathway that establishes the anteroposterior axis in Xenopus. Development 122, 3881-3892. Quinlan, G. A., Williams, E. A., Tan, S. S., and Tam, P. P. L. (1995). Neuroectodermalfate of epiblast cells in the distal region of the mouse egg cylinder: Implication for body plan organization during early embryogenesis.Development 121,87-98. Rabl, C. (1915). Edouard van Beneden und der gegenwmige Stand der wichtigsten von ihm behandelten Probleme. Arch. Mikrosk. Anat. 88,3-470. Raff, M. C. (1996). Size control: The regulation of cell numbers in animal development. Cell 86, 173-175. Rauber,. A. (1 876). Primitivrinne und Urmund. Beitrag zur Entwicklungsgeschichtedes Hiihnchens. Gegenbaurs Jahrb. 2,550-576. Rhinn, M., Dierich, A., Shawlot, W., Behringer, R. R., Le Meur, M., and Ang, S. L. (1998). Sequential roles for Otx2 in visceral endoderm and neuroectoderm for forebrain and midbrain induction and specification. Development 125,845-856. Rosenquist, G. C. (1970). Location and movements of cardiogenic cells in the chick embryo: The heart-forming portion of the primitive streak. Dev. Biol. 22,461-475. Rosenquist, T. A,, and Martin, G. R. (1995). Visceral endoderm-1 (VE-I): An antigen marker that distinguishes anterior from posterior embryonic visceral endoderm in the early post-implantation mouse embryo. Mech. Dev. 49,117-121. Rubenstein, J. L., Shimamura, K., Martinez, S., and Puelles, L. (1998). Regionalizationof the prosencephalic neural plate. Annu. Rev. Neurosci. 21,445-477. Salamat, M., Miosge, N., and Herken, R. (1995). Development of Reichert’s membrane in the early mouse embryo. Anat. Embryol. 192,275-281. Schauinsland,H. (1903). Beitrage zur Entwickelungsgeschichteund Anatomie der Wirbeltiere. II. Studien zur Entwickelungsgeschichteder Sauropsiden.Zoologica 16, 119-143. Schmitz, B., and Campos-Ortega,J. A. (1994). Dorso-ventral polarity of the zebrafish embryo is distinguishable prior to the onset of gastrulation. Roux’s Arch. Dev. Biol. 203,374-380. Schneider. S., Steinbeisser, H., Warga, R. M., and Hausen, P.(1996). P-catenin translocation into nuclei demarcates the dorsalizing Centers in frog and fish embryos. Mech. Dev. 57,191-198. Seessel, A. (1877). Zur Entwicklungsgeschichtedes Vorderdarms. Arch. Anat. Entw. Band. 1 (Arch. Anat. Phys. Anat. Abt.) 449-467. Seifert, R., Jacob, M., and Jacob, H. J. (1993). The avian prechordal head region-A morphological study. J. Anat. 183,75-89. Seleiro, E. A. P., Connolly, D. J., and Cooke, J. (1996). Early developmentalexpression and experimental axis determination by the chicken Vgl gene. Cur,: Biol. 6,1476-1486. Selenka, E. (1 883). “Studien iiber Entwickelungsgeschichteder Thiere. 1. Keimblatter und Primitivorgane der Maus.” Kreidel, Wiesbaden. Selenka, E. (1884). “Studien iiber Entwickelungsgeschichteder Thiere. 3. Die Blatterumkehr im Ei der Nagethiere.” Kreidel, Wiesbaden. Serbedzija, G. N., Chen, J. N., and Fishman, M. C. (1998). Regulation in the heart field of zebrafish. Development 125,1095-1101. Shawlot, W., and Behringer, R. R. (1995). Requirement for liml in head-organizer function. Nature (London) 374,425-430. Shawlot, W., Deng, J. M., and Behringer, R. R. (1998). Expression of the mouse cerberus-related gene, Cerrl, suggests a role in anterior neural induction and somitogenesis.Proc. Natl. Acad. Sci. 95,6198-6203. Simon, D. (1960). Contribution a l’ttude de la circulation et du transport des gonocytes primaires dans les blastodermes d‘oiseau cultivks in vitro. Arch. Anat. Microsc. Morphol. Exp. 49,92176. Smith, L. J. (1985). Embryonic axis orientation in the mouse and its correlation with blastocyst rela-

102

Christoph Viebahn

tionships to the uterus. 11. Relationships from 4 t to 94 days. J. Embryol. Exp. Morphol. 89, 1535. Smith, S. C., Graveson, A. C., and Hall, B. K. (1994). Evidence for a developmental and evolutionary link between placodal ectoderm and neural crest. J. Exp. Zool. 270,292-301. Snow, M. H. L. (1977). Gastrulation in the mouse: Growth and regionalization of the epiblast. J. Embryol. Exp. Morphol. 42,293-303. Sobotta, J. (1903). Die Entwicklung des Eies der Maus vom Schlusse der Furchungsperiode bis zum Auftreten der Amniosfalten. Arch. Mikrosk. Anar. 61,274-330. Sobotta, J. (191 1). Die Entwicklung des Eies der Maus vom ersten Auftreten des Mesoderms an bis zur Ausbildung der Embryonalanlage und dem Auftreten der Amniosfalten. I. Teil: Die Keimblase. Arch. Mikrosk. Anat. 78,271-352. Spemann, H. (1938). “Embryonic Development and Induction.” Yale Univ. Press, New Haven, Connecticut (republished, Garland, New York, 1988). Spemann, H., and Mangold, H. (1924). ijber Induktion von Embryonalanlagen durch Implantation artfremder Organisatoren. Rou’s Arch. Dev. Biol. 100,599-638. Spratt, N. T. (1946). Formation of the primitive streak in the explanted chick blastoderm marked with carbon particles. J. Exp. Zool. 103,259-304. Spratt, N. T., and Haas, H. (1965). Germ layer formation and the role of the primitive streak in the chick. I. Basic architecture and morphogenetic tissue movements. J. Exp. Zool. 158,9-38. Starck, D. (1975). “Embryologie. Ein Lehrbuch auf allgemein biologischer Grundlage. 3. Autlage.” Thieme, Stuttgart. Stem, C. D. (1990). The marginal zone and its contribution to the hypoblast and primitive streak of the chick embryo. Development 109,667-682. St. Johnston, D. (1995). The intracellular localization of messenger RNAs. Cell 81, 161-170. St. Johnston, D., Driever, W., Berleth, T., Richstein, S., and Nusslein-Volhard, C. (1989). Multiple steps in the localization of bicoid RNA to the anterior pole of the Drosophila oocyte. Developmenf (Suppl. 1989), 13-19. Sulik, K., Dehart, D. B., Inagaki, T., Carson, J. L., Vrablic, T., Gesteland, K., and Schoenwolf, G. C. (1994).Morphogenesis of the murine node and notochordal plate. Dev. Dynam. 201,260-278. Swift, C. H. (1914). Origin and early history of the primordial genh-cells in the chick. Am. J. Anar. 15,483-516. Tam,P. P. L., and Zhou, S. X. (1996).The allocation of epiblast cells to ectodermal and germline lineages is influenced by the position of the cells in the gastrulating mouse embryo. Dev. Biol. 178, 124-132. Tam, P. P. L., Parameswaran, M., Kinder, S. J., and Weinberger, R. P. (1997). The allocation of epiblast cells to the embryonic heart and other mesodermal lineages: The role of ingression and tissue movement during gastrulation. Development 124,1631-1642. Tang, S . J., Hoodless, P. A., Lu, Z., Breitman, M. L., McInnes, R. R., Wrana, J. L., and Buchwald, M. (1998).The 7%-2 homeobox gene is a downstream target of BMP signalling and is required for mouse mesoderm development. Development 125,1877-1887. Thomas, P. Q., and Beddington, R. (1996).Anterior primitive endoderm may be responsible for patterning the anterior neural plate in the mouse embryo. Curr: Biol. 6, 1487-1496. Thomas, P. Q., Brown, A., and Beddington, R. S. P. (1998). Hex: A homeobox gene revealing periimplantation asymmetry in the mouse embryo and an early transient marker of endothelial cell precursors. Development 125,8594. Thomsen, G. H., and Melton, D. A. (1993). Processed vgl protein is an axial mesoderm inducer in Xenopus. Cell 74,433-441. Trinkaus, J. P. (1984). “Cells into Organs: The Forces that Shape the Embryo.” Prentice-Hall, New York. Vakaet, L. (1962). Some new data concerning the formation of the definitive endoblast in the chick embryo. J. Embryol. Exp. Morphol. 10,38-57.

3. Anterior Margin of the Mammalian Gastrula

103

van Beneden, E. (191 1). Recherches sur I’embryologie des Mammifires. De la segmentation, de la formation de la cavitt blastodermique et de I’embryon didermique chez le murin. Arch. Biol. 26, 1-63. van Beneden, E. (1912). Recherches sur I’embryologie des mammifires. 11. De la ligne primitive, du prolongement ctphalique, de la notocorde et du mtsoblaste chez le lapin et chez le murin. Arch. Biol. 27, 191-401. van der Stricht, 0. (1923). L’tvolution du blystocyste de chienne. C. R. Ass. Anat. (Lyon),491-494. Varlet, I., Collignon, J., and Robertson, E. J. (1997). nodal expression in the primitive endoderm is required for specification of the anterior axis during mouse gastrulation. Developmenr 124, 10331044. Venvoerd, C. D. A., and van Oostrom, C. G. (1979). Cephalic neural crest and placodes. Adv. A n d . Embryol. 58, 1-75. Viebahn, C. (1995). Epitbelio-mesenchymal transformation during formation of the mesoderm in the mammalian embryo. AcfaAnat. 154,79-97. Viebahn, C., and Downs, K. M. (1999). In preparation. Viebahn, C., and Hrabi de Angelis, M. (1995). Formation of the primitive streak in the mammalian embryo: A case of planar induction? Dev. Biol. 170,773. Viebahn, C., Mayer, B., and Hrabi de Angelis, M. (1995). Signs of the principal body axes prior to primitive streak formation in the rabbit embryo. Anat. Embryol. 192, 159-169. Vogt, W. (1925). Gestaltungsanalyse am Amphibienkeim mit ortlicher Vitalfirbung. I. Teil. Methodik und Wirkungsweise der ortlichen Vitalfdrbung mit Agar als Farbtrager. RouxS Arch. Entwickl. Mech. 106,542-610. Waddington, C. H. (1932). Experiments on the development of chick and duck embryos, cultivated in virro. Phil. Trans. R. SOC.Lond. Ser: B 221, 179-228. Waddington, C. H. (l933a). Induction by the endoderm in birds. Roux’s Arch. Dev. Biol. 128,502521. Waddington, C. H. (1933b). Induction by the primitive streak and its derivatives in the chick. J. Exp. Biol. 10,38-46. Waddington, C. H. (1934). Experiments on embryonic induction. Part 111. A note on inductions by chick primitive streak transplanted to the rabbit embryo. J. Exp. Biol. 11,224-229. Waddington, C. H. (1956). “Principles of Embryology.” Macmillan, New York. Waldrip, W. R.,. Bikoff, E. K.,Hoodless, P. A,, Wrana, J. L., and Robertson, E. J. (1998). Smad2 signaling in extraembryonic tissues determines anterior-posterior polarity of the early mouse embryo. Cell 92,797-808. Weinstein, M., Yang, X., Li, C., Xu, X., Gotay, J., and Deng, C. X. (1998). Failure of egg cylinder elongation and mesoderm induction in mouse embryos lacking the tumor suppressor smad2. Proc. Natl. Acad. Sci. U.S.A. 95,9378-9383. Wetzel, R. (1929). Untersuchungen am Huhnchen. Die Entwicklung des Keims wahrend der ersten beiden Bebriitungstagen. Roux’s Arch. Entw. 119, 188-320. Williams, B. S., and Biggers, J. D. (1990). Polar trophoblast (Rauber’s layer) of the rabbit blastocyst. Anat. Rec. 227,211-222. Wolpert, L. (1992). Gastrulation and the evolution of development. Developmenf (Suppl. 1992), 713. Yoon, C., Kawakami, K., and Hopkins, N. (1997). Zebrafish vusa homologue RNA is localized to the cleavage planes of 2- and 4-cell-stage embryos and is expressed in the primordial germ cells. Development 124,3157-3166. Zhang, H. B., and Bradley, A. (1996). Mice deficient for BMP2 are nonviable and have defects in amnion/chorion and cardiac development. Developmenr 122,2977-2986. Zhou, X., Sasaki, H., Lowe, L., Hogan, B. L. M., and Kuehn, M. R. (1993). Nodal is a novel TGFp-like gene expressed in the mouse node during gastrulation. Nature (London) 361,543-547.

This Page Intentionally Left Blank

4 The Other Side of the Embryo: An Appreciation of the Non-D Quadrants in Leech Embryos David A. Weisblat, FranCoise Z. Huang, Deborah E. Isaksen, Nai-Jia L. Liu, and Paul Chang Department of Molecular and Cell Biology University of California Berkeley, California 94720

I. Introduction and Overview of Leech Development 11. Macromere Behavior during Cleavage

III. Syncytial Yolk Cell Formation N. Regulation of Macromere Fusion V. Epiboly A. Embryos without Germinal Bands B. Embryos with Reduced Epithelium C. A Possible Role for the Macromeres in Epiboly VI. Conclusions A. Cleavage B. Syncytial Yolk Cell Formation C. Regulation of Macromere Fusion D. Epiboly References

1. Introduction and Overview of Leech Development In animals that develop by spiral cleavages (including annelids, mollusks, and several other protostome taxa), the first two cleavage planes include or lie parallel to the animalhegetal axis, generating four blastomeres designated as A, B, C, and D. A general feature of spiralian development is that one quadrant (widely designated as the D quadrant) contributes the bulk of the mesoderm to the embryo. Thus, studies on the mechanisms of cell fate determination in spiralian embryos have tended to focus on the question of how the D quadrant is determined to be different from the A, B, and C quadrants, either by the segregation of developmental determinants, for species in which initial cleavages are unequal, or by cell-cell interactions, for species in which initial cleavages are equal (Goldstein and Freeman, 1997; Freeman and Lundelius, 1992; Pilon and Weisblat, 1997; Boyer et al., 1996). Analysis of segmentation in the leech as an example of embryonic pattern formation also focuses attention on the D quadrant derivaCurrent Topics in Drvelopmrntol Biology, Vol.46 Copyright 0 1999 by Academic Press. All rights of reproduction in any form reserved 0070-2153/99 $30.00

105

106

David A. Weisblat et al.

tives because the D quadrant is the precursor of segmental ectoderm and mesoderm (Wedeen, 1995; Savage and Shankland, 1997). In contrast, the goal of this article is to summarize findings that reveal a phenomenologically rich developmental cell biology in the A, B, and C quadrants in embryos of glossiphoniid leeches, such as Helobdella robusru (class Hirudinea; phylum Annelida). In glossiphoniidleeches, these non-D quadrants are precursors of definitive endoderm (Whitman, 1878; Bychowsky, 1921; Nardelli-Haefliger and Shankland, 1993) among other tissues (Nardelli-Haefliger and Shankland, 1993; Smith and Weisblat, 1994; Huang er al., 1999). To appreciate the roles of these cells in embryogenesis, however, it is necessary first to summarize glossiphoniid leech development in terms of the contributions of the D quadrant. Leeches are examples of “unequal cleavers,” in which factors influencing subsequent cell fates (i.e., determinants) are partitioned unequally during the first two cleavages; in such embryos, the second embryonic axis is therefore recognizable at first cleavage (with the first, animalhegetal axis being recognizable upon the migration of the female pronucleus just prior to polar body formation). In the leech, some determinants are contained within animal and vegetal domains of yolk-free cytoplasm (teloplasm) that arise prior to first cleavage and are distributed unequally during the first two cleavages into the D quadrant (Fernandez and Olea, 1982; Astrow et al., 1987; Nelson and Weisblat 1991, 1992). The eight-cell Helobdella embryo exhibits an apparently typical spiralian configuration of four large vegetal cells (macromeres A’-D’) and four small animal cells (micromeres a’-d’), with the animal quartet skewed clockwise by roughly 45” relative to the vegetal quartet (Fig. IA, stage 4a; Fig. 1C). [In fact, the orientation of the B quadrant spiral cleavages are actually opposite those of the other three macromeres, i.e., levorotatory at the third and fifth cleavages and dextrorotatory at the fourth (Sandig and Dohle, 1988; Huang et al., 1999); the significance of this deviation from typical spiral cleavage will be discussed later.] Subsequently, macromere D’ exhibits an extensively modified pattern of cleavage, giving rise to 5 bilateral pairs of medium-sized blastomeres called teloblasts and 16 additional small blastomeres (all of which we designate as micromeres) during the early phase of development (Fig. lA, stages 4b-7). The teloblasts are embryonic stem cells; through repeated divisions, they each generate a coherent column (bandlet) of segmental founder cells called blast cells; four pairs of teloblasts (designated N, O/P, O/P, and Q) give rise to ectoderm, and one pair (designated M) gives rise to mesoderm (Fig. 1C). On each side of the embryo, the five bandlets come together in a parallel array, forming left and right germinal bands (not to be confused with the Drosophilu germ band); the germinal bands are connected at their distal ends in the vicinity of the animal pole of the embryo (Fig. 1A and B, early stage 8). This is the prospective anterior end of the animal and contains the first blast cells produced from the teloblasts. The germinal bands and the animal territory between them are covered by a squamous epithelium derived from many of the micromeres.

4. Non-D Quadrants in Leech Embryos

107

Continuing divisions of the teloblasts add more blast cells to the proximal, more posterior ends of the germinal bands. As they do so, the germinal bands move over the surface of the embryo and gradually coalesce like a zipper from anterior to posterior, forming the germinal plate along the ventral midline (Fig. 1A and B, mid-stage 8 to stage 9). The movements of the germinal bands are accompanied by an expansion of the overlying squamous epithelium, which thus is undergoing epiboly. The basal surface of the squamous epithelium contacts a sparse network of mesodermally derived muscle fibers (Weisblat et al., 1984). Together, these tissues make up the provisional integument, a temporary body covering for the embryo. The bulk of the embryo consists of the non-D quadrants and the yolky remnants of the teloblasts. Prior to the completion of epiboly, the germinal bands have disconnected from the columns of blast cells proximal to the teloblasts. Both mesodermal and ectodermal teloblasts produce more blast cells than are used to found the segmental tissues. It has previously been assumed that these supernumerary blast cells die (Shankland, 1984), but this now seems unlikely (Desjeux, 1995; Shankland, 1998; Desjeux and Price, 1999). As cell division continues within the germinal plate, it spreads laterally and dorsally over the surface of the yolk (Fig. 2, mid-stage 9). Eventually its lateral edges meet along the dorsal midline, closing the tube that constitutes the main body of the leech (Fig. 2, stage 10).During this period, definitive epidermal and muscle cells replace the provisional integument and enclose the yolk. Also during this period, the “yolk mass,” which marks the prospective midgut, changes from its initially spherical form, first to a pear-shaped mass and then gradually to the multilobed structure that is the midgut of the worm (Fig. 2, late stage 8 to stage 11).We use the ambiguous term “yolk mass” intentionally to defer the problem of defining it in terms of specific cells. When the proboscis has matured and the yolky contents of the midgut have been digested, the juvenile leech is ready for its first meal, in the case of H.robusta, a small freshwater snail. The 11 stages of glossiphoniid leech embryogenesis referred to above were originally defined by Fernandez and Stent (1980) and later refined slightly (Stent et al., 1992). Leech eggs are fertilized internally but arrest in first meiosis until after they are laid (zygote deposition), the beginning of stage 1; thus, we can also time developmental events in terms of hours after zygote deposition (hours AZD) . In standard accounts of leech embryogenesis, the A, B, and C quadrant macromeres are cast in passive and interchangeable roles, initially providing the substrate on which the morphogenetic cell movements of gastrulation are played out, and then later giving rise to or being enveloped by the gut and digested. But as shall be described in the following sections, we are now discovering that the A, B, and C quadrants play active and complex roles during cleavage, gastrulation, and gut formation. Moreover, in at least some of these processes, there is evidence of specific roles for each of the three quadrants.

108

David A. Weisblat et al.

Fig. 1 Summary of leech development. (A) Relevant stages as seen from the animal pole (prospective dorsal views; posterior toward the bottom). The A, B, and C quadrant macromeres are labeled; micromeres (small contours) and the proteloblasts and teloblasts arising from the D quadrant are not. (B) Left equatorial views of stages 8 and 9 highlight the epibolic movements of the germinal bands and micromere-derived epithelium during gastrulation. By early stage 8, the germinal bands (grey) are joined at their anterior (Ant) ends and elongate through the addition of blast cells from the teloblasts at their posterior (Post) ends. During stage 8, they move ventrovegetally over the surface of the embryo (arrows)and gradually coalescefrom anterior to posterior, forming the germinal plate (grey) along the ventral (Vn) midline. By stage 9, germinal plate formation is complete and C"' has fused with A/B to form the syncytial yolk cell A/B/C; dorsal (Dl) territory is indicated. (C) Partial cell lineage diagram for Helobdella, emphasizing micromere production and the cell fusions entailed in formation of the gut precursor. The corresponding developmental stages and hours after zygote deposition are indicated on the time line at left; breaks in the time line denote changes in scale. The macromeres, proteloblast,and teloblasts are indicated in capital letters, as are the fusion products (A/B, A/B/C, and SYC). Blast cells are denoted by lowercaseletters; micromeres are denoted by lowercase letters and primes (e.g., nopq'). Documented cell fusions are denoted by the merger of the various cell lines; dotted lines indicate the omission of bilaterally symmetric lineages (M, and M,; NOPQ, and NOPQ,), the continuing production of blast cells from the teloblasts (M, N, O/P, OIP, and Q), and uncertainties in the timing of fusions of supernumerary blast cells and teloblasts other than the M and N.

II. Macromere Behavior during Cleavage During stage 4, the A, B, and C blastomeres each undergo three highly unequal divisions, forming a macromere and three micromeres (Fig. 1C). The terminology used here to refer to the non-D quadrant cells is modified from the standard

109

4. Non-D Quadrants in Leech Embryos

1b' 1b" 1b"

....

:5 Fig. 1 (continued).

....

Fig. 2 Midgut morphogenesis in HeZobdeZla. Lateral and dorsal views (stages 10 and 11) of later development. The yolk mass and developing midgut are denoted by stippling.

4. Non-D Quadrants in Leech Embryos

111

terminology applied to most annelid and other spiralian embryos (see Smith and Weisblat, 1994). At each highly unequal cleavage, the large and small daughter cells are designated by capital and lowercase versions of the parent cell, respectively and a prime (’) is added to each. Thus, the A quadrant macromere is designated A’, A“, and A”’ after third, fourth, and fifth cleavages, respectively (Fig. 1C). During these micromere-forming divisions, several differences between the A, B, and C quadrant cells emerge. First, cell C is slightly older than cells A and B, because cell CD divides before cell AB at second cleavage and divides ahead of the A and B quadrants throughout the three rounds of micromere-forming divisions. The A and B lineages divide in synchrony, slightly after the C lineage. Another difference, first reported by Sandig and Dohle (1988), is that the B quadrant macromere divisions proceed with the opposite handedness of the A and C quadrants, in violation of canonical spiral cleavage. Thus, the A and B quadrant cells exhibit mirror symmetric divisions with respect to the AB cleavage plane formed at second cleavage. Another potential difference between the C quadrant and the other two non-D quadrants emerges from investigations into the mechanisms of unequal cleavage in Tubifex hattai. Ishii and Shimizu (1995; Shimizu, 1996a) showed that the first mitotic spindle in this oligochaete annelid contains only one centrosome, as judged by y-tubulin immunoreactivity, and is monastral by that criterion. This asymmetry causes the mitotic apparatus and cytokinetic furrow to be shifted to one side and thus leads to the unequal first cleavage; the cell inheriting the aster is the larger, CD cell, whereas cell AB lacks the aster. In CD, the centrosome replicates normally prior to the second cleavage, so that both the C and D daughters inherit one. [Thus, it seems likely that the unequal division of blastomere CD (at second cleavage) is established by a different mechanism than the unequal division of the zygote at first cleavage (Shimizu, 1996b); see also Symes and Weisblat (1992).] In contrast, the AB cell and its A and B progeny exhibit no y-tubulin immunoreactivity and are classified as anastral in Tubifex by Ishii and Shimizu (1995; Shimizu, 1996a). Whether these features will apply to glossiphoniid leeches remains to be determined. When the macromeres arise at third cleavage (Fig. lA, stage 4A), they have relatively simple shapes, like the sections of an orange with only four sections. During stages 4-7, however, the A, B, and C quadrant macromeres undergo extensive and complex changes in shape and position, as cleavages in the D quadrant generate teloblasts (Fig. 3). We have been able to document these changes more accurately by injecting cells with a histochemically detectable enzyme, pgalactosidase. The intensely colored precipitate formed when this enzyme acts on a synthetic indolyl substrate remains insoluble even when the embryos are cleared with organic solvents such as benzyl benzoate/benzyl alcohol. This allows us to see the three-dimensional shapes of the cells by examining the fixed, stained, and cleared embryos at different stages, using transmitted light under the dissecting or compound microscope (Liu et al., 1998) (Fig. 3).

112

David A. Weisblat et al.

Fig. 3 Shape and positional changes of the A, B, and C quadrant cells during cleavage. From the time of their birth (stage 3) through stage 7, the A and B quadrant macromeres undergo little change in position or shape, while the C quadrant macromere spreads clockwise to engulf most of the teloblasts arising from the D quadrant.

As cleavages proceed within the D quadrant to form the 10 teloblasts and 16 micromeres during stages 4-6, the medium-sized teloblasts remain roughly spherical. The space around and between them in the embryo is taken up by the macromeres, chiefly macromere C”’, which consequently shifts clockwise relative to the D quadrant cells and assumes a highly complex shape (Fig. 3). During this time, macromere C”’ also seems to shift along the animalhegetal axis, so as to occupy more territory at the vegetal pole of the embryo. Concurrently, the A”’ and B’”macromeres come to lie in apposition across the midline of the embryo. A” and B”’also shift to take up more territory at the animal pole beneath the micromere cap and withdraw from the vegetal pole. The foregoing description of the movements of the A”’, B’f’, and C macromeres resolves the following paradox regarding the establishment of the anteroposterior (AP) axis and the bilateral symmetry of the early embryo. Since the time of Whitman (1878), four- and eight-cell leech embryos have been illustrated as in Fig. 4A/4B, with the AP axis bisecting the B and D quadrants. Representing the preceding, two-cell stage in this “D-centric” orientation, the first cleavage plane lies at an oblique angle with respect to the AP axis and the corresponding two-cell embryo lies as in Fig. 4A/4B. But it is aesthetically and perhaps even scientifically appealing to assume that the first cleavage is transverse to the AP axis (Fig. 4C). This “CD-centric” orientation leads to a four- and eightcell embryo in which the B-D axis is oblique to the AP axis. Is either view, the D-centric or CD-centric, a better representation of biological reality? In the CD-centric model (Fig. 4C), the A and B quadrant cells are sit-

4. Non-D Quadrants in Leech Embryos

113

uated in apposition at the embryonic midline from the time of their birth onward, in which case, the reverse handedness of the B quadrant cleavages casts the A and B quadrant cells and their progeny as mirror symmetric lineages with respect to the embryonic midline. In this reference system, the unequal second cleavage of cell CD forces a temporary displacement of the sister endodermal (C quadrant) and segmental (D quadrant) lineages. As C and D derivatives shift clockwise and counterclockwise, respectively, during later cleavages, this displacement is “corrected” with both C and D lineages straddling the midline. The alternative, D-centric view (Fig. 4B) requires first that the AP axis is established with a fixed offset with respect to the first cleavage plane, because embryos can develop normally from zygotes in which the first cleavage plane was redirected by compression (Nelson and Weisblat, 1992). From this, it seems that the AP axis is determined by the first cleavage plane, rather than vice versa, or at least that the AP axis is not established irreversibly prior to first cleavage. Moreover, this reference system, with B initially straddling the midline, also entails a concerted clockwise shift of all three non-D quadrant macromeres, to bring A and B into bilaterally symmetric positions by stage 7. Thus, the D-centric view seems like a rather roundabout means for establishing embryonic axes in the leech, but Okham was almost certainly not an embryologist! [The question of how bilateral symmetry is established in spirally cleaving leech embryos is discussed further by Weisblat (1998).] Another process in which the macromeres are involved during cleavage is the envelopment of the D quadrant derivatives by the non-D cells, primarily macromere C (Fig. 3). It seems that this process must require substantial alteration of the cytoskeleton in the enveloping cell, perhaps triggered by the expression of new surface proteins on the nascent proteloblasts and teloblasts. We also observed that macromere C spreads more than A”’ or B’”, enveloping most of the teloblasts and contacting all 10. By contrast, macromere A”‘ contacts only three teloblasts by the beginning of stage 7, even though the A and C quadrant cells start out with equal access to the cleaving D quadrant cells. This apparent difference between the A”’ and C” cells is quantitative rather than qualitative, however. In embryos from which the C quadrant cell is cut away at stages 3-4a, teloblasts arise from the D quadrant and are enveloped by cells A”’and Br”(Isaksen et al., 1999). These results should serve as a reminder that descriptions of cell-specific traits derived solely from normal development do not reveal whether the observed behaviors result from intrinsic differences between the cells or from extrinsic influences.

111. Syncytial Yolk Cell Formation Three developmental pathways can be imagined by which yolky endodermal precursor cells could give rise to a definitive gut epithelium with the concomi-

114

David A. Weisblat et al.

Fig. 4 Origins of bilateral symmetry in the Helobdellu embryo. All embryos are depicted as viewed from the animal pole (with anterior up, according to each representation). Grey shading indicates teloplasm in the top three rows, and teloblasts and germinal bands in the bottom two rows (compare with Fig. 1). (A) The D-centric view with classical spiral third cleavage that is not seen in the leech. (B) The D-centric view indicating reversed B quadrant cleavage (Sandig and Dohle, 1988). (C) The ABcentric view with reversed B quadrant cleavage. In the D-centric representations, the first cleavage (top row) is oblique to the A-P axis, so that the D quadrant lies at the posterior pole. If the spiral third cleavage was completely dextrorotatory (A, second row), then the primary quartet micromeres (small circles) would arise with a’ and b‘ as one left-right pair of cells, and d’ and c’ as another, with respect to the germinal bands, which indicate the bilateral plane of the adult (bottom row). This orientation is consistent with the distribution of their definitive progeny, as indicated schematically by the hatched triangles in the bottom two rows [a’ and b‘ progeny, left and right diagonal hatching, respectively; d’ and c’, horizontal and vertical hatching, respectively; for more accurate representations on the positions of these cells, see Nardelli-Haefliger and Shankland (1993) and Smith and Weisblat (1994)l. But because the B quadrant cleaves with reverse handedness (B, second row), maintaining the D-centric representation requires positional shifts (arrows, third row) from the a’ and b’ micromeres as well as from the A, B, and C quadrant cells. In the AB-centric representation (C), the first cleavage is transverse to the A-Paxis of the embryo, and the lateral displacements of the C and C quadrant cells are corrected when

4. Non-D Quadrants in Leech Embryos

115

Fig. 5 Three pathways of midgut formation from yolky precursor cells. (For simplicity, surrounding mesodermal and ectodermal lineages are omitted from this schematic.) In one pathway, yolky precursor cells (left) cleave directly to form a hollow tube of similar, yolk-rich cells around the prospective gut lumen (center top); the yolk within them is gradually digested, giving rise to the midgut epithelium (right). Alternative pathways entail the division of the yolky precursor cells into distinct epithelial and yolk lineages (bracketed cells, light grey and dark grey, respectively). These lineages could then develop with the yolk either on the outside (center middle) or on the inside (center bottom) of the prospective gut epithelium.

tant absorption of the yolk (Fig. 5). In one pathway, the initially solid mass of yolk cells gives rise to the epithelium directly, cleaving to form a cellular tube that defines the lumen of the gut; in this path, the yolk is maintained intracellularly by the nascent gut epithelial cells (Fig. 5 , top). In the other pathways, the endodermal precursors give rise at some point to separate epithelial and yolk cell sublineages; within this general scheme, the yolk cells could be either inside (Fig. 5, bottom) or outside (Fig. 5, middle) of the prospective gut, as defined by the epithelial cells. As reviewed by Anderson (1973), the midgut epithelium arises by variations of all of these routes in different annelid species. For example, in Tubifex (an oligochaete annelid), the definitive gut arises more or less directly

C"' envelopes the nascent teloblasts during late cleavage (C, third row). In this representation, micromeres a' and b' arise as a left-right pair and only cells c' and d' (or their progeny) must shift to reach their definitive positions (C, third row). By the time cleavage is complete (fourth row), most of the teloblasts are completely enveloped by macromere C"', and macromeres A"' and B"' occupy portions near the animal pole of the embryo that were originally occupied by macromere C"'. This symmetry is maintained through germinal band formation (fifth row), by which point macromeres A"' and B"' have fused, forming cell A/B (from Hydrobiologiu, 1999;Cellular origins of bilateral symmetry in glossiphoniid leech embryos; D. A. Weisblat; Fig. 3,O 1999 Kluwer Academic Publishing; reproduced with kind permission from Kluwer Academic Publishers).

116

David A. Weisblat et al.

from the yolk cells, so the yolk remains inside of the prospective gut epithelial cells and the gut lumen arises as a hollowing out of this mass of dividing yolk cells. By contrast, in Rhynchelmis (oligochaete) and Cupitella (a polychaete annelid), the definitive gut arises as a hollow epithelial ball surrounded by the yolk cells, so the yolk lies outside the lumen of the nascent gut. Glossiphoniid leeches are an example within the annelids of the third pathway, in which the epithelium arises outside the yolk cells and surrounding them. Many of these routes to forming the gut epithelium involve syncytial yolk cell intermediates. Syncytial cells can arise by cell fusion, as in the formation of mammalian striated muscle by myoblast fusion, and/or cytokinesis without karyokinesis, as in the Drosophilu blastoderm. In Drosophilu, the wave of cellularization that occurs after the thirteenth round of nuclear proliferation forms precursors for all the gut tissues simultaneously with the mesodermal and ectodermal precursors. These cells arise at the surface of the embryo and are internalized later during gastrulation (Fig. 6). In glossiphoniid leeches, by contrast, holoblastic cell divisions separate mesodermal, ectodermal, and endodermal pre-

Fig. 6 Syncytial yolk cells in the leech, the fly, and wingless insects. In Dmsophila, the nucleus of the zygote (far left) proliferates directly to form a yolky syncytial blastoderm (top left). Cellularization of the syncytial blastoderm (top right) puts precursors of the midgut epithelium (light grey) at the surface of the embryo. These midgut precursor cells move inside and surround the residual yolk (far right) as a result of cell movements during gastrulation. In this simplified version of leech development,complete cleavages (lower left) lead to early segregation of mesodermal and ectodermal lineages (white) from midgut precursors (dark grey). A syncytial yolk cell (lower right; dark grey) arises later as the result of cell-cell fusions among the midgut precursors (plus spent teloblasts and supernumerary blast cells from the mesodermal and ectodermal lineages). Cellularization of the syncytial yolk cell results in the formation of the definitive midgut epithelium surroundingthe residual yolk. Wingless insects exhibit hybrid pathways of gut formation. In Thysanura (silverfish and bristletails), the zygote forms a syncytial blastoderm, but in the initial cellularization, precursors of the midgut epithelium remain in the syncytial yolk cell, resulting in an intermediate stage that is similar to the leech embryo. Collembola (springtails)resemble leeches even more closely, in that initial cleavages are complete and a syncytial yolk cell forms by cell fusion interior to the blastoderm.

4. Non-D Quadrants in Leech Embryos

117

cursors during cleavage and cell fusion is part of the later process by which endodermal precursors give rise to the definitive midgut epithelium (Fig. 6). Endodermal precursor cells (A”’,B”’, and (2”‘) fuse to form a syncytial yolk cell (Liu et al., 1998), within which nuclei continue to proliferate. Eventually, some of the syncytial nuclei migrate to the periphery and cellularize to form the midgut epithelium (Nardelli-Haefliger and Shankland, 1993). By this means, the definitive gut forms around the yolk cells, which therefore lie within the lumen of the nascent gut, as stated above. A similar process of midgut formation by cellularization of yolk nuclei that were never part of the blastoderm has been described for wingless insects such as the Thysanura (bristletails and silverfish). We might consider these insects as undergoing two waves of cellularization of syncytial yolk nuclei. The first wave forms the cellular precursors of the segmental ectoderm and mesoderm, plus the foregut and hindgut. The second wave forms the cellular precursors of the midgut epithelium (Fig. 6). Another group of wingless insects, the Collembola (springtails), are even more similar to leech, in that they undergo holoblastic cleavages initially and then later form a syncytial yolk cell by the fusion of precursor cells within the blastoderm (Fig. 6). These parallels between the process of midgut formation in annelids and basal arthropods raise the possibility that the process of gut formation from a syncytial yolk cell is ancestral to both groups and that the syncytial blastoderm of the more derived insects has evolved from the syncytial yolk cell of the ancestor. This area seems ripe for further analysis and comparisons, applying the more sophisticated lineage tracing techniques and molecular markers now available. Whitman (1878) concluded that the epithelial lining of the gut in glossiphoniid leech arises from multinucleate A”’, B”’, and C”’macromeres, but he seemed unaware that these cells fuse with one another, stating, “I have found that these blastomeres preserve their individuality during the entire period of invagination and neurulation. . . .” A clear appreciation of the fusion process was provided by Bychowsky (1921), however, who reported a gradual loss of visible outlines for the yolk cells, beginning at the anterior end of the prospective midgut. Judging from his drawings, he was refemng to embryos at early stage 9, by which time the germinal plate has completely formed. In our laboratory, we have used a more sensitive assay to observe cell fusion, namely, the visualization of readily detectable macromolecules that diffuse between cells that have fused. For this purpose, we have employed P-galactosidase and fluorescently labeled dextran molecules as microinjected lineage tracers, with equivalent results (Liu et al., 1998). We find that the three macromeres fuse in a stepwise manner to initiate formation of the syncytial yolk cell. The first step is the fusion of macromeres A”’ and B”’ to form a cell we designate as A/B (Fig. 1C). When we injected P-galactosidase into either the A or the B quadrant cell during stages 4 to 5 and then stained at progressively later times, we detected diffusion of the enzyme from one cell into the other (as determined by the distrib-

118

David A. Weisblat et al.

ution of the histochemical reaction product) as easily as midstage 7; thus, A’”-B”’ fusion occurs during the time interval corresponding to 51-63 hr of AZD. In Helobdella, this is approximately 65 hr before the stage at which Bychowsky (1921) inferred fusion by the loss of visible cell outlines within the yolk mass. At about the end of stage 7 (approximately 12 hr later), macromeres A”’ and B’” had fused in virtually all embryos in a batch. Macromere C”‘ fuses with A/B at the end of stage 8 (87-92 hr AZD), approximately 25 hr after the A”‘-B”’ fusion, yielding a cell we designate as A/B/C (Fig. 1C). Despite the fact that fusion has occurred, the cleavage furrow between what were separate macromeres remains distinct for about 24 hr after fusion is first evident by the diffusion assay. We observe, as did Bychowsky (1921), that the disappearance of the furrows proceeds from anterior to posterior. Electron microscopic analysis of the fusion process confirmed the perdurance of the apical junction structures between the nominal A’” and B”’ cells well after fusion had occurred (Liu et al., 1998). Specifically, we analyzed the membranes of A”’ and B’” where they contacted one another and where each contacted C”’, in embryos fixed at least 14 hr after fusion had occurred, as judged by observing the diffusion of fluorescent lineage tracers from one cell to the other. The first obvious breakdown of membranes between the fusing A”’ and B”’ cells consisted of gaps up to 3 p,m wide. No such gaps were observed in embryos fixed prior to fusion. These gaps extend through multiple serial sections and are bounded by islands of flattened double membrane, presumably formed by the joining of the nominal A’” and B”’ membranes. Even in the embryos that had initiated fusion many hours before, the gaps were seen only at the animal end of the embryo, near the surface of the embryo where the macromeres are also in contact with blast cells and micromeres. The tight junctions and other structures associated with the apical junctions of the macromeres still appear normal at this point, which explains the perdurance of the cleavage furrows relative to the actual cell fusion events. The observation that the membrane breakdown initiates at the animal end of the embryo is consistent with Bychowsky’s (1921) report that loss of visible yolk cell boundaries occurs progressively from anterior to posterior within the yolk mass. A somewhat similar description of cell-cell fusion in the nematode hypodermis has been published (Mohler et al., 1998). The appearance of the fusion pores specifically at the animal end of the embryo suggests the possibility that macromere fusion may be regulated or induced by signals emanating from the blast cells and/or micromeres. This possibility is explored in the next section. Schmidt (1939) proposed that the M teloblasts of glassiphoniid leech embryos fused with the macromeres on the basis of light microscopic examination of sectioned embryos of somewhat indeterminate age. We used the diffusion assay to confirm and extend Schmidt’s findings. M or N teloblasts were injected with a fluorescent lineage tracer during stage 6, and the resultant embryos were fixed and cleared at various times, beginning at the end of stage 8. We designate the

4. Non-D Quadrants in Leech Embryos

119 cell resulting from the fusion of teloblasts with A/B/C as the syncytial yolk cell (SYC; Fig. 1C). We find that M teloblast fusion occurs during the period 89-1 18 hr AZD. The N teloblast fusion occurs over an even later time period, during the interval of 118-141 hr AZD. This result is consistent with the fact that the M teloblasts finish producing their full complement of segmental founder cells 35 hr earlier than the N teloblasts (Zackson, 1984; Weisblat and Shankland, 1985; Lans et al., 1993). Near the end of blast cell production, each teloblast generates “supemumerary” blast cells that are not incorporated into the germinal bands or germinal plate. It was originally supposed that these cells die (Shankland, 1984), but recent experiments suggest that, like the teloblasts, these cells fuse with the endodermal lineage instead of dying (Desjeux, 1995; Shankland, 1998; Desjeux and Price, 1999). This would mean that the teloblasts and their supernumerary blast cells contribute to the gut epithelium, an endodermal tissue, as well as to segmental mesoderm and ectoderm, an observation that further muddies the significance of the classical germ layer designations [as do the precise lineage analyses in animals such as the nematode Caenorhabditis elegans (Sulston et al., 1983)l. Thus, the syncytial yolk cell receives contributions of nuclei (and cytoplasm) from three embryonic lineages-the macromeres, the teloblasts, and the supernumerary blast cells. Whether these different components contribute equally to the definitive gut epithelium and the remaining yolk nuclei remains to be determined.

IV. Regulation of Macromere Fusion The B quadrant cell and its descendant B’-B”’ macromeres are in constant contact with their counterparts in the A and C quadrants from stage 3 onward (Figs. 1 and 3), yet macromere B”’ fuses selectively with A’” only after 45-57 hr, with C”’ only after 81-86 hr, and with the teloblasts yet later (Fig. 1C). The regulation of the timing and specificity of the fusion process presents a fascinating cell biological puzzle that we have only just started to investigate (Isaksen, 1997; Isaksen et al., 1999). The first evidence regarding the source of the regulation of the fusion process came from embryos that were removed from the fertilization membrane, so as to be amenable to microsurgical cell ablation after injecting the A and B cells with fluorescent lineage tracer to monitor fusion. When cell C is removed from such embryos, the remaining quadrants undergo apparently normal subsequent divisions and macromeres A”’and B’” fuse as in controls. Moreover, in such embryos, the nascent teloblasts became embedded by the remaining macromeres; thus, the process of teloblast envelopment seems unlikely to be the explanation for why C”’ fusion is delayed relative to A”’-B”’ fusion in intact embryos. In contrast to the foregoing results, when macromere D’ is removed from the

120

David A. Weisblat et al.

embryos, macromeres A”’ and B” fail to fuse. This result suggests that signals from the D quadrant are required for Aff’-B” fusion to proceed normally. But to control for the possibility that the trauma of removing macromere D’ blocked macromere fusion artifactually, we sought to interfere with the biochemical activity of selected cells in situ by microinjecting them with the ricin Achain, a potent inhibitor of eukaryotic translation (Endo and Tsurugi, 1988), or with RNase A. Cells injected with either of these reagents round up and undergo at most one further division, but do not lyse (Nelson and Weisblat, 1992; Isaksen et al., 1999); we therefore refer to the injected cells as being “biochemically arrested” (Smith et al., 1996). Consistent with the cell ablation results, biochemical arrest of the C quadrant cell has no effect on A”’-€3”’fusion, whereas arresting the D cell blocks A”‘-B” fusion (Fig. 7). The relative simplicity of the biochemical arrest technique also allows us to extend these experiments to later stages of development, for which the microsurgical approach is not feasible. Thus, in subsequent experiments, we arrested progeny of macromere D ’ at progressively later times in development. If we wait until macromere D’ has cleaved and then arrest the two daughter cells DM and DNOPQ at stage 4b, fusion is still blocked in almost all of the embryos (Fig. 7). However, by delaying the arrest by a few hours and then arresting DMf and DNOPQ” by ricin injection at stage 4c, then most embryos fuse (Fig. 7).

Fig. 7 Biochemical arrest of D lineage blastomeres blocks a”’-B”’ fusion, with an early critical period. ”he D quadrant was arrested by microinjectionof ricin A chain into blastomere D, macromere D‘, or the descendantproteloblasts(DM and DNOPQ) at times indicated by the hatched bars; the time line at left is as in Fig. 1C. Results (assessed at about 75 hr after zygote deposition) are indicated schematically below each lineage diagram. Note that the cells injected with rich no longer divide. Biochemical arrest of D, D’, or early DM and DNOPQ lineages block a”’-B”’ fusion, whereas later injections of DNOPQ and D M do not. Other experiments are. described in the text (from Weisblat et al., 1998).

4. Non-D Quadrants in Leech Embryos

121

Between stages 4b and 4c, the D quadrant cells have given rise to five additional micromeres (dm’, dm”, dnopq‘, dnopq”, and dnopq”’), which were not injected with ricin in the experiments described above. Does this mean that some or all of these micromeres are required for the fusion signal?Apparently not, because fusion is observed in a substantial fraction (20-30%) of embryos in which cell DM and all three dnopq micromeres were arrested by ricin injection, so that cell D N O P Q is allowed to continue cleaving. Fusion is also seen in most embryos if either DM or DNOPQ is allowed to continue cleaving. We conclude that the capacity to induce fusion is distributed among the D quadrant progeny and that A”’-B”’ fusion is largely immune to the effects of D quadrant arrest by the end of stage 4. There are two main interpretations for these observations. One is that the putative D-derived signal has already been sent by stage 5. The other is that the signal is still in the D quadrant derivatives, but that its transmission is immune to the ricin-induced biochemical arrest after the end of stage 4. For example, assuming that the injected ricin Achain acts by inhibiting protein synthesis, one possibility is that the required D lineage signal involves a protein that is synthesized during stage 4, but becomes active only in response to a posttranslational modification much later in development. In any case, it is striking that the critical period of ricin sensitivity ends at least 40 hr before fusion actually occurs. In another series of experiments, we examined the effects on A”’-B” fusion of arresting A and B quadrant cells with ricin or RNase injections. We found a time course of sensitivity similar to that obtained for D quadrant arrest. That is, arresting either cell prior to stage 4b with ricin injection effectively blocked fusion, whereas both cells were immune to the effects of ricin injection after stage 4c. Within the transition period, we found (much to our surprise) that macromeres of the A lineage become resistant to the effects of rich injection several hours prior to B lineage macromeres, despite the fact that the lineages arise via synchronous divisions from A and B blastomeres that arise as sisters by an equal division at second cleavage, and had been presumed by us to be identical. We note that the A quadrant cell has a significantly greater area of cell surface contact with cell D and its derivatives than does cell B. It is also noteworthy that here, too, the fusion process becomes immune to the effects of ricin injection many hours prior to the actual fusion. Moreover, because arrest of either A or B can block fusion, we conclude that both A’” and B”’ play an active role in fusion.

V. Epiboly The most dramatic and visible aspect of gastrulation in glossiphoniid leech embryos is the ventrovegetal movements of the germinal bands that lead to germinal plate formation during stage 8. The movements of the germinal bands are accompanied by epiboly of the micromere-derived epithelium, which, together

122

David A. Weisblat et al.

with a network of underlying muscle fibers of mesodermal origin, comprise the provisional integument of the embryo. Little is known about how epibolic cell movements come about in any embryo, but superficially, the cell movements in leech resemble the epiboly seen in teleost fish, where the process is probably better studied than in any other organism (Trinkaus, 1984; Keller and Trinkaus, 1987; Solnica-Krezel and Driever, 1994). To understand the possible role(s) that macromeres might play in this process, we first need to review what is known about the contributions of other cells. To distinguish the contributions of the micromere-derived epithelium and germinal bands in the epibolic movements of gastrulation, ricin-mediated biochemical arrest was used to ask whether the epithelium is able to undergo its normal gastrulation movements in the absence of the germinal bands and vice versa (Smith et al., 1996). The results of these experiments have led us to appreciate the possibility that the macromeres may well be actively involved in the cell movements, rather than just providing a passive substrate for the movements of the overlying cells. Embryos with a complete epithelium but lacking germinal bands can be generated by injecting teloblasts and proteloblasts with ricin A chain after all the various micromeres have been produced. The converse experiment, arresting the epithelial precursors while leaving the teloblasts intact, is impossible on several grounds, not the least of which is that it would require the successful injection of 17 small cells in each embryo (Smith and Weisblat, 1994). But it is possible to reduce the population of epithelial cells by roughly one-third by deleting just five micromeres, because there is no regulative replacement of the missing epithelial cells, at least in early development (Smith et al., 1996).

A. Embryos without Germinal Bands

In embryos without germinal bands, the epithelium still undergoes epiboly but its expansion is abnormal in certain respects. First, the epithelial expansion is delayed, in that the vegetal translation of the leading edge is delayed with respect to control embryos (Fig. 8A). Though the expansion is delayed, the epithelial surface area seems to increase as in controls, with the result that the epithelium buckles and folds over itself. It seems that the epithelium is committed to an early expansion of its surface area (presumably by flattening of the epithelial cells) independent of any movement by its leading edge. Second, in embryos without germinal bands, the leading edge of the epithelium is typically very irregular. Contours of individual cells protrude from the margin, as in a mosaic of irregularly shaped tiles. In control embryos (Fig. 8B), the leading edge is quite continuous, as if parts of the tiles at the edge had been cut off to form a smooth border. Perhaps a closer analogy would be to imagine a mosaic of elastic tiles. In its relaxed state the cells at the edge would protrude, but placing the edge under tension would tend to straighten it, just as stretching a piece of wrinkled clothing

123 flattens it. Thus, both of these abnormalities of the micromere-derived epithelium in embryos without germinal bands can be explained by assuming that, in normal development, the epithelium is placed under tension in early epiboly. The forces generating this tension must arise from the germinal bands and/or the macromeres, because those are the only other cells in the embryo at this point.

4. Non-D Quadrants in Leech Embryos

B. Embryos with Reduced Epithelium

In embryos with reduced numbers of micromere-derived epithelial cells, the germinal plate forms more or less normally, but by late epiboly, the apical surfaces of individual epithelial cells are greatly expanded relative to those in controls

Fig. 8 Experimental perturbations of epiboly (Smith et al., 1996). Schematic representations of embryos at early stage 8 (top row) and midstage 8 (bottom row), as viewed from the dorsal-posterior side of the animal pole. (A) Embryos in which teloblasts (medium circles) have been arrested by microinjection with ncin A chain lack germinal bands. In such embryos, the micromeres (small contours) still proliferate and spread from animal to vegetal territory, but somewhat more slowly than in normal embryos; moreover, the leading edge of the micromere-derived epithelium is often irregular. (B and C) When teloblasts are allowed to divide normally, germinal bands (grey) originate in animal territory and then move vegetally over the surface of the embryo. (B) In normal embryos, a cross-section through the germinal bands (middle drawing; taken at the position indicated by the arrow in the upper drawing) shows that the boundary between the macromere and the leading edge of the micromere-derived epithelium (arrow) is roughly in register with the leading edge of the germinal band. (C) By contrast, in embryos with reduced numbers of micromere-derivedepithelial cells, the leading edge of the germinal band initially advances ahead of the micromere-derivedepithelium (upper drawing); the corresponding cross-sectional view (middle drawing) reveals that the macromere is deformed to remain in contact with the micromere derivatives.By midstage 8, the micromere-derived epithelial cells have expanded dramatically relative to those in control embryos and their leading edges are back in register with the germinal bands.

124

David A. Weisblat et al.

(Fig. 8C). The most interesting deviation from controls is seen in early epiboly, however. Whereas the leading edges of the epithelium and germinal bands normally remain in close proximity throughout epiboly (Fig. 8B), in embryos with reduced numbers of epithelial cells, the germinal bands are well out in front of the leading edge of the epithelium throughout much of their length (Fig. 8C). The germinal bands are not exposed to the perivitelline space, however, nor has the leading edge of the epithelium detached from the underlying yolk cell. Instead, a broad, thin portion of the macromere extends over the germinal bands in such embryos, thereby retaining the continuity of the connection between the macromere and the epithelial margin (Fig. 8; cross-sectional views). It seems that something causes the germinal bands to translocate vegetally and the epithelium normally expands concomitantly with this movement. But when the epithelium contains fewer cells, its ability to expand is reduced and the resulting imbalance of forces is resolved by deforming the macromere instead. What is the source of the force? The fact that the germinal band moves out ahead of the epithelial margin in the experimental embryos argues against the possibility that the germinal bands are being pushed or pulled by the epithelium. Another possibility is that the force for the early movements of the germinal bands and epithelial expansion could originate in the germinal bands, either from active movements of individual blast cells or through bending forces resulting from the increasing length of the germinal bands (as teloblasts keep adding new blast cells to their posterior ends). In either case, the proposed force originating from the germinal bands is not the only one acting during epiboly. In late epiboly, the epithelial margin is in advance of the germinal bands, even in the embryos that have reduced numbers of epithelial cells and that are severely expanded relative to the situation in normal embryos. At that stage, after the epithelial margin has passed the equator of the embryo, we can imagine that contractile forces around the epithelial perimeter could act like a purse string, drawing it toward the vegetal pole. The remaining candidate for contributing locomotive forces, both in early and late epiboly, is the macromeres. In fish, it appears that cortical contraction of the yolk cell cortical cytoplasm provides a force that essentially tows the leading edge of the blastoderm epithelium vegetally during epiboly (Trinkaus, 1984). Perhaps a similar process is at work within the yolk cell of the leech. Of course, these possibilities for the sources of the forces driving epiboly are not mutually exclusive. Indeed, the fact that the epithelial margin trails the edge of the germinal bands during early epiboly and leads it during late epiboly makes it seem likely that different combinations of forces are acting during different phases of epiboly. C. A Possible Role for the Macromeres in Epiboly

The fact that the macromeres are so deformed in the embryos with reduced numbers of micromere-derived epithelial cells indicates that there is a strong me-

4. Non-D Quadrants in Leech Embryos

125

chanical connection between the yolk cell and the epithelial margin. Examining this boundary by electron microscopy in normal embryos, we find the typical structures associated with an apical epithelialjunction, such as tight junctions and anchoring junctions (Isaksen, 1997). Presumably, the cytoskeletons in adjoining cells are linked via transmembrane proteins such as cadherins. Similar structures are seen at the surface of the embryo where the macromeres meet one another. Thus, we may also view the macromeres as giant epithelial cells, whose apical surfaces are lost during epiboly. If, as it seems, the leading edges of the micromere-derived epithelium are firmly tethered to the macromeres, then apical constriction of the macromeres could provide the initial force required to expand the micromere-derived portion of the epithelium toward the vegetal pole (Fig. 9). Consistent with this notion, we find that epiboly is not blocked by treating the embryos with nocodazole, but is blocked by treating the embryos with 1,4butanedionemonoxime (BDM) (E. Cheng, personal communication). BDM inhibits muscle and nonmuscle myosin ATPase activity (Backx et al., 1994; Cramer and Mitchison, 1995). Because of its low molecular weight and good solubility, BDM, like nocodazole, diffuses throughout the embryo; therefore, these experiments do not reveal whether these pharmacological agents block epiboly by effects on macromeres, or on other cells.

Fig. 9 A model of epiboly driven by apical constrictions of yolk cells. Schematic cross-sections through HeLobdelLu embryos at stages corresponding to early and late epiboly. The animal pole is up. At early stage 8 (left) germinal bands (grey) occupy animal territory, covered by micromere-derived epithelial cells (white triangular contours) and lying atop the yolky macromeres (large irregular contours). In this model, we view the entire surface of the embryo, including both the micromere derivatives and the macromeres, as the apical surface of an epithelium (thick line), with the cortical cytoskeleton linked from cell to cell via the specializations at the apical junctions. The apposing faces of these cells (thinner lines) comprise the basolateral surfaces of the epithelial cells. During epiboly the germinal bands move across the surface of the embryo from animal to vegetal territory, and their movement is accompanied by the expansion of the micromere-derived epithelium (light arrows). Apical constrictions (heavy arrows) within the three macromeres could provide the force needed to tow the germinal bands and the leading edge of the micromere derivatives vegetally. The contribution from this source might be especially important during early epiboly, when the germinal bands are still on the animal side of the embryo and purse string forces at the leading edge of the micromere derivatives would act against epiboly.

126

David A. Weisblat et al.

In hopes of blocking epiboly using a cell-specific technique, we examined the effects on epiboly of “biochemically arresting” individual macromeres in the A, B, or C quadrants by intracellular microinjection of ricin A chain, which is confined to the injected cell (Smith, 1994; Liu er al., 1998).As described above, the ricin-injected macromere rounds up, and consequently becomes partially enveloped by the other macromeres, but does not lyse. The remaining cells in the embryo seem unaffected; for example, the rate at which teloblasts produce blast cells is not affected when ricin is injected into any of the macromeres. Of course, the biochemical actions of ricin and BDM are not at all the same, but in undertaking these experiments, it was hoped that ricin would block all biochemical processes in the affected cell through its general inhibition of protein synthesis. In fact, when either the A or B quadrant cells were injected, the coalescence of the germinal bands proceeded slightly slower than in uninjected controls, as judged by the numbers of left and right mesodermal hemisomites that had coalesced within a set time, but did not cease. By contrast, when the C quadrant macromere was poisoned, -30% of the embryos exhibited a substantial increase in germinal band coalescence relative to controls ! These results are subject to a variety of interpretations. Perhaps the A and B cells play only a passive role in the epibolic cell movements of gastrulation. Or it may be that when either the A”’ or the €3”’ macromere is poisoned the other one can take its place. Or perhaps the role of the A”’ and B”‘ macromeres in epiboly is not appreciably affected by ricin poisoning (presumably an inhibition of protein synthesis). We suspect that the acceleration of germinal plate formation after biochemical arrest of the C’” macromere results secondarily as the teloblast and bandlets of blast cells are forced to the surface of the embryo by the rounding up of the C”’ macromere in which they are normally embedded. Clearly, further experiments, using more specific and selective reagents, are needed to pursue these questions.

VI. Conclusions In the adult forms of leeches and other annelid worms, segmental mesoderm and ectoderm comprise most of the animal. Thus, it is not surprising that it is the genesis of these tissues that has garnered the most attention from developmental biologists studying leeches. This tendency has been reinforced in recent years by the tantalizing prospect of comparing the sequential, holoblastic segmentation process in leeches with the simultaneous, syncytial segmentation process in flies (e.g., Weisblat er al., 1994). But the purpose of this article has been to highlight what is known, and how much remains to be learned, about the contributions of the A, B, and C quadrant macromeres to the development of the leech embryo. These cells exhibit a variety of complex and fascinating behaviors that merit that attention of cell, developmental, and evolutionary biologists.

4. Non-D Quadrants in Leech Embryos

127

A. Cleavage

It is curious that the B quadrant macromere cleaves with the opposite handedness of the A and C (and D) quadrants during three rounds of micromere production (Sandig and Dohle, 1988). Given that this feature of glossiphoniid leech development escaped notice for over a century, we are led to wonder whether the “classical” unidirectional spiral cleavage (in which all four quadrants cleave in the same direction, with dextral and sinistral cleavages alternating) is as general as it appears. There are clear and detailed descriptions of true classical spiral third cleavages even within the annelids, e.g., the polychaete (Wilson, 1892). Thus, this reversal of the handedness of the B quadrant cleavages with respect to the A, C, and D quadrants may be a relatively new feature, perhaps unique to the glossiphoniid leeches. In any case, the pattern of B quadrant cleavage in leech reinforces the notion that A and B serve as contralateral homologs in these embryos and therefore that the first cleavage plane is oriented transverse to the prospective A-P axis. We have also noted here the dramatic deformation of the non-D quadrant macromeres to accommodate the cleavages occurring in the D quadrant and to envelop the teleoblasts and their precursors. In normal embryos this behavior is normally most pronounced in the C quadrant macromere, which might be taken as evidence of cell-specific properties. But we note that in embryos from which the C quadrant macromere has been deleted, teloblasts are enveloped by processes of the remaining macromeres.

B. Syncytial Yolk Cell Formation That the formation of the leech midgut epithelium proceeds by cellularization of a multinucleate syncytium suggests a possible homology at the cellular level between the embryogenesis of glossiphoniid leeches and Drosophila. This possibility is strengthened by comparison with previous descriptions of gut formation in other arthropods, as reviewed by Anderson ( 1 973). Wingless insects, in which the developmental processes are more likely to resemble those in the ancestral form, also develop via a syncytial blastoderm. But in these insects, it seems that syncytial yolk cell nuclei migrate to the periphery of the yolk cell and become cellularized at two separate times in development. The first wave gives rise to a blastoderm containing mesodermal and ectodermal precursors, but the midgut precursors arise only from the second wave of cellularization. In contrast, in winged insects such as Drosophila, the blastoderm arising from the syncytial yolk nuclei already contains midgut precursor cells. The midgut epithelium is formed by epithelia of blastodermal origin that spread over the yolk from the anterior and posterior ends of the embryo. Combining these observations with the data from leech, it is tempting to spec-

128

David A. Weisblat et al.

ulate that formation of the midgut epithelium by cellularization of a multinucleate syncytium was a feature of the common ancestor of annelids and arthropods, and may have arisen early in the protostome lineage, while the derivation of mesoderm and ectoderm from the syncytium arose more recently, within the arthropod lineage. The mollusks are the other major protostome group, and they also undergo spiral cleavage. But although the A,B, and C quadrant macromeres also contribute to endoderm in this group, we find no evidence in the literature for the production of a syncytial precursor (e.g., Damen, 1994).

C. Regulation of Macrornere Fusion

Granted that the endoderm arises from a syncytial yolk cell (SYC),we then face the question of how that SYC arises. A priori, syncytial cells can arise either by karyokinesis without cytokinesis, or by cell fusion as in leech. In some insects, such as Drosophilu, the syncytial yolk cell arises by nuclear division without cytokinesis. But in others there are initial holoblastic cleavages followed by cell fusion, as in the leech Helobdellu, and it seems likely that the latter process would be ancestral to the former. As with any other embryonic cell-cell fusion process, the stepwise fusion of macromeres, teloblasts, and (probably) supernumerary blast cells must be carefully regulated both in terms of cell specificity and in the timing of the fusion process. Having investigated the regulation of the fusion of the A”’ and B”’ macromeres in H. robustu, we find that the process is not autonomous to the fusing cells, but rather requires a signal from D quadrant cells. Both the signaling process and the fusion process become immune to the effects of microinjected ricin A chain many hours before fusion can be detected. In leech, the SYC ultimately receives contributions of nuclei and cytoplasm from macromeres, teloblasts, and blast cells. Presumably, the mechanism of cellcell fusion is the same for all the cells contributing to the SYC.But it seems likely that the regulation of the various fusion steps will be more complex, given the very different times and classes of embryonic cells involved. D. Epiboly

The orchestrated movements of populations of cells are among the most fascinating and least understood aspects of embryogenesis. Epiboly is one dramatic example of such morphogenetic cell movements, and is relatively accessible for analysis because it occurs on the surface of the embryo and involves relatively simple sets of cells. The fact that the same term is applied to processes in animals so highly diverged as leeches and fish begs the question of what, if any-

4. Non-D Quadrants in Leech Embryos

129

thing, the two processes have in common beyond the operationally equivalent spreading of one layer of cells over another. Our investigations of epiboly in glossiphoniid leech embryos are at an early stage. But our electron microscopic studies of the cell-cell contacts lead us to appreciate the important fact (of which we should probably already have been well aware), namely, that the entire surface of the embryo, both the macromeres and the micromere-derived cells, constitutes a single epithelium. Thus, contrary to the implications of the commonly used definition of epiboly, there is no “free edge” of the advancing portion of the epithelium. This realization, along with time-lapse video observations (P. Chang, unpublished observations), rules out a notion that otherwise seemed reasonable, namely, that the micromere-derived epithelium spreads by lamellipodial movements at its leading edge, as do epithelia in culture (see Bray, 1992).Now, viewing the whole embryonic surface as a single epithelium, we can view epiboly as comprising two parallel and complementary processes. In one, the micromere-derived portion of the epithelium expands continually during epiboly, due to cell division and cell flattening. This expansion is also accompanied by cell shape changes and rearrangements to accommodate the increase and decrease in the perimeter of the micromere-derived portion of the epithelium as it reaches and then passes the equator of the embryo. In the second process, that portion of the embryonic epithelium derived from the macromeres undergoes a progressive loss of apical surface area, in what we can now visualize as another example of apical contraction of a large, yolk-filled epithelial cell. This view of epiboly should apply equally to glossiphoniid leeches and teleost fish, such as Fundulus, wherein the epithelial nature of the junctions between the yolk cell and the margin of the blastoderm epithelium has been described by Betchaku andTrinkaus (1978).But in some other embryos, epibolic process have been described in which there is a free edge to the advancing epithelium, as in ventral enclosure of the nematode hypodennis (Williams-Masson et al., 1997). We suggest that this modification of epiboly will be seen only in animals that have evolved impermeable egg cases or other protective strategies (e.g., the uterus) so that they can afford the luxury of a partially open embryo. Less wellprotected embryos must maintain the integrity of their surface epithelium in order to preserve their internal environment and are thereby constrained to carry out epiboly by the apical constriction process proposed here. This seems likely to be the ancestral condition for metazoan embryos.

Acknowledgments We thank Bob Goldstein and Ray Keller for helpful discussions. Work presented here is supported by the National Science Foundation.

130

David A. Weisblat et al.

References Anderson, D. T. (1973). “Embryology and Phylogeny in Annelids and Arthropods.” Pergamon, Oxford. Astrow, S. H., Holton, B., and Weisblat, D. A. (1987). Centrifugation redistributes factors determining cleavage patterns in leech embryos. Dev. Biol. 120,270-283. Backx, P.H., Gao, W. D., Azan-Backx, M. D., and Marban, E. (1994). Mechanism of force inhibition by 2,3-butanedione monoxime in rat cardiac muscle: Roles of [Ca2+]iand cross-bridge kinetics. J. Physiol. 476,487-500. Betchaku, T., and Trinkaus, J. P. (1978). Contact relations, surface activity, and cortical microfilaments of marginal cells of the enveloping layer and of the yolk syncytial and yolk cytoplasmic layers of Fundulus before and during epiboly. J. Exp. Zool. 206,381-426. Boyer, B. C., Henry, J. Q., and Martindale, M. Q. (1996). Dual origins of mesoderm in a basal spiralian: Cell lineage analyses in the polyclad turbellarian Hoploplanu inquilina. Dev. Biol. 179, 329-38. Bray, D. (1992). “Cell Movements.” Garland Publ., New York. Bychowsky, A. (1921). Ueber die Entwicklung der Nephridien von Clepsine sexoculata Bergmann. Rev. Suisse Zool. 29,41-131. Cramer, L. P., and Mitchison, T. J. (1995). Myosin is involved in postmitotic cell spreading. J. Cell Biol. 131, 179-89. Damen, P. (1994). Cell-lineage, and specification of developmentalfate and dorsoventral organization in the mollusc Patella vulgata. Ph.D. Thesis. University of Utrecht, Utrecht, Netherlands. Desjeux, I. (1995). An investigation into the regulation of segment number in the leech. Ph.D. Thesis. University Medical School, Edinburgh, Scotland, UK. Desjeux, I., and Price, D. J. (1999). The production and elimination of supernumerary blast cells in the leech embryo. Dev. Genes Evol. (in press). Endo, Y.,and Tsurugi, K. (1988). The WAN-glycosidase activity of ricin A-chain: The characteristics of the enzymatic activity of ricin A-chain with ribosomes and with rRNA. J. Biol. Chem. 263, 8735-8739. Fernandez, J., and Olea, N. (1982). Embryonic developmentof glossiphoniid leeches. In “Developmental Biology of Freshwater Invertebrates” (F. W. Harrison and R. R. Cowden, eds.), pp. 317361. Alan R. Liss, New York. Fernandez, J., and Stent, G. S. (1980). Embryonic development of the glossiphoniid leech Themmyzon rude: Structure and development of the germinal bands. Dev. B i d . 78,407-434. Freeman, G., and Lundelius, J. W. (1992). Evolutionary implicationsof the mode of D quadrant specification in coelomates with spiral cleavage. J. Evol. B i d . 5,205-247. Goldstein, B., and Freeman, G. (1997). Axis specification in animal development. BioEssays 19, 105- 16. Huang, F. Z., Ramirez-Weber, F. A., and Weisblat, D. A. (1999). In preparation. Isaksen, D. E. (1997). The identification of a TGF-fl class gene and the regulation of endodermal precursor cell fusion in the leech. Ph.D. Thesis. University of California, Berkeley, California. Isaksen, D. E., Liu, N.-J. L., and Weisblat, D. A. (1999). In preparation. Ishii, R., and Shimizu, T. (1995). Unequal first cleavage in the Tubifex egg: Involvement of a monastral mitotic apparatus. Dev. Growth Difler 37,687-701. Keller, R. E., and Trinkaus, J. P.(1987). Rearrangement of enveloping later cells without disruption of the epithelial permeability barrier as a factor in Fundulus epiboly. Dev. Biol. 120, 1224. Lans, D., Wedeen, C . J., and Weisblat, D. A. (1993). Cell lineage analysis of the expression of an engrailed homolog in leech embryos. Development 117,857-871. Liu, N.-J. L., Isaksen, D. E., Smith, C. M., and Weisblat, D. A. (1998). Movements and stepwise fusion of endodermal precursor cells in leech. Dev. Genes Evol. 208, 117-127.

4. Non-D Quadrants in Leech Embryos

131

Mohler, W. A., Simske, J. S., Williams-Masson, E. M., Hardin, J. D., and White, J. G. (1998). Dynamics and ultrastructure of developmental cell fusions in the Caenorhabditis elegans hypodermis. Curr: Biol.8, 1087-1090. Nardelli-Haefliger, D., and Shankland, M. (1993). Lox 10, a member of the NK-2 homeobox gene class, is expressed in a segmental pattern in the endoderm and in the cephalic nervous system of the leech Helobdella. Development 118,877-892. Nelson, B. H., and Weisblat, D. A. (1991). Conversion of ectoderm to mesoderm by cytoplasmic extrusion in leech embryos. Science 253,435-438. Nelson, B. H., and Weisblat, D. A. (1992). Cytoplasmic and cortical determinants interact to specify ectoderm and mesoderm in the leech embryo. Developrnenr 115, 103-115. Pilon, M., and Weisblat, D. A. (1997). Ananos homolog in leech. Development 124, 1771-1780. Sandig, M., and Dohle, W. (1988). The cleavage pattern in the leech Theromyzon tessulatum (Hirudinea, Glossiphoniidae).J. Morphol. 196,217-252. Savage, R. M., and Shankland, M. (1997). Identification and characterizationof a hunchback Orthologue, L$, and its expression during leech ernbryogenesis. Dev. Biol. 175,205-217. phylogenitique des modes de developpementdes organes. Schmidt, G. A. (1939). D&g.g&n.g&rescence Arch. Zool. Exp. Gen. 81,317-370. Shankland, M. (1984). Positional determinationof supernumerary blast cell death in the leech embryo. Nature (London) 307,541-543. Shankland, M. (1998). Anteroposteriorpattern formation in the leech embryo. In “Cell Lineage and Fate Determination”(S. A. Moody, ed.), pp. 207-224. Academic Press, San Diego. Shimizu, T. (1996a). Behavior of centrosomes in early Tubifex embryos: Asymmetric segregation and mitotic cycle-dependent duplication. Rouxk Arch. Dev. Biol. 205,290-299. Shimizu, T. (1996b). The first two cleavages in Tubifex involve distinct mechanisms to generate asymmetry in mitotic apparatus. Hydrobiologia 334,269-276. Smith, C. M. (1994). Developmental fates of micromeres in leech embryos and their roles in morphogenesis. Ph.D. Thesis. University of California, Berkeley, California. Smith, C. M., and Weisblat, D. A. (1994). Micromere fate maps in leech embryos: Lineage-specific differences in rates of cell proliferation. Development 120,3427-3438. Smith, C. M., Lans, D., and Weisblat, D. A. (1996). Cellular mechanisms of epiboly in leech embryos. Development 122, 1885- 1894. Solnica-Krezel,L., and Dnever, W. (1994). Microtubule arrays of the zebrafish yolk cell: Organization and function during epiboly. Development 120,2443-2455. Stent, G. S., Kristan, W. B., Jr., Torrence, S. A,, French, K. A,, and Weisblat, D. A. (1992). Development of the leech nervous system. Znt. Rev. Neurobiol. 33, 109-193. Sulston, J. E., Schierenberg,E., White, J. G., and Thornson, J. N. (1983). The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev. Biol. 100,64-119. Symes, K., and Weisblat, D. A. (1992). An investigation of the specification of unequal cleavages in leech embryos. Dev. Biol. 150,203-218. Trinkaus, J. P. (1 984). Mechanism of Fundulus epiboly-A current view. Am. Zool. 24,673688. Wedeen, C. J. (1995). Regionalization and segmentation of the leech. J. Neurobiol. 27,277-293. Weisblat, D. A. (1999). Cellular origins of bilateral symmetry in glossiphoniid leech embryos. Hydrobiologia (in press). Weisblat, D. A,, and Shankland, M. (1985). Cell lineage and segmentation in the leech. Phil. Trans. R. SOC.Lond. Ser: B 312,39-56. Weisblat, D. A,, Kim, S. Y.,and Stent, G. S. (1984). Embryonic origins of cells in the leech Helobdella triserialis. Dev. Biol. 104, 65-85. Weisblat, D. A,, Wedeen, C. J., and Kostriken, R. G. (1994). Evolution of developmental mechanisms: Spatial and temporal modes of rostrocaudal patterning. Cum Top. Dev. Biol. 29, 101134.

132

David A. Weisblat et al.

Weisblat, D. A,, Huang, F.Z., and Isaksen, D. E. (1998). Cell fate specification in glossiphoniid leech: Macromeres, micromeres and proteloblasts. In “Cell Lineage and Fate Determination” S. A. Moody, ed.), pp. 185-196. Academic Press, San Diego. Whitman, C. 0. (1878). The embryology of Clepsine. Q. J. Microsc. Sci. 18,215-315. Williams-Masson, E. M., Malik, A. N., and Hardin, J. (1997).An actin-mediated two-step mechanism is required for ventral enclosure of the C. elegans hypodermis. Development 124,28892901. Wilson, E. B. (1892). The cell lineage of Nereis. J. Morphol. 6,361-480. Zackson, S. L. (1984). Cell lineage, cell-cell interaction, and segment formation in the ectoderm of a glossiphoniid leech embryo. Dev. Biol. 104, 143-160.

Sperm Nuclear Activation during Fertilization Shirley J. Wright Department of Biology University of Dayton Dayton, Ohio 45469

I. Introduction 11. The Sperm Nucleus

A. Sperm Nuclear Architecture B. Sperm Nuclear Events Prior to Fertilization III. Egg Stage at Time of Fertilization N.Transformation of the Sperm Nucleus into a Male Pronucleus A. Removal of the Sperm Nuclear Envelope B. Sperm Nucleoprotein Remodeling C. Sperm Nuclear Decondensation D. Remodeling of the Sperm Nuclear Matrix E. Male Pronuclear Envelope Formation F. Higher Order Structure of Chromatin in the Male Pronucleus G. Factors in the Egg Cytoplasm Responsible for Sperm Nuclear Decondensation H. Activation of Transcription and Replication V. Asynchronous Behavior of the Paternal and Maternal Chromatin VI. Technological Advances to Combat Human Infertility VILConclusions and Future Directions References

The delivery of the paternal genome to the egg is a primary goal of fertilization. In preparation for this step, the nucleus of the developing spermatozoon undergoes extensive morphological and biochemical transformations during spermatogenesis to yield a tightly compacted sperm nucleus. These modifications are essentially reversed during fertilization. As a result, the incorporated sperm nucleus undergoes many steps in the egg cytoplasm as it develops into a male pronucleus. The sperm nucleus (1) loses its nuclear envelope, (2) undergoes nucleoprotein remodeling, (3) decondenses and increases in size, (4) becomes more spherical, (5) acquires a new nuclear envelope., and (6) becomes functionally competent to synthesize DNA and RNA. These changes are coordinate with meiotic processing of the maternal chromatin, and often result in behaviors asynchronous with the maternal chromatin. For example, in eggs fertilized during meiosis, the sperm nucleus decondenses while the maternal chromatin remains condensed. A model is presented that suggests some reasons why this puzzling behavior exists. Defects in any of the processes attending male pronuclear development often result in infertility. New assisted reproductive technologies have been developed that ensure delivery of the sperm nucleus to the egg cytoplasm so that a healthy embryo is produced. An emerging challenge is to furCurrent Topics in Developmental Blology. Val 46

Coovneht Q 1999 bv Academic Press All nzhhts of reprducuon In any form reserved OO'k-s53/99 $30.60

133

134

Shirley J. Wright

ther characterizethe molecular mechanisms that control sperm nuclear transformations and link these to causes of human infertility. Further understanding of this basic process promises to revolutionize our understanding of the mystery of the beginning of new life. o 1999Academic press.

1. Introduction The interaction of the sperm and egg results in the multistep process of fertilization that involves both nuclear and cytoplasmic transformations. During fertilization, the sperm fuses with the egg plasma membrane and delivers the condensed sperm nucleus into the egg cytoplasm. This sets in motion a series of events leading to egg activation, nuclear activation, and embryonic development. Fertilization has been defined at least three different ways: (1) the moment of sperm-egg fusion, (2) the period from sperm-egg fusion to pronuclear development, and (3) the period from sperm-egg fusion through the first cell cycle and mitosis to interphase of the two-cell stage. Here the emphasis is primarily on the second definition of fertilization. This article focuses on sperm nuclear structure, and the morphological and biochemical transformations the sperm nucleus undergoes during fertilization. How these relate to the activity of the maternal chromatin is also discussed. Also reviewed are the new assisted reproductive techniques that ensure delivery of the sperm nucleus to the egg cytoplasm as a means to combat infertility.

II. The Sperm Nucleus In most species, the nuclei of mature sperm are highly condensed and genetically inactive. This facilitates transmission of the paternal genome to the egg at fertilization. During fertilization, the sperm nucleus becomes reactivated in the egg cytoplasm. To aid in our understanding of sperm nuclear activation during fertilization, it is useful to briefly review alterations the sperm nucleus undergoes during spermatogenesis, because these events are essentially reversed during fertilization. Important features to note are that during spermiogenesis, the phase of spermatogenesis in which spermatids differentiate into sperm, the paternal genome undergoes extensive morphological and biochemical changes that lead to a tightly packed, genetically quiescent sperm nucleus.

A. Sperm Nuclear Architecture

1. Sperm Chromatin During spermatogenesis, the paternal chromatin undergoes biochemical changes. These include total or partial replacement of somatic histones of nu-

5. Sperm Nuclear Activation during Fertilization

135

cleosomal DNA with highly basic, nonnucleosomal protamines, sperm-specific histones, or other basic proteins (Poccia, 1986; Risley, 1990; Wouters-Tyrou et al., 1998). For example, approximately 85% of human sperm chromatin is bound with protamines, and the rest is nucleosomal (Tanphaichitr et al., 1982). In mouse sperm, all of the somatic histones are replaced with protamines (Pogany et al., 1981). In sea urchin sperm, only the somatic histones HI and H2B are replaced with the sperm-specific histones Sp H1 and Sp H2B, whereas somatic histones H2A, H3, and H4 are unchanged (Poccia, 1986). The sperm-specific histones are very similar to somatic histones except they contain additional, tandemly repeated SPKK motifs in their N-terminal regions (Sp H1, Sp H2B), and also in the C-terminal region of Sp H1 (Poccia, 1986; Poccia and Collas, 1996; Suzuki, 1989). To aid in chromosome condensation during spermatogenesis, sperm nuclear basic proteins are posttranslationally modified: protamines and sperm-specific histones are phosphorylated and histones H3 and H4 are hyperacetylated (Gatewood et d.,1987, 1990; Poccia and Collas, 1996; Wouters-Tyrou et d.,1998). These posttranslational modifications are absent from mature sperm, with some exceptions; human sperm retain acetylated isoforms of H4, whereas mouse sperm do not (Adenot et al., 1997). Protamines are characterized by high levels of arginine and cysteine, and reside in the minor (Balhorn, 1982) and/or major (Prieto et al., 1997) grooves of sperm DNA, depending on the species. The cysteine residues of the protarnines are involved in intra- and intermolecular disulfide bonds that appear during the transit of sperm through the epididymis (Wouters-Tyrou et al., 1998). The disulfide bonds of protamines provide mechanical and chemical stability to the sperm nucleus. This leads to extreme condensation of the sperm chromosomes, which are at least six times more condensed than mitotic somatic chromosomes (Pogany et al., 1981; Ward and Coffey, 1991). Sperm DNA that is packaged with protamines is not negatively supercoiled as is observed in nucleosomal DNA (Risley et al., 1986; Ward, 1994; Ward et al., 1989). Because of the condensed state of sperm chromatin, mature sperm nuclei do not synthesize DNA or RNA (Stewart et al., 1984), even though DNA polymerase and RNA polymerase activity may be present in the mature sperm nucleus (Kramer and Krawetz, 1997). Several specific mRNAs have been found in mature sperm nuclei, including Zfp59 mRNA, which is associated with the sperm nuclear matrix (Kramer and Krawetz, 1997; Passanti et al., 1995). Whether this mRNA serves a role in reactivating the paternal genome in the zygote or is a remnant of transcription during spermatogenesis has not been established.

2. Sperm Nuclear Matrix Morphological changes of the sperm nucleus during spermiogenesis include not only condensation of the sperm nucleus, but also changes in the sperm nuclear matrix. During spermiogenesis, the sperm nucleus of most species undergoes

136

Shirley J. Wright

Fig. 1 Three-dimensional view of sperm nuclei undergoing decondensation into male pronuclei in representative species. Sperm nuclei (first column) come in a variety of shapes and sizes. Once exposed to egg cytoplasm during fertilization,sperm nuclei enlarge and change shape (second column). Despite the variation in sperm nuclear shape among species, all sperm nuclei shown decondense into a spherical male pronucleus (sphere on right). Sperm nuclei of the different species are not drawn to scale with respect to each other, but do show the correct relative size as they initially decondense. Sperm nuclei depicted: (A) human, (B) bull, (C) mouse, (D) hamster, (E) frog (Xenopus), (F) seaurchin, and ( G )surf clam (Spisula).

dramatic shape changes from spherical to hook-shaped in rodents, barrel-shaped in clams, paddle-shaped in humans and dogs, and conical in sea urchins (Fig. 1). These shape changes reflect remodeling of the sperm nuclear matrix, which supports the distinct shape of mature sperm nuclei. Evidence for a sperm nuclear matrix has been produced by extracting from sperm nuclei their nucleic acids and soluble proteins using detergents, buffered salts, acids, or bases. For example, when detergent-isolated sea urchin sperm nuclei are extracted with 0.5 N HC1 or 2 M NaCl and subsequently digested with DNase I, most of the proteins and nucleic acids are lost, but the sperm nuclei retain their conical shape (Kunkle, 1984). a. Perinuclear Matrix. The sperm nuclear matrix of mammals has been called the perinuclear matrix, and consists of a network of fibers distributed

5. Sperm Nuclear Activation during Fertilization

137

throughout the nuclear interior, and the perinuclear theca, a protein meshwork immediately outside the sperm nuclear envelope (Fig. 2). The perinuclear theca is further characterized by a subacrosomal layer and the calyx region. In the bovine perinuclear theca, the calyx region is further subdivided into three layers: the postacrosomal layer, which is directly adjacent to the outer nuclear membrane of the nuclear envelope, the paracrystalline sheet, which is associated with the plasma membrane, and the postacrosomal sheath, which is sandwiched between the postacrosomal layer and paracrystalline sheet (Hess et al., 1993;Longo et al., 1987). Proteins of the calyx region include cylicin (-75 m a ) and calicin (-60 ma).Calicin is thought to be critical for shaping the sperm head because an altered distribution of calicin has been found in defective, round-headed sperm from infertile men (Courtot, 1991; Escalier, 1990). The perinuclear matrix of mammalian sperm is highly resistant to treatment with buffers of high salt concentration and with nonionic detergents (BellvC and O’Brien, 1983; Longo et al., 1987). When mammalian sperm are exposed to detergents, DNase I, buffered salts, and/or dithiothreitol, the perinuclear matrix is revealed as soluble proteins and nucleic acids are released from the sperm nucleus. The perinuclear matrix conforms to the original shape and size of the sperm nucleus (BellvC and

Fig. 2 Idealized, schematic diagram of the pennuclear matrix of a mammalian sperm. For clarity, the plasma membrane is not shown. Like all nuclei, the sperm nucleus is surrounded by a nuclear envelope. The perinuclear matrix consists of a fibrous meshwork in the nuclear interior and the pennuclear theca, which surrounds the sperm nuclear envelope. The perinuclear theca is composed of two components: the subacrosomal layer found immediately beneath the acrosomal cap, and the calyx region located outside the nuclear envelope at the base of the sperm. The implantation fossa houses the sperm centrioles, which nucleate the sperm flagellum. The pennuclear matrix is important in shaping the sperm nucleus.

138

Shirley J. Wright

O’Brien, 1983; BellvC etaf.,1992;Tanphaichitr et al., 1981;Tsanev andAvramova, 1981).

b. Nuclear Lamina. The nuclear matrix of somatic cells contains the nuclear lamina and a fibrous internal meshwork (Nelson et al., 1986). The nuclear lamina of somatic cells is composed of intermediate filament proteins, lamins A, B, and C. The lamins form a fibrous lamina located adjacent to the inner nuclear membrane of the nuclear envelope. Heterochromatin attaches to the nuclear lamina, which provides an anchoring site for chromatin. Several male germ-line-specific lamins have been identified, including lamin L,, of Xenopus sperm nuclei (Benavente and Krohne, 1985). Sea urchin sperm nuclei contain B-type lamins (Collas et al., 1995; Schatten et al., 1985). However, the nuclear lamina of mammalian sperm is absent, and instead the perinuclear theca surrounds the outer nuclear membrane of the nuclear envelope. Thus, different species have different mechanisms to provide structural integrity to the nucleus. As insoluble as these structures are, they must be easily removed so that the paternal genome can be accessed during fertilization. In somatic cells, the nuclear matrix not only acts as a structural element, but also plays a role in genome organization and activation. Actively transcribed genes are associated with the nuclear matrix to which DNA is bound into discrete loops (Ciejek ef al., 1983). Sperm chromatin is also characterized by DNA loops (Section II,A,3) bound at the base to areas of the sperm nuclear matrix (Kramer and Krawetz, 1996; Ward and Coffey, 1990). The specific organization of the sperm DNA, which is mediated by the sperm nuclear matrix, may help prime the sperm DNA for rapid activation of the paternal genome during fertilization. 3. Higher Order Structure in the Sperm Nucleus a. Somatic Chromatin. To understand the higher organization of DNA in sperm, we must first review the most accepted model of the structural organization of DNA in somatic cells. The structure of eukaryotic somatic chromatin is based on repeating subunits, or nucleosomes, linked by structurally distinct spacer DNA. Core nucleosomes usually contain 146 base pairs (bp) of nuclease-resistant DNA organized by a protein octamer containing histones H2A, H2B, H3, and H4 (Kornberg, 1977). Approximately every 200 bp of somatic cell DNA is coiled twice around an octamer of histones to form a nucleosome with a “beadson-a-string appearance” and a diameter of 1 1 nm (Thoma et al., 1979). DNA of the nucleosome is negatively supercoiled. The DNA is further compacted into a 30-nm fiber in a solenoid configuration whereby six nucleosomes are coiled about each other to form one turn of the solenoid. The 30-nm fiber is then organized into -300-nm loops containing 30 to 100 kbp attached to the nuclear matrix by scaffold-associated regions (SARs), also known as matrix attachment re-

5 . Sperm Nuclear Activation during Fertilization

139

gions (MARS), which are AT-rich segments of DNA several hundred base pairs in length. Hart and Laemmli (1998) have proposed that if the SARs define the bases of the loops of mitotic chromosomes and are juxtaposed by the chromosome scaffolding, then native chromosomes should have an AT-rich subregion where the SARs “queue up.” Using confocal fluorescence microscopy and image reconstruction of specially stained giant chromosomes of the Indian muntjac, Saitoh and Laernmli (1994) have shown that the AT-queue proceeds down the chromosome in a helical path. The AT-queue and protein scaffold of the nuclear matrix lead to the characteristic banding patterns of mitotic chromosomes, because Q-bands are AT-rich and have shorter DNA loops, whereas R-bands are AT-poor and contain longer DNA loops (Hart and Laemmli, 1998).

b. Nucleosomal Sperm Chromatin. During sperm nuclear morphogenesis, the higher order structure of sperm chromatin becomes packaged differently than somatic chromatin. Sea urchin sperm nuclei possess an extremely long nucleosome repeat length: 239 bp for Lytechinus pictus, 240 bp for Arbacia lixula, 243-250 bp for Strongylocentrotus purpuratus, and 260 bp for Arbacia punctulata (Keichline and Wasserman, 1979; see Poccia, 1986; Savic et al., 1981). In addition, sperm chromatin that is packaged by nucleosomes contains thick chromatin fibers; however, the fibers are more tightly packed in the sperm nucleus than is observed for somatic chromatin (Risley, 1990). The thick fibers of sea urchin sperm contain 40- to 50-nm large granular superbeads that are regularly arranged along the sperm chromatin. Each superbead contains as many as 20 to 48 nucleosomes (Aboukarsh and Kunkle, 1985; Zentgraf and Franke, 1984; Zentgraf et al., 1980). The large size of the superbeads may be due to internucleosomal cross-links of sperm-specific histones Sp H1 and Sp H2B (Zentgraf and Franke, 1984). c. Nucleoprotamine Sperm Chromatin. Sperm chromatin that is packaged by protamines also has a higher order structure that differs from that of somatic chromatin. Several models of mammalian sperm chromatin packaging have been proposed. According to one model (Balhorn, 1982), protamines bind to DNA by lying lengthwise inside the minor groove of DNA. The positive charge of the arginine residues of the protamines completely neutralizes the negatively charged phosphate groups on the DNA. This allows the protamine-DNA complex of one strand to fit into the major groove of a neighboring DNA strand. The neighboring DNA strands may then bind to each other by Van der Waal’s forces. The DNA strands become packed side-by-side into linear arrays that are not supercoiled.

d. Chromatin Fibers and Loops. The next level of organization of mammalian sperm chromatin is characterized by chromatin fibers that have regularly spaced nodular structures 50- 100 nm in diameter (Allen et al., 1995; Hud et al.,

140

Shirley J. Wright

1995; Koehler et al., 1983). Based on in vitro studies, these nodular structures have been proposed to have a toroid shape in vivo (Allen et al., 1995).Ward has proposed a working model of protamine-packaged,mammalian sperm chromatin that shows that the doughnutlike structures of the nucleoprotaminetoroid include one or more DNAloop domains of -50 kb (Ward, 1993;Ward and Coffey, 1991; Ward and Zalensky, 1996). In contrast to DNA loops of somatic chromatin, DNA loops of hamster (-47 kb per loop) and human (-27 kb per loop) sperm nuclei are 50-60% smaller than DNA loops of somatic nuclei (Barone et al., 1994; Kramer and Krawetz, 1997; Ward et al., 1989). The doughnut structures attach to the nuclear matrix via MARS. In the hamster sperm nucleus induced to decondense fully in vitro, DNA is additionally anchored to the nuclear annulus, a structure located at the implantation fossa (Ward and Coffey, 1989). The DNA loop domains of the sperm nucleus are thought to have important functional relationships with both DNA replication and RNA transcription.

e. Order of Chromosome Packaging. One of the fundamental questions regarding sperm nuclear morphogenesis is whether the chromosomes have a specific spatial arrangement within the condensed sperm nucleus. This could have implications in reactivating the paternal genome at fertilization. Several researchers have addressed this question using fluorescence in situ hybridization (FISH) and/or confocal laser scanning microscopy, and their efforts have yielded intriguing results (de Lara et al., 1993; Nadel et uZ., 1995; Ward et al., 1996; Zalensky et al., 1993, 1995, 1997). By mapping the position of three different genes (the 5s rRNA gene cluster, the CAD gene involved in uridine biosynthesis, and the MHC class I 1.6 gene) in mature, hamster sperm nuclei, Ward and colleagues (1996) have shown that the precise position of these genes in the hook-shaped sperm nucleus is variable, but there are preferred areas within which 80% of the FISH signals were located. The relative positions of genes on the same chromosome (5s rRNA gene and the class I 1.6 gene) were variable. When arrangements of genes on different chromosomes were compared to each other, there was no preferred location for one gene with respect to another. Thus the sperm nucleus offers some flexibility with regard to the positions of chromosomes with respect to each other. f. Telomeres and Centromeres. Several studies have addressed the topological arrangement of the telomeres and centromeres in the condensed nucleus of mature, mammalian sperm (Ward and Zalensky, 1996; Zalensky et al., 1995, 1997). During spermatogenesis in humans, telomeres are clustered and peripherally localized in spermatocytes.This arrangement persists in mature sperm, in that telomeric sequences are peripherally located, whereas centromeres are more centrally localized. In addition, telomeric sequences are clustered and positioned toward the posterior part of the hook-shaped mouse and rat sperm nucleus, whereas in bovine, porcine, equine, and human sperm, the telomeric sequences

5. Sperm Nuclear Activation during Fertilization

141

are spread over the periphery of the nucleus. Also in several different mammalian species, the centromeres are clustered in a central region of the sperm nuclear volume into a specific structure called the chromocenter (Ward and Zalensky, 1996;Zalensky et al., 1995).Higher order architecture of the sperm nucleus may be achieved in part by telomere-telomere interactions (as intrachromosomal dimers, or interchromosomal tetramers) that are mediated partly by sperm telomere binding proteins (Zalensky et al., 1997).The chromosomes may be arranged in loops bound by their telomeres. This is consistent with a model of sperm chromatin packaging showing that protamine-bound DNA is organized into loop domains in the sperm nucleus (Nadel er al., 1995; Ward, 1993,1994). If the chromosomes of the hamster sperm nucleus are arranged nonspecifically (Ward et al., 1996), and if the telomeres are also peripherally localized in this nucleus as they are in other mammalian sperm (Zalensky et al., 1997), no matter what position each chromosome takes within the sperm nucleus, each chromosome must be oriented so that its centromere is in the nuclear center while its telomeres are at the nuclear periphery. More experimentation is required to determine if this is a general principle of mammalian sperm chromatin organization. The significance of this organization is unclear. The architecture of the sperm nucleus may have important functions in mechanisms that mediate unpacking and reactivation of the paternal genome at fertilization. That the free ends of the chromosomes are peripherally localized in the sperm nucleus is suggestive. 4. Summary

In summary, sperm-specific basic proteins such as protamines and sperm-specific histones tightly package sperm chromatin. Telomeres and centromeres have specific arrangements within the sperm nucleus. The higher order structure of sperm chromatin must not only permit tight packaging of the sperm nucleus during spermatogenesis, but also must be dynamic enough to allow for rapid reactivation of the paternal genome during fertilization. These activation events include chromosome decondensation, DNA replication, and RNA transcription.

B. Sperm Nuclear Events Prior to Fertilization On its journey to meet the egg, the mature sperm experiences a series of environments different from the testicular environment. One of the major changes observed is that the sperm become vigorously motile, once released from the testicular environment. In marine organisms, sperm gain motility soon after exposure to seawater. In mammals, motility is gradually acquired in the epididymis. Human sperm exhibit -60% motility in the cauda epididymis, and become fully motile in the female reproductive tract (Moore et al., 1983).Mammalian sperm

142

Shirley J. Wright

also undergo capacitation in the female reproductive tract, which leads to the ability of sperm to acrosome react and fertilize the egg (Yanagimachi, 1988). The effects of these different environments on the nucleus of mature sperm have not been extensively investigated. During transit in the epididymis, cysteine residues of mammalian sperm protamines form additional disulfide crossbridges between free thiols (Curry and Watson, 1995; Wouters-Tyrou et al., 1998). This aids in further condensation and stabilization of the sperm nucleus (Curry and Watson, 1995; Wouters-Qrou et al., 1998). Posttesticular events in "the sea urchin sperm nucleus are different than those observed in mammalian sperm. The sea urchin egg is surrounded by a relatively thick, jelly coat that the sperm must penetrate before fusing with the egg during fertilization. When sperm reach the egg jelly, a CAMP-dependent kinase in the sperm becomes activated when sperm surface receptors bind the fucose sulfate-rich glycoconjugate of the egg jelly (Garbers et al., 1983; SeGall and Lennarz, 1979). This sperm kinase phosphorylates a single serine residue in the N-terminal end of sperm histone Sp H1 (Porter et al., 1988). This phosphorylation event is thought to help prime the sperm nucleus for fertilization events in the egg cytoplasm. Therefore, before the sperm fertilizes the egg, there are signal transduction mechanisms that translate signals to the sperm nucleus prior to sperm-egg fusion.

111. Egg Stage at Time of Fertilization Eggs of various species are normally fertilized at different stages of meiosis (Austin, 1965; Dalcq, 1957; Wilson, 1925). Dalcq (1957) categorized eggs into four different classes based on the stage of meiosis at the time of fertilization (Fig. 3). Class I eggs (germinal vesicle stage) are fertilized at meiotic prophase (dictyate stage or diplotene) when a germinal vesicle is present. This pattern of fertilization is characteristic of nematodes (Rhabdidtis, Parascaris), bivalves (Spisula,Barnea), echiuroids (Urechis, Thalassema),some annelids (Myzostoma), and crustaceans. Class I1 eggs are fertilized during first meiotic metaphase and are characteristicof some annelids (Chaetopterus,Arenicolu, Subellaria), insects (Drosophila), ascidians, ophiuroids (Amphipholis),and some molluscs (Patella, Mytilus,Dentalium).Class I11 eggs are fertilized during second meiotic metaphase and are characteristic of most vertebrates, including amphibians (Xenopus,Rana, Bufo),and most mammals (mice, hamsters, rats, cats, rabbits, sheep, pigs, cows, humans). Therefore fertilization of Class I-Class 111eggs triggers not only male pronuclear development, but also resumption of meiosis. In contrast, Class IV eggs are fertilized after meiosis is complete, when a female pronucleus (pronuclear stage) is present. This egg type is characterized by some echinoderms (sea urchins) and some cnidarians (sea anemones). Eggs of some species can be fertilized at more than one stage of meiosis and therefore fit into more than one egg class. For example, eggs of the starfish As-

5. Sperm Nuclear Activation during Fertilization

143

Fig. 3 Diagram showing the stage of meiosis at fertilization in different egg types. Eggs of different species are normally fertilized at either the germinal vesicle stage (Class I), rnetaphase of first meiosis (Class It), rnetaphase of second meiosis (Class III), or the pronuclear stage (Class IV).Polar bodies are also shown in Class 111 and Class IV eggs. See text for species representing each egg type.

terias can be fertilized at the germinal vesicle stage, during germinal vesicle breakdown, and at metaphase I (Masui, 1985). In addition, eggs of the starfish Asterina miniata (bat star) can be fertilized at any time following germinal vesicle breakdown (Costello et al., 1957; Longo et al., 1991a).

IV. Transformation of the Sperm Nucleus into a Male Pronucleus During fertilization, the sperm fuses to the egg and activates it. As the sperm fuses to the egg plasma membrane by the tip of the acrosomal process, as in echinoderms, or by the equatorial segment, as in mammals, the sperm nucleus is delivered to the cortical egg cytoplasm along with one or more sperm mitochondria and the sperm centrioles. Reawakening of the quiescent sperm nucleus into an active male pronucleus during fertilization is a multistep process. Generally, the steps are similar in most species. The main goals are to (1) remove the sperm nuclear envelope and associated structures to expose the sperm chromatin directly to factors in the egg cytoplasm, (2) remove sperm basic nucleoproteins and repackage the sperm chromatin with somatic histones, (3) decondense the tightly compacted sperm nucleus, (4) surround the paternal chromatin with a new nuclear envelope, and (5) activate transcription and DNA replication. These processes result in a striking transformation in the shape, volume, nucleoprotein composition, and activity of the sperm nucleus (Longo, 1973, 1981; Poccia, 1986; Poccia and Collas, 1996). All this occurs while the sperm chromatin migrates from the egg cortex by a microtubule-based mechanism toward the maternal chromatin, or toward the egg center where the maternal chromatin will eventually reside in most species. Several new methods and technologies, including the use of DNA fluorochromes and confocal microscopy, have been used to analyze eggs and transformation of the sperm nucleus into a male pronucleus (Brandriff and Gordon, 1992; Luttmer and Longo, 1986; Wright etal., 1993). This process has been ob-

144

Shirley J. Wright

served in vivo by light and electron microscopy, as well as in vitro using various chemical conditions. As early as 1978, the ability of egg cytoplasm to decondense sperm nuclei in vitro was examined using a powerful method that employs cell-free extracts prepared from eggs or embryos. With this method, sea urchin sperm nuclei (Strongylocentrotus, Lytechinus) were induced to decondense in vitro (Eng and Metz, 1980; Kunkle et al., 1978b), andXenopus sperm nuclei decondensed in vitro using cell-free extracts prepared from Rana eggs (Lohka and Masui, 1983). Since then, egglembryo extract systems have been developed for clams (Spisula solidissima), sea urchins (Lytechinus pictus, Strongylocentrotus purpuratus), fruit flies (Drosophilamelanogaster), and the amphibians Xenopus laevis and Bufo japonicus (Almouzni and Wolffe, 1993; Berrios and Avilion, 1990; Cameron and Poccia, 1994; Kawasaki et al., 1994; Longo et al., 1994; Ohsumi and Katagiri, 1991; Philpott et al., 1991; Poccia and Collas, 1996; Raskin et al., 1997; Ulitzer and Gruenbaum, 1989). As a result of these in vivo and in vitro studies, much has been discovered about male pronuclear development. Described below are highlights of this research. A. Removal of the Sperm Nuclear Envelope

Sperm nuclei are surrounded by a nuclear envelope that lacks pores. The sperm nuclear envelope is associated with nuclear lamins in sea urchins and frogs, and a perinuclear theca in mammals (Section II,A,2). These are rapidly removed from the sperm nucleus and are released into the egg cytoplasm. Breakdown of the sperm nuclear envelope occurs rapidly after sperm-egg fusion. The inner and outer laminae of the incorporated sperm nuclear envelope fuse at multiple sites, thereby forming vesicles. Initially, these vesicles outline the condensed sperm chromatin, but later are scattered and are indistinguishable from other membranous elements in the zygote cytoplasm (Longo, 1973, 1976; Poccia and Collas, 1996).Eventually, fewer vesicles are associated with the dispersing sperm chromatin. As a result of the dissolution of the sperm nuclear envelope, the condensed sperm chromatin becomes directly exposed to the zygote cytoplasm. In some organisms, such as sea urchins (Arbacia,Lytechinus),surf clams (Spisula), annelids (Hydroides), and rodents (hamsters and mice), breakdown of the sperm nuclear envelope is incomplete (Longo, 1973; Poccia and Collas, 1996; Szollosi et al., 1990;Yanagimachi, 1988). Portions of the sperm nuclear envelope lining the acrosomal and centriolar or implantation fossae are retained and are eventually incorporated into the male pronuclear envelope. Little is known about the molecular mechanisms governing removal of the sperm nuclear envelope. 1. Nuclear Lamina Dissolution of the sperm nuclear lamina has been observed using a cell-free system derived from sea urchin eggs (Collas et al., 1995, 1997; Collas and Poccia,

5. Sperm Nuclear Activation during Fertilization

145

1998).Although the nuclear envelope is removed during sperm nuclear isolation, the nuclear lamina, which contains a major 65-kDa B-type lamin, is retained. Within 1 min of incubation in egg extract, sea urchin sperm nuclear lamin B is phosphorylated in a calcium-dependent manner. Nuclear lamina dissolution first occurs laterally in the sperm nuclear periphery and last at the acrosomal and centriolar fossae. Within 10 min, the nuclear lamina is completely disassembled, and phosphorylated lamins are released into the egg extract. Egg protein kinase C may mediate phosphorylation and disassembly of the nuclear lamina of sea urchin sperm during fertilization. Lamina dissolution precedes sperm nuclear decondensation, but is not sufficient to promote decondensation of the sperm nucleus (Collas et al., 1997).

2. Perinuclear Theca Many mammalian sperm nuclei are covered with a perinuclear theca that must be removed during fertilization (Section II,A,2). Few studies have analyzed its fate during fertilization. However, with the advent of specific antibodies, the fate of this structure during fertilization has been traced in bovine and human sperm (Paranko and Salonen, 1995; Sutovsky et al., 1997). During sperm incorporation, the sperm plasma membrane is removed from the sperm nucleus. This exposes the perinuclear theca directly to the egg cytoplasm. In bovine sperm, removal of the perinuclear theca appears to be mediated by microvilli (Sutovsky et al., 1996a, 1997). Three lines of evidence support this hypothesis. First, microvilli emanating from the oocyte surface appear to intercalate between the layers of the perinuclear theca and sperm membranes (inner acrosoma1 membrane and plasma membrane). The microvilli become closely apposed to the perinuclear theca. Second, removal of the perinuclear theca is blocked by cytochalasin B treatment, which disrupts microfilaments of the microvilli. Third, disassembly of the perinuclear theca does not occur in unfertilized eggs that were injected at metaphase I1 with intact sperm, even though the injected sperm underwent partial removal of the sperm plasma membrane by 20 hr after sperm injection. In recently incorporated sperm of both humans and bovines, the perinuclear theca is first seen as a ring in the egg cortex near the condensed sperm nucleus (Paranko and Salonen, 1995; Sutovsky et al., 1997). The perinuclear theca then appears to split apart before the sperm nucleus decondenses. Immunolabeling with perinuclear theca antibodies disappears from the oocyte cortex at the time of initial decondensation of the sperm nucleus, suggesting that the perinuclear theca is removed before the sperm nucleus can fully decondense. This loss of immunostaining suggests that the perinuclear theca may be degraded soon after its removal from the sperm nucleus. When perinuclear theca removal was blocked by cytochalasin B, sperm nuclei did not decondense, suggesting that dissolution of the perinuclear theca is a prerequisite for sperm nuclear decondensation (Sutovsky et al., 1997). In summary, the perinuclear theca is rapidly removed from

146

Shirley J. Wright

the condensed sperm nucleus by a microvillus-mediated event, and its dissolution is a prerequisite for sperm nuclear decondensation. B. Sperm Nucleoprotein Remodeling

After the nuclear envelope and associated structures are removed from the sperm nucleus, the sperm basic nucleoproteins are removed from sperm chromatin and are replaced with somaticlike histones found in the egg cytoplasm. Sperm nucleoprotein remodeling during fertilization has been the focus of many in vivo and in vitro studies on a variety of species. Several excellent reviews have described this process in both nucleosomal and nucleoprotamine sperm chromatin of several species (Green el al., 1995; Poccia, 1986; Wolgemuth, 1983). Therefore, a brief overview is provided here using sea urchins and mammals as models.

1. Nucleosomal Sperm Chromatin During fertilization, the sperm chromatin of sea urchins, which is packaged with sperm-specific histones (Sp H1, Sp H2B) and somatic histones (H2A, H3, H4), undergoes several changes in nucleoprotein composition during development of the male pronucleus. Remodeling of sea urchin sperm chromatin occurs after removal of the sperm nuclear envelope and nuclear lamina (Collas et al., 1997; Poccia, 1986). This exposes the sperm chromatin to egg cytoplasmic factors responsible for the nucleoprotein alterations. In the sea urchin Strongylocentrotus purpurutus, the serines of the SPKK motifs on the N-terminal extensions of sperm-specific histones Sp HI and Sp H2B are phosphorylated by a sperm histone kinase (Green and Poccia, 1985; Poccia and Collas, 1996; Poccia et al., 1990). Although necessary, the sperm histone H1 and H2B phosphorylation event is not sufficient to trigger sperm chromatin dispersion (Poccia et al., 1990). Phosphorylated sperm histone Sp H1 is then removed from sperm chromatin almost immediately after fertilization (Green and Poccia, 1985; Poccia et al., 1981). It is replaced by cleavage-stage histone CS H1 from the egg cytoplasm, which also becomes phosphorylated soon after fertilization (Green and Poccia, 1985). The exchange of histone Sp H1 for histone CS H1 immediately precedes sperm nuclear decondensation (Green and Poccia, 1985; Poccia et al., 1981). Phosphorylated sperm histone Sp H2B is also removed from sperm chromatin, but unlike sperm histone H1, this occurs during sperm nuclear decondensation. In the sea urchin Tetrapygus niger, sperm histones Sp H3 and Sp H4 are lost 515 min postinsemination whereas Sp H2A and Sp H2B disappear 20-40 min postinsemination, and Sp H1 is completely absent by 30 min after fertilization (Imschenetzky et al., 1991). Comparable to sperm basic proteins, the nonbasic sperm nucleoproteins also appear to be replaced during sperm chromatin disper-

5. Sperm Nuclear Activation during Fertilization

147

sion by ones similar to those found within the female pronucleus (Kunkle et al., 1978a). Eventually the composition of the histones on the male pronuclear chromatin is similar to that of the female pronucleus. Later sperm nucleoprotein changes include the accumulation of cleavagestage histones CS H2A, CS H2B, H3”, and H4’ from the egg cytoplasm (Green et al., 1995). By the time DNA synthesis is completed, most of the sperm-specific histones are absent from the paternal chromatin. The cleavage-stage histone variants persist in the embryo up to the 16-cell stage (Poccia, 1986). These are replaced by the a-histone variants during the morula and blastula stages. As development proceeds, the a-histone variants are eventually replaced by P-, y-, and &-histonevariants (Poccia, 1986).

2. Nucleoprotamine Sperm Chromatin

a. Disulfide Bond Reduction. Mammalian sperm packaged with protamines also undergo several steps before the male pronuclear chromatin is packaged with somatic-type histones. First, the disulfide bonds of the sperm protamines are reduced before the sperm nucleus decondenses. Several studies have shown that reduction of protamine disulfide bonds induced chemically in vitro by disulfide-reducingagents such as P-mercaptoethanol or dithiothreitol, or with reducing agents and detergents, high salt, or polyanions, leads to swelling of the sperm nucleus (Jager et al., 1990; Perreault, 1990, 1992; Reyes et al., 1989; Wogelmuth, 1983; a k i n et al., 1985, 1989). Reduction of protamine disulfide bonds in vivo is accomplished by the tripeptide y-glutamyl-cysteinyl-glycine, also known as glutathione, which is synthesized during egg maturation (Perreault, 1990; Perreault et al., 1988; Sutovsky and Schatten, 1997; Wiesel and Schultz, 1981). When oxidized, glutathione reduces the disulfide bonds of the sperm protamines (Perreault, 1990). This event is thought to promote exchange of sperm protamines for egg-derived histones stored in the egg cytoplasm (Perreault, 1990). Depleting glutathione levels during oocyte maturation with the glutathione inhibitor buthionine sulfoximineblocks decondensation of the sperm nucleus, indicating that glutathione is necessary for development of the male pronucleus in hamsters and cows (Perreault et al., 1988; Sutovsky and Schatten, 1997). The synthesis of glutathione during oocyte maturation is also a prerequisite for sperm nuclear decondensation in mice and pigs (Calvin et al., 1986; Yoshida et al., 1993). Although reduction of protamine disulfide bonds is a prerequisite, it is not sufficient for decondensation in vivo,because other factors are required for male pronuclear development (Section IV,G,1). b. Protamine Removal. The next step of nucleoprotein alterations is the removal of protamines from the sperm chromatin. By analyzing sperm nucleoproteins labeled with [3H]arginine,several studies have suggested that protamines are removed during initial stages of sperm nuclear decondensation (Betzalel et

148

Shirley J. Wright

aL, 1986; Ecklund and Levine, 1975; KopeEnf and Pavlok, 1975). No radioactivity was associated with the dispersing sperm chromatin after anaphase 11, suggesting that sperm nucleoproteins are completely removed at this time. These results were later confirmed using the Feulgen reaction (Garagna and Redi, 1988), monobromobimane, which recognizes sulfhydryl groups of the protamines (Perreault, 1990,1992), and antiprotamine antibodies, which no longer stained sperm chromatin by telophase I1 (Rodman et al., 1981). In each case, protamines were lost from the dispersing sperm chromatin by the completion of second polar body formation. The mechanism responsible for protamine removal has not been identified in mammals; however, proteolysis and charge changes induced by protamine phosphorylation have been proposed (see Perreault, 1990). In amphibians, the acidic protein nucleoplasmin has been identified as the agent responsible for protamine removal in egg extracts in vitro (Ohsumi and Katagiri, 1991; Philpott et aL, 1991). To date, nucleoplasmin has not been identified in mammalian oocytes.

c. Protamine Replacement with Histones. The next step in sperm nucleoprotein remodeling is the replacement of sperm protamines with egg histones. Several studies have shown that sperm protamine loss is accompanied by appearance of histones on the sperm chromatin (Garagna and Redi, 1988; KopeCnf and Pavlok, 1984;McClay and Clarke, 1997; Nonchev and Tsanev, 1990; Rodman et aL, 1981). Using immunofluorescence microscopy with both antiprotamine and antihistone antibodies, Nonchev and Tsanev (1990) have shown that some protamines remain associated with the paternal chromatin as histones accumulate on the male pronucleus of mouse zygotes. Antiprotamine immunofluorescenceis still detected up to 8 hr postinsemination. In hamster eggs labeled with either [3H]tryptophan or [3H]arginineand subsequently fertilized by human sperm, [3H]argininelabeled proteins become associated with both male and female pronuclei, whereas those labeled with [3H]tryptophando not (KopeEnyetal., 1986).This suggests that the arginine-rich histones, which lack tryptophan, become associated with dispersing sperm chromatin (KopeEnf et al., 1986). Therefore, protamine removal is concomitant with initial stages of sperm chromatin dispersion. Protamine removal and histone replacement occur before the onset of DNA synthesis in several species, including the spoonworm (Urechis),mice, and hamsters (Das et al., 1975; Naish et aL, 1987; Nonchev and Tsanev, 1990). Histones are susceptible to a wide variety of posttranslational modifications, including acetylation, which is correlated with the potential for transcriptional activity (Turner, 1991). Using immunofluorescence confocal microscopy of mouse zygotes coupled with antibodies specific for histone H4 acetylated at lysine 5, Adenot et al. (1997) have followed the incorporation of hyperacetylated histone H4 on the sperm nucleus and maternal chromosomes. Hyperacetylated histone H4 staining was undetectable in nuclei of mature sperm. However, immediately after fertilization, while the maternal chromatin was still in metaphase

5. Sperm Nuclear Activation during Fertilization

149

11, hyperacetylated histone H4 staining was associated only with the incorporated sperm nucleus. Intense hyperacetylated histone H4 staining remained on the sperm chromatin during all phases of male pronuclear development (Section IV,C,2), while the maternal chromatin slowly acquired staining. During G,, the male pronucleus exhibited higher levels of hyperacetylated histone H4 compared to the female pronucleus, but the levels became equivalent during S phase and G,. In addition, histone H4 acetylation not only precedes sperm nuclear decondensation, but also DNA replication and transcription. As with sea urchins, histones in mammals also undergo subtype switches during oogenesis and embryogenesis (Clarke et al., 1998). In addition to acetylation changes, histone subtypes such as H1 accumulate on embryonic chromatin after the embryo becomes transcriptionally active. This is thought to aid in reprogramming of the embryonic genome and allow for totipotency of embryonic nuclei.

C. Sperm Nuclear Decondensation 1. Morphology

The sperm nucleus decondenses after its nuclear envelope and associated structures are removed. Figure 1 shows the initial stages of sperm nuclear decondensation in a variety of species. Sperm nuclei dramatically increase in size during decondensation (Figs. 1 and 4). Numerous transmission electron micrographs have shown that sperm nuclei from a variety of species decondense from the periphery toward the center of the nucleus (Longo, 1973, 1985). This is readily observed because the electron opacity of the condensed sperm nucleus decreases as sperm chromatin disperses. In sea urchins, the pattern of decondensation follows that of nuclear lamina removal: sperm chromatin first disperses laterally along the periphery of the sperm nucleus. In mammals, dispersion of the sperm chromatin also usually starts along the periphery of the midregion of the sperm nucleus and progresses anteriorly and posteriorly (Bedford, 1968; Longo, 1973; Stefanini et al., 1969; Yanagimachi and Noda, 1970a-c). Dispersion continues until all of the entire nucleus is decondensed. The implantation fossa, which contains the sperm nuclear annulus, is the last area to decondense in mammals (Garagna and Redi, 1988). Eventually all of the sperm chromatin becomes a morphologically homogeneous mass of dispersed chromatin and may increase in volume as much as 50 to 175 times or more (Adenot et al., 199 1; de Roeper and Barry, 1976; Graham, 1966; Longo, 1973, 1977, 1981; Luttmer and Longo, 1987, 1988; Wright and Longo, 1988). Based on the pattern of sperm nuclear decondensation, Longo (1973, 1981, 1985) has suggested that the agent(s) responsible for sperm chromatin dispersion during fertilization “migrates” from the periphery to the center of the sperm nucleus.

150

Shirley J. Wright

Fig. 4 Fluorescence micrograph of a mouse zygote stained with a DNA fluorochrome (4’,6-diamidino-2-phenylindole dihydrochloride) to show the second polar body (top; small, bright sphere), female pronucleus (middle; bright sphere), and the larger male pronucleus (bottom; bright sphere). Two falciform-shaped mouse sperm nuclei are located outside the zygote (upper left) for reference. Sperm nuclei undergo extensive nuclear swelling and shape changes during male pronuclear development (compare sperm nuclei to male pronucleus). Bar, 10 pm.

2. Kinetics

The rate at which sperm nuclei decondense in several species has been carefully quantified using video-intensified fluorescence microscopy (Fig. 5 ) . In two sea urchins, Lytechinus variegatus and Arbacia punctulata, sperm nuclei decondense at a linear rate of 0.68 and 0.69 pm2/min, respectively (Luttmer and Longo, 1987). However, sperm nuclear enlargement does not always consist of a single, uniform rate (Table I). For example, sperm nuclei of the hamster initially decondense at a rapid rate of 272 pm2/min (phase B), then undergo a period of condensation at a rate of - 102 pm2/min (phase C), followed by a second rapid rate of decondensation of 106 pm2/min (phase D), which later slows to 29 pm2/ min (phase E) (Wright and Longo, 1988).As shown in Fig. 5, the mouse follows a similar pattern of decondensation (Adenot et al., 1991; McClay and Clarke,

151 1997). Several phases of sperm nuclear decondensation have been observed in humans (Lassalle and Testart, 1991; Tesafik and KopeCny, 1989a). The sperm nucleus also first decondenses, then partly recondenses before final decondensation into a male pronucleus. The rate(s) at which a sperm nucleus decondenses as it transforms into a male pronucleus is coordinate with meiotic processing of the maternal chromatin (Section V). Clues about how sperm nuclear decondensation is regulated come from analyses of the different egg types (Fig. 3). When sperm nuclear decondensation is an5. Sperm Nuclear Activation during Fertilization

250

1

GV MI MI1 AII PB PNI PN2 Fig. 5 Changes in sperm nuclear area with progression of meiotic maturation in zygotes of representative species. GV, Germinal vesicle stage; MI, metaphase of first meiosis; MII, metaphase of second meiosis; AII, anaphase of second meiosis; PB, extrusion of the second polar body during telophase 11; PNI, small (early) female pronucleus in mammals (in sea urchins, PNl refers to the female pronucleus at the time of fertilization); PN2, large (late) female pronucleus in mammals (female pronucleus just before pronuclear fusion in sea urchins). Sperm nuclei of Class I-Class III eggs decondense in a characteristic pattern of an initial decondensation during meiosis, condensation during completion of meiosis, and a final period of decondensation during female pronuclear development, which begins after second polar body formation. Initially sperm chromosomes are asynchronous with the maternal chromosomes in that they decondense while the maternal chromosomes are condensed. As meiosis proceeds, sperm chromosomes become synchronous with the maternal chromosomes in that they condense and then decondense into a male pronucleus in concert with female pronuclear development. At the time of fertilization in sea urchins, the maternal chromatin is at the pronuclear stage and sperm nuclei decondense in only one phase. A, The sea urchin, Arbaciapuncrulara, Class IV egg [data from Luttmer and Longo (1987)l. C, The surf clam, Spisula solidissima, Class I egg [data from Luttmer and Longo (1988)l. H, Hamster, Class 111 egg [data from Wright and Longo (1988)l. L, The sea urchin, Lytechinus variegafus, Class IV egg [data from Luttmer and Longo (1987)l. M, Mouse, Class 111 egg [data from Adenot et al. (1991)l. S, The bat starfish, Asterina miniata, Class II egg [data from Longo et al. (1991a)l. In each species, the pattern of sperm nuclear decondensation is correlated with the stage of meiosis at the time of fertilization.

152

Shirley J. Wright

Table I Phases of Sperm Nuclear Decondensation in Representative Species

Egg typeu

Phase of sperm nuclear decondensation

Surf clam, Spisula

I

Multiple

Coot clam, M u h i a Oyster, Crassostrea Spoonworm, Urechis Fruit fly, Drosophilia Starfish,Asterina Mouse, Mus

I I I

II-IV III

Multiple Multiple Multiple Multiple Mu1tiple Multiple

Hamster, Mesocricerus Human, Homo Sea urchin, Arbacia Sea urchin, Lyrechinus

111 IV IV

Organism

n

UI

Multiple Multiple Single Single

Referenceb Da-Yuan and Longo (1983), Luttmer and Longo (1988) Longo and Scarpa (1991) Longo er al. (1993) Das and Barker (1976) Ulitzer and Gruenbaum (1989) Longo et al. ( 1991a) Adenot et al. (1991), McClay and Clarke ( 1997) Wright and Longo (1988) Lassalle and Testart (1991) Luttmer and Longo (1987) Luttmer and Longo (1987)

%ee Section III for a description of egg types. bSee text for full reference.

alyzed in Class I-Class IV eggs, sperm nuclei decondense in phases that correlate with the stages of activity of the maternal chromatin (Fig. 5; Table I). For example, during fertilization in the sea urchin (Class IV), the maternal chromatin is in one phase (pronuclear stage) and the sperm nucleus decondenses in one linear phase (Fig. 5; Table I). Moreover, the sperm nucleus of the surf clam (Class I, germinal vesicle stage) transforms into a male pronucleus in four phases coordinate with the four phases of meiosis of the maternal chromatin (Fig. 5). The sperm nucleus is unchanged during the germinal vesicle stage, decondenses during dissolution of the germinal vesicle, partially recondenses during metaphase I and 11, and decondenses into a male pronucleus during female pronuclear development (Fig. 5). Therefore, in the species that are fertilized during meiosis (Class I-Class 111), three general phases of sperm nuclear decondensation can be observed: (1) initial phase of sperm chromatin dispersion during or after germinal vesicle breakdown, (2) partial recondensation of sperm chromatin during extrusion of the polar bodies, and (3) a second phase of sperm nuclear decondensation that occurs concomitant with female pronuclear development (Fig. 5; Table I). In species fertilized after meiosis (Class IV eggs), the sperm nucleus decondenses in one phase (Fig. 5; Table I). D. Remodeling of the Sperm Nuclear Matrix

As shown in Fig. 1, the pattern of sperm nuclear decondensation appears to determine the initial shape of the male pronucleus (Longo, 1973).For example, be-

5. Sperm Nuclear Activation during Fertilization

153

cause the conical sperm nucleus of sea urchins decondenses laterally at first, and the apical and basal portions decondense last, a heart-shaped chromatin mass results that later becomes spherical on further decondensation (Cameron and Poccia, 1994; Longo and Anderson, 1968; Luttmer, 1987; Luttmer and Longo, 1987; Raskin et al., 1997). In clams, the barrel-shaped sperm nucleus rapidly decondenses into a spherical male pronucleus. The corkscrew shape of the Xenopus sperm nucleus initially decondenses into an elongated, elliptical mass that later becomes spherical. Moreover, the hook-shaped hamster and mouse sperm nuclei are initially elliptical during sperm chromatin dispersion, reminiscent of the shape of the sperm nucleus. The sperm nuclear matrix is responsible for the shape of the sperm nucleus (Section II,A,2). During sperm nuclear decondensation, the nuclear matrix must be remodeled to accommodate the changes observed in nuclear shape as the sperm nucleus becomes a spherical male pronucleus. Few studies have analyzed the sperm nuclear matrix during male pronuclear development due to the difficulty in distinguishing sperm and egg proteins. The shape changes of the sperm nucleus during male pronuclear development in vivo have been quantified using computer-assisted video microscopy of sea urchins, clams, and hamsters (Luttmer, 1987; Luttmer and Longo, 1987, 1988; Raskin et al., 1997; Wright and Longo, 1988). Based on the equation for shape, whereby a sphere equals one and a line equals zero, these studies have shown that shape changes parallel sperm nuclear decondensation. The conical sea urchin and hook-shaped hamster sperm nuclei become elliptical prior to developing into spherical male pronuclei (Luttmer and Longo, 1987; Wright and Longo, 1988). Clam sperm nuclei are essentially spherical (Fig. l), and they remmain spherical during all phases of male pronuclear development in vivo (Luttmer, 1987; Luttmer and Longo, 1988). Shape changes of sea urchin sperm nuclei induced in vitro by egg extract are sensitive to serine protease inhibitors, whereas expansion of the sperm nucleus is unaffected (Raskin et al., 1997). This suggests that shape changes and decondensation are distinct processes and can be uncoupled from each other. Therefore, in the sea urchin, serine proteases are required to alter the sperm nuclear matrix, allowing shape transformations and subsequent remodeling of the sperm nuclear matrix. Whether the mechanism exists in vivo in sea urchins and other species has not been determined. However, proteases have been postulated to play a role in male pronuclear development in mammals (Zirkin et al., 1989).

E. Male Pronuclear Envelope Formation

The next step in formation of the male pronucleus is the assembly of the nuclear envelope. Once the decondensing sperm chromatin is surrounded by a new nuclear envelope it is considered a male pronucleus. Formation of the male pronuclear envelope has been studied in a variety of organisms (Longo, 1973; Poccia

154

Shirley J. Wright

and Collas, 1996). Membrane vesicles can be found associated with dispersing sperm chromatin even before the sperm nucleus fully decondenses (Longo and Anderson, 1968; Terasaki and Jaffe, 1991). Therefore, the male pronuclear envelope forms concurrent with decondensation of the sperm chromatin. In sea urchins, vesicles peripheral to the dispersing sperm chromatin coalesce to form elongated cisternae, which surround the dispersing sperm chromatin and fuse to form the male pronuclear envelope (Longo and Anderson, 1968). Eventually the cisternae develop nuclear pores. In areas where the sperm nuclear envelope persists, the elongated vesicles derived from maternal membranous components fuse with the sperm-derived membranes. It has been estimated that in sea urchins the male pronuclear envelope is composed of at least 80% egg membranes and no more than 20% sperm nuclear envelope (Longo, 1976, 1985). Therefore, the male pronuclear envelope is composed of membranous elements derived from both the sperm and the egg. The time at which the male pronuclear envelope forms appears to differ in the various organisms studied (Longo, 1973, 1985; Poccia and Collas, 1996). In sea urchins and hamsters, the male pronuclear envelope forms during sperm chromatin dispersion (Longo and Anderson, 1968; Yanagimachi and Noda, 1970a,c). However, in Spisula it forms during female pronuclear development, and in Mytilus it forms prior to the completion of meiosis (Longo, 1973; Longo and Anderson, 1969, 1970). The reason for the difference in timing is unclear but may be due to the period required for development of the male pronucleus, which varies among species (Longo, 1973, 1985). Table I1 shows the approximate time required for formation of the male pronuclear envelope in several species. The basis for the differences is not understood, but may be related to the size of the sperm nucleus or concentration and activity of the agents involved in decondensation and envelope formation.

Table I1 Timing of Male Pronuclear Envelope Formation

Organism Surf clam, Spisula Coot clam, Mulinia Oyster, Crassostrea Brittle star, Amphipholis Fruit fly, Drosophilia Frog, Xenopus Domestic fowl, Gallus Mammals Sea urchin, Arbacia

Egg type"

I I I I1 I1 111 111 111 IV

Appearance of male pronuclear envelopeb 45 min -45 min -45 min 15 min <20 min 40 min 25 min 3-4 hr 8 min

9 e e Section III for a description of egg types. %me postinsemination. =See text for full reference.

Reference' Longo and Anderson (1970) Longo and Scarpa ( 199 1) Longo et al. (1993) Yamashita (1985) Fitch et al. (1998) Graham et al. (1966) Okamura and Nishiyama (1978) Longo (1973, 1985) Longo and Anderson (1968)

5. Sperm Nuclear Activation during Fertilization

155

Formation of the male pronuclear envelope in vitro involves several steps (reviewed in Collas and Poccia, 1998; Poccia and Collas, 1996). A cell-free system consisting of detergent-permeabilized sperm nuclei incubated in extract prepared from fertilized sea urchin eggs was used to study male pronuclear envelope formation in vitro. Nuclear envelope formation begins by the ATP-dependent binding of vesicles to dispersing sperm chromatin. Subsequent fusion of the membrane vesicles requires GTP. Targeting of the membranous elements to the dispersing sperm chromatin is mediated by a lamin B receptorlike integral membrane protein. Once the male pronuclear envelope forms in vivo, the male pronucleus still undergoes further enlargement as it continues to migrate toward the female pronucleus and/or the egg center. Male pronuclear swelling in vitro has been shown to be associated with import of soluble lamin B into the nucleus and subsequent growth of the nuclear envelope by fusion of additional vesicles. In the male pronucleus, the lamin B associates with the lamin B receptor, which has been proposed to tether the nuclear envelope to the nuclear lamina. Association of lamins A/C and B with the nascent nuclear envelope of male and female pronuclei has also been shown by immunofluorescence microscopy coupled with antilamin antibodies (Schatten et al., 1985). Lamins surrounding the male pronucleus in mice are thought to be derived from maternal sources (Stricker et al., 1989). In Drosophila, lamins are also recruited from maternal stores to the periphery of the male pronucleus (Liu et al., 1997). Nuclear lamins also accumulate on male and female pronuclei of bovine and porcine zygotes, and later embryogenesis revealed that the nuclear lamins may be developmentally regulated (Prather et al., 1989).

F. Higher Order Structure of Chromatin in the Male Pronucleus

The compact sperm chromatin appears to have a specific structure in the sperm nucleus of a variety of species (Section II,A,3). Exactly how the unpacking of the sperm nucleus is mediated has not been demonstrated to date. Male pronuclei also have a specific arrangement of the sperm chromatin. Two models are presented below.

1. Nucleosomal Sperm Chromatin After fertilization, the extremely long nucleosome repeat lengths of sea urchin sperm chromatin (Section II,A,3), are reduced by 30-50 bp as sperm chromatin loses sperm-specific histones and accumulates cleavage-stage histones and later embryo histone variants (Savic et ul., 1981). The nucleosome repeat length is 195 bp for Strongylocentrotuspurpurutus 8-cell embryos, -205 bp for Lytechinus pictus 4-cell embryos, 218 bp for Arbacia lixulu gastrulae, and 220 bp for 32-cell embryos and pluteus larvae of Arbacia punctulata and Strongylocentro-

-

156

Shirley J. Wright

tuspurpuratus (Keichline and Wasserman, 1979; see Poccia, 1986; Savic et al., 1981). These differences are due to the normal amount of DNA in the nucleosome core (-140 bp), but a different length of linker DNA (Keichline and Wasserman, 1979; see Poccia, 1986; Savic et al., 1981). This reduction in base pairs begins 30 min after fertilization in sea urchins (Lytechinus pictus, Strongylocentrotus purpuratus).It is accompanied by DNA synthesis and accumulation of histones CS H2A and CS H2B (Poccia eta!., 1984). Whether the chromosomes have a higher spatial organization in the male and female pronuclei has not been demonstrated for sea urchins.

2. Nucleoprotamine Sperm Chromatin Brandriff and Gordon (1992) have examined the spatial distribution of chromatin in zygote pronuclei using fluorescence in situ hybridization of zygotes created by fertilizing hamster eggs with human sperm. Pronuclear chromatin was extended three- to fourfold compared to somatic interphase chromatin. In partially decondensed sperm heads and fully decondensed male pronuclei, a whole chromosome probe for human chromosome 4 was confined to one area of the nucleus. Thus, each chromosome remains in a spatially limited domain, starting from sperm chromatin dispersion to the fully developed male pronuclear stage. Results suggested that as a chromosome decondenses from the sperm nucleus into the egg cytoplasm during male pronuclear development, the chromosome occupies one domain of the nucleus rather than the entire nuclear volume. It is unclear whether this is due to the humadhamster hybrid system used and whether human chromosomes keep this arrangement in male pronuclei induced to form in human oocytes. The significance of the spatial arrangement of specific chromosomes into a restricted domain in the male pronucleus is not clear. It could be very important for subsequent functioning of the male pronucleus in activating transcription and DNA replication. In human sperm nuclei induced to decondense in vitro, the centromeres appear to be centrally localized, and the telomeres are found at the nuclear periphery (Zalensky et al., 1993, 1997). Using antikinetochore antibodies to stain mouse zygotes, Schatten et al. (1988) have shown that the kinetochores of the sperm nucleus are undetectable, but can be observed initially throughout the developing male pronucleus. When mature pronuclei are present, the kinetochore antibodies label a distinct patch surrounding the nucleoli of both male and female pronuclei. Thus, in mouse zygotes, the kinetochores, which indicate the positions of the centromeres, are centrally localized in the fully mature male and female pronuclei. This pattern is consistent with that shown for the mature sperm nucleus (Haaf and Ward, 1995) (Section II,A3). The implications for this spatial arrangement may pertain to mechanisms that mediate unpacking the sperm nucleus during fertilization, and may influence the initiation and regulation of paternal gene activity.

5. Sperm Nuclear Activation during Fertilization

157

G. Factors in the Egg Cytoplasm Responsible for Sperm Nuclear Decondensation 1. Characteristics of Sperm Nuclear Decondensing Factors Several in vivo and in vitro studies have characterized the factors in the egg cytoplasm responsible for male pronuclear development in a variety of species (Cameron and Poccia, 1994; Collas and Poccia, 1998; Longo, 1990; Longo et al., 1991b, 1994; Luttmer and Longo, 1987, 1998; Poccia and Collas, 1996; Raskin et al., 1997; Wright and Longo, 1988). In general, sperm nuclear decondensation requires ATP and is temperature sensitive. Several studies using polyspermic or polygynic conditions show that sperm nuclear decondensing factors are present in limited quantities in the egg cytoplasm (Luttmer and Longo, 1987, 1988; Raskin et af.,1997; Wright and Longo, 1988). In sea urchin zygotes, cytoplasmic alkalinization that normally occurs during fertilization is also required to promote decondensation.In addition, protein kinases, including tyrosine kinases, are also required (Cameron and Poccia, 1994;Raskin et al., 1997). However, nascent proteins, microtubules, and microfilaments are not involved in this process (Luttmer and Longo, 1987).In species that are normally fertilized during meiosis (Class IClass I11 eggs; Section 111), the two phases of sperm nuclear expansion (phases B and D; Section IV,C,2) have different requirements and may be regulated by different factors. For example, in hamster zygotes, phase B is unaffected by polyploidy (polyspermy, polygyny), protein synthesis inhibition by puromycin, and metabolic inhibition by antimycin A, whereas phase D is sensitive to these agents (Wright and Longo, 1988).A similar situation has been reported for clam zygotes in that phase B is refractory to many of the treatments to which phase D is sensitive, including protein synthesis inhibition (Longo et al., 1991b; Luttmer and Longo, 1988). Therefore, factors that decondense chromosomes during phase D are newly synthesized and require ATP (Longo et al., 1991b; Luttmer and Longo, 1988; McClay and Clarke, 1997; Wright and Longo, 1988). In vitro experiments with amphibians (Xenopus, Bufo),have shown additionallythat sperm nuclear decondensation requires nucleoplasmin and topoisomerase II activity (Newport, 1987; Ohsumi and Katagiri, 1991; Philpott and Leno, 1992;Philpott et al., 1991). Topoisomerase II is not required for sperm nuclear expansion in clams, sea urchin, mice, and hamsters (Perreault et al., 1991;Wright and Schatten, 1988,1990). Despite this difference, other studies have demonstrated that sperm nuclear decondensing factors are evolutionarily conserved because sperm nuclei can decondense in heterologous species egg cytoplasm (Brown et al., 1987; Lolika and Masui, 1983; Longo, 1977,1985; Perreault, 1990; Raskin et al., 1997).

2. Oocyte Competence to Form a Male Pronucleus The ability of the oocyte to initiate sperm nuclear decondensation depends on the meiotic state of the maternal chromosomes, and is acquired during meiotic mat-

158

Shirley J. Wright

uration (reviewed in Longo, 1985; Perreault, 1992; Wolgemuth, 1983; Zirkin et al., 1989). For example, eggs of animals normally fertilized at a later stage of meiosis (Class I11 eggs) can be fertilized at the germinal vesicle stage, but the incorporated sperm nucleus remains condensed in the egg cytoplasm (reviewed in Longo, 1985; McClay and Clarke, 1997; Zirkin et al., 1989). In many mammals with Class I11 eggs (Section In), the ability of the egg cytoplasm to decondense the sperm nucleus appears after germinal vesicle breakdown, reaches a maximum at metaphase 11, disappears soon after fertilization, and reappears during first cleavage (Longo, 1985; Perreault, 1992; Usui and Yanagimachi, 1976; Wolgemuth, 1983; Zirkin et al., 1989). Therefore, the sperm nuclear decondensing activity of the egg cytoplasm is cell cycle dependent (Maeda et al., 1998). Although the egg cytoplasm may not yet acquire sperm nuclear decondensing ability, as in germinal vesicle eggs, or has lost this ability, as in pronucleate eggs, the activity may still be present in oocytes and eggs. Maeda et al. (1998) have shown that demembranated sperm nuclei injected into the cytoplasm of immature mouse oocytes containing a germinal vesicle do not decondense; however, if they are injected into the germinal vesicle, they do decondense. Moreover, demembranated sperm nuclei injected into the cytoplasm of pronucleate mouse eggs or two-cell embryos do not decondense; however, if injected into the male or female pronucleus, or two-cell interphase nucleus, the sperm nuclei decondense. These results indicate that sperm nuclear decondensing factors in mice are sequestered in interphase nuclei rather than in the cytoplasm, and are released into the cytoplasm during nuclear envelope breakdown. A candidate for such a factor could be nucleoplasmin, but it has been identified only in amphibian eggs. H. Activation of Transcription and Replication

As the male pronucleus develops it is also migrating toward the female pronucleus or egg center in anticipation of first mitosis. In sea urchins (Class IV eggs), the male and female pronuclei fuse within 15-20 min after fertilization and form a zygote nucleus (Longo and Anderson, 1968; Luttmer and Longo, 1987). This is known as the “sea urchin type of fertilization” (Wilson, 1925). Paternal DNA is synthesized within the resulting zygote nucleus; however, the male pronucleus is capable of synthesizing DNA by 16 min postinsemination if pronuclear fusion is prevented (Longo and Plunkett, 1973). In Class I-Class 111eggs, the male and female pronuclei do not fuse once they become associated with each other. Therefore DNA replication and transcription occur within the individual male and female pronuclei similar to those shown in Fig. 4.During mitotic prophase, each pronucleus eventually gives rise to a group of chromosomes that intermingle on the metaphase plate. The first time the paternal and maternal chromosomes are found within the same nucleus is at the two-cell stage. This is known as the “Ascaris type of fertilization” (Wilson, 1925).

5. Sperm Nuclear Activation during Fertilization

159

1. DNA Replication DNA replication is one of the first functional tasks performed by the male pronucleus. The timing of DNA replication differs in various species. DNA replication begins at 16 min after fertilization in sea urchins, at 20 min in Xenopus, at 4 hr in hamsters, and at 12 hr in humans (Gurdon and Woodland, 1968; Longo and Plunkett, 1973; Naish et al., 1987; Tesafik and KopeEny, 1989b). The ability of pronuclei to replicate DNA is acquired late during meiotic maturation in frogs and mice (McClay and Clarke, 1997; Zhao and Benbow, 1994). In mice, DNA replication may take place as the pronuclei migrate to the egg center, as well as when the nuclei are closely apposed (Luthardt and Donahue, 1973). Nucleoprotein remodeling and male pronuclear envelope formation are prerequisite for DNA synthesis. The initiation of DNA synthesis in hamster zygotes occurs only after pronucleus formation in mature, midsized, mammalian pronuclei (Naish et al., 1987;Zirkin et al., 1989).Decondensed sperm nuclei and newly formed male pronuclei will not synthesize DNA until they are fully mature, even when they reside in the cytoplasm of an oocyte known to support DNA synthesis (Naish et al., 1987). The duration of S phase is 6 hr in mice (Howlett, 1986), and the female pronucleus exhibits higher levels of DNA replication than the male pronuclew (Table 111). In the human, DNA synthesis may be preceded by a small amount of RNA synthesis (Tesdik and KopeCnny’, 1989b).

2. Transcription In many species, RNA synthesis occurs later during embryogenesis-at the midblastula transition in Xenopus and during cycle 10 in Drosophila (see Poccia and Collas, 1996). Therefore, male pronuclei in these species are involved primarily with replication rather than transcription. In contrast, the male pronucleus of Ascaris zygotes actively synthesizes large quantities of RNA, while the maternal chromatin is occupied with meiotic maturation (see Longo, 1985). Sea urchin RNA synthesis is initiated in the zygote nucleus at approximately 30 min after fertilization and is independent of DNA synthesis (Longo and Kunkle, 1977; Poccia et al., 1985). In the mouse embryo, activation of transcription, also known as zygotic gene activation, begins during S/G, phase of the first cell cycle when male and female pronuclei are present (Adenot et al., 1997; Ram and Schultz, 1993). The male pronucleus exhibits greater transcriptional activity than the female pronuclew (Table 111).This was confirmed by studies showing immunolocalization of RNA polymerase I1 in a speckled pattern in pronucleate eggs and in the nuclear periphery of two-cell embryos (Worrad et al., 1995). Although paternally derived mRNA is observed in late one-cell embryos, the major transcriptional activation begins at the two-cell stage approximately 2-4 hr after completion of the first mitosis (Bouniol et al., 1995; Flach et al., 1982; Henery et al., 1995;

160

Shirley J. Wright

Table 111 AsynchronousBehavior of the Paternal and Maternal Genomes during the First Cell Cycle Activity of paternal genome

Activity of maternal genome Female pronuclear envelope is unaffected Unchanged

Removal of sperm nuclear envelope Nucleoprotein remodeling Decondensation (Phase B) Decondensation Shape changes during decondensation

Chromosome condensation Unchanged Unchanged

Formation of male pronuclear envelope

Species

Referencea

Sea urchin

Longo and Anderson (1968)

Most species

Poccia (1986)

surf clam,

Luttmer and Longo (1988). hamster Wright and Longo (1988) Sea urchin Luttmer and Longo (1987) Sea urchin, Longo (1973),Luttmer and most mammals Longo (1987), Wright and Longo (1988) Sea urchin Longo (1973)

Female pronuclear envelope is already formed Large female pronucleus Small female pronucleus

Sea urchin

Luttmer and Longo (1987)

Mouse, hamster, human

None or low levels

Mouse

Brandriff and Gordon (1992). Payne er al. (1997). Worrad et al. (1999, Wright and Longo (1988) Adenot et al. (1997)

Low concentration in female pronucleus

Mouse

Worrad et al. ( 1994)

Not present yet in the female pronucleus

Mouse

DNA replication has not begun

Female pronucleus is actively synthesizing DNA

Mouse, hamster

Long, less decondensed chromosomes at first metaphase

Shorter and more condensed chromosomes at first metaphase Rapid and more advanced response

Mouse

Aoki et al. (1997). Bouniol et al. (1995). Ram and Schultz (1993). Wiekowski et al. ( 1993) Abramczuk and Sawicki (1975), Luthardt and Donahue (1973), Naish et al. (1987), Siracusa et al. (1975) Donahue (1972), Dyban and Sorokin (1983), Kaufman (1973)

Small male pronucleus Large male pronucleus

High levels of hyperacetylated histone H4 High concentrationof transcription factors in male pronucleus Transcriptionin the male pronucleus

Slow response to premature chromosome condensation DNA methylation and genomic imprinting

Different pattern present

%See text for full reference.

Mouse

Ciemerych and Czolowska (1993)

Mouse

Monk er al. (1987). Surani et al. ( 1990)

5 . Sperm Nuclear Activation during Fertilization

161

Matsumoto er al., 1994). In fact, the total amount of transcription in the zygote in G, is -20% that of the two-cell embryo (Aoki et al., 1997). Cleavage of the two-cell stage does not require transcription, whereas it is critical for later cleavages (Poueymirou and Schultz, 1989). Some of the first zygotic genes to be activated in mouse embryos encode the heat-shock protein HSP 70.1, the translation initiation factor eIF-4C, the Tri-C-P5 subunit of the cytosolic chaperonin TriC, and the sex-determining region genes, Sty and Zfy (Christians et al., 1995; Davis et al., 1996; Sevigny et al., 1995; Temeles et al., 1994; Thompson et al., 1995; Zwingman et al., 1993). Transcriptional activation of the embryonic genome is marked by synthesis of a-amanitin-sensitive proteins called transcription-requiring proteins (Conover et al., 1991). Phosphorylation plays a role in the regulation of transcription, because the protein kinase inhibitor, H8, can block transcription in the early mouse embryo (Poueymirou and Schultz, 1989). Although prominent nucleoli are present in mouse zygotes, rRNA synthesis does not begin until the late two-cell stage (see Cuadros-Fernindez and Esponda, 1996).

V. Asynchronous Behavior of the Paternal and Maternal Chromatin Several perplexing problems emerge when one analyzes the pattern of male pronuclear development in Class I-Class IV eggs. The paternal and maternal chromatin exhibits different activities (Table 111).Prior to pronuclear development, the maternal and paternal chromatin exhibits asynchronous behaviors with respect to the state of condensation of the chromosomes. For example, how can the sperm nucleus decondense when maternal chromosomes are condensing (or are condensed), as is seen in Class I-Class I11 eggs? And later, how can both maternal and paternal chromosomes decondense simultaneously to form female and male pronuclei? What coordinates this behavior? In Class I eggs, the sperm nuclear envelope is removed concomitant with dissolution of the nuclear envelope of the germinal vesicle, but in Class IV eggs, the sperm nuclear envelope is removed and the sperm nucleus decondenses while the female pronuclear envelope and maternal chromatin are unchanged. How are these activities coordinated? Based on the experiments described in the previous sections, a model can be constructed to account for sperm nuclear behavior during fertilization of different egg types. This model is shown in Fig. 6. Maternal chromatin is decondensed or partially condensed in the germinal vesicle egg. After germinal vesicle breakdown, maternal chromatin condenses due to activity of factor 2 (F2) in the model. Candidates for F2 are maturation-promoting factor (MPF, p34cdc2/cyclinB kinase) and cytostatic factor (CSF, p39c-mos/cyclin-dependent kinase 2), which induce and maintain chromosome condensation (Minishull, 1993; Motlik and

162 162

ShirleyJ.J. Wright Wright Shirley MEIOSIS MEIOSIS GV

Female

D-

GVBD

F2

C

MI1

MI

F2

Fig. 6 Amodel representing factors that act on the maternal (top row) and paternal (bottomrow) chromatin during fertilization of different egg types (Class I-Class IV). Key stages of meiosis are shown: germinal vesicle (GV) stage, germinal vesicle breakdown (GVBD), metaphase of first meiosis (MI), metaphase of second meiosis (MII), anaphase of second meiosis (AII), and the pronuclear (PN) stage, which occurs after completion of meiosis. This model proposes the presence of three factors: factors 1 (Fl) and 3 (F3) decondense (D) chromatin, while factor 2 (F2) condenses (C) it. Maternal chromatin (female), which is partially decondensed at the GV stage, is condensed by factor 2 during meiosis through AII. Soon after AII, factor 2 is inactivated. This allows factor 3, which requires ATP and is newly synthesizedduring meiosis, to decondense the maternal chromatin during female pronuclear development. Because eggs of various species are fertilized at different stages of meiosis, sperm nuclei exhibit differentpatterns of decondensation(male). In Class I eggs, which are fertilized at the GV stage, factor 1 initially decondenses (phase B) the sperm nucleus (SN I) and remodels the sperm chromatin to resemble maternal chromatin. Factor 2 then condenses the remodeled paternal chromatin (phase C). On inactivation of factor 2, factor 3 decondenses the paternal chromatin into a male pronucleus (phase D) in concert with female pronuclear development. In Class II eggs, which are fertilized at MI, and Class III eggs, which are fertilized at MI, factor 1 also initially decondenses sperm nuclei of Class I1 (SN II) and Class 111 (SN 111) eggs, respectively. As nucleoprotein remodeling occurs, factor 2 condenses the remodeled paternal chromatin (phase C). When factor 2 becomes inactivated after anaphase 11, factor 3 decondenses the paternal chromatin into a male pronucleus (phase D) along with female pronuclear development. At the time of fertilization in Class IV eggs, factors 1 and 2 are inactive, and active factor 3 has decondensed the maternal chromatin into a prominent female pronucleus. During fertilization of this egg type, the sperm nucleus (SN IV) decondenses in one phase into a male pronucleus, while the female pronucleusis unchanged.Thus activities of the paternal chromatin parallel those of the maternal chromatin. See text for candidates for factors 1-3.

Kubelka, 1990; Parrish et al., 1992). Another putative candidate for F2 is condensin, which is a protein complex of Xenopus egg extract that is responsible for chromosome condensation (Hirano et al., 1997). Condensin may exert its action by binding to the topoisomerase 11that is bound to the SARs of the nuclear matrix. Whether condensin actively condenses the maternal and paternal chromatin during phase C has not been established. Mixing of the germinal vesicle nucleoplasm with the egg cytoplasm is a prerequisite for sperm nuclear decondensation in many species. Thus, in the germinal vesicle egg cytoplasm, sperm nuclear decondensing factors (factor 1 in the

5. Sperm Nuclear Activation during Fertilization

163

model) are not active, but become active after germinal vesicle breakdown. Candidates for factor 1 (Fl) are glutathione to reduce protamine disulfide bonds, and nucleoplasmin to aid in the sperm nucleoprotein remodeling that occurs during phase B of decondensation (Section IV,B,2). As nucleoprotein remodeling is completed in Class I-Class I11 eggs, the basic proteins of sperm chromatin are replaced by somatic histones found in the egg cytoplasm. The remodeled paternal chromatin, which now has a nucleoprotein content similar to that of the maternal chromatin, falls under the influence of F2 and condenses. In light of this idea, Iwao and Elinson (1990) have shown that supernumerary sperm nuclei are degraded in physiological polyspermy of the newt, Cynops pyrrhogaster, due to a lack of a local concentration of MPF, suggesting that MPF can interact with paternal chromatin. As calpain I1 protease degrades the CSF signal and MPF activity disappears, factor 3 (F3) becomes active and decondenses the maternal and paternal chromatin in concert to form a female and male pronucleus, respectively. Factor 3 is a newly synthesized protein that requires ATP and either is a kinase or requires kinase activity. In type IV eggs, meiosis is complete and factors 1 and 2 are inactive, whereas F3 is active at the time of fertilization. Factor 3 maintains the maternal chromatin in a decondensed state and decondenses sperm chromatin. During mitosis, F3 is inactivated or destroyed, but reappears at interphase. The asynchronous activities of the paternal and maternal chromatin after pronuclear formation (Table 111) may relate in part to differences in histone acetylation patterns. Although transcriptional activity was higher in the mouse male pronucleus than in the female pronucleus, levels of histone hyperacetylation were equivalent, suggesting that the transcriptional differences may be due to differential evolution of acetylation in the pronuclei prior to DNA replication (Adenot et al., 1997). The larger size of the male pronucleus in mice (Fig. 4) may also be a result of the higher transcription rate and greater extent of decondensation of the paternal genome. Whether the dissimilar size of the male and female pronuclei in other species, including humans (Table 111), is also due to differences in the extent of decondensation and transcription remains to be established. Differences in DNA methylation of the chromatin may also contribute to the asynchronous activities of the paternal and maternal chromatin (Table 111). Pronuclear transplantation experiments show that the male and female pronuclei of the mouse are not equivalent (McGrath and Solter, 1984; Surani and Barton, 1983; Surani et al., 1986). When the male pronucleus of a mouse zygote was replaced with a donor female pronucleus, the resulting digynic embryo developed but had a small placenta and eventually died. When the female pronucleus of a normal mouse zygote was replaced with a donor male pronucleus to produce a diandric zygote, a poorly formed embryo with a well-developed placenta resulted, and the embryo eventually died. This may be due in part to activity of the im-

164

Shirley J. Wright

printed gene, insulinlike growth factor 11, which is only paternally expressed, probably because of a higher level of DNA methylation of the maternal gene (DeChiara et al., 1991; Swain et al., 1987). The DNA methylation differences in the parental genomes may specify whether a gene came from the mother or father. Some human genes are also imprinted and expression of the mutated or missing allele can lead to affected phenotypes (e.g., Prader-Willi syndrome, Angelman’s syndrome), even though a normal allele is present from the other parent (Nicholls et al., 1989). Thus chromatin differences in the paternal and maternal genomes may lead to the asynchronous behaviors of these genomes. These asynchronous behaviors are part of the normal developmental process. This model is useful in determining the relationship of factors that regulate the behavior of the parental genomes within the egg cytoplasm. This is of paramount importance to know because of problems with infertility of domestic animals and humans. By using this model, mechanisms governing fertilization and early development can be identified. This has important applications for human assisted reproduction.

VI. Technological Advances to Combat Human Infertility In order to fertilize an egg and contribute the paternal genome to the embryo, the sperm undergoes a series of morphological and biochemical changes. If any of these critical steps are disrupted, infertility can result, and thus, the causes of infertility are many. To combat this problem, assisted reproductive technologies have been developed (see Sutovsky et al., 1996b). In some cases of infertility, sperm are unable to reach the egg due to male dysfunction or blockage of the female reproductive tract. Artificial insemination and in vitro fertilization have been used to alleviate these problems successfully. The in vitro fertilization technique does have its problems because many polyspermic embryos are produced. Researchers are attempting to solve this in mice and humans by microsurgically removing the extra male pronucleus. This technique has led to normal development in mice (Chida, 1996; Feng and Gordon, 1996; Gordon et al., 1989). In other cases, infertility results from improper interaction of the sperm and egg. For example, in cases where sperm fail to acrosome react and cannot penetrate the zona pellucida, techniques called zona drilling, zona cutting, or subzonal sperm injection have been successfully used (Payne et al., 1994). In these techniques, a hole is placed in the zona pellucida through which sperm pass or are injected, allowing the sperm to fuse to the egg. In cases in which antisperm antibodies prevent sperm egg fusion, intracytoplasmic sperm injection (ICSI), whereby a sperm is microinjected into the egg, has bypassed this problem and resulted in normal human embryos and births (Sutovsky et al., 1996b). In cases in which only immature sperm are available, the ICSI technique has been successfully used with round spermatids (Sutovsky et al., 1996b). The

5. Sperm Nuclear Activation during Fertilization

165

ICSI technique is unsuccessful with mature sperm that do not decondense in the egg cytoplasm or do not form functional male pronuclei (Asch et al., 1995). One way to test the suitability of human sperm for the ICSI technique is to microinject them into Xenopus, mouse, or hamster eggs so that their extent of nuclear decondensation can be assessed (Ohsumi et al., 1986; Paranko and Salonen, 1995; Rybouchkin ef al., 1995). Another way to test the ability of sperm nuclei to decondense prior to using the ICSI technique has been developed using Xenopus egg extract in the human sperm activation assay (Brown et al., 1987, 1995). This in vitro assay, which incubates human sperm nuclei in Xenopus egg extract, has been used to identify some cases of infertility that would otherwise be unexplained. In cases in which defective sperm nuclei do not decondense in human eggs, a donor male pronucleus is needed, assuming the egg cytoplasm can successfully support sperm nuclear decondensation. Male pronuclei can be produced in vitro in egg extracts and in fragments of human oocytes (Levron et al., 1995; Montag et al., 1992). These in vitro-developed male pronuclei could then be transplanted to needed embryos produced by artificially activating eggs. As an alternative to induce sperm nuclear decondensation, ICSI-injected eggs or inseminated eggs could be treated with either glutathione or cysteamine (a glutathione precursor). In addition, sperm nuclei (or eggs) could be treated with thi01s (dithiothreitol, P-mercaptoethanol) to induce sperm nuclear decondensation prior to ICSI. Because not all of the mechanisms involved in normal embryonic development have been discovered, cases of unexplained infertility remain. The sperm aster has been discovered to play a critical role in interaction of the male and female pronuclei of bovine and primate embryos (Schatten, 1994; Sutovsky et al., 1996b). Mature sperm deliver a proximal centriole to the egg cytoplasm during fertilization (Palermo et al., 1997; Sathananthan et al., 1991, 1996; Simerly et al., 1995). This contributes to the centrosome, which organizes the sperm aster in the egg cytoplasm and is responsible for pronuclear apposition. Sperm of some men with idiopathic infertility may not contain a functional proximal sperm centriole (Asch et al., 1995). Performing ICSI with these sperm would never lead to a normal embryo unless a functional sperm centriole/sperm aster could also be injected into the egg and tethered to the sperm nucleus. Other causes of unexplained infertility may be due to chromosomal defects of the sperm and/or egg. Chromosomal abnormalities are the major cause of pre- and postimplantation embryo loss (Martini et al., 1997). Sperm with chromosome aberrations would also be unsuitable for ICSI. With the advent of FISH and sensitive probes, it is possible to observe the presence of mutated genes or other chromosome anomalies such as aneuploidy or duplication in sperm and/ or eggs (Downie et af.,1997; Egozcue et al., 1997; Hassold, 1998; Martini et al., 1997). The FISH technique can be applied to large numbers of sperm. Several chromosome-specific probes can be distinguished simultaneously in the same sperm using different fluorescent detection systems so that chromosome

166

Shirley J. Wright

aberrations are readily detected (Downie et al., 1997; Martini et al., 1997). Moreover, with the FISH technique, sperm bearing the X chromosome can be readily distinguished from those carrying the Y chromosome (Flaherty and Matthews, 1996; Flaherty et al., 1997; Han et al., 1993; Hassold, 1998). However, a major problem with FISH is that sperm are fixed during the procedures. Therefore, it is not possible to perform the technique on live sperm. Although FISH can be used as an excellent diagnostic tool for the general sperm population, the actual live sperm chosen for ICSI cannot be assessed using this technique.

VII. Conclusions and Future Directions The sperm nucleus undergoes remarkable morphological and biochemical transformations during male pronuclear development. These occur in concert with meiotic processing of the maternal chromatin. Initially the behavior of the maternal and paternal genomes is asynchronous during fertilization, but later becomes synchronous. One of the challenges of the future will be to discover the molecular basis for this asynchrony. The model presented here, which is based on compelling data from many researchers, argues that several factors control the paternal chromatin during male pronuclear development. It will be especially important and challenging to determine the identity of these factors, their mode of interaction, and how they rearrange the higher order structure of chromatin to produce a fully functional male pronucleus. Another challenge of the future will be to investigate when, where, and how the first genes are expressed in the newly formed male pronucleus. The identity of these genes may give new insights into their role during fertilization and early embryogenesis. The tremendous progress that has been made toward understanding processes attending male pronuclear development places the study of human fertilization in an exciting era. Advances in assisted reproductive technology have already led to new understanding of the causes of infertility.Another challenge of the future involves designing new ways to test for genetic defects and subsequently manipulate the genome so that eventually defective genes could be targeted for gene transplants (Simoni, 1994). ICSI could then be performed to transplant the microsurgically altered male pronucleus into the activated egg. Then, too, as new technologies emerge for genetically altering the human genome, new ethical guidelines will be required to lead us along the way. Although much still remains to be done, many of the tools are already at hand to provide further comprehension of the molecular mechanisms underlying fertilization. Therefore, research in this field holds out the prospect of understanding fundamental aspects of fertilization and development, and at the same time provides new approaches to combat infertility.

5 . Sperm Nuclear Activation during Fertilization

167

Acknowledgments I thank David Wright for his valuable discussions, critical reading of the manuscript, and excellent assistance in creating the computer graphics for the figures. I also thank Jason %.Pierre for his technical help. This review was funded in part by faculty seed grant awards from the Research Council of the University of Dayton.

References Aboukarsh, N., and Kunkle, M. (1985). Ultrastructural organization of heterochromatin within sea urchin sperm nuclei. Gamete Res. 12,55-64. Abramczuk, J., and Sawicki, N. (1975). Pronuclear synthesis of DNA in fertilized and parthenogenetically activated mouse eggs. Exp. Cell Res. 92,361-372. Adenot, P. G., Sziillosi, M. S., Geze, M., Renard, J.-P., and Devey, P. (1991). Dynamics of paternal chromatin changes in live one-cell mouse embryo after natural fertilization.Mol. Reprod. Dev. 28,23-34. Adenot, P. G., Mercier, Y., Renard, J.-P., and Thompson, E. M. (1997). Differential H4 acetylation of paternal and maternal chromatin precedes DNA replication and differential transcriptional activity in pronuclei of 1-cell mouse embryos. Development 124,4615-4625. Allen, M. J., Bradbury, E. M., and Balhorn, R. (1995). The natural subcellular structure of the bovine sperm cell. J. Struct. Biol. 114, 197-208. ALmouzni, G., and Wolffe, A. P. (1993). Nuclear assembly, structure, and function. The use of Xenopus in vitro systems. Exp. Cell Res. 205, 1-15. Aoki, F., Worrad, D. M., and Schultz, R. M. (1997). Regulation of transcriptional activity during the first and second cell cycles in the preimplantation mouse embryo. Do.Biol. 186,296-307. Asch, R., Simerly, C. S., Ord, T., Ord, V. A., and Schatten, G. (1995). The stages at which human fertilization arrests: Microtubule and chromosome configurationsin inseminated oocytes which fail to complete fertilization and development in humans. Human Reprod. 10,1897-1906. Austin, C. R. (1965). “Fertilization” Prentice-Hall, New York. Balhorn, R. (1982). A model for the structure of chromatin in mammalian sperm. J. Cell Biol. 93, 298 - 305. Barone, J. G., de Lara, J., Cummings, K. B., and Ward, W. S. (1994). DNA organization in human sperm. J. Andml. 15,139-144. Bedford, J. M. (1968). Ultrastructuralchanges of the sperm head during fertilization in the rabbit. Am. J. Anat. 123,329-358. BellvC, A. R., and O’Brien, D. A. (1983). The mammalian spermatozoon: Structure and temporal assembly. In “Mechanism and Control of Animal Fertilization” (J. F. Hartmann, ed.), pp. 55-137. Academic Press, New York. Bellvb, A. R., Chandrika, R., Martinova, Y. S., and Barth, A. H. (1992). The perinuclear matrix as a structural element of the mouse sperm nucleus. Biol. Reprod. 47,45 1-465. Benavente, R., and Krohne, G. (1985). Change of karyoskeleton during spermatogenesisof Xenopus. Expression of lamin L,, a nuclear lamina protein specific for the male germ line. Proc. Natl. Acad. Sci. U.S.A. 82,6176-6180. Berrios, M., and Avilion, A, A. (1990). Nuclear formation in a Drosophila cell-free system. Exp. Cell Res. 191,64-70. Betzalel, M., Shalgi, R., and Moav, B. (1986). Protamine in the rat; its fate in vivo, and its degradation in vitro by egg homogenate. Gamete Res. 14,293-306. Bouniol, C., Nguyen, E., and Debey, P. (1995). Endogenous transcription occurs at the 1-cell stage in the mouse embryo. Exp. Cell Res. 218,57-62.

168

Shirley J. Wright

Brandriff, B. F., and Gordon, L. A. (1992). Spatial distribution of sperm-derivedchromatin in zygotes determined by fluorescence in situ hybridization.Mutut. Res. 296,33-42. Brown, D. B., Blake, E. J., Wolgemuth, D. J., Gordon, K., and Ruddle, F. H. (1987). Chromatin decondensation and DNA synthesis in human sperm activated in vitro by using Xenopus luevis egg extracts. J. Exp. Zool. 242,215-231. Brown, D. B., Hayes, E. J., Uchida, T., and Nagamani, M. (1995). Some cases of human male infertility are explained by abnormal in vitro human sperm activation. Fertil. Steril. 64,612-622. Calvin, H. I., Grosshans, K., and Blake, E. J. (1986). Estimation and manipulation of glutathione levels in prepuberal mouse ovaries and ova: Relevance to sperm nucleus transformation in the fertilized egg. Gamete Res. 14,265-275. Cameron, L. A., and Poccia, D. L. (1994). In vitro development of the sea urchin male pronucleus. Dev. Biol. 162,568-578. Chida, S. (1996). Micromanipulation on oocytes for the defect of monospennic fertilization: Intracytoplasmic sperm injection and extra male pronuclear extraction. Jpn. J. Fertil. Steril. 41, 100105. Christians, E., Campion, C., Thompson, E. M., and Renard, J.-P. (1995). Expression of the HSP 70.1 gene, a landmark of early zygotic gene activity in the mouse embryo, is restricted to the first burst of transcription. Development 112, 113-122. Ciejek, E., Tsai, M.-J., and O’Malley, B. W. (1983). Actively transcribed genes are associated with the nuclear matrix. Nature (London) 306,607-609. Ciemerych, M. A., and Czolowska, R. (1993). Differentialchromatin condensation of female and male pronuclei in mouse zygotes. Mol. Reprod. Dev. 34,73-80. Clarke, H. J., McClay, D. W., and Mohamed, 0.A. (1998). Linker histone transitions during mammalian oogenesis and embryogenesis. Dev. Genet. 22,17-30. Collas, P., and Poccia, D. (1998). Remodeling the sperm nucleus into a male pronucleus at fertilization. Theriogenology 49,67-81. Collas, P., Pinto-Correia, C., and Poccia, D. L. (1995). Lamin dynamics during sea urchin male pronuclear formation in vitro. Exp. Cell Res. 219,687-698. Collas, P., Thompson, L., Fields, A., Poccia, D. L., and Courvalin, J. (1997). Protein kinase C-mediated interphase lamin B phosphorylation and solubilization.J. Biol. Chem. 272,21274-21280. Conover, J. C., Temeles, G. L., Zimmermann, J. W., Burke, B., and Schultz, R. M. (1991). Stagespecific expression of a family of proteins that are major products of zygotic gene activation in the mouse embryo. Dev. Biol. 144,392-404. Costello, D. P., Davidson, M. E., Eggress, A., Fox, M. H., and Henley, C. (1957). “Methods for Obtaining and Handling Marine Eggs and Embryos.” Lancaster Press, Lancaster, Pennsylvania. Courtot, A. M. (1991). Presence and localization of the 60 kD calicin in human spermatozoapresenting postacrosomal sheath defects: Preliminary results. Mol. Reprod. Dev. 28,272-279. Cuadros-Fernhdez, J. M., and Esponda, P. (1996). Immunocytochemical localization of the nucleoloar protein fibrillarin and RNA polymerase I during mouse early embryogenesis.Zygote 4, 49-58. Curry, M. R., and Watson, P. F.(1995). Sperm structure and function. In “Gametes-The Spermatozoon” (J. G. Grudzinskasand J. L. Yovich, eds.), pp. 45-69. Cambridge Univ. Press, London and New York. Dalcq, A. M. (1957). “Introduction to General Embryology.” Oxford Univ. Press, London. Das, N. K., and Barker, C. (1976). Mitotic chromosome condensation in the sperm nucleus during postfertilization maturation division in Urechis eggs. J. Cell Biol. 68, 155-159. Das, N. K., Micou-Eastwood,J., and Alfert, M. (1975). Cytochemical studies on the protarnine-type protein transition in sperm nuclei after fertilization and the early embryonic histones of Urechis cuupo. Dev. Biol. 43,333-339. Davis, W., DeSousa, P. A., and Schultz, R. M. (1996). Transient expression of translation initiation factor eIF-4C during the 2-cell stage of the preimplantation mouse embryo: Identification by

5. Sperm Nuclear Activation during Fertilization

169

mRNA differential display and the role of DNA replication in zygotic gene activation. Dev. Biol. 174,190-201. Da-Yuan, C., and Longo, F. J. (1983). Sperm nuclear dispersion coordinate with meiotic maturation in fertilized Spisula solidissima eggs. Dev. Biol. 99,217-224. DeChiara, T. M., Robertson, E. J., and Efstradiatis,A. (1991). Parental imprinting of the mouse insulin-like growth factor I1 gene. Cell 64,849-859. de Lara, J., Wydner, K. L., Hyland, K. M., and W. S. Ward (1993). Fluorescent in situ hybridization of the telomere repeat sequence in hamster sperm nuclear structures.J. Cell. Biochem. 53,213221. de Roeper, A. D., and Barry, J. M. (1976). Nuclear swelling and chromatin decondensation can occur without cortical granule explosion in eggs of Xenopus laevis. Exp. Cell Res. 100,41-415. Donahue, R. P. (1972). Fertilization of the mouse oocyte: Sequence and timing of nuclear progression to the two cell stage. J. Exp. Zool. 180,305-318. Downie, S. E., Flaherty, S. P., and Matthews, C. D. (1997). Detection of chromosomes and estimation of aneuploidy in human spermatozoa using fluorescence in-situ hybridization.Mol. Human Reprod. 3,585-598. Dyban, A. P., and Sorokin,A. V. (1983). A comparison of the size of paternal and maternal homologous chromosomesduring the first two cleavage divisions in mice. Ontogenez 14,238-245. Ecklund, P. S., and Levine, L. (1975). Mouse sperm basic nuclear protein, electrophoreticcharacterization and fate after fertilization.J. Cell Biol. 66,251-262. Egozcue, J., Blanco, J., and Vidal, F. (1997). Chromosome studies in human sperm nuclei using fluorescence in-siru hybridization (FISH). Human Reprod. Update 3,441-452. Eng, L. A., and Metz, C. B. (1980). Sperm head decondensation by a high molecular weight fraction of sea urchin egg homogenate. J. Exp. Zool. 212, 159-167. Escalier, D. (1990). Failure of differentiationof the pennuclear skeletal complex in the round-headed human spermatozoa.Inr. J. Dev. Biol. 34,287-297. Feng, Y.-L., and Gordon, J. W. (1996). Birth of normal mice after removal of the supernumerary male pronucleus from polyspermic zygotes. Human Reprod. 11,341-344. Fitch, K.R., Yasuda, G. K., Owens, K. N., and Wakimoto, B. T. (1998). Paternal effects in Drosophila: Implications for mechanisms of early development. Curr: Top. Dev. Biol. 38, 1-34. Flach, G., Johnson, M. H., Braude, P. R., Taylor, R. A. S., and Bolton, V.N. (1982). The transition from maternal to embryonic control in 2-cell mouse embryo. EMBO J. 1,681-686. Flaherty, S. P.,and Matthews, C. D. (1996). Application of modern techniques to evaluate sperm sex selection methods. Mol. Human Reprod. 2,937-942. Flaherty, S. P., Michalowska,J., Swann, N. J., Dmowski, W. P., Matthews, C. D., andAitken, R. J. (1997). Albumin gradients do not enrich Y-bearing human spermatozoa.Human Reprod. 12, 938-942. Garagna, S., and Redi, C. A. (1988). Chromatin topology during the transformation of the mouse sperm nucleus into pronucleus in vivo. J. Exp. Zool. 246, 18-193. Garbers, D. L., Kopf, G. S., Tubb, D. J., and Olson, G. (1983). Elevation of sperm adenosine 3’5’monophosphateconcentrations by a fucose-sulfate-richcomplex associated with eggs. 1. Structural characterization.B i d . Reprod. 29, 1211- 1220. Gatewood, J. M., Cook, G. R., Balhorn, R., Bradbury, E. M., and Schmid, S. W. (1987). Sequencespecific packaging of DNA in human sperm chromatin. Science 236,962-964. Gatewood, J. M., Cook, G. R., Balhorn, R., Schmid, S. W., and Bradbury, E. M. (1990). Isolation of four core histones from human sperm chromatin representing a minor subset of somatic histones. J. Biol. Chem. 265,20662-20666. Gordon, J. W., Grunfeld, L., Garrisi, G. J., Navot, D., and Laufer, N. (1989). Successful microsurgical removal of a pronucleus from tripolar human zygotes. Fertil. Steril. 52,367-372. Graham, C. F. (1966). The regulation of DNA synthesis and mitosis in multinucleated frog eggs. J. Cell Sci. 1, 363-374.

170

Shirley J. Wright

Graham, C. F., Arms, K., and Gurdon, J. B. (1966). The induction of DNA synthesis by frog egg cytoplasm. Dev. Biol. 14,349-381. Green, G. R., and Poccia, D. (1985). Phosphorylationof sea urchin sperm HI and H2B histones precedes chromatin decondensation and HI exchange during pronuclear formation. Dev. Biol. 108, 235-245. Green, G. R., Collas, P., Burrell, A., and Poccia, D. L. (1995). Histone phosphorylation during sea urchin development.Sem. Cell Biol. 6,219-217. Gurdon, J. B., and Woodland, H. R. (1968). The cytoplasmic control of nuclear activity in animal development. Biol. Rev. 43,233-267. Haaf, T., and Ward, D. C. (1995). Higher order nuclear structure in mammalian sperm revealed by in situ hybridization and extended chromatin fibers. Exp. Cell Res. 219,604-61 I . Han, T. L., Ford, J. H., Webb, G. C., Flaherty, S . P., Correll, A., and Matthews, C. D. (1993). Simultaneous detection of X- and Y-bearing human sperm by double fluorescence in situ hybridization. Mol. Reprod. Dev. 34,308-313. Hart, C. M., and Laemmli, U. K. (1998). Facilitation of chromatin dynamics by SARs. Curt Opin. Genet. Dev. 8,519-525. Hassold, T. J. (1998). Nondisjunction in the human male. Cum Top. Dev. Biol. 37,383-406. Henery, C. C., Miranda, M., Wiekowski, M., Wilmut, I., and DePamphilis, M. L. (1995). Repression of gene expression at the beginning of mouse development.Dev. Biol. 169,448-460. Hess, H., Heid, H., and Franke, W. W. (1993). Molecular characterization of mammalian cylicin, a basic protein of the sperm head cytoskeleton. J. Cell Biol. 122, 1043-1052. Hirano, T., Kobayashi, R., and Hirano, M.(1997). Condensins, chromosome condensation protein complexes containing XCAP-C, XCAP-E and a Xenopus homolog of the Dmsophila Barren protein. Cell 89,511-521. Howlett, S . K. (1986). The effect of inhibiting DNA replication in the one-cell mouse embryo. RouxS Arch. Dev. Biol. 195,499-505. Hud, N. V., Downing, K. H., and Balhorn, R. (1995). Aconstant radius of curvature model for the organization of DNA in toroidal condensates. Proc. Natl. Acad. Sci. V.S.A. 92,3581-3585. Imschenetzky, M., Puchi, M., Pimentel, C., Bustos, A., and Gonzales, M. (1991). Immunobiochemical evidence for the loss of sperm specific histones during male pronucleus formation in monospedc zygotes of sea urchins. J. Cell. Biochem. 47, 1-10, Iwao, Y., and Elinson, R. P. (1990). Control of sperm nuclear behavior in physiologically polyspermic newt eggs: Possible involvement of MPF. Dev. Biol. 142,301-312. Jager, S., Wijchman, J., and Kremer, J. (1990). Studies on the decondensationof human, mouse, and bull sperm nuclei by heparin and other polyanions. J. Exp. Zool. 256,315-322. Kaufman, M. H. (1973). Timing of the first cleavage division of the mouse and the duration of its component stages: A study of living and fixed eggs. J. Cell Sci. 12,799-808. Kawasaki, K., Philpott, A., Avilion, A. A., Bemos, M., and Fisher, P. A. (1994). Chromatin decondensation in Drosophila embryo extracts. J. Biol. Chem. 269,10169-10176. Keichline, L. D., and Wasserman, P. M. (1979). Structure of chromatin in sea urchin embryos, sperm, and adult somatic cells. Biochemistry 18,214-219. Koehler, J. K., Wurschmidt, U., and Larsen, M. P. (1983). Nuclear and chromatin structure in rat spermatozoa. Gamete Res. 8,357-370. Kopetn);, V., and Pavlok, A. (1975). Autoradiographicstudy of mouse spermatozoonarginine-rich nuclear protein in fertilization.J. Exp. Zool. 191,85-96. Kopetny, V., and Pavlok, A. (1984).Association of newly synthesized 3H-arginine labelled proteins with chromatin structure in fertilization. Gamete Res. 9,399-408. KopeEnq, V., BabuHik, P., Tesaiik, J., and Pavlok, A. (1986). Selective association of some hamsteregg synthesized proteins with decondensing human sperm chromatin. Histochemistry 84, 197200. Komberg, R. D. (1977). Structure of chromatin. Annu. Rev. Biochem. 46,931-954.

5. Sperm Nuclear Activation during Fertilization

171

Kramer, J. A., and Krawetz, S. A. (1996). Nuclear matrix interactions within the sperm genome. J. Biol. Chem. 271, 11619-11622. Kramer, J. A., and Krawetz, S. A. (1997). RNA in spermatozoa: Implications for the alternative haploid genome. Mol. Human Reprod. 3,473-478. Kunkle, M. (1984). Structural components of sea urchin sperm nuclei. Gamete Res. 9,469-479. Kunkle, M., Longo, F. J., and Magun, B. E. (1978a). Nuclear protein changes in the maternally and paternally derived chromatin at fertilization. J. Exp. 2001. 203,371 -380. Kunkle, M., Magun, B. E., and Longo, F. J. (1978b). Analysis of isolated sea urchin sperm nuclei incubated in egg cytosol. J. Exp. Zool. 203,381-390. Lassalle, B., and Testart, J. (1991). Sequential transformations of human sperm nucleus in human egg. J. Reprod. Ferril. 91,393-402. Levron, J., Willadsen, S., Munn6, S., and Cohen, J. (1995). Formation of male pronuclei in partitioned human oocytes. Biol. Reprod. 53,209-213. Liu, J., Lin, H., Lopez, J. M., and Wolfner, M. F. (1997). Formation of male pronuclear lamina in Drosophilu melanogastex Dev.Biol. 184, 187- 196. Lohka, M. J., and Masui, Y. (1983).Formation in virro of sperm pronuclei and mitotic chromosomes induced by amphibian ooplasmic components. Science 220,7 19-721, Longo, F. J. (1973). Fertilization: A comparative ultrastructural review. Biol. Reprod. 9, 149-215. Longo, F. J. (1976). Derivation of the membrane comprising the male pronuclear envelope in inseminated sea urchin eggs. Dev. Biol. 49,347-368. Longo, F. J. (1977). An ultrastructural study of cross-fertilization (Arbacia X Myrilus). J. Cell Biol. 73,14-26. Longo, F. J. (1981). Regulation of pronuclear development. In “Bioregulators of Reproduction” (G. Jagiello and C. Vogel, eds.), pp. 529-557. Academic Press, New York. Longo, F. J. (1985). Pronuclear events during fertilization. In “Biology of Fertilization” (C. B. Metz and A. Monroy, eds.), Vol. 3, pp. 25 1-298. Academic Press, New York. Longo, F. J. (1990). Dynamics of sperm nuclear transformations at fertilization. In “Fertilization in Mammals” (B. D. Bavister, J. Cummins, and E. R.S. Roldan, eds.), pp. 297-307. Serono Symposia, Norwell, U.S.A. Longo, F. J., and Anderson, E. (1968). The fine structure of pronuclear development and fusion in the sea urchin, Arbacia puncruluta. J. Cell Biol. 39,335-368. Longo, F. J., and Anderson, E. (1969). Cytological aspects of fertilization in the lamellibranch, Mytilus edulis. 11. Development of the male pronucleus and the association of the maternally and paternally derived genomes. J. Exp. Zool. 172,97-120. Longo, F. J., and Anderson, E. (1970). An ultrastructural analysis of fertilization in the surf clam, Spisula solidissima. 11. Development of the male pronucleus and the association of the maternally and paternally derived chromosomes. J. Ulrrasrrucr. Res. 33,5 15-527. Longo, F. J., and Kunkle, M. (1977). Synthesis of RNA by male pronuclei of fertilized sea urchin eggs. J. Exp. Zool. 201,431-437. Longo, F. J., and Plunkett, W. (1973). The onset of DNA synthesis and its relation to morphogenetic events of the pronuclei in activated eggs of the sea urchin, Arbacia puncfuluru. Dev. Biol. 30,5667. Longo, F. J., and Scarpa, J. (1991). Expansions of the sperm nucleus and association of the maternal and paternal genomes in fertilized Mulinia lateralis eggs. Biol. Bull. 1 8 0 , 5 6 4 4 . Longo, F. J., Krohne, G., and Franke, W. W. (1987). Basic proteins of the pennuclear theca of mammalian spermatozoa and spermatids: A novel class of cytoskeletal elements. J. Cell Biol. 105, I105 -1 120. Longo, F. J., Cook, S., and Mathews, L. (1991a). Pronuclear formation in starfish eggs inseminated at different stages of meiotic maturation: Correlation of sperm nuclear transformations and activity of the maternal chromatin. Dev.Biol. 147,62-72. Longo, F. J., Cook, S., Mathews, L., and Wright, S. (1991b). Nascent protein requirement for com-

172

Shirley J. Wright

pletion of meiotic maturation and pronuclear development: Examination of fertilized and A232187-activated surf clam (Spisula solidissima) eggs. Dev. Biol. 148,75586. Longo, F. J., Mathews, L., and Hedgecock, D. (1993). Morphogenesis of maternal and paternal genomes in fertilized oyster eggs (Crassostrea gigas): Effects of cytochalasin B at different periods during meiotic maturation. Biol. Bull. 185, 197-214. Longo, F. J., Mathews, L., and Palazzo, R. E. (1994). Sperm nuclear transformationsin cytoplasmic extracts from surf clam (Spisula solidissima) oocytes. Dev. Biol. 162,245-258. Luthardt, F. W., and Donahue, R. (1973). Pronuclear DNA synthesis in mouse eggs. Enp. Cell Res. 82,143-151. Luttmer, S . J. (1987). Relationship of sperm nuclear enlargement to the stage of egg meiotic maturation at fertilization. Ph.D. Thesis, University of Iowa. Luttmer, S. J., and Longo, F. J. (1986). Examination of living and fixed gametes and early embryos stained with supravital fluorechromes(Hoechst 33342 and 3,3‘-dihexyloxacarboxyanineiodide). Gamete Res. 15,267-283. Luttmer, S. J., and Longo, F. J. (1987). Rates of male pronuclear enlargement in sea urchin zygotes. J. EXP.2.001. 243,289-298. Luttmer, S. J., and Longo, F. J. (1988). Sperm nuclear transformationsconsist of enlargement and condensation coordinate with stages of meiotic maturation in fertilized Spisula solidissima oocytes. Dev. Biol. 128,86-96. Maeda, Y., Yanagimachi, H., Tateno, H., Usui, N., and Yanagimachi, R. (1998). Decondensation of the mouse sperm nucleus within the interphase nucleus. Zygote 6,39-45. Martini, E., Flaherty, S. P., Swann, N. J., Payne, D., and Matthews, C. D. (1997). Analysis of unfertilized oocytes subjected to intracytoplasmic sperm injection using two rounds of fluorescence insitu hybridization and probes to five chromosomes. Human Reprod. 12,2011-2018. Masui, Y. (1985). Meiotic arrest in animal oocytes. In “Biology of Fertilization” (B.B. Metz, and A. Monroy, eds.), Vol. 1, pp. 189-219. Academic Press, New York. Matsumoto, K., Anzai, M., Nakagata, N., Takahashi, A., Takahashi, Y., and Miyata, K. (1994). Onset of paternal gene activation in early mouse embryos fertilized with transgenic mouse sperm. Mol. Reprod. Dev. 39, 136-140. McClay, D. W., and Clarke, H. J. (1997). The ability to organize sperm DNA into functional chromatin is acquired during meiotic maturation in murine oocytes. Dev. Biol. 186,73-84. McGrath, J., and Solter, D. (1984). Completion of mouse embryogenesis requires both the maternal and paternal genome. Cell 37,179-183. Minishull, J. (1993). Cyclin synthesis: Who needs it? BioEssays 15, 149-155. Monk, M., Boubelik, M., and Lehnert, S. (1987). Temporal and regional changes in DNA methylation in the embryonic, extraembryonicand germ cell lineages during mouse embryo development. Development 99,371-382. Montag, M., Tok, V., Liow, S.-L., Bongso, A., and Ng, S.-C. (1992). In vitro decondensation of mammalian sperm and subsequent formation of pronuclei-likestructures for micromanipulation. Mol. Reprod. Dev. 33,338-346. Moore, H. D. M., Hartman, T. D., and Pryor, J. P. (1983). Development of the oocyte penetrating capacity of spermatozoa in the human epididymis. Int. J. Androl. 6,310-31 8. Motlii, J., and Kubelka, M. (1990). Cell-cycle aspects of growth and maturation of mammalian oocytes. Mol. Reprod. Dev. 27,366-375. Nadel, B., de Lara, J., Finkernagel, S. W., and Ward, W. S. (1995). Cell specific organization of the 5 s rRNA gene cluster DNA loop domains in spermatozoa and somatic cells. Biol. Reprod. 53, 1222- 1228. Naish, S. J., Perreault, S. D., and Zirkin, B. R. (1987). DNA synthesis in the fertilizing hamster sperm nucleus: Sperm template availability and egg cytoplasmic control. Biol. Reprod. 36,245253. Nelson, W. G., Pienta, K. J., Barrack, E. R., and Coffey, D. S. (1986). The role of the nuclear matrix in the organization and function of DNA. Annu. Rev. Biophys. Biophys. Chem. 15,457-475.

5. Sperm Nuclear Activation during Fertilization

173

Newport, J. W. (1987). Nuclear reconstitution in vitro. Stages of assembly around protein-free DNA. Cell 48,205-217. Nicholls, R. D., Kroll, J. H. M., Butler, M. G., Karma, S., and Lalande, M. (1989). Genetic imprinting suggested by maternal heterodisomy in non-deletion Prader-Willi syndrome. Nature (London) 342,28 1-285. Nonchev, S . , and Tsanev, R. (1990). Protamine-histone replacement and DNA replication in the male mouse pronucleus. Mol. Reprod. Dev. 25,72-76. Ohsumi, K., and Katagiri, C. (1991). Characterization of the ooplasmic factor inducting decondensation of and protarnine removal from toad sperm nuclei: Involvement of nucleoplasmin. Dev. Biol. 148,295-305. Ohsumi, K., Katagiri, C., and Yanagimachi, R. (1986). Development of pronuclei from human spermatozoa injected microsurgically into frog (Xenopus) eggs. J. Exp. 2001. 237,3 19-325. Okamura, F., and Nishiyama, H. (1978). Penetration of spermatozoon into the ovum and transformation of the sperm nucleus into the male pronucleus in the domestic fowl, Gallus gallus. Cell 2%sue Res. 190,89-98. Palermo, G. D., Colombero, L. T., and Rosenwaks, Z. (1997). The human sperm centrosome is responsible for normal syngamy and early embryonic development. Rev. Reprod. 2,19-27. Paranko, J., and Salonen, I. (1995). Calicin in human sperm fertilizing zona-free hamster eggs in vitro. Reprod. Fertil. Dev. 7,97- 105. Panish, J. J., Kim, C. I., and Bae, I. H. (1992). Current concepts of cell-cycle regulation and its relationship to oocyte maturation, fertilization and embryo development. Theriogenology 38,277296. Passanti, C., Corbi, N., Paggi, M. G., Russo, M. A,, Perez, M., Cotelli, F., Stefanini, M., and Amati, P. (1995). The product of Zfp59 (Mfg2), a mouse gene expressed at the spermatid stage of spermatogenesis accumulates in the spermatozoa nuclei. Cell Growth Difer: 6, 1037-1044. Payne, D., Warnes, G. M., Flaherty, S. P., and Matthews, C. D. (1994). Local experience with zona drilling, zona cutting and sperm microinjection. Reprod. Fertil. Dev. 6,45-50. Payne, D., Flaherty, S. P., Bany, M. F., and Matthews, C. D. (1997). Preliminary observations on polar body extrusion and pronuclear formation in human oocytes using time-lapse video cinematography. Human Reprod. 12,532-541. Perreault, S . D. (1990). Regulation of sperm nuclear reactivation during fertilization. In “Fertilization in Mammals” (B. D. Bavister, J. Cummins, and E. R. S. Roldan, eds.), pp. 285-296. Serono Symposia, Norwell, U.S.A. Perreault, S. D. (1992). Chromatin remodeling in mammalian zygotes. Murat. Res. 296,43Perreault, S . D., Barbee, R. R., and Slott, V. L. (1988). Importance of glutathione in the acqu and maintenance of sperm nuclear decondensing activity in maturing hamster oocytes. Dev. Biol. 125,181-186. Perreault, S. D., Barbee, R., and Stone, C. (1991). Topoisomerase I1 inhibitor (VM-26) effects on cell cycle events from oocyte maturation through first cleavage in the hamster. J. Cell Biol. 115, 462a. Philpott, A., and Leno, G. H. (1992). Nucleoplasmin remodels sperm chromatin in Xenopus egg extract. Cell 69,759-767. Philpott, A., Leno, G. H., and Laskey, R. A. (1991). Sperm decondensation in Xenopus egg cytoplasm is mediated by nucleoplasmin. Cell 65,56-578. Poccia, D. (1986). Remodeling of nucleoproteins during gametogenesis, fertilization, and early development. Znt. Rev. Cytol. 105, 1-65. Poccia, D., and Collas, P. (1996). Transforming sperm nuclei into male pronuclei in vivo and in vitro. Cum Top. Dev. Biol. 34,25-88. Poccia, D., Salik, J., and Krystal, G. (1981). Transitions in histone variants of the male pronucleus following fertilization and evidence for a maternal store of cleavage-stage histones in the sea urchin egg. Dev. Biol. 82,287-296. Poccia, D., Greenough, T., Green, G. R., Nash, E., Erickson, J., and Gibbs, M. (1984). Remodeling

174

Shirley J. Wright

of sperm chromatin following fertilization: Nucleosome repeat length and histone variant transitions in the absence of DNA synthesis. Dev. Biol. 104,274-286. Poccia, D. L., Wolff, R., Kragh, S., and Williamson, P. (1985). RNA synthesis in male pronuclei of the sea urchin. Biochim. Biophys. Acra 824,349-356. Poccia, D., Pavan, W., and Green, G. R. (1990). 6DMAP inhibits chromatin decondensation but not sperm histone kinase in sea urchin male pronuclei. Exp. Cell Res. 188,226-234. Pogany, G. C., Corzett, M., Weston, S., and Balhorn, R. (1981). DNA and protein content of mouse sperm. Implicationsregarding sperm chromatin structure. Exp. Cell Res. 136, 127-136. Porter, D. C., Moy, G. W., and Vacquier, V. D. (1988). CAMP-dependent protein kinase of sea urchin sperm pbosphorylates sperm histone H1 on a single site. J. Biol. Chem. 263,2750-2755. Poueymirou, W. T., and Schultz, R. M. (1989). Regulation of mouse preimplantation development: Inhibition of synthesis of proteins in the two-cell embryo that require transcription by inhibitors of CAMP-dependent protein kinase. Dev. Biol. 133,588-599. Prather, R. S., Sims, M. M., Maul, G. G., First, N. L., and Schatten, G. (1989). Nuclear lamin antigens are developmentally regulated during porcine and bovine embryogenesis.Biol. Reprod. 40, 123-132. Prieto, M. C., Maki, A. H., and Balhorn, R. (1997). Analysis of DNA-protamine interactions by optical detection of magnetic resonance. Biochemistry 36, 11944-11951. Ram, P. T., and Schultz, R. M. (1993). Reporter gene expression in G2 of the I-cell mouse embryo. Dev. Biol. 156,552-556. Raskin, D. M., Wright, D. J., and Wright, S. J. (1997). Sea urchin sperm nuclear enlargement and shape transformationsare differentiallyregulated in vitro. J. Exp. Zool. 277,401-416. Reyes, R.,Rosado, A., Hernandez, O., and Delgado, N. M. (1989). Heparin and glutathione: Physiological decondensing agents of human sperm nuclei. Gamete Res. 23,39-47. Risley, M. S. (1990). Chromatin organization in sperm. In “Chromosomes:Eukaryotic, Prokaryotic and Viral” (K. W. Adolph, ed.), Vol. II, pp. 61-85. CRC Press, Boca Raton, Florida. Risley, M. S., Einheber, S., and Bumcrot, D. A. (1986). Changes in DNA topology during spermatogenesis. Chromosoma 94,2 17-227. Rodman, T. C., Pruslin, F. H., Hoffman, H. P., and Allfrey, B. G. (1981). Turnover of basic chromosomal proteins in fertilized eggs: A cytoimmunochemical study of events in vivo. J. Cell Biol. 90, 35 1-361. Rybouchkin, A., Dozortsev, D., De Sutter, P., Qian, C., and Dhont, M. (1995). Intracytoplasmicinjection of human spermatozoa into mouse oocytes: A useful model to investigate the oocyte-activating capacity and the karyotype of human spermatozoa. Human Reprod. 10, 1130-1135. Saitoh, Y.,and Laemmli, U. K. (1994). Metaphase chromosome structure: Bands arise from a differential folding path of the highly AT-rich scaffold. Cell 76,609-622. Sathananthan, A. H., Kola, I., Osborne, J., Trounson, A., Ng, S. C., Bongso, A., and Ratnam, S. S. (1991). Centrioles in the beginning of human development.Proc. Natl. Acad. Sci. U.S.A. 88, 4806-4810. Sathananthan, A. H., Ratnam, S. S., Ng, S. C., Tarin, J. J., Gianaroli, L., andTrounson, A. (1996). The sperm centriole: its inheritance,replication and perpetuation in early human embryos. Human Reprod. ll, 345-356. Savic, A., Richman, P., Williamson, P., and Poccia, D. (1981).Alterations in chromatin structure during early sea urchin embryogenesis. Proc. Natl. Acad. Sci. U.S.A. 78,3706-3710. Schatten, G. (1994). The centrosome and its mode of inheritance:The reduction of the centrosome during gametogenesis and its restoration during fertilization.Dev. Biol. 165,299-335. Schatten, G., Maul, G. G., Schatten, H., Chaly, N., Simerly, C., Balczon, R., and Brown, D. L. (1985). Nuclear lamins and peripheral nuclear antigens during fertilization and embryogenesis in mice and sea urchins. Proc. Natl. Acad. Sci. U.S.A. 82,4727-4731. Schatten, G., Simerly, C., Palmer, D. K., Margolis, R. L., Maul, G., Andrews, B.. and Schatten, H. (1988). Kinetochore appearance during meiosis, fertilization and mitosis in mouse oocytes and zygotes. Chromosoma 96,341-352.

5. Sperm Nuclear Activation during Fertilization

175

SeGall, G. K., and Lennarz, W. J. (1979). Chemical characterization of the component of the jelly coat from sea urchin eggs responsible for induction of the acrosome reaction. Dev. Biol. 71,3348. Sevigny, G., Kothary, R., Tremblay, E., DeRepentigny, Y., Joly, E. C., and Bibor-Hardy, V. (1995). The cytosolic chaperonin subunit of TRiC-P5 beings to be expressed at the two-cell stage in mouse embryos. Biochem. Biophys. Rex Comm. 216,279-283. Simerly, C., Wu,G. J., Zoran, S., Ord, T., Rawlins, R., Jones, J., Navara, C., Geniti, M., Rinehart, R., Binor, Z., Asch, R., and Schatten, G. (1995). The paternal inheritance of the centrosome, the cell’s microtubule-organizingcenter, in humans, and the implications for infertility.Nature Med. 1,47-52. Simoni, M. (1994). Transgenic animals in male reproduction research. Exp. Clin. Endocrinol. 102, 419-433. Siracusa, G., Coletta, M., and Monesi, V. (1975). Duplication of DNA during the first cell cycle in the mouse embryo. J. Reprod. Fertil. 42,395-398. Stefanini, M., Oura, C., and Zamboni, L. (1969). Ultrastructureof fertilization in the mouse. 2. Penetration of sperm into the ovum. J. Submicrosc. Cytol. 1, 1-23. Stewart, T. A., Bellve, A. R., and Leder P. (1984). Transcription and promoter usage of the myc gene in normal somatic and spermatogeniccells. Science 226,707-710. Stricker, S., Rather, R., Simerly, C., Schatten, H., and Schatten, G. (1989). Nuclear architectural changes during fertilization and development. In “The Cell Biology of Fertilization” (H. Schatten, and G. Schatten, eds.), pp. 225-250. Academic Press, New York. Surani, M. A. H., and Barton, S. C. (1983). Development of gynogenetic eggs in the mouse: Implications for parthenogeneticembryos. Science 222,1034- 1036. Surani, M. A. H., Barton, S. C., and Norris, M. L. (1986). Nuclear transplantation in the mouse: Heritable differences between parental genomes after activation of the embryonic genome. Cell 45, 127-136. Surani, M. A,, Kothary, R., Allen, N. D., Singh, P. B., Fundele, R., Ferguson-Smith,A. C., and Barton, s. C. (1990). Genome imprinting and development in the mouse. Development (1990 Suppl.), 89-98. Sutovsky, P., and Schatten, G. (1997). Depletion of glutathione during bovine oocyte maturation reversibly blocks the decondensationof the male pronucleus and pronuclear apposition during fertilization. Biol. Reprod. 56, 1503-1512. Sutovsky, P., Navara, C. S., and Schatten, G. (1996a). Fate of the sperm mitochondria, and the incorporation, conversion, and disassembly of the sperm tail structures during bovine fertilization. Biol. Reprod. 55, 1195-1205. Sutovsky, P., Hewitson, L., Simerly, C., and Schatten, G. (1996b). Molecular medical approaches for alleviating infertility and understanding assisted reproductive technologies. Proc. Assoc. Am. Phys. 108,432-443. Sutovsky, P., Oko, R., Hewitson, L., and Schatten, G. (1997). The removal of the sperm perinuclear theca and its association with the bovine oocyte surface during fertilization. Dev. Biol. 188,7584. Suzuki, M. (1989). SPKK, a new nucleic acid-bindingunit of protein found in histone. EMBO J. 8, 797-804. Swain, J. L., Stewart, T. A., and Leder, P. (1987). Parental legacy determines methylation and expression of an autosomal transgene: A molecular mechanism for parental imprinting. Cell 50, 719-727. Szollosi, D., Szollosi, M. S., Czolowska, R., and Tarkowski, A. K. (1990). Sperm penetration into immature mouse oocytes and nuclear changes during maturation:An EM study. Biol. Cell 69, 53-64. Tanphaichitr, N., Sobhon, P., Chalermisarachai,P., and Patilantakamkool,M. (1981). Acid-extracted nuclear proteins and ultrastructure of human sperm chromatin as revealed by differential extraction with urea, mercaptoethanol and salt. Gamete Res. 4,297-315.

176

Shirley J. Wright

Tanphaichitr, N., Sobhon, P., Chalermisarachai,P., and Chutatape, C. (1982). Biochemical and ultrastructural characterizationsof nucleoprotamine in human sperm heads treated with micrococcal nuclease and salt. Gamete Res. 6,235-255. Temeles, G. L., Ram, P. T., Rothstein, J. L., and Schultz, R. M. (1994). Expression patterns of novel genes during mouse preimplantation embryogenesis.Mol. Reprod. Dev. 37, 121-129. Terasaki, M., and Jaffe, L. A. (1991). Organization of the sea urchin egg endoplasmicreticulum and its reorganization at fertilization.J. Cell Biol. 114,929-940. Tesafik, J., and KopEcny’,V. (1989a). Development of male pronucleus: Ultrastructureand timing. Gamete Res. 24, 135-150. Tesafik, J., and KopEcnf, V. (1989b). Nucleic acid synthesis and development of human male pronucleus. J. Reprod. Fen. 86,549-558. Thoma, F., Koller, T., and Mug, A. (1979). Involvement of histone H1 in the organization of the nucleosome and the salt-dependentsuperstructuresof chromatin. J. Cell Biol. 83,402-427. Thompson, E. M., Legouy, E., Christians, E., and Renard, J. P. (1995). Progressive maturation of chromatin structure regulates HSWO.1 gene expression in the preimplantationmouse embryo. Development 121,3425-3437. Tsanev, R., and Avramova, Z. (1981). Nonprotamine nucleoprotein ultrastructures in mature ram sperm nuclei. Eur: J. Cell Biol. 24, 139-145. Turner, B. M. (1991). Histone acetylation and control of gene expression. J. Cell Sci. 99, 13-20. Ulitzer, N., and Gruenbaum, Y. (1989). Nuclear envelope assembly around sperm chromatin in cellfree preparations from Drosophila embryos. FEES Lett. 259, 113-1 16. Usui, N., and Yanagimachi, R. (1976). Behavior of hamster sperm nuclei incorporated into eggs at various stages of maturation, fertilization and early development: The appearance and disappearance of factors involved in sperm chromatin decondensation in egg cytoplasm. J. Ultrustruct. Res. 57,276-288. Ward, W. S. (1993). DNA loop domain tertiary structure in mammalian spermatozoa.Biol. Reprod. 48,1193-1201. Ward, W. S . (1994). The structure of the sleeping genome: Implications of sperm DNA organization for somatic cells. J. Cell. Biochem. 55,77-82. Ward, W. S., and Coffey, D. S. (1989). Identification of a sperm nuclear annulus: A sperm DNA anchor. Biol. Reprod. 41,361-370. Ward, W. S., and Coffey, D. S. (1990). Specific organization of genes in relation to the sperm nuclear matrix. Biochem. Biophys. Res. Commun. 173,20-25. Ward, W. S., and Coffey, D. S. (1991). DNA packaging and organization in mammalian spermatozoa: Comparison with somatic cells. Biol. Repmd. 44,569-574. Ward, W. S., and Zalensky, A. 0.(1996). The unique, complex organization of the transcriptionally silent sperm chromatin. Crit. Rev. Eukaryot. Gene Enpr: 6,139-147. Ward, W. S., Partin, A. W., and Coffey, D. S. (1989). DNAloop domains in mammalian spermatozoa. Chromosomu 98,153-159. Ward, W. S., McNeil, J., de Lara, J., and Lawrence, J. (1996). Localization of three genes in the hook-shaped hamster sperm nucleus by fluorescent in situ hybridization. Biol. Repmd. 54, 12711278. Wiekowski, M., Miranda, M., and DePamphilis,M. L. (1993). Requirements for promoter activity in mouse oocytes and embryos distinguish paternal pronuclei from maternal and zygotic nuclei. Dev. Biol. 159,366-378. Wiesel, S., and Schultz, G. A. (1981). Factors which may affect removal of protamine from sperm DNA during fertilization in the rabbit. Gamete Res. 4,25-34. Wilson, E. B. (1925). “The Cell in Development and Heredity.” Macmillan, New York. Wogelmuth, D. J. (1983). Synthetic activities of the mammalian early embryo: Molecular and genetic alterations following fertilization.In “Mechanism and Control of Animal Fertilization” (J. F. Hartmann, ed.), pp. 415-452. Academic Press, New York.

5. Sperm Nuclear Activation during Fertilization

177

Worrad, D. M., Ram, P. T., and Schultz, R. M. (1994). Regulation of gene expression in the mouse oocyte and early preimplantationembryo: Developmental changes in Spl and TATA box-binding protein, TBP. Development 120,2347-2357. Worrad, D. M., Turner, B. M., and Schultz, R. M. (1995). Temporally restricted spatial localization of acetylated isoforms of histone H4 and RNA polymerase II in the 2-cell mouse embryo. Development 121,2949-2959. Wouters-Tyrou, D., Martinage, A., Chevaillier,P., and SautiBre, P. (1998). Nuclear basic proteins in spermiogenesis.Biochemie 80, 117-128. Wright, S. J., and Longo, F. J. (1988). Sperm nuclear enlargement in fertilized hamster eggs is related to meiotic maturation of the maternal chromatin. J. Exp. Zool. 247, 155-165. Wright, S. J., and Schatten, G.(1988). Chromosome condensation but not pronuclear decondensation is blocked by a topoisomeraseII inhibitor during fertilization and mitosis in surf clams, mice and sea urchins. J. Cell B i d . 107, 172a. Wright, S. J., and Schatten, G. (1990). Teniposide, a topoisomerase I1 inhibitor, prevents chromosome condensation and separation but not decondensation in fertilized surf clam (Spisula solidissima) oocytes. Dev. Biol. 142,224-232. Wright, S. J., Centonze, V. E., Stricker, S. A., DeVries, P., Paddock, S. W., and Schatten, G.(1993). An introduction to confocal microscopy and three-dimensionalreconstruction.Methods Cell Biol. 38, 1-45. Yamashita, M. (1985). Electron microscopic analysis of the sperm nuclear changes in meiosis inhibited eggs of the brittle-starAmphipholis kochii. J. Exp. Zuol. 235, 105-117. Yanagimachi, R. (1988). Mammalian fertilization.In “The Physiology of Reproduction” (E. Knobil, J. D. Neill, L. L. Ewing, C. L. Markert, G. S. Greenwald, and D. W. Pfaff, eds.), pp. 135-186. Raven Press, New York. Yanagimachi, R., and Noda, Y.D. (1970a). Ultrastructuralchanges in the hamster sperm head during fertilization. J. Ultrastruct. Res. 31,465-485. Yanagimachi, R., and Noda, Y. D. (1970b). Fine structure of the hamster sperm head. Am. J. Anat. 128,367-387. Yanagimachi, R., and Noda, Y. D. (1970~).Electron microscopic studies of sperm incorporation into the golden hamster egg. Am. J. Anat. 128,429-462. Yoshida, M., Ishigaki, K., Nagai, T., Chikyu, M., and Pursel, V. G. (1993). Glutathione concentration during maturation and after fertilization in pig oocytes: Relevance to the ability of oocytes to form male pronucleus. Biol. Reprod. 49,89-94. Zalensky, A. O., Breneman, J. W., Zalenskaya, I. A., Brinkley, B. R., and Bradbury, E. M. (1993). Organization of centromeres in the decondensed nuclei of mature human sperm. Chromosoma 102,509-518. Zalensky, A. O., Allen, M. J., Kobayahi, A., Zalenskaya, I. A., Balhom, R., and Bradbury, E. M. (1995). Well-defined genome architecture in the human sperm nucleus. Chmnosoma 103,577-590. Zalensky, A. O., Tomilin, N. V., Zalenskaya, I. A., Teplitz, R. L., and Bradbury, E. M. (1997). Telomere-telomere interactions and candidate telomere binding protein(s) in mammalian sperm cells. Exp. Cell Res. 232,29-41. Zentgraf, H., and Franke, W. W. (1984). Differences of supranucleosomalorganization in different kinds of chromatin: Cell type-specific globular subunits containing different numbers of nucleosomes. J. Cell Biol. 99,272-286. Zentgraf, H., Miiller, U., and Franke, W. W. (1980). Supranucleosomalorganization of sea urchin sperm chromatin in regularly arranged 40 to 50 nm large granular subunits. Eu,: J. Cell Biol. 20, 254-264. Zhao, J., and Benbow, R. M. (1994). Inhibition of DNA replication in cell-free extracts of Xenopus laevis eggs by extracts of Xenopus laevis oocytes. Dev. Biol. 164,6242. Zirkin, B. R., Soucek, D. A., Chang, T. S. K., and Perreault, S. D. (1985). In vitro and in vivo studies of mammalian sperm nuclear decondensation. Gamete Res. 11,349-366.

178

Shirley J. Wright

Zirkin, B. R., Perreault, S. D., and Naish, S. J. (1989). Formation and function of the male pronucleus during mammalian fertilization.In “The Molecular Biology of Fertilization” (H. Schatten and G . Schatten, eds.), pp. 91-114. Academic Press, New York. Zwingman, T.,Erickson, R. P., Boyer, T., and Ao, A. (1993). Transcriptionof the sex-determiningregion genes Sry and Zfy in the mouse preimplantationembryo. Proc. Nutl. Acud. Sci. U.S.A. 90, 814-817.

6 Fibroblast Growth Factor Signaling Regulates Growth and Morphogenesis at Multiple Steps during Brain Development' Flora M. Vaccarino,* ( j Michael L. Schwartz,j Rossana Raballo, * Julianne Rhee, * and Richard Lyn-Cook* *Child Study Center and +Sectionof Neurobiology Yale University School of Medicine New Haven. Connecticut 06520

I. Overview 11. DevelopmentalControl Genes and Morphogenesis within the CNS

A. Specification of Compartments B. Control of Cell Interactions C. Morphogens and Pattern Formation 111. The Fibroblast Growth Factor Family A. FGF Ligands B. FGF Receptors IV. FGFs Regulate Patterning of the Neuroepithelium A. FGFs Act as Inducers and Local Growth Regulators B. Early Actions of FGFs: Regional Patterning C. Kinetics of Growth within the Dorsal TelencephalicCompartment D. Later Actions of FGFs: Regulation of the Size of the Cerebral Cortex E. Mechanism of Action of FGFs V. Effect of FGFs on Cell Fate and Cortical Lineages A. Neuronal and Glial Progenitor Cells Are Specified Early within the Cortical Neuroepithelium B. Time-Dependent Action of FGF2 on Different Cell Lineages C. The Lack of FGF2 Decreases the Number of Neurons and Glia in the Cerebral Cortex VI. Conclusions and Future Prospects References

~~

The fibroblast growth factor (FGF) family comprises several members with distinct patterns of expression in the developing central nervous system. FGFs regulate the early specification and the subsequent growth of central nervous system regions. These different actions require the coordinated activation of distinct sets of target genes by FGFs at the appropriate stage of development. The role of FGF2 in the growth and morphogenesis of the cerebral cortex is reviewed in detail. The cellular and molecular mechanisms that underlie the action of FGF2 on cortical development are discussed. Q 1999 Academic Press. 'This work represents a collaboration between the laboratories of the first two authors. Current Topics in Developmental Bialngy, Vol. 46 Copyright Q 1999 by Academic Press. All rights of repduction in any form reserved.

0070-2153/99 $30.00

179

Flora M. Vaccarino et al.

180

1. Overview Morphogenesis in the vertebrate nervous system progresses through hierarchical steps. Polarization is followed by a subdivision of the central nervous system (CNS) into compartments, the specification of cell number and fate, migration, and finally, the establishment of connections (Fig. 1).Combinations of regulatory genes are used initially to specify regional differences within the CNS in a manner that is reminiscent of gap and segment polarity genes in segments and parasegments of the fruit fly [for reviews, see Krumlauf (1994) and Lumsden and Krumlauf (1996)l. Many of these genes code for transcription factors that bind

Polarization

Compartments

Cellular Specification

\

/ number

type

Y

44 Connectivity

-0-4 Fig. 1 Overview of the major steps in CNS development. Each step involves the functioning of stagespecific developmental control genes and an array of effector genes with more restricted functions.

6. Fibroblast Growth Factors and Brain Morphogenesis

181

to DNA through an evolutionarily conserved region called the homeodomain (Gehring et al., 1994). Following regional specification, a cascade of genes controls more restricted functions. These include the determination of cell number and fate and position within each region. Our general view is that in most cases, these executive functions are regulated by cell-to-cell interactions mediated by morphogens and different membrane receptors. The morphogenetic functions of the fibroblast growth factor (FGF) family will be reviewed in more detail.

II. Developmental Control Genes and Morphogenesis within the CNS A. Specification of Compartments

In both vertebrates and invertebrates, the CNS can be divided into compartments. Compartments are made up of groups of cells sharing a common set of genes that define their regional identity (Lawrence and Struhl, 1996). All descendants from the founder cell group of a given compartment form a histogenetic field of cells that share a genetic ancestry that coordinates their growth and fate. In the neural tube, segmentlike units called rhombomeres are morphologically evident, each rhombomere expressing a set of homeodomain-containinggenes of the Hox family (Capecchi, 1997). In the rostral part of the neuraxis, however, regional specification is more complex. Dorsal regions of the telencephalic vesicles give rise to the cerebral cortex, ventrolateral regions give rise to the basal ganglia, and the most ventral areas develop into the hypothalamus (Rubenstein and Shimamura, 1997). E m ,Dlx, Otx, Otp, PM, and other divergent homeobox genes are differentially expressed in these compartments (Simeone et al., 1992a,b, 1994;Callaerts et al., 1997; Rubenstein and Shimamura, 1997). Null mutations of these genes in vertebrates have shown that they are required for either regional or cellular specification within their domain of expression (Acampora et al., 1995; Anderson et al., 1997; Stoykova et al., 1996). For example, Otx2 is an early gene involved in regional specification of the forebrain as distinct from lower regions of the neuraxis (Acampora ef al., 1995; Ang et al., 1996; Rhinn et al., 1998). Later on, Otxl, another member of this family, is expressed in neuronal progenitor cells of the forebrain and is required for the proper cellular specification within the cerebral cortex (Acampora et al., 1996;Frank et al., 1994).Thus, it is clear that these genes “control” development in various ways. But how do they execute this function?

B. Control of Cell Interactions The different functions and complex spatiotemporal dynamics of expression of developmental control genes make it difficult to understand the basic rules

182

Flora M. Vaccarino et al.

through which they orchestrate CNS development. Here, we will argue that homeodomain genes in vertebrates, similar to gap genes in Drosophilu, coordinate various sets of interlaced developmental functions within their domain of expression. These “master control genes” fulfill this function by controlling the expression of sets of molecules that are expressed at the cell surface and allow cells to interact with one another (Fig. 2). For example, progenitor cells intermix within a compartment, but do not cross compartment boundaries, in part due to the presence of region-specific surface properties of cells. By directing the expression of specific sets of adhesion molecules, homeobox genes may regulate the sorting of cells into different compartments and prevent their diffusion across compartment boundaries (Edelman and Jones, 1993; Stoykova ef ul., 1997). In addition, homeodomain genes could control cell number and fate by directing the expression of growth factors and other receptors on the cell surface. A specific example is the direct regulation of decapentaplegic (a TGF-P homolog) by the homeodomain gene Ulfrubithoraxin Drosophilu (Capovilla et ul., 1994). Finally, by affecting the differential expression of recognition and guidance molecules, homeodomain genes could influence growth core navigation and target recognition (Fig. 2). In essence, the genetic ancestry of a compartment would provide a basic layout and allow certain groups of cells to interact with one another. Cell-to-cell communication causes new waves of gene expression, through which cells would carry out the different tasks of sorting themselves out, proliferating, differentiating, migrating, and so on. These events put new cells in contact with one another, create novel interactions, and cause a further refining of the pattern.

C. Morphogens and Pattern Formation

Specialized groups of cells within the CNS generate diffusible signals, morphogens, that may act on distant cell populations through extracellular diffusion and the formation of concentration gradients. Morphogens act on receptors located at the surface of responding cells; the activation of these receptors is then thought to influence target cells by activating gene expression. The precise mechanisms of these interactions with the genome and target genes are not well known. It has been hypothesized that morphogenetic molecules may induce the expression of different sets of genes at different steps in their concentration along

Fig. 2 Scheme of the possible relationships between cell-intrinsic genetic factors and cell surface signals. Regulatory genes that exert a master control role over development organize a variety of cellular functions. These include sets of molecules that are expressed the cell surface and control cell adhesion, cell growth and proliferation, and connectivity. The effects of these perturbations and cell-to-cell interactions are then relayed back into the nucleus through specific intracellular transduction systems.

This Page Intentionally Left Blank

183 the gradient of diffusion. This property would explain how a linear gradient of these molecules could be translated into qualitative differences in body pattern. An alternative hypothesis is that these morphogens change gene expression in neighboring cells and changes are subsequently propagated indirectly,by a chain of cell-to-cell interactions. Both mechanisms of action have been reported in limited experimental cases; however, their general validity remains to be demonstrated (Reilly and Melton, 1996; Lecuit et al., 1996). Morphogens are generated either within or at the interface between different compartments, and may influence cells at either one or both sides of the compartment boundary (Meinhardt, 1983). Morphogens have different roles during development; some appear to direct tissue patterning and induce new cell fates and some regulate the growth and development of a whole compartment (see Sections IV,A-IV,D). We will see that these different actions may be explained by the secondary induction of different sets of genes in target cells. Concentration gradients of morphogenetic molecules are thought to model the size and shape of an organ (Lawrence and Struhl, 1996).A number of factors may influence the steepness and shape of their diffusion gradient, including the extent to which they bind to cell surface molecules (receptors) and to components of the extracellular matrix. For example, members of the FGF family bind to ubiquitously expressed sulfated heparin-like proteoglycans and this binding promotes their receptor-mediated actions, limits their diffusion, and protects the peptides from degradation [for reviews, see Basilico and Moscatelli (1992), Burgess and Maciag (1989), and Baird (1994)l. 6. Fibroblast Growth Factors and Brain Morphogenesis

111. The Fibroblast Growth Factor Family A. FCF Ligands

At present, the FGF family is composed of 17 ligands and 4 receptors that have been shown to play an important role in the development of bone, skin, and the CNS. This large family of factors, and related receptors, appear to have evolved through phases of gene duplication, one of which may have coincided with the emergence of the vertebrate body scaffold (Coulier et al., 1997). FGFs produce a variety of biological effects in vitro, most prominently cell proliferation, migration, and survival (Baird, 1994; Eckenstein, 1994; Basilico and Moscatelli, 1992) through mechanisms that are still largely unknown. FGFl and FGF2 are the first two members of this family to have been isolated and are expressed within the adult brain (Bohen et al., 1985; Stock et al., 1992;Woodward et al., 1992; Gonzalez et al., 1995). FGF2 may have several molecular forms, including an 18-kDa variant that is exported extracellularly to the cell membrane and a higher molecular weight form found in the nucleus that may act as an autocrine regulator of gene expression (Woodward et al., 1992; Amalric et al., 1994).

184

Flora M. Vaccarino et al.

FGFl and FGF2 are among the most potent mitogens for neuroectodermal cells, including precursor cells of the CNS (Murphy et al., 1990,1994; Kilpatrick and Bartlett, 1995; Gensburger et al., 1987; Cattaneo and McKay, 1990; Ray et al., 1993; Engele and Churchill Bohn, 1992; Vaccarino et al., 1995). Additionally, several studies have shown that FGFl and FGF2 mRNA and proteins are expressed in the brain neuroepithelium at early stages development (E9-Ell in the mouse) (Nurcombe et al., 1993; Hebert et al., 1990; Gonzalez et al., 1990; Powell et al., 1991; Giordano et al., 1991; Ernfors el al., 1990; Weise et al., 1993). Thus, the proliferative action of these factors may have a profound impact on the morphogenesis of the CNS. Several other FGFs are differentially localized within the developing neuroepithelium (Smallwood et al., 1996; Crossley and Martin, 1995; Mason et al., 1994; Mahmood et al., 1995). It is to be expected that the actions of the various FGF ligands are constrained by their differential pattern of expression and their time course (see Section IV, A).

B. FCF Receptors

FGFs act on a particular class of tyrosine kinase receptors located at the surface of responding cells. Four FGF receptors are known, FGFRI-FGFR4 (Partanen et al., 1993). The activation of these receptors is thought to generate signaling cascades that eventually alter the cellular pattern of gene expression. Remarkably, several lines of evidence have shown that the activation of FGFRl causes diverse actions such as mesoderm induction, DNA synthesis, cell differentiation, and neurite outgrowth (McFarlane et al., 1996; Amaya et al., 1991; Zhan et al., 1993;Robinson et al., 1995; Saffel et al., 1997). How does the action of FGFR1 give rise to qualitatively different results? One possibility is that post-receptor tyrosine kinase transducing systems bear this responsibility, either because they are differentially activated by different levels of ligand or because of regional or developmental differences of the responding cells. For example, one of the earliest actions of FGF family members is the formation and patterning of posterior mesoderm and the direct patterning of neural tissue (Kengaku and Okamoto, 1995; Lamb and Harland, 1995; Cox and Hemmati-Brivanlou, 1995;Amaya et al., 1991,1993; Pownall et al., 1996).The induction of posterior mesoderm by FGF is thought to require signaling mediated by the cellular protooncogene Ras, the serine/threonine protein kinase Raf-1 (MacNicol et al., 1993), and may also involve an Src-type protein kinase intermediate (Weinstein et al., 1998).The same RaslRaf-1 pathway, in conjunction with other intracellular factors, is involved in FGF-mediated cell proliferation. The action of FGFl and FGF2 on the cell cycle involves the passage from a quiescent phase (GI) into the DNA-synthetic phase ( S phase) (DeHamer et al., 1994). FGFl and FGF2 are able to promote DNA synthesis in quiescent cells through p90/FRS2, a molecule that specificallycouples activated FGFRl to the RasIMAP kinase pathway (Goh

6. Fibroblast Growth Factors and Brain Morphogenesis

185

et al., 1996; Rabin et al., 1993; Ong et al., 1996; Kouhara et al., 1997; Huang et al., 1995) (Fig. 3). Downstream molecules of this FGF-activated mitogenic pathway might include cyclin D/cdk4 and the retinoblastoma (Rb) protein, regulators of the G, phase of the cell cycle (Peeper et al., 1997; Takauwa et al., 1993) (Fig. 3). Conversely, the stimulation of neurite outgrowth by FGFs requires phospholipase C (PLC-y)/PKC signaling (William et al., 1994; Saffel et al., 1997). Some of these cellular factors are able to translocate to the nucleus, where they activate transcription factors such as c-jun, p90Rsk, ~ 6 2 ~c-myc, f , and regulate gene expression. Indeed, FGFl and FGF2 ligands per se have been reported to translocate to the nucleus, probably by receptor-mediated internalization, and DNA synthesis has been linked to the persistence of FGF ligands in the nucleus (Prudovsky et al., 1994; Mehta et al., 1998). In conclusion, even though a given intracellular factor may be necessary for an effect, it may not be sufficient, and therefore it is the entire spectrum of intracellular signaling molecules and pathways that confers functional specificity. The intracellular ramifications of FGF signaling and ultimately the biological function depend on the developmental stage and cell type, which provide a “context” of available signal transducers and target genes.

IV. FGFs Regulate Patterning of the Neuroepithelium A. FGFs Act as Inducers and local Growth Regulators

During development, cell proliferation is coupled with regional and cellular specification. We have hypothesized that these functions are coordinated via the reciprocal interactions among regulatory genes of the homeobox family and signaling molecules at the cell surface (Fig. 2). Evidence suggests that cell-to-cell signaling mediated by FGFs is conveyed to the nucleus, where it stimulates the expression of homeobox genes (Robe1 et al., 1995; Pownall et al., 1996; Partanen et al., 1998). We will distinguish early from late actions of FGFs on developing tissue. Early in development, FGFs affect both growth and regional specification. For example, FGF signaling is required for the initiation and the outgrowth of the limb (Cohn et al., 1995; Niswander and Martin, 1993; Yang and Niswander, 1995). The capacity of certain cells to instruct adjacent cells and recruit them into the coordinated growth of a region or body segment is called induction. FGFs display this capacity within the CNS, as we will see in Section IV,B. Later, this inductive capacity is lost, but FGFs still retain their capacity to regulate growth. Thus, the functional response of cells to FGFs is dictated, in part, by the genes targeted by these factors at specific stages of development. Several FGFs are expressed within the proliferative neuroepithelium. FGF8 is localized within boundary regions both in the brain stem and in the forebrain (see below), whereas FGFZ is more diffusely distributed within the neuroepithelium.

186

Flora M. Vaccarino et al.

U

cytoskeletal proteins: cortactin, vinculin etc

MEK

\

nucleus

c-jun c-myc

.

?

:

? :’ A*’

I

master control genes homeodomain genes

I

Fig. 3 FGF-mediated intracellular transduction systems. The binding of FGF to cell surface receptors triggers the activation of signaling cascades both in the cytoplasm and in the nucleus. In the cytoplasm, FGFs activate the RaslRaf-1IMAPK cascade, as well as the PLC-y and c-Src cascades. Each of these pathways is involved in a set of specific functions (see text). Some of these molecules, including MAPK, PKC, and the FGF receptor-ligand complex, translocate to the nucleus, possibly affecting the phosphorylation and activity of transcription factors and triggering cascades of gene expression.

6. Fibroblast Growth Factors and Brain Morphogenesis

187 We have hypothesized that members of the FGF family may be important for the regulation of cerebral size and shape during vertebrate evolution. In this respect, it is intriguing that FGFs participate in cranial suture and cranial bone morphogenesis and in the regulation of cellular proliferation within the cerebral cortex. Thus, FGFs may act to coordinate the growth and size of the skull and brain (Kim et al., 1998).

B. Early Actions of FCFs: Regional Patterning

FGF8 is highly concentrated within a specialized morphogenetic center at the mesencephalic/rhomboencephalic boundary, an area also designated as the isthmus (Crossley and martin, 1995). This region is at the caudal boundary of expression of homeobox genes of the Otx and En families, which are involved in the determination of a mesencephalic fate. Through the secretion of FGF8 into the adjacent compartment, cells of the isthmus induce a mesencephalic fate. First, the overexpression of FGF8 within the dorsal mesencephalon leads to marked mesencephalon overgrowth (Lee et al., 1997). Second, if a bead soaked in FGF8 is placed in a more anterior diencephalic location, it induces the ectopic growth of a new mesencephalon of opposite polarity (mirror image) (Crossley et al., 1996). This duplication replaces the normal diencephalic tissue, and thus it appears to be a true transformation of cell and tissue identity. This is apparently achieved through the FGF-induced ectopic peak of expression of En-2 and other homeobox genes involved in mesencephalic cell fate (Crossley et al., 1996). These experiments suggest a crucial role of a localized source of FGF8 in the induction and specification of the mesencephalon. Confirming this hypothesis, mice with a reduction in FGF8 mRNA levels show a deletion of posterior midbrain and cerebellar structures (Meyers et al., 1998). In the neuroepithelium of the dorsal telencephalon, similar gene-morphogen interactions may occur. FGF8 is expressed within the forebrain in the ventral region between the telencephalic vesicles, the commissural plate or anterior neural ridge (Crossley and Martin, 1995). This area also expresses Wnt-1, a secreted factor also present in the isthmus, and possibly other members of the FGF family (Mahmood et al., 1995). In this location, FGF8 induces the expression of BF-1, a transcription factor essential for the growth of the cerebral cortex (Shimamura and Rubenstein, 1997). Mice carrying hypomorphic alleles of FGF8 lack the olfactory bulbs, demonstrating that FGF8 is required for the specification of the olfactory placodes. These mice also show a reduction in the size of the telencephalon, suggesting that, in addition to the regional specification of olfactory bulbs, FGF8 may regulate cerebral cortical growth. In this respect, the action of FGF8 may be complemented by other FGFs expressed diffusely throughout the dorsal telencephalic region. For example, FGF2 is required for the proper growth of the telencephalic epithelium (see Section IV,C). Both

188

Flora M. Vaccarino et al.

FGF8 and FGF2 act on the same type of FGF receptors, suggesting a complex interaction of several FGF ligands during the morphogenesis of the telencephalon. In conclusion, a morphogenetic center may exist in the rostro/anterior region of the forebrain, a source of FGFs and Wnt proteins. The winged helix transcription factor BF-I may be one of the genes that act downstream of FGF8. However, more needs to be learned concerning the integration of the function of the various factors present in this location during forebrain early growth and patterning. C. Kinetics of Growth within the Dorsal Telencephalic Compartment

Before we describe the actions of FGFs on the growth of the cerebral cortex, we will briefly review the kinetics of proliferation of the dorsal telencephalic neuroepithelium, also called the pseudostratified ventricular epithelium (PVE). Progenitor cells of the PVE line the lateral ventricles and are the founder cell population that will give rise to all the neurons and glia of the cerebral cortex. These founder cells proliferate symmetrically, doubling their number at each completed cell cycle, and accounting for the rapid growth of the forebrain vesicles. At the beginning of neurogenesis, about E l l in the mouse and E l 3 in the rat, some of the founder cells begin to divide asymmetrically. In asymmetrical divisions, one of the daughter cells exits the cycle and starts migrating toward the cortical plate, where it will complete its differentiation. Symmetric terminal mitoses also occur, in which both daughter cells become postmitotic neurons (Kornack and Rakic, 1995). It is generally accepted that the future phenotype of the progenitor is determined some time during the earliest mitotic cycles (McConnell and Kaznowski, 1991; Frantz and McConnell, 1996). The regulation of neuronal fate will be discussed in Section V, A, B. In addition to becoming neurons, cells of the PVE give rise to more specialized glial progenitors that migrate to a distinct zone of the cerebral cortical wall called the subventricular zone or secondary proliferative population (SPP). Thus, cells of the PVE “seed” glial progenitor cells of the subventricular zone, which will continue to proliferate locally, generating glial cells well into the postnatal period (Takahashi et al., 1995b). The number of neuronal progenitor cells, their rate of proliferation, and the number of mitotic cycles they undergo are thought to be major determinants of the final size of the cerebral cortex and the total number of cerebral cortical neurons (Takahashi et al., 1994). The length of the cell cycle increases during neurogenesis from 8 hr to more than 18 hr and this increase is due to a prolongation of the G, phase. The factors that affect this progressive lengthening of the G, phase of the cell cycle are unknown (Takahashi et al., 1995a). Another important factor in the regulation of cortical size and neuron number

6. Fibroblast Growth Factors and Brain Morphogenesis

189 is the number of mitotic cycles that progenitor cells undergo. The decision of a cell to reenter or exit the cycle occurs in early G, and results in either the progression to the DNA synthetic (S) phase, or the progression into the postmitotic phase (Fig. 4, cycle A; see color plate). The kinetics of the progression from a PVE containing 100% founder cells to the complete exhaustion of the PVE is nonlinear. The proportion of proliferative cells exiting the cell cycle (called the quiescent fraction, or Q) is initially very small, and gradually increases to 50% approximately two-thirds of way through the period of neurogenesis. Q rapidly increases during the last third of neurogenesis, as most progeny exit the cell cycle during this period (Takahashi et al., 1996). It should be noted that in most studies examining the proportion of proliferating and Q fraction cells, the eventual fate of Q cells is unknown. Many of these cells become postmitotic and continue to differentiate, while a fraction will undergo programmed cell death (Fig. 4, cycle B). The size of the pool of cells which will eventually sucumb to programmed cell death is currently under debate. In addition to the loss of cells of the Q fraction to cell death, apoptotic death may also influence the size of the progenitor pool as well. A recent estimate suggests that the proportion of progenitor cells of the PVE that die is in the neighborhood of 2 and 7% (Thomaidou et al., 1997) (see Section IV,E).

D. later Actions of FGFs: Regulation of the Size of the Cerebral Cortex

Members of the FGF, epidermal growth factor (EGF), and platelet-derived growth factor (PDGF) families have been shown to influence the growth of progenitor cells of the forebrain in primary culture (Cattaneo and McKay, 1990; Kenisberg et al., 1992; Reynolds et al., 1992; Ray et al., 1993; Vescovi et al., 1993; Kilpatrick and Bartlett, 1995; Williams et al., 1997). However, the extension of these effects to an understanding of their role in vivo is problematic. The first question is whether any of these factors are present within the neuroepithelium and at what stage of development. EGF receptors have been identified in the postnatal period, after neurogenesis has ended; they are localized in the subependimal layer and in glial progenitor cells (Gomez-Pinilla et al., 1988; Morshead et al., 1994). In contrast, ligands and receptors of the FGF family are present within the PVE during early stages of cortical neurogenesis. FGF2 mRNA shows a dorsoventral gradient within the proliferative neuroepithelium, with peak levels in dorsal regions of the telencephalon (Fig. 5a; see color plate). FGFRl and FGFR2 are expressed in a spatiotemporal pattern similar to that of FGF2, suggesting an autocrine/paracrine function for FGF2 (Fig. 5b). These receptors are localized within dividing cells of the PVE, indicating that FGF-like factors target progenitor cells directly (Fig. 5e,f). There is a down-regulation of FGF2, FGFRl, and FGFR2 in progenitor cells with the progression of neurogenesis. For example, FGFRl and FGF2 are high-

Flora M. Vaccarino et al.

190

ly expressed in the rat PVE at E14.5, when the majority of cells are still proliferating, and are almost absent at E17.5, when many progenitor cells are exiting the mitotic cycle (Vaccarino et al., 1999). These observations suggest that a decrease in FGF signaling may be necessary to allow cells to exit the cycle to become neurons. Alternatively, the decrease in FGF2 signaling may affect the length of the cell cycle in cortical progenitor cells. To test whether FGF signaling has an impact during neurogenesis, we have microinjected FGF2 into the cerebral ventricles of E15.5 rat fetuses, a time when infragranular neurons of the cerebral cortex (layers VI and V) are being generated. The microinjection of FGF2 at this time resulted in a significant increase in both cortical volume (53%) and cell number (67%) by the end of neurogenesis (Vaccarino et al., 1999). Further, these effects persisted into maturity-embryos injected at E15.5 and allowed to grow to maturity (P90)showed similar increases in cortical volume and neuron number (Table I). In contrast to the dramatic increase in cortical volume, no differences were found in cortical thickness between control and FGF2-treated animals. Thus, FGF-induced changes in volume are likely to be attributed to an overall increase in the surface area of the cortical mantle. These experiments have thus clearly shown that the overexpression of FGF2 by a single microinjection at the beginning of neurogenesis has a major impact on the number of neurons generated. This is consistent with the expression of FGFRl in neuronal progenitor cells of the PVE. Table I FGF2 Treatment during EmbryogenesisO Total number of cells (lo6) Cell type Neurons Glia Glutamate positive GABA positive

Control

FGF2 Injection

Increase

36.6 f 2.5 14.9 f 1.9 14.4 f 0.6 8.3 f 0.4

68.5 f 3.3*** 22.8 f 3.1 34.2 f 3.2**' 15.7 f 1.6*

87% 53% 136.5% 87.8%

Note: * p < 0.05; ***p < 0.0001. OData show that FGF2 treatment during embryogenesis increases the number of neurons in the cerebral cortex (mean values f standard error). FGF2 (75 ng) was injected in the cerebral ventricles of rat embryos at E15.5 and brains were analyzed in their tenth postnatal week. Values were obtained by stereological analysis of serial sections encompassing the entire cerebral hemisphere and spaced 100-500 papart; N = 3 to 4 animals per group. Cresyl violet-stained sections were used to obtain total number of neurons and glia; glutamate- and GABA-positivecells were evaluated after specific immunostaining (Vaccarino et aL, 1999). Analysis of variance (ANOVA) revealed that FGF2treated animals were significantlydifferent from controls in total cell number ( p < 0.001; F = 51.5) and in cell density ( p < 0.001; F = 39.5). There was a differential effect of FGF2 treatment on glutamate and GABA neurons with respect to both cell number ( p < 0.01; F = 11.3) and cell density ( p < O.OOO1; F = 70.3). Post hoc pairwise comparisons (Sheffe post hoc test) revealed that FGF2 treatment affected total number and density of neurons, as indicated (*).

Fig. 4 Regulatory events in the cell cycle. The progenitor A divides to give rise to two daughter cells undergoing different fates. At left, one daughter reenters the cycle to divide again (cell A’) and the other permanently exits the cycle (postmitotic cell). At right, cell A’ reenters the cycle and cell B’ dies by apoptosis.

Fig. 5 Distribution of FGF2 and FGFRl mRNA and proteins during cerebral cortical development. In siru hybridization for FGF2 (a, c) and FGFRl (b, d) at E14.5, showing the localization of FGF2 and FGFRl within the PVE. Double immunocytochemistry for BrdU (e) and FGFRl (0 at E16.5. Arrows indicate corresponding double-labeled cells of the PVE (red arrows) and SPP (purple arrows). Ventricle side is down. Scale bar, 400 pm in a and b; SO pm in c-f.

Fig. 6 FGF2 treatment does not affect the number of apoptotic cells. Parietal area of the cerebral cortex showing TUNEL-positive cells (brown) in control (a) and FGF2-treated brain (b). Arrows indicate stained pyknotic cells. FGF2 or vehicle was microinjected at E15.5 and 22 hr after the microinjection embryos were fixed with 70% ethanol/5% acetic acid and embedded in parafin. TUNEL assay was carried out as described (Gavrieli ef a/., 1992; Thomaidou rt a / . , 1997). Scale bar, 50 pm.

Fig. 7 Effect of delayed FGF2 treatment on the number of neurons and glia in the cerebral cortex. FGF2 (300 ng) was injected in the cerebral ventricles at E20.5 and frozen brain sections were analyzed at P21. Values are obtained from a total of six animals, injected with either PBS ( N = 3) or with FGF2 ( N = 3). Analysis of variance (ANOVA) revealed that FGF2-treated animals were significantly different from controls in total cell number 0,< 0.01; F = 13.3) and in cell density (p c 0.05; F = 6.4). This effect was entirely due to an increase in glial cells, as shown in the graph (Sheffe post hoc test).

6. Fibroblast Growth Factors and Brain Morphogenesis

191

E. Mechanism of Action of FGFs The above observations suggest that FGF2 increases the neuronal output of the PVE. As we have seen, this effect may be produced by at least three separate mechanisms: a shortening of the cell cycle length, an increase in progenitor cell survival, or an increase in the number of divisions of progenitor cells. Cell cycle studies have shown that FGF2 treatment in vivo has no effect on the length of the cell cycle. Nevertheless, these studies have shown that FGF2 induced a 12% increase in the growth fraction, which is the proportion of cells within the population that are able to divide (Vaccarino et ul., 1999). Two mechanisms that may be responsible for this increase in growth fraction induced by FGF2 are summarized in Fig. 4. One is a promotion of DNA synthesis due to reentry of progenitors into S phase. In fact, several lines of evidence suggest that in the absence of mitogenic factors such as FGF2, the default path for neuronal and glial progenitors is to exit the cell cycle, become postmitotic, and differentiate (Raffet al., 1988; Vaccarino et al., 1995). A second possibility is an enhancement of cell survival of progenitor cells. Cells undergo programmed cell death, occurring sometime during the GI phase of the cell cycle. The rapid clearance (on the order of a few hours) and engulfment of cell debris by macrophages may mask the importance of this phenomenon (Thomaidou et ul., 1997). Studies have emphasized that progenitor cell death may somehow be related to the defective replication of a cell’s DNA. If indeed FGFs or other growth factors are able to modify the magnitude of this phenomenon, a decrease in cell death may affect the growth fraction and generate data quite similar to those obtained after FGF2 microinjection. The resolution of this problem resides in the correct and reliable identification of dying (apoptotic) cells. Using the TUNEL method (terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling) to label the free end of fragmented DNA in apoptotic cells, we have been unable to identify reliable differences in the number of dying cells between control and FGF-treated embryos 22-72 hr after in vivo FGF treatment (Fig. 6; see color plate). In sum, our data fail to provide any evidence suggesting an effect of FGF2 either on the length of the cell cycle or on apoptosis. The first possibility referred to above is that the effect of FGF2 is to increase the growth fraction by increasing the number of cell cycles of cortical progenitors. These additional rounds of cell division may also prolong the duration of neurogenesis. Despite a large volume of data on the molecules that regulate the cell cycle, little is known about how FGF2 might increase cell divisions. The G I to S restriction is a major “checkpoint” of the cell cycle, regulated by two major classes of cyclin-dependent kinases (cdks), cdk2/cdk4, that complex with cyclins E and D (Ross, 1996). This phase is also regulated by an array of cdk inhibitors such as P15, P21, and P19. However, how growth factors such as FGFs impact on these dynamics is not well understood. Possible mechanisms for influencing the activity of the cdk/

192

Flora M. Vaccarino et al.

cyclin complexes include changing cyclin content or cyclin activity (Koff et al., 1993), changing the expression of cdk inhibitors (Hannon and Beach, 1994), or affecting the phosphorylation of cdk targets such as the tumor suppressor gene retinoblastoma (Rb) (Takauwa et al., 1993). It should be also considered that FGFs may exert a direct or indirect action on DNA-binding transcription factors, which in turn affect cell proliferation. In vitro and in vivo studies suggest that FGF signaling induces homeobox genes (Robe1 et al., 1995; Partanen et al., 1998). It would be interesting to test this hypothesis at the molecular level.

V. Effect of FGFs on Cell Fate and Cortical Lineages A. Neuronal and Clial Progenitor Cells Are Specified Early within the Cortical Neuroepithelium

It is generally thought that both environmental and genetic factors influence cell fate in the cerebral cortex. Progenitors of the dorsal telencephalic neuroepithelium progressively restrict their fate into separate lineages. Neuronal progenitors remain within the PVE (the cell population adjacent to the ventricles) and glial precursor cells settle into the SPP, away from the ventricular surface. Cells of the PVE stop dividing by the end of gestation in rat and mouse, whereas the production of glial cells continues through the early postnatal period. Each of these two progenitor populations is limited to producing cells of their respective lineages. The presence of multipotential progenitors in vivo is controversial, because they have not been detected by retrovirus-mediated lineage analysis (Grove et al., 1993; Williams et al., 1991; L u s h et al., 1988, 1993). Cells of the PVE include distinct progenitors for glutamate and GABA neurons, and distinct astrocyte and oligodendrocyte progenitor cells populate the SPP. How do progenitor cells restrict themselves into these lineages in just the right proportion, and how irreversible is the commitment of progenitor cells? Do cells with multipotential fate persist in the cortex until late in development? B. Time-Dependent Action of FCF2 on Different Cell Lineages

FGF2 has been shown to increase the proliferation of multipotential progenitors in culture, giving rise to the enhanced generation of both neurons and glia (Kilpatrick and Bartlett, 1995; Vescovi et al., 1993). However, multipotential progenitors do not appear to be affected by FGF2 treatment in vivo. The microinjection of FGF2 at E15.5 increases the number of neurons but does not affect the

6. Fibroblast Growth Factors and Brain Morphogenesis

193

production of glial cells to any significant extent, even if the animals are examined as adults (Table I). Thus, FGF2 appears to affect neuronal progenitors selectively at this stage of development. Within neuronal progenitors, FGF2 differentially affected neuronal sublineages at E l 5.5. The total number of excitatory neurons containing glutamate (pyramidal neurons) was increased significantly more than inhibitory (nonpyramidal) neurons after FGF treatment (Table I). An identical pattern was observed in primary mixed cultures of early cortical progenitors treated with FGF2 (Vaccarino et al., 1995). In addition to being expressed by neuronal progenitor cells, FGFR1 is expressed by glial progenitor cells both in v i m and in situ, within the subventricular layer (Fig. 5) (Bansal et al., 1996). To investigate whether exogenous FGF2 targets glial-restricted progenitors, and as a consequence affects the genesis of glial cells, we have performed microinjections of recombinant FGF2 at E20.5. At this stage of development, neurogenesis has ended, but glial progenitors are actively proliferating within the secondary proliferative population of the subventricular layer. A single administration of FGF2 elicited a large increase in the total number of glial cells in the absence of an increase in neurons (Fig. 7; see color plate). Thus, FGF2 is able to target glial progenitor cells independently of its effect on neuronal progenitors. However, this action becomes significant only at stages of development when substantial populations of glial progenitors are present within the SPP. These data virtually exclude the possibility that FGF2 targets bipotential glialheuronal progenitors in vivo.

C. The lack of FGF2 Decreases the Number of Neurons and Glia in the Cerebral Cortex

The coordinated action of exogenous FGF2 on neuronal and glial precursor cells suggests that FGF2 may intervene in the mechanisms that couple the number of neurons with the number of glia during development. This idea is confirmed by the phenotype of FGF2 knockout mice, in which the FGF2 gene is inactivated by homologous recombination. Mice lacking the FGF2 gene had an approximately 50% decrease in the total number of cells in the cerebral cortex as compared to wild-type littermates (Table 11), suggesting that FGF2 plays a role during normal cortical development. This decrease is due to a lack of approximately equal numbers of neurons and glia. No qualitative change in cortical cytoarchitecture is evident in FGF2-deficient mice, except the decrease in cortical cell density (Fig. 8). Preliminary studies suggest that these changes emerge very early during embryonic development and may be due to an altered proliferation of progenitor cells. In conclusion, these data, combined with the effect of exogenous FGF2, strongly suggest that FGF2 plays a critical role in the regulation of the number of both neurons and glia within the cerebral cortex.

194

Flora M. Vaccarino et al.

Table I1 Cell Number in the Cerebral Cortex of FGF2 Null Mutant Micea Total number of cells (lo6) Cell type Neurons Glia NeuN-positive cells S- 100-positive cells

FGF2+‘+ 16.3 f 0.9 6.30 i 0.7 7.07 i 0.08 8.77 i 0.61

FGF~-/8.57 i 0.3**** 3.95 f 0.4* 5.38 i 0.33** 5.41 f 0.34***

Percentage of control 52 63 76 62

Note: * p < 0.05; **p < 0.01; ***p< 0.001; ****p < 0.0001.

5erial sections from the brain of FGF2-’‘ and FGF2+/+(wild-type) mice (aged 8-17 weeks) were used for stereological analysis as described in the legend of Table I (Vaccarino et al., 1999). Total number of neurons and glia were first estimated in cresyl violet-stained sections (N = 6 animals per group). In these samples, FGF2-I- mice were significantly different from FGF2+/+mice in total cell number ( p < 0.0001; F = 63.0) and in cell density ( p < 0.0001; F = 40.0). A subset of the previous mice ( N = 3 per group) was then analyzed for the number of NeuN- and S-100-positivecells in frozen sections processed free-floating for immunocytochemistry (Vaccarino et al., 1999). FGF2-Imice were again significantly different from FGF2+/+ mice in total cell number ( p < 0.001; F = 41.9) and in cell density ( p < 0.001; F = 26.7) in these immunostained sections (ANOVA). Post hoc pairwise comparisons revealed significant effects of genotype in total number and density of both neurons and glia as indicated (*, Sheffe post hoc test).

VI. Conclusions and Future Prospects The combined analysis of FGFZtreated and FGF2-deficient animals leads us to propose that FGF2 does not affect cell fate during cortical development. The action of FGF2 is thus different than that of FGF8, which plays an important role in cell specification within areas of the brain stem. The increase in neuron number due to microinjected FGF2 is not due to a switch in lineage (from glial to neuronal), because the total number of glia was not decreased by FGF treatment. Similarly, in FGF2-’- knockout mice, the total number of both neurons and glia was reduced in FGF2 null mutants compared to wild-type mice. Thus, it seems that FGF exerts a proliferative action within both neuronal and glial lineages at the appropriate developmental stages. In conclusion, FGF2 appears to control the number of divisions of neuronal and glial progenitors in a temporally restricted fashion. FGF2 independently affects neuronal and glial progenitors, yet these actions may be coordinated, because the lack of FGF2 affects the number of neurons and glia to an equal extent. Within neuronal cells, FGF2 preferentially targets glutamate rather than GABA progenitors. Future experiments will further explore these data at the molecular level. Important areas to investigate are the differential expression of FGF receptor or other FGF-triggered signaling molecules among distinct progenitor cells, and the genes that are linked to the FGF pathway in these progenitors.

6. Fibroblast Growth Factors and Brain Morphogenesis

195

Fig. 8 Decrease in cell number in the cerebral cortex of FGF2 null mutant mice. Cresyl violet-stained sections showing the frontal and the occipital cortex of wild-type (wt) and FGF2 null mutant mice (ko) at 8-17 weeks. Note the decrease in density of cortical cells. Quantification of total cell number by stereology at high magnification revealed that both neurons and glia are decreased in FGF2 knockout mice (Vaccarino et al., 1999). Scale bar, 100 p n .

References Acampora, D., Mazan, S., Lallemand, Y., Avantaggiato, V., Maury, M., Simeone, A., and Brulet, P. (1995). Forebrain and midbrain regions are deleted in Otx-’- mutants due to a defective anterior neuroectoderm specification during gastrulation. Development 121,3279-3290. Acampora, D., Mazan, S., Avantaggio, V., Barone, P.,Tuorto, F., Lallemand, Y., Brulet, P.,and Simeone, A. (1996). Epilepsy and brain abnormalitiesin mice lacking the Otxl gene. Nature Genet. 14,218-222. Amalric, F., Bouche, G., Bonnet, H., Brethenou, P., Roman, A. M., Truchet, I., and Quarto, N. (1994). Fibroblast growth factor-2 (FGF-2) in the nucleus: Translocation process and targets. Biochem. P h a m c o l . 47,111-115. Amaya, E., Musci, T.J., and Kirschner, M. W. (1991). Expression of a dominant negative mutant of the FGF receptor disrupts mesoderm formation in Xenopus embryos. Cell 66,257-270. Amaya, E., Stein, P. A., Musci, T. J., and Kirschner, M. W. (1993). FGF signaling in the early specification of mesoderm in Xenopus. Development 118,477-487. Anderson, S. A., Qiu, M., Bulfone, A., Eisenstat, D. D., Meneses, J., Pedersen, R., and Rubenstein, J. L. R. (1997). Mutations of the homeobox genes DLr-1 and DLx-2 disrupt the striatal subventricular zone and differentiationof late born striatal neurons. Neuron 19,l-20. Ang, S.-L., Jin, O., Rhinn, M., Daigle, N., Stevenson, L., and Rossant, J. (1996). A targeted mouse Otx2 mutation leads to severe defects in gastrulation and formation of axial mesoderm and to deletion of rostra1 brain. Development 122,243-252. Baird, A. (1994). Fibroblast growth factors: Activities and significance of nonneurotrophin neurotropic growth factors. Cur,: Opin. Neurobiol. 4,78-86. Bansal, R., Kumar, M., Murray, K., Morrison, R., and Pfeiffer, S. E. (1996). Regulation of FGF receptors in the oligodendrocytelineages. Mol. Cell. Neurosci. 7,263-275. Basilico, C., and Moscatelli, D. (1992). The FGF family of growth factors and oncogenes. Adv. CancerRes. 59, 115-165.

196

Flora M. Vaccarino et al.

Bohen, P., Esch, F., Baird, A,, Jones, K. L., and Gospodarowicz, D. (1985). Human brain fibroblast growth factor. Isolation and partial chemical characterization.FEBS Lett. 185,177-181. Burgess,W. H., and Maciag, T. (1989). The heparin-binding (fibroblast) growth factor family of proteins. Annu. Rev. Biochem. 58,575-606. Callaerts, P., Halder, G., and Gehring, W. J. (1997). PAX-6 in development and evolution. Annu. Rev. Neurosci. 20,483-532. Capecchi, M. R. (1997). “The Role of Hox genes in Hindbrain Development.” Oxford Univ. Press, London. Capovilla, M., Brandt, M., and Botas, J. (1994). Direct regulation of decapentuplegic by Ultrabithorax and its role in Drosophila midgut morphogenesis. Cell 76,461-475. Cattaneo, E., and McKay, R. D. G. (1990). Proliferation and differentiationof neuronal stem cells regulated by nerve growth factor. Nature (London) 347,762-765. Cohn, M. J., Izpisua-Belmonte, J.-C.,Abud, H., Heath, J. K., andTickle, C. (1995). Fibroblast growthfactors induce additional limb development from the flank of chick embryos. Cell 80,739-746. Coulier, F., Pontarotti, P., Roubin, R., Hartung, H., Goldfarb, M., and Birnbaum, D. (1997). Of worms and men: An evolutionary perspective on the fibroblast growth factor (FGF) and FGF receptor families. J. Mol. Evol. 44,43-56. Cox, W. G. and Hemmati-Brivanlou, A. (1995). Caudalization of neural fate by tissue recombination and bFGF. Development 121,4349-4358. Crossley, P. H., and Martin, G. R. (1995). The mouse F@ gene encodes a family of polypeptides and is expressed in regions that direct outgrowth and patterning in the developing embryo. Development 121,439-451. Crossley, P. H., Martinez, S., and Martin, G.R. (1996). Midbrain development induced by FGF8 in the chick embryo. Nature (London) 380,66-68. DeHamer, M. K., Guevara, J. L., Hannon, K., Olwin, B. B., and Calof, A. L. (1994). Genesis of olfactory receptor neurons in vitro: Regulation of progenitor cell divisions by fibroblast growth factor. Neuron 13,1083-1097. Eckenstein, F. P. (1994). Fibroblast growth factors in the nervous system. J. Neurobiol. 25, 14671480. Edelman, G. M., and Jones, F. S. (1993). Outside and downstream of the homeobox. J. Biol. Chem. 268,20683-20686. Engele, J., and Churchill Bohn, M. (1992). Effects of acidic and basic fibroblast growth factors (aFGF, bFGF) on glial precursor cell proliferation:Age dependency and brain region specificity. Dev. Biol. 152,363-372. Ernfors, P., Lonnerberg, P., Ayer, L. L. C., and Persson, H. (1990). Developmental and regional expression of basic fibroblast growth factor mRNA in the rat central nervous system. J. Neurosci. Res. 27,lO-15. Frantz, G . D., and McConnell, S. K. (1996). Restriction of late cerebral cortical progenitors to an upper-layer fate. Neumn 17,55-61. Frantz, G. D., Weimann, J. M., Levin, M. E., and McConnell, S. K. (1994). Otxl and Otx2 define layers and regions in developing cerebral cortex and cerebellum.J. Neurosci. 14,5725-5740. Gavrieli,Y.,Sherman,Y., and Ben-Sasson, S. A. (1992). Identification of programmed cell death in situ via specific labelling of nuclear fragmentation.J. Cell Biol. 119,493-501. Gehring, W. J., Qian, Y.Q., Billeter, M., Furukubo-Tokunaga, K., Schier, A. S., Resendez-Perez, D., Affolter, M., Otting, G., and Wuthrich, K. (1994). Homeodomain-DNA recognition. Cell 78,211223. Gensburger, C., Labourdette, G., and Sensenbrenner, M. (1987). Brain basic fibroblast growth factor stimulates the proliferation of rat neuronal precursor cells in vitro. FEBS Lett. 217, 1-5. Giordano, S., Sherman, L., and Morrison, R. (1991). Multiple molecular weight forms of basic fibroblast growth factor are developmentally regulated in the rat central nervous system. Ann. N.l! Acad. Sci. 638,420-423.

6. Fibroblast Growth Factors and Brain Morphogenesis

197

Goh, K. C., Lim, Y.P., Ong, S. H., Siak, C. b., Cao, X., Tan, Y.H., and Guy, G. R. (1996). Identification of p90, a prominent tyrosine-phosphorylatedprotein in fibroblast growth factor-stimulated cells, as 80K-H. J. Biol. Chem. 271,5832-5838. Gomez-Pinilla, F., Knauer, D. J., and Nieto-Sampedro, M. (1988). Epidermal growth factor receptor immunoreactivityin rat brain. Development and cellular localization.Bruin Res. 438,385-390. Gonzalez,A. M., Buscaglia, M., Ong, M., and Baird, A. (1990). Distribution of basic fibroblast growth factor in the 18-day rat fetus. J. Cell Biol. 110,753-765. Gonzalez,A.-M., Berry, M., Maher, P. A,, Logan, A., and Baird, A. (1995). Acomprehensive analysis of the distribution of FGF-2 and FGF-I in the rat brain. Bruin Res. 701,201-226. Grove, E. A., Williams, B. P., Li, Q.-D., Hajihosseini, M., Friedrich, A., and Price, J. (1993). Multiple restricted lineages in the embryonic rat cerebral cortex. Development 117,553-561. Hannon, G. J., and Beach, D. (1994). ~ 1 5 ' " 'is~a~potential ~ effector of TGF-P-induced cell cycle arrest. Nature (London) 371,257-261. Hebert, J. M., Basilico, C., Goldfarb, M., Haub, O., and Martin, G. R. (1990). Isolation of cDNAs encoding four mouse FGF family members and characterizationof their expression patterns during embryogenesis.Dev. Biol. 138,454-463. Huang, J., Mohammadi, M., Rodrigues, G. A,, and Schlessinger,J. (1995). Reduced activation of RAF-I and MAP kinase by a fibroblast growth factor receptor mutant deficient in stimulation of phosphatidylinositolhydrolysis. J. Biol. Chem. 270,5065-5072. Kengaku, M., and Okamoto, H. (1995). bFGF as a possible morphogen for the anteroposterioraxis of the central nervous system in Xenopus. Development 121,3121-3130. Kenisberg, R. L., Mazzoni, I. E., Collier, B., and Cuello, A. C. (1992). Epidermal growth factor affects both glia and cholinergic neurons in septa1 cell cultures. Neuroscience 50,85-97. Kilpatrick, T. J., and Bartlett, P. F. (1995). Cloned multipotentialprecursors from the mouse cerebrum require FGF-2, whereas glial restricted precursors are stimulated with either FGF-2 or EGF. J. Neurosci. 15,3653-3661. Kim, H.-J., Rice, D. P. C., Kettunen, P. J., and Thesleff, I. (1998). FGF-, BMP- and Shh-mediated signalling pathways in the regulation of cranial suture morphogenesis and calvarial bone development. Development 125, 1241-1251. Koff, A,, Ohtsuki, M., Polyak, K., Roberts, J. M., and Massague, J. (1993). Negative regulation of G1 in mammalian cells: Inhibition of cyclin E-dependent kinase by TGF-P. Science 260,536539. Kornack, D. R., and Rakic, P. (1995). Radial and horizontal deployment of clonally related cells in the primate neocortex: Relationship to distinct mitotic lineages. Neuron 15,311-321. Kouhara, H., Hadari, Y. R., Spivak-Kroizman,T., Schilling, J., Bar-Sagi, D., Lax, I., and Schlessinger, J. (1997). A lipid-anchored Grb2-bindingprotein that links FGF-receptor activation to the Ras/MAPK signaling pathway. Cell 89,693-702. Krumlauf, R. (1994). Hox genes and vertebrate development. Cell 78,191-201. Lamb, T.M., and Harland, R. M. (1995). Fibroblast growth factor is a direct neural inducer, which combined with noggin generates anterior-posterior neural pattern. Development 121,3627-3636. Lawrence, P. A,, and Struhl, G. (1996). Morphogens, compartments, and pattern: Lessons from Drosophila? Cell 8 5 9 5 1-961. Lecuit, T., Brook, W. J., Ng, M., Calleja, M., Sun, H., and Cohen, S. M. (1996). Two distinct mechanisms for long-range patterning by Decapentaplegic in the Drosophila wing. Nature 381, 387394. Lee, S. M. K., Danielian, P. S., Fritzsch, B., and McMahon, A. P. (1997). Evidence that FGF8 signaling from the midbrain-hindbrain junction regulates growth and polarity in the developing midbrain. Development 124,959-969. Lumsden, A., and Krumlauf, R. (1996). Patterning the vertebrate neuraxis. Science 274,1109-1115. Luskin, M. B., Pearlman, A. L., and Sanes, J. R. (1988). Cell lineage in the cerebral cortex of the mouse studied in vivo and in vitro with a recombinant virus. Neuron 1,635-647.

198

Flora M. Vaccarino et al.

Luskin, M. B., Parnavelas, J. G., and Barfield, J. A. (1993). Neurons, astrocytes, and oligodendrocytes of the rat cerebral cortex originate from separate progenitor cells: An ultrastructural analysis of clonally related cells. J. Neumsci. 13, 1730-1750. MacNicol,A. M., Muslin, A. J., and Williams, L. T. (1993). Raf-1 kinase is essential for early Xenopus development and mediates the induction of mesoderm by FGF. Cell 73,571-583. Mahmood, R., Kiefer, P., Guthrie, S., Dickson, C., and Mason, I. (1995). Multiple roles of FGF3 during cranial neural development in the chicken. Development 121, 1399-1410. Mason, I. J., Fuller-Pace, F., Smith, R., and Dickson, C. (1994). FGF-7 (keratinocyte growth factor) expression during mouse development suggests roles in myogenesis, forebrain regionalization and epithelial-mesenchymal interactions.Mech. Dev. 45, 15-30. McConnell, S. K., and Kaznowski, C. E. (1991). Cell cycle dependence of laminar determination in developing neocortex. Science 254,282-285. McFarlane, S., Cornel, E., Amaya, E., and Holt, C. E. (1996). Inhibition of FGF receptor activity in retinal ganglion cell axons causes emors in target recognition. Neuron 17,245-254. Mehta, V. B., Connors, L., Wang, H . 4 . R., and Ciu, I. M. (1998). Fibroblast variants nonresponsive to fibroblast growth factor 1 are defective in its nuclear translocation. J. Biol. Chem. 273,41974205. Meinhardt, H. (1983). Cell determination boundaries as organizing regions for secondary embryonic fields. Dev. Biol. 96,375-385. Meyers, E. N., Lewandoski, M., and Martin, G. R. (1998). An Fgfs mutant allelic series generated by Cre- and Flp-mediated recombination.Nature Genet. 18, 136-141. Morshead, C. M., Reynolds, B. A., Craig, C. G., McBurney, M. W., Staines, W. A., Morassutti, D., Weiss, S., and van der Koy, D. (1994). Neural stem cells in the adult mammalian forebrain: A relatively quiescent subpopulationof subependimal cells. Neuron 13, 1071-1082. Murphy, M., Drago, J., and Bartlett, P. F. (1990). Fibroblast growth factor stimulates the proliferation and differentiation of neural precursor cells in vitro. J. Neumsci. Res. 25,463-475. Murphy, M., Reid, K., Ford, M., Furness, J. B., and Bartlett P. F. (1994). FGF2 regulates proliferation of neural crest cells, with subsequent neuronal differentiation regulated by LIF or related factors. Development 120,3519-3528. Niswander, L., and Martin, G. R. (1993). Mixed signals from the AER: FGF-4 and Bmp-2 have opposite effects on limb growth. Prog. Clin. B i d . Res. 383B, 625-633. Nurcombe, V., Ford, M. D., Wildschut, J. A., and Bartlett, P. F. (1993). Developmental regulation of neural response to FGF-I and FGF-2 by heparan sulfate proteoglycan. Science 260, 103-106. Ong, S. H., Goh, K. C., Lim, Y. P., Low, B. C., Klint, P.,Claesson-Welsh,L., Cao, X., Tan, Y., and Guy, G. R. (1996). SUC1-associatedneurotrophic factor target (SNT) protein is a major FGFstimulated tyrosine phosphorylated90-kDa protein which binds the SH2 domain of GRB2. Biochem. Biophys. Res. Commun. 225,1021-1026. Partanen, J., Vainikka, S., and Alitalo, K. (1993). Structural and functional specificityof FGF receptors. Philos. Trans. R. SOC. Lonu'. Set B: Biol. Sci. 340,297-303. Partanen, J., Schwartz, L., and Rossaur, J. (1998). Opposite phenotypes of hypomorphic and Y766 phosphorylationsite mutations reveal a function for FGFRI in anteroposterior patterning of mouse mesoderm. Genes Dev. 12,2332-2344. Peeper, D. S., Upton, T. M., Ladha, M. H., Neuman, E., Zalvide, J., Barnards, R., DeCaprio, J. A., and Ewen, M. E. (1997). Ras signaling linked to the cell-cycle machinery by the retinoblastoma protein. Nature (London)386, 177- 181. Powell, P.P., Finklestein, S. P., Dionne, C. A., Jaye, M., and Klagsbrun, M. (1991). Temporal, differential and regional expression of mRNA for basic fibroblast growth factor in the developing and adult rat brain. Bruin Res. Mol. Bruin Res. 11,71-77. Pownall, M. E., Tucker, A. S., Slack, J. M. W., and Isaacs, H. V. (1996). eFGE Xcad3 and Hox genes form a molecular pathway that establishes the anteroposterior axis in Xenopus. Development 122, 3881-3892.

6. Fibroblast Growth Factors and Brain Morphogenesis

199

Prudovsky, I., Savion, N., Zhan, X., Friesel, R., Xu, J., Hou, J., McKeehan, W. L.,and Maciag, T. (1994). Intact and functional fibroblast growth factor (FGF) receptor-1 trafficks near the nucleus in response to FGF-1. J. Biol. Chem. 269,31720-31724. Rabin, S. J., Cleghon, V., and Kaplan, D. R. (1993). SNT, a differentiation-specifictarget of neurotrophic factor-induced tyrosine kinase activity in neurons and PC12 cells. Mol. Cell. Eiol. 13, 2203-2213. Raff, M. C., Lillien, L. E., Richardson, W. D., Burne, J. F., and Noble, M. D. (1988). Platelet-derived growth factor from astrocytes drives the clock that times oligodendrocytedevelopment in culture. Nature (London) 333,562-565. Ray, J., Peterson, D. A., Schinstine, M., and Gage, F. H. (1993). Proliferation, differentiation, and long-term culture of primary hippocampal neurons. Proc. Nat. Acad. Sci. U.S.A. 90,3602-3606. Reilly, K. M., and Melton, D. A. (1996). Short-range signaling by candidate morphogens of the TGFP family and evidence for a relay mechanism of induction. Cell 86,743-754. Reynolds, B. A., Tetzlaff, W., and Weiss, S. (1992). A multipotent EGF-responsive striatal embryonic progenitor cells produces neurons and astrocytes. J. Neurosci. 12,4565-4574. Rhinn, M., Dierich,A., Shawlot, W., Behringer, R. R., LeMeur, M., and Ang, S.-L. (1998). Sequential roles for Otx2 in visceral endoderm and neuroectoderm for forebrain and midbrain induction and specification. Development 125,845-856. Robel, L., Ding, M., James, A. J., Lin, X., Leckman, J. F., and Vaccarino, F. M. (1995). Fibroblast growth factor 2 increases Otn2 expression in precursor cells from the mammalian telencephalon. J. Neumsci. 15,7879-7891. Robinson, M. L., MacMillan-Crow, L. A., Thompson, J. A., and Overbeek, P. A. (1995). Expression of a truncated FGF receptor results in defective lens development in transgenic mice. Development 121,3959-3967. Ross, E. M. (1996). Cell division and the nervous system: Regulating the cycle from neural differentiation to death. Trends Neurosci. 19,62-68. Rubenstein, J. L. R., and Shimamura, K. (1997). Regulation of patterning and differentiation in the embryonic vertebrate forebrain. In “Molecular and Cellular Approaches to Neural Development” (W. M. Cowan, T. M. Jessel, and S. I. Zipursky, Eds.). Oxford Univ. Press, London. Saffel, J. L., Williams, E. J., Mason, I. J., Walsh, F. S., and Doherty, P. (1997). Expression of a dominant negative FGF receptor inhibits axonal growth and FGF receptor phosphorylation stimulated by CAMS.Neuron 18,231-242. Shimamura, K., and Rubenstein, J. L. R. (1997). Inductive interactions direct early regionalization of the mouse forebrain. Development 124,2709-2718. Simeone, A,, Acampora, D., Gulisano, M., Stomaiuolo,A., and Boncinelli, E. (1992a). Nested expression domains of four homeobox genes in developing rostral brain. Nature (London) 358, 687-690. Simeone, A., Gulisano, M., Acampora, D., Stomaiuolo, A., Rambaldi, M., and Boncinelli, E. (1992b). Two vertebrate homeobox genes related to the Dmsophila empty spiracles gene are expressed in the embryonic cerebral cortex. EMEO J. 11,2541-2550. Simeone, A., D’ Apice, M.-R., Casanova, J., Graziani, F., Acampora, D., Nigro, V., and Avantaggiato, V. (1994). Orthopedia, a novel homeobox-containinggene expressed in the developing central nervous system of both mouse and Dmsophila. Neuron 13,83-101. Smallwood, P. M., Munoz-Sanjuan,I., Tong, P., Macke, J. P., Hendry, S. H. C., Gilbert, D. J., Copeland, N. G., Jenkins, N. A., and Nathans, J. (1996). Fibroblast growth factor (FGF) homologous factors: New members of the FGF family implicated in nervous system development. Proc. Nat. Acad. Sci. U.S.A. 93,9850-9857. Stock, A,, Kuzis, K., Woodward, W. R., Nishi, R., and Eckenstein, F. P. (1992). Localization of acidic fibroblast growth factor in specific subcortical neuronal populations. J. Neumsci. 12, 4688-4700. Stoykova,A,, Fritsch, R., Walther, C., and Gruss, P. (1996). Forebrain patterning defects in small eye

200

Flora M. Vaccarino et al.

mutant mice. Development 122,3453-3465. Stoykova, A., Gotz, M., Gruss, P., and Price, J. (1997). Pax6-dependent regulation of adhesive patterning, R-cadherin expression and boundary formation in developing forebrain. Development 124,3765-3777. Takahashi, T., Nowakowski, R. S., and Caviness, V. S., Jr. (1994). Mode of cell proliferation in the developing mouse neocortex. Proc. Natl. Acad. Sci. U.S.A. 91,373-379. Takahashi, T., Nowakowski, R. S.,and Caviness, V. S., Jr. (1995a). The cell cycle of the pseudostratified ventricular epithelium of the embryonic murine cerebral wall. J. Neurosci. 15,6046-6057. Takahashi, T., Nowakowski, R. S., and Caviness, V. S., Jr. (1995b). Early ontogeny of the secondary proliferative population of the embryonic murine cerebral wall. J. Neurosci. 15,6058-6068. Takahashi, T., Nowakowski, R. S., and Caviness, V. S. (1996). The leaving or Q fraction of the murine cerebral proliferative epithelium:A general model of neocortical neuronogenesis. J. Neurosci. 16,6183-6196. Takauwa, N., Zhou, W., Kumada, M., and Takuwa, Y.(1993). CA2-dependent stimulation of retinoblastoma gene product phosphorylation and p34cdc2 kinase activation in serum-stimulated human fibroblasts. J. Biol. Chem. 268, 138-145. Thomaidou, D., Mione, M. C., Cavanagh, J. F. R., and Parnavelas, J. G.(1997). Apoptosis and its relation to the cell cycle in the developing cerebral cortex. J. Neurosci. 17,1075-1085. Vaccarino, F. M., Schwartz, M. L., Hartigan, D., and Leckman, J. F. (1995). Effect of basic fibroblast growth factor on the genesis of excitatory and inhibitory neurons in primary cultures of cells from the mammalian telencephalon. Cereb. Cortex 1, 1047-321 1. Vaccarino, F. M., Raballo, R., Nilsen, J., Hung, E. E., Wyland, J. J., and Schwartz, M. L. (1999). Changes in cerebral cortex size are governed by fibroblast growth factor during embryogenesis. Nature Neurosci. 1,246-253. Vescovi, A. L., Reynolds, B. A., Fraser, D. D., and Weiss, S. (1993). bFGF regulates the proliferative fate of unipotent (neuronal) and bipotent (neuronallastroglial)EGF-generated CNS progenitor cells. Neuron 11,951-966. Weinstein, D. C., Marden, J., Carnevali, F.,and Hemmati-Brivanlou, A. (1998). FGF-mediated mesoderm induction involves the Src-family kinases Laloo. Nature (London) 394,904-908. Weise, B., Janet, T., and Grothe, C. (1993). Localizationof bFGF and FGF-receptorin the developing nervous system of the embryonic and newborn rat. J. Neumsci. Res. 34,442-453. William, E. J., Fumess, J., Walsh, F. S., and Doherty, P. (1994).Activation of the FGF receptor underlies neurite outgrowth stimulated by L1, N-CAM, and N-cadherin. Neuron 13,583-594. Williams, B. P., Read, J., and Price, J. (1991). The generation of neurons and oligodendrocytesfrom a common progenitor cell. Neuron 7,685-693. Williams, B. P., Park, J. K., Alberta, J. A., Muhlebach, S. G.,Hwang, G.Y.,Roberts, T. M., and Stiles, C. D. (1997). A PDGF-regulated immediate early gene response initiates neuronal differentiation in ventricular zone progenitor cells. Neuron 18,553-562. Woodward, W. R., Nishi, R., Meshul, C. K., Williams, T. E., Coulombe, M., and Eckenstein, F. P. (1992). Nuclear and cytoplasmic localization of basic fibroblast growth factor in astrocytes and CA2 hippocampal neurons. J. Neurosci. 12,142-152. Yang, Y., and Niswander, L. (1995). Interaction between the signaling molecules WNT7a and SSH during vertebrate limb development: Dorsal signals regulate anteroposteriorpatterning. Cell 80, 939-947. Zhan, X., Hu, X., Friesel, R.,and Maciag, T. (1993). Long term growth factor exposure and differential tyrosine phosphorylation are required for DNA synthesis in Balbk3T3 cells. J. Biol. Chem. 268,961 1-9620.

Index

A Amnion, formation, 6 7 4 8 Amphibians, gastrula anterior margin, 84-85 Annelids, unequal cleavage pattern generation, 17-20 first cleavage, 18-20 micromere formation, 20 second cleavage, 20 Arabidopsis, endo-1,4-P-D-glUCanaSe genes, 39-58 cell growth relations, 54-55 cellulose interactions, 4 2 4 3 expression, 52-53 molecular characterization, 45-52 mutant genes, 55-56 overview, 3941,57-58 types, 4 3 4 5 Ascidians, unequal cleavage pattern generation, 11-17 centrosome-attractingbody, 11-16 posterior egg cytoplasm role, 16-17 Aves, gastrula anterior margin, 81-86 axial inversion, 84 hypoblast, 82-83 Koller’s sickle, 81-82 primordial germ cells, 84 Axial inversion, bird-mammal pre-streak-stage embryos compared, 84

epiboly, 124-126, 128-129 fusion regulation, 119-121, 128 overview, 105-108, 126-129 syncytial yolk formation, 114-119, 127-128 Brain development, fibroblast growth factor role, 179-195 cell fate, 192-194 glial progenitor cells, 192-194 neuronal progenitor cells, 192-194 time-dependent actions, 192-193 central nervous system morphogenesis, 181-1 83 cell interaction regulation, 181-182 compartment specification, 181 morphogens, 182-183 pattern formation, 182-183 developmental control genes, 181-183 fibroblast growth factor ligands, 183-184 fibroblast growth factor receptors, 184-185 future research directions, 194-195 neuroepithelium patterning regulation, 185-192 action mechanisms, 191-192 cerebral cortex size regulation, 189-190 dorsal telencephalic compartment growth kinetics, 188-189 regional specification, 185-1 88 overview, 179-181, 194-195

B

C

Birds, see Aves Blastocyst, morphology, 65-66 Blastomeres, non-D quadrants in leech, 105-129 cleavage, 127 epiboly, 121-126 epithelium embryo reduction, 123-1 24 germinal band deficient embryos, 122 macromere role, 124-126, 128-129 macromeres cleavage behavior, 108-1 14, 127

Caenorhabditis elegans, par gene mutants, unique cleavage pattern generation, 20-3 1 asymmetric cleavage, 21-23 par gene characteristics, 23-31 Cell division cleavage patterns, see Cleavage single-layered epithelium, 5 Cell fate, cortical lineages, fibroblast growth factor role, 192-194 glial progenitor cells, 192-194 neuronal progenitor cells, 192-194

201

202 Cell fate (con?.) time-dependentactions, 192-193 Cell growth, endo-1,4-P-D-g1UCanaSeinteractions, 54-55 Cellulase genes, Arabidopsis, 39-58 cell growth relations, 54-55 cellulose interactions, 4243 expression, 52-53 molecular characterization, 45-52 mutant genes, 55-56 overview, 3941,57-58 types, 4345 Cellulose, endo-1.4-P-D-glUCanaSegenes interactions, Arabidopsis, 42-43 Central nervous system, morphogenesis,developmental control genes, 181-183 cell interaction regulation, 181-182 compartment specification, 181 morphogens, 182-183 pattern formation, 182-183 Centromeres, sperm nuclear activation during fertilization, 140-14 1 Centrosome,cleavage plane positioning role, 3 4 Centrosome-attractingbody, unequal cleavage pattern, ascidian embryos, 11-16 Chromatin, sperm nuclear activation during fertilization asynchronousbehavior, paternal-maternal patterns compared, 161-164 nucleus characteristics chromatin fibers, 139-140 nucleoprotaminesperm chromatin, 139 nucleosomal sperm chromatin, 139 somatic chromatin, 138-139 sperm chromatin, 134-135 sperm nucleus to male pronucleus transformation nucleoprotaminesperm chromatin, 147-149, 156 nucleosomal sperm chromatin, 146-147, 155-156 pronuclear chromatin structure, 155-156 Cleavage macromere behavior in leech, non-D quadrants, 108-114, 127 unique pattern generation, 1-32 cleavage plane positioning, 3-5 centrosome, 3 4 furrow formation, 3-4 mitotic apparatus, 3-4

Index orthogonal pattern, 4-5 single-layered epithelial cell division, 5 extrinsic cues, 3 intrinsic cues, 3 invariant pattern role, 2-3 micromere formation in sea urchin embryos, 5-8 first cleavage, 7-8 maternal factors, 5 4 maternally localized substance, 7-8 vegetal cortex, 6-7 overview, 2-3,31-32 par mutants in nematode embryos, 20-31 asymmetric cleavage, 21-23 par gene characteristics, 23-31 spiral cleavage in gastropod embryos, 8-1 1 cleavage clock, 1 1 description, 8-9 handedness determination, 9 maternal factor role, 9 mechanisms, 10 unequal cleavage in annelid embryos, 17-20 first cleavage, 18-20 micromere formation, 20 second cleavage, 20 unequal cleavage in ascidian embryos, 11-17 centrosome-attracting body, 11-16 posterior egg cytoplasm role, 16-17 Cytoplasm, posterior egg cytoplasm, unequal cleavage in ascidian embryos, 16-17

D DNA, replication, sperm nucleus to male pronucleus transformation. 159

E Ectoderm, differentiation,78 Eggs cytoplasm sperm nuclear decondensation factors, 157-158 unequal cleavage in ascidian embryos, 16-17 fertilization male pronucleus formation, 157-158 meiotic stages, 142-143 Embryogenesis,cleavage patterns, see Cleavage

Index Embryonic disc, morphology, 69 Endoderm, differentiation, 79-80 Endo-1,4-p-o-glucanasegenes, Arabidopsis, 39-58 cell growth relations, 54-55 cellulose interactions,4 2 4 3 expression, 52-53 molecular characterization,45-52 mutant genes, 55-56 overview, 3941,57-58 types, 4 3 4 5 Epiblast, differentiation, 74 Epiboly, non-D quadrants in leech, 121-126 epithelium embryo reduction, 123-124 germinal band deficient embryos, 122 macromere role, 124-126, 128-129 Epithelium epiboly in leech, 123-124 neuroepithelium pattern regulation, fibroblast growth factor role action mechanisms, 191-192 cerebral cortex size regulation, 189-190 dorsal telencephalic compartment growth kinetics, 188-189 neuronal progenitor cell lineages, 192-194 regional specification, 185-188 single-layered cell division, 5 Evolution, gastrula anterior margin in mammals, 94

F Fertilization infertility treatment, 164-166 sperm nuclear activation, 133-166 chromatin asynchronous behavior, 161-164 egg stages, 142-143 future research directions, 166 nucleus characteristics, 134-142 centromeres, 140-141 chromatin fibers, 139-140 chromosome packaging order, 140 nuclear lamina, 138 nuclear matrix, 135-138 nucleoprotamine sperm chromatin, 139 nucleosomal sperm chromatin, 139 perinuclear matrix, 136-137 pre-fertilization events, 141-142 somatic chromatin, 138-139 sperm chromatin, 134-135 telomeres, 140-141

203 overview, 133-134.166 sperm nucleus to male pronucleus transformation, 143-161 cytoplasmic decondensation factors, 157-158 disulfide bond reduction, 147 DNA replication, 159 histones, 148-149 male pronuclear envelope formation, 153-155 nuclear decondensation, 149-1 52 nuclear envelope removal, 144-146 nuclear lamina, 144-145 nuclear matrix remodeling, 152-1 53 nucleoprotamine sperm chromatin, 147-149, 156 nucleoprotein remodeling, 146-149 nucleosomal sperm chromatin, 146-147, 15.5-156 oocyte competence, 157-158 perinuclear theca, 145-146 pronuclear chromatin structure, 155156 protamine removal, 147-148 protamine replacement, 148-149 transcription activation, 158-1 6 1 Fibroblast growth factor, brain development role, 179-195 cell fate, 192-194 glial progenitor cells, 192-194 neuronal progenitor cells, 192-194 time-dependent actions, 192-193 central nervous system morphogenesis, 181-1 83 cell interaction regulation, 181-1 82 compartment specification, 181 morphogens, 182-183 pattern formation, 182-183 developmental control genes, 181-183 fibroblast growth factor ligands, 183-184 fibroblast growth factor receptors, 184-185 future research directions, 194-195 neuroepithelium patterning regulation, 185-192 action mechanisms, 191-192 cerebral cortex size regulation, 189-190 dorsal telencephalic compartment growth kinetics, 188-189 regional specification, 185-1 88 overview, 179-181, 194-195 Fish, gastrula anterior margin, 85-86

204 G Gastropods, spiral cleavage pattern generation, 8-11 cleavage clock, 11 description, 8-9 handedness determination, 9 maternal factor role, 9 mechanisms, 10 Gastrula anterior margin amphibians, 84-85 birds, 8 1-86 axial inversion, 84 hypoblast, 82-83 Koller’s sickle, 81-82 primordial germ cells, 84 bony fish, 85-86 mammals, 63-94 amnion formation, 67-68 anterior margin, 69-74 blastocyst, 65-66 developmentalchanges, 7 6 8 1 early gastrulation stages, 87-89 ectoderm, 78 embryonic disc, 69 endoderm, 79-80 epiblast differentiation, 74 evolution, 94 extraembryonicmembrane generation, 91-93 gastrulation organization, 93 gastrulation staging, 93-94 gene expression, 87-89 historical perspectives, 74-75 hypoblast differentiation, 74 implantation, 67 late gastrulation stages, 89 lower layer homology, 90 marginal crescent formation, 76-77 mesoderm, 78-79 morphology, 65-75 overview, 64-65,93-94 phylogenetic implications, 89-93 post-crescent formation changes, 77-8 1 prechordal plate, 80-81 size, 68 topography, 77-78 vegetal hemisphere yolk accumulation, 90-9 1 Glial cells, progenitor cell lineages, fibroblast growth factor role, 192-194

Index H Halocynthia roretzi, unequal cleavage pattern generation, 11-17 centrosome-attractingbody, 11-16 posterior egg cytoplasm role, 16-17 Helobdella robusta, blastomere non-D quadrants, 105-129 cleavage, 127 epiboly, 121-126 epithelium embryo reduction, 123-124 germinal band deficient embryos, 122 macromere role, 124-126, 128-129 macromeres cleavage behavior, 108-114, 127 epiboly, 124-126, 128-129 fusion regulation, 119-121, 128 overview, 105-108, 126-129 syncytial yolk formation, 114-119, 127-128 Histones, sperm nucleus to male pronucleus transformation, protamine replacement, 148-149 Hypoblast, differentiation birds, 82-83 mammals, 74

I Infertility, treatment, technological advances, 164-166

K Koller’s sickle, bird gastrula anterior margin, 81-82

L Leech, blastomere non-D quadrants, 105-129 cleavage, 127 epiboly, 121-126 epithelium embryo reduction, 123-124 germinal band deficient embryos, 122 macromere role, 124-126, 128-129 macromeres cleavage behavior, 108-114, 127 epiboly, 124-126, 128-129 fusion regulation, 119-121, 128 overview, 105-108, 126-129 syncytial yolk formation, 114-119, 127-128 Ligands, see specific types

Index Lineages, see Cell fate Lymnaea, spiral cleavage pattern generation, 8-1 1 cleavage clock, 11 description, 8-9 handedness determination,9 maternal factor role, 9 mechanisms, 10

M Macromeres, blastomere non-D quadrants in leech cleavage behavior, 108-114, 127 epiboly, 124-126, 128-129 fusion regulation, 119-121, 128 Male pronucleus, sperm nucleus transformation, 143-1 61 cytoplasmic decondensation factors, 157158 DNA replication, 159 male pronuclear envelope formation, 153-155 nuclear decondensation, 149-152 kinetics, 150-152 morphology, 149-150 nuclear envelope removal, 144-146 nuclear lamina, 144-145 pennuclear theca, 145-146 nuclear matrix remodeling, 152-153 nucleoprotein remodeling disulfide bond reduction, 147 histones, 148-149 nucleoprotamine sperm chromatin, 147-149,156 nucleosomal sperm chromatin, 146-147, 155-156 protamine removal, 147-148 protamine replacement, 148-149 oocyte competence, 157-158 pronuclear chromatin structure, 155-156 transcription activation, 158-161 Mammals, gastrula anterior margin, 63-94 developmental changes, 76-8 1 ectoderm, 78 endoderm, 79-80 marginal crescent formation, 7 6 7 7 mesoderm, 78-79 post-crescent formation changes, 77-81 prechordal plate, 8C-8 1 topography, 77-78

205 evolution, 94 gastrulation organization,93 gastrulation staging, 93-94 gene expression, 87-89 early gastrulation stages, 87-89 late gastrulation stages, 89 morphology, 65-75 amnion formation, 67-68 anterior margin, 69-74 blastocyst, 65-66 embryonic disc, 69 epiblast differentiation, 74 historical perspectives, 74-75 hypoblast differentiation,74 implantation, 67 size, 68 overview, 64-65,93-94 phylogenetic implications, 89-93 extraembryonic membrane generation, 91-93 lower layer homology, 90 vegetal hemisphere yolk accumulation, 90-9 1 Maternal factors, unique cleavage pattern generation micromere formation in sea urchin embryos, 5-8 overview, 2-3,3 1-32 spiral cleavage in gastropod embryos, 9 Meiosis, fertilization, egg stages, 142-143 Mesoderm, differentiation, 78-79 Micromeres, formation annelid embryos, unequal cleavage, 20 sea urchin embryos, 5-8 first cleavage, 7-8 maternal factors, 5-6 maternally localized substance, 7-8 vegetal cortex, 6-7 Mitotic apparatus, cleavage plane positioning, 3-4 Morphogens, central nervous system pattern formation, 182-183

N Nematodes, par gene mutants, unique cleavage pattern generation, 20-3 1 asymmetric cleavage, 21-23 par gene characteristics, 23-31 Nervous system, see Brain development; Central nervous system

206 Neuroepithelium,pattern regulation, fibroblast growth factor role action mechanisms, 191-192 cerebral cortex size regulation, 189-190 dorsal telencephalic compartment growth kinetics, 188-189 glial progenitor cell lineages, 192-194 neuronal progenitor cell lineages, 192-194 regional specification, 185-188 Neurons, progenitor cell lineages, fibroblast growth factor role, 192-194 Nuclear lamina, sperm nuclear activation during fertilization characteristics, 138 sperm nucleus to male pronucleus transformation, 144-145

0 Oocytes cytoplasm sperm nuclear decondensation factors, 157-1 58 unequal cleavage in ascidian embryos, 16-17 fertilization male pronucleus formation, 157-158 meiotic stages, 142-143

P par gene, mutations in nematode embryos, unique pattern generation, 20-3 1 asymmetric cleavage, 21-23 par gene characteristics, 23-3 1 Pattern formation central nervous system, morphogen role, 182-183 gastrulation, see Gastrula neuroepitheliumpatterning regulation, fibroblast growth factor role, 185-192 action mechanisms, 191-192 cerebral cortex size regulation, 189-190 dorsal telencephalic compartment growth kinetics, 188-189 regional specification, 185-1 88 unique cleavage pattern generation, 1-32 cleavage plane positioning, 3-5 centrosome, 3 4 furrow formation, 3-4 mitotic apparatus, 3-4

Index orthogonal pattern, 4-5 single-layered epithelial cell division, 5 extrinsic cues, 3 intrinsic cues, 3 invariant pattern role, 2-3 micromere formation in sea urchin embryos, 5-8 first cleavage, 7-8 maternal factors, 5-6 maternally localized substance, 7-8 vegetal cortex, 6-7 overview, 2-3,31-32 par mutants in nematode embryos, 20-31 asymmetric cleavage, 21-23 par gene characteristics,23-31 spiral cleavage in gastropod embryos, 8-1 1 cleavage clock, 11 description, 8-9 handedness determination, 9 maternal factor role, 9 mechanisms, 10 unequal cleavage in annelid embryos, 17-20 first cleavage, 18-20 micromere formation, 20 second cleavage, 20 unequal cleavage in ascidian embryos, 11-17 centrosome-attracting body, 11-16 posterior egg cytoplasm role, 16-17 Perinuclear matrix, sperm nuclear activation during fertilization characteristics, 136-137 sperm nucleus to male pronucleus transformation, 145-146 Prechordal plate, differentiation, 80-81 Primordial Germ cells, birds, mammal cells compared, 84 Pronucleus, See Male pronucleus Protamine, sperm nucleus to male pronucleus transformation nucleoprotamhe sperm chromatin, 139, 147-149, 156 removal, 147-148 replacement, 148-149 S

Sea urchin, micromere formation, 5-8 first cleavage, 7-8 maternal factors, 5-6

Index matemally localized substance, 7-8 vegetal cortex, 6-7 Sperm infertility treatment, 164-166 nuclear activation during fertilization, 133-166 chromatin asynchronous behavior, 161-1 64 egg stages, 142-143 future research directions, 166 nucleus characteristics, 134-142 centromeres, 140-141 chromatin fibers, 139-140 chromosome packaging order, 140 nuclear lamina, 138 nuclear matrix, 135-138 nucleoprotamine sperm chromatin, 139 nucleosomal sperm chromatin, 139 pennuclear matrix, 136-137 pre-fertilization events, 141-142 somatic chromatin, 138-139 sperm chromatin, 134-135 telomeres, 140-141 overview, 133-134, 166 sperm nucleus to male pronucleus transformation, 143-161 cytoplasmic decondensation factors, 157-1 58 disulfide bond reduction, 147 DNA replication, 159 histones, 148-149 male pronuclear envelope formation, 153-155 nuclear decondensation, 149-152 nuclear envelope removal, 144-146 nuclear lamina, 144-145 nuclear matrix remodeling, 152-153 nucleoprotamine sperm chromatin, 147-149, 156

207 nucleoprotein remodeling, 146-149 nucleosomal sperm chromatin, 146-147, 155-156 oocyte competence, 157-158 pennuclear theca, 145-146 pronuclear chromatin structure, 155-156 protamine removal, 147-148 protamine replacement, 148-149 transcription activation, 158-161 Syncytial yolk, formation, non-D quadrants in leech, 114-119, 127-128

T Telomeres, sperm nuclear activation during fertilization, 140-141 Transcription, activation, sperm nucleus to male pronucleus transformation, 158-161 DNA replication, 159 transcription, 159-161 Tubifex, unequal cleavage pattern generation, 17-20 first cleavage, 18-20 micromere formation, 20 second cleavage, 20

V Vegetal cortex, micromere formation role, 6-7

Y Yolk syncytial yolk formation in leech, 114-1 19, 127-1 28 vegetal hemisphere yolk accumulation in mammals, 90-91

This Page Intentionally Left Blank

Contents of Previous Volumes

1 Pattern Formation in Zebrafish-Fruitful ology and Genetics

liaisons between Embry-

Lilianna Solnica-Krezel

2 Molecular and Cellular Basis of Pattern Formation during Vertebrate Limb Development Jennifer K. Ng, Koji Tamura, Dirk Biischer, andJuan Carlos Izpislia-Belrnonte

3 Wise, Winsome, or Weird? Mechanisms of Sperm Storage in Female Animals Deborah M. Neubaum and Mariana F. Wolfner

4 Developmental Genetics of Caenorhabditis elegans Sex Determination Patricia f. Kuwabara

5 Petal and Stamen Development Vivian F. Irish

6 Gonadotropin-InducedResumption of Oocyte Meiosis and MeiosisActivating Sterols Claus Yding Andersen, Mogens Baltsen, and Anne Grete Byskov

Cumulative Subject Index, Volumes 20 through 41

209

Contents of Previous Volumes

21 0

1 Epigenetic Modification and Imprinting of the Mammalian Genome during Development Keith E. Latham

2 A Comparison of Hair Bundle Mechanoreceptors in Sea Anemones and Vertebrate Systems Glen M. Watson and Patricia Mire

3 Development of Neural Crest in Xenopus Roberto Mayor, Rodrigo Young, and Alexander Vargas

4 Cell Determination and Transdeterminationin Drosophila Imaginal Discs Lisa Maves and Gerold Schubiger

5 Cellular Mechanisms of Wingless/Wnt Signal Transduction Herman Dierick and Amy Bejsovec

6 Seeking Muscle Stem Cells )effrey Boone Miller, Laura Schaefer, andlanice A. Dominov

7 Neural Crest Diversification Andrew K. Groves and Marianne Bronner-Fraser

8 Genetic, Molecular, and Morphological Analysis of Compound leaf Development Tom Goliber, Sharon Kessler,]u-)iun Chen, Geeta Bharathan, and Neelima Sinha

1 Green Fluorescent Protein (GFP) as a Vital Marker in Mammals Masahito Ikawa, Shuichi Yamada, Tomoko Nakanishi, and Masaru Okabe

2 Insights into Development and Genetics from Mouse Chimeras )ohn D. West

3 Molecular Regulation of Pronephric Development Thomas Carroll, )ohn Wallingford, Dan Seufert, and Peter D. Vize

Contents of Previous Volumes

21 1

4 Symmetry Breaking in the Zygotes of the Fucoid Algae: Controversies and Recent Progress Kenneth R. Robinson, Michele Wozniak, Rongsun Pu, and Mark Messerli

5 Reevaluating Concepts of Apical Dominance and the Control of Axillary Bud Outgrowth Carolyn A. Napoli, Christine Anne Beveridge, and Kimberley Cathryn Snowden

6 Control of Messenger RNA Stability during Development Aparecida Maria Fontes, Jun-itsu Ito, and Marcelo Jacobs-Lorena

7 ECF Receptor Signaling in Drosophila Oogenesis Laura A. Nilson and Trudi Schupbach

1 Development of the leaf Epidermis Philip W. Becraft

2 Genes and Their Products in Sea Urchin Development Giovanni Giudice

3 The Organizer of the Castrulating Mouse Embryo Anne Camus and Patrick I? L. Tam 4 Molecular Genetics of Gynoecium Development in Arabidopsis John L. Bowman, Stuart F. Baum, Yuval Eshed, Joanna Putterill, and John Alvarez

5 Digging out Roots: Pattern Formation, Cell Division, and Morphogenesis in Plants Ben Scheres and Renze Heidstra

This Page Intentionally Left Blank