ADVANCES IN DEVELOPMENTAL BIOLOGY
Volume4
0
1996
This Page Intentionally Left Blank
ADVANCES IN DEVELOPMENTAL BIOLOGY Editor:
PAUL M. WASSARMAN Department o f Cell Biology and Anatomy Mount Sinai School o f Medicine New York, New York
VOLUME 4
1996
0
JAl PRESS INC.
Greenwich, Connecticut
London, England
Copyright
0 1996 hy / A / PRESS /NC
5 5 Old Post R w d No 2 Greenwich, Connet ticut 068J’O
)A/ PRESS 1TD. 38 Tavisroc-k Streef Coven/ Garden London, England WC2E 7PB
All rights rcserved. No p a r t of this /)uh/ica/ioni m y Oc rcprodoc-ed, s t o r d on
rotr/cv:~/ system, o r transniitted i n any w ~ y ,of. b y J I J Y riiean5, e/rrt/-on/c.,n ~ e c l i a .nI /~, photocopying, recording, filming o r otherwise w i / h o ~ i111 t ior ~ ~ e r i i i i si n~ writing ~ ~ i i f r o i i i the poblislie,.
ISBN: 1 -5.5938-969-9 Manufactured in the United S t d e s ot Anieric
CONTENTS
LIST OF CONTRIBUTOKS
vii
PREFACE Paul M. Wassarm?ii
ix
THE IN VlVO FUNCTION OF MU LLERIAN-INHIBIT1NG SU BSTANCE r l U RI NG MAMMAL IA N SE X UA L DEVEL O PME NT Yuji Mishina and Richnrd R. Behringer
1
THE ROLE OF TIHE dpp-GROUI’ GENES IN DORSOVENTRAL PATTCKNING OF THE DROSOPHILA EMBKY(1 Christine Rushlow , ? i d Sicgirieti Rotli
27
THE TEKMINAL GENE HIERAKCHY O F DROSOPHILA AND THE GENETIC CONTROL O F TISSUE SPEC IF ICAT I 0N AN D MOR I’ t i OG E NESIS Man Lun R. Yip and H o w ~ r dD. Lipshitr
83
ANTERIOR-POSTERIOR POLARIZATION AND MESODERM INDUCING FACTORS IN THE PREGASTRULA MOUSE EMBRYO: COMPARISON TO CHICK AND FROG EMBRYOS Rosemary E Bac-hvnrova
147
HUMAN Y CHROMOSOME FUNCTION IN MALE GERM CELL DEVELOPMENT Peter H. Vogt
191
INDEX
259 V
This Page Intentionally Left Blank
LIST OF CONTRIBUTORS
Rosemary E Bachvarova
Department of Cell Biology and Anatomy Cornell University Medical College
Richard R. Behringer
Department of Molecular Genetics University of Texas M . D . Anderson Cancer Center
Howard D. Lipshitz
Division of Biology California Institute of Technology and Research Institute The Hospital for Sick Children
Yuji Mishiiia
Department of Molecular Genetics University of Texas M.D. Anderson Cancer Center
Siegfried Roth
Max Planck Institut fur Entwickunsgbiologie Tubingen, Germany
Christine Rushlow
Department of Biology N e w York University
Peter H. Vogt
Institute of Human Genetics ancl Anthropology University of Heidelberg
Man Lun R. Yip
Division of Biology California Institute of Technology
vi i
This Page Intentionally Left Blank
PREFACE Advurices in Developizeritul Biology was launched as a series by JAI Press in 1992 with the appearance of Volume 1 . The series is inextricably linked to thecompanion series, Advances iiz Developzeritul Biochemistry, that was launched at the same time. As stated in the Preface to Volume 1 : Together the two series will provide annual reviews of research topics in developmental biologyibiochernistr): written from the pcrspectives of leading investigators in these fields. It is intended that each review draw heavily from the author's own research contributions and perspective. Thus, the presentations are not necessarily encyclopedic in coverage, nor do they necessarily reflect all opposing views of the subject.
Volume 4 of the series follows thcse same guidelines. Volume 4 of Advances in Developrrierital Biologv consists of five chapters that review specific aspects of fly and mammalian development. In Chapter I , Y. Mishina and R. Behringer discuss various aspects of Mullerian-inhibiting substance (MIS) in mammals, from a brief history of its discovery to recent studies of the MIS gene in transgenic and knock-out animals. I n Chapter 2, C. Rushlow and S. Roth discuss the rolc of the +/?-group genes in dorsoventral patterning of the D l l ~ s o p h i ~embryo. ~z In Chapter 3, M. Yip and H. Lipshitz discuss the terminal (asegmental termini) gene hierarchy of Drmophila and the genetic control of tissue specification and morphogenesis. I n Chapter 4, R. Bachvarova discusses induction ix
X
PREFACE
of mesoderm and the origin of anterior-posterior polarity in the mouse embryo, using the frog embryo as a paradigm. I n Chapter 5 , P. Vogt discusses human Y chromosome function in male germ cell development. Finally, I am grateful to the authors for their excellent contributions, as well as for their cooperation and great patiericc during the preparation of this volume. Paul M. Wassarman Set.1'~~ Editor
THE IN VIVO FUNCTION OF MULLERIAN-INHIBITING SUBSTANCE DURING MAMMALIAN SEXUAL DEVELOPMENT
Yuji Mishina and Richard R. Behringer
I. Introduction . . . . . . . . . . . . . . , . . . . . . . . . . 11. Mullerian-inhibiting Substance (MIS) . , . . , . . . . . . . . . A. The Mullerian Inhibitor: A Historical Perspective . . . . . . B. MIS is a Member of the TGF-13 Superfamily . . . . . . . . C. Sexual Dimorphic Expression Pattern of MIS . . . . . . . 111. MIS-Related Abnormalities . . . . . . . . . . . . . . . . . . A. Gain of Function: The Freemartin . . . . . . . . . . . . . . B. Loss of Function: The Persistent Miillerian Duct Syndrome IV. In vivo Functional Analysis of MIS . . . . . . . . . . . . . . . . A. MIS Transgenic M i c e . . . . . . . . . . . . . . . . . . . . B. MIS Knock-Out Mice . . . . . . . . . . . . . . . . . . . C. MlSiTfiii Double Mutant Mice . . . . . . . . . . . . . . . . ,
,
,
,
,
,
,
,
Advances in Developmental I$iology Volume 4, pages 1-25. Copyright 0 1996 by J A l Press Inc. All rights of reproduction in any form resened. ISBN: 1-55938-969-9
I
. . 2
. .3 . . 3 . .6 . . 7 . . 8 . . 8 . . 8 . . 9 . . 9 . 10 . 12
2
YUJI MlSHlNA a t i d RICHAKD R. BEHRINGER
V. Regulatory Factors for MIS . . . . . . . . . . . . . . . . . . . . . A. Does SRY Regulate the Expression of MIS? . . . . . . . . . B. SF-I as a Trans-Regulatory Factor for MIS . . . . . . . . . . C. Loss of Function of SF-1 . . . . . . . . . . . . . . . . . . . VI. Receptor for MIS . . . . . . . . . . . . . . . . . . . . . . . . . . A. Identification of a MIS Receptor . . . . . . . . . . . . . . . B. Molecular Cloning of a MIS Receptor . . . . . . . . . . . . V11. Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.
13 13 14 14 15
IS 17 18
20 20
INTRODUCTION
In mammals, the presence of the Y chromosome acts as a dominant male determinant because of a gene located on this chromosome tenned SRY in humans and Sry in mouse. SRYISq) encodes an HMG box-type transcription factor that presumably regulates the expression ot' a set of' genes that directs the differentiation ofthe pair of indifferent fetal gonads (genital ridges) into testes (Sinclair et al., 1990; Gubbay et a]., 1990; Koopman et al., 1991). In the absence of this gene, as in the case of XX individuals, the undifferentiated gonads develop as ovaries. Thus, during the process of sex determination in mammals the activity or lack ofactivity of Sty specifies the fate of a singlc paired primordium to either a inale or female phenotype. Once sex detennination is established (i.e.,testes form), the production of hormones secreted by the fetal gonads controls the differentiation of the genital duct systems and external genitalia. Both XX and XY individuals develop two pairs of genital ducts associated with the mesonephroi and undifferentiated gonads during development (Figure 1). The Miillerian ducts (also known as the paramesonephric ducts) have the potential to differentiate into female reproductive organs including the uterus, oviducts, and upper portion of the vagina. The Wolffian ducts (also known as the mesonephric ducts) are the primordia of male reproductive organs that include the vas deferens, epididymides, and seminal vesicles. Since each individual, regardless of sex chromosome genotype, has the potential to develop both male and female reproductive organs, a mechanism has evolved to differentiate only one sexual type of reproductive organs. Two hormones, Mullerianinhibiting substance (MIS) and testosterone, have been found to play essential roles during the differentiation of the genital duct system (Figure 1) (Donahoe et al., 1987; Cate et al., 1990; Cate and Wilson, 1993; Josso et al., 1993).
MIS function During S e m i / Developnient
3
indifferent gonads
Mullerian duct
urogenital sinus
Fetus ovary
stis
oviduct
epididymis vas deferens
uterine horn
seminal vesicle
body of uterus
Female
Male
Figure 1. Schematic representation of mammalian sexual differentiation. The Mullerian ducts (paramesonephric ducts) give rise to the uterus, oviducts, and the upper portion of the vagina. The Wolffian ducts (mesonephric ducts) give rise to the epiclidymides, vas deferens, and seminal vesicles. MIS produced hy the Sertoli cells of the fetal testes causes the regression of the Mullerian ducts and testosterone produced by Leydig cells induces the differentiation of the Wolffian duct system. The absence of both hormones during female fetal development permits the development of the Mullerian duct system while the Wolffian ducts passively regress.
I I. MU L L E R IA N- INHIB IT I NG
s uBSTANCE (MI s)
A. The Mullerian Inhibitor: A Historical Perspective
In the middle of this century, Alfred Jost performed pioneering experiments that investigated the influence of fetal hormones during development (Jost, 1947, 1953). He surgically manipulated the gonads of fetal rabbits during various stages of somatic sexual differentiation. Removal of the ovaries from female fetuses before or at the beginning of somatic sexual differentiation resulted in female development (Figure 2). Castration of male fetuses at this stage also resulted in differentiation to thc female phenotype (Figure 2). The Wolffian duct system regressed and the Mullerian ducts persisted and differentiated. Thus, the testes impose a male pattern ofdifferentiation upon a genetic program that is inherently
4
YUJl MlSHlNA and RICHARD R. BEHRINGER
Female
Maie
ovaries
testes
I
ovx
I
testx
Figure 2. Summary of fetal rabbit experiments by Jost.When the ovaries or testes of a fetal rabbit are removed, female development occurs. The Mullerian ducts (MD)differentiate and the Wolffian ducts (WD) regress.
female. Jost also noted that development began towards the male phenotype but reversed upon removal of the gonads. These observations led to the hypothesis that the fetal testis secreted two types of hormones: the first hormone possessed stimulatory activity to induce the differentiation of the urogenital sinus and Wolffian duct system and the second hormone possessed inhibitory activity that caused the regression of the Mullerian ducts. Three additional lines of evidence supported the idea of a Mullerian duct regression hormone secreted by the fetal testis. First, unilateral removal of the fetal testis yielded individuals that possessed a normal Wolffian duct on the unoperated side, however, a not fully normal Wolffian duct and remnants of a uterine horn or a well-developed uterine horn developed on the operated side (Figure 3 ) . Second, when a testicular graft was introduced into a female fetus, a local inhibition of the Mullerian duct and stabilization of the Wolffian duct resulted (Figure 3). Finatly, when a crystal of synthetic androgen was introduced into the abdominal cavity of a castrated male fetus, the development of Wolffian duct-derived structures was stimulated but Mullerian duct differentiation
MIS Function During Sexual Development
ovaries
5 testis
Female
Male
I
I
testis graft
hemi-castrate
I
I
unilaternal
MDWD+
unilateral MD+ WD-
Figure 3. Summary of fetal rabbit experiments by Jost. When a testicular graft is placed unilaterally in a female fetus there is a local regression of the Mullerian duct and differentiation of the Wolffian duct on the side of the fetus where the graft was placed. Hemicastration of a male fetus results in a local persistence of the Mullerian duct.
was not inhibited (Figure 4). These results demonstrated that a Mullerian duct inhibitory activity which had a limited distance of action was produced by the fetal testis and this activity was separable from the masculinizing activities of androgens. Jost initially termed this novel inhibitory activity 1'hormone inhibitrice or the Mullerian inhibitor (Josso et al., 1993). Currently, the hormone is most frequently termed antiMullerian hormone (AMH) or Mullerian-inhibiting substance (MIS) (Josso et al., 1993). In summary, the pathways of male or female sexual differentiation are controlled by the presence or absence of hormones produced by the fetal gonads. During male development, XY fetuses usually develop testes that initially produce MIS. MIS in turn actively induces the regression of the Mullerian ducts, thereby preventing the development of female reproductive organs. Subsequently, testosterone is produced by the Leydig cells of the testes to induce the differentiation of the Wolffian ducts and masculinization of the external genitalia. During female develop-
YUJl MlSHlNA and RICHARD R. BEHRINGER
6
testes
3 syrit hetic
androgen
Figure 4. Summary of fetal rabbit experiments by Jost.When a male fetus is castrated (which should lead to female differentiation, see Figure 2) and crystals of synthetic androgen are added, the Mullerian duct system persists and the Wolffian duct system differentiates. This experiment definitively separated the Mullerian duct regression activity from the Wolffian duct differentiation activity of androgens.
ment, XX fetuses usually develop ovaries that do not express MIS, which creates a permissive environment for the differentiation of the Mullerian ducts. In addition, the lack of testosterone leads to the passive regression of the Wolffian ducts. Thus, MIS and testosterone mediate a switch between the differentiation of the male and female extragonadal reproductive organs. B. MIS is a Member of the TCF-fl Superfamily
In 1969, Picon developed an organ culture system to assay for MIS activity. This assay was instrumental in the purification of the hormone and the subsequent molecular cloning of the gene. In this assay, morphological changes of the Mullerian duct of day 14.5 fetal rat urogenital ridges induced by MIS are scored histologically. A more quantitative assay for MIS activity has been developed that is based upon the ability of MIS to repress aromatase activity in the fetal ovary (Vigier et al., 1989; di Clemente et al., 1992).
MIS Function During Sexual Development
7
The MIS cDNA was initially cloned by two independent groups (Cate et al., 1986; Picard et al., 1986). One approach was to sequence tryptic peptides of bovine MIS and subsequently use degenerate oligonucleotides as probes to screen a newborn bovine testis cDNA library (Cate et al., 1986). The other approach was to screen a newborn bovine testis cDNA expression library using antisera raised against bovine MIS (Picard et al., 1986). Since then, the MIS gene has been isolated in human, cow, rat, and mouse (Cate et al., 1986; Picard et al., 1986; Haqq et al., 1992; Munsterberg and Lovell-Badge, 1991). The MIS gene is subdivided into five exons encompassing approximately 2.75 kb. In humans, MIS maps to the short arm of chromosome 19 and in mouse to chromosome 10 between phenylalanine hydroxylase and mast cell growth factor (Cohen-Haguenauer et al., 1987; King et al., 1991). DNA sequencing of MIS revealed a similarity between the C-terminal portion of MIS and members of the transforming growth factor-p (TGF-P) gene superfamily of growth and differentiation factors (Cate et al., 1986). The other members of this large gene family include activins, inhibins, bone morphogenetic proteins (BMP), and growth differentiation factors (GDF; Massague, 1990). MIS is a homodimeric gylcoprotein (Budzik et al., 1983). Like other members of the TGF-P family, the MIS homodimer which is 140-kDa requires proteolytic processing to generate an N-terminal domain I 10-kDa homodimer and a C-terminal domain 25-kDa homodimer (Pepinsky et al., 1988; Cate et al., 1990). The C-terminal domain of MIS possesses Mullerian duct regression activity which is enhanced in the presence of the N-terminal domain (Wilson et al., 1993). This is in contrast to TGF-P in which the N-terminal domain is dispensible for biological activity (Pircher et al., 1986).
C. Sexual Dimorphic Expression Pattern of MIS MIS expression is restricted to Sertoli cells of the fetal and adult testis and granulosa cells of the postnatal ovary (Cate et al., 1990). It is initially detected at the time of seminiferous tubule formation (Picon, 1970; Tran et al., 1977; Vigier et al., 1983; Munsterberg and Lovell-Badge, 1991). The highest levels of MIS in males are detected during the period of Mullerian duct regression. Although the Mullerian ducts become insensitive to MIS at latter stages (Picon, 1969; Josso et al., 1977), the levels of MIS remain high until birth, when they precipitously drop during pubertal maturation. In the ovary after birth, MIS is found in granulosa
8
YUJIMlSHlNA and RICHARD R. BEHRINGER
cells of preantral and antral follicles and is not detectable in primary and growing follicles or corpus lutea (Takahashi et al., 1986; Bezard et al., 1987; Ueno et al., 1989; Munsterberg and Lovell-Badge, 1991). MIS protein is detected most abundantly in granulosa cells that contact the oocyte and line the antrum. The levels of MIS in the ovary after birth are 0.1% of the levels produced by the fetal testes (Josso, 1986). The expression pattern observed in male and female gonads suggests that, in addition to its Mullerian duct inhibitory activity, MIS may regulate gonadal function and gametogenesis.
111. MIS-RELATED ABNORMAL1TIES A. Gain of Function: The Freemartin In certain species, such as the bovine, exposure of a female fetus to a male twin’s blood by chorioallantoic anastomosis results in regression of the Mullerian ducts, a condition known as freemartinism (Jost et al., 1972). In addition, freemartin ovaries cease to grow, become depleted of germ cells, and may develop seminiferous tubules containing Sertoli cells. Some aspects ofthe freemartin effect, including inhibition of germ cell proliferation and the development of seminiferous cord-like structures, can be reproduced in vitro when fetal rat ovaries are exposed to purified MIS (Vigier et al., 1987). Therefore, as Jost first suspected, a proportion of the abnormal phenotypes associated with the freemartin are likely to be due to the ectopic exposure of a female fetus to MIS (Jost et al., 1972). B. Loss of Function: The Persistent Miillerian Duct Syndrome
In humans, persistent Miillerian duct syndrome (PMDS) is a rare form of male pseudohermaphroditism characterized by the presence of a uterus and Fallopian tubes in XY individuals that are overtly male in phenotype (Guerrier et al., 1989; Josso et al., 1993). Mutations in the MIS gene have been found in a proportion of the PMDS cases with undetectable or low levels of serum MIS (Knebelmann et al., 1991; Carre-Eusebe et al., 1992; Imbeaud et al., 1994). MIS-positive forms of PMDS have been reported in dogs (Meyers-Wallen et al., 1989) and humans (Guerrier et al., 1989). Presumably, these PMDS cases are due to the insensitivity of target tissues to the hormone, for example, a mutation in the receptor for MIS.
MIS Function During Sexual Development
9
Thus, mutations in the MIS gene of humans can result in PMDS and confirm the requirement of this fetal hormone for the regression of the Mullerian ducts during male development. However, to fully examine the requirements of this hormone during development and fertility, an experimental animal model for MIS deficiency is required (see below).
IV. IN VlVO FUNCTIONAL ANALYSIS OF MIS A. MIS Transgenic Mice
To investigate the potential functions of MIS during mouse development, the human MIS gene was ectopically expressed in transgenic mice by means of the mouse metallothionein (MT) promoter (Behringer et al., 1990). The MT promoter was chosen because it can direct the expression of heterologous genes to a variety of fetal and adult tissues in transgenic mice (Palmiter and Brinster, 1986). As a result, lines of transgenic mice were established that possessed circulating levels of human MIS in plasma ranging on average from 40 to 4,400 ng/ml. Female transgenic mice highly expressing the human MIS gene lacked a uterus and oviducts. Ovaries were present in newborn transgenic females but germ cells were subsequently lost and the somatic components of the ovary reorganized into structures reminiscent of the seminiferous tubules of the male gonad. These virilized ovaries subsequently degenerated since they were never found in adult female transgenic mice. However, some transgenic females, predominantly those in the lower-expressing lines, lacked a uterus but retained one or both ovaries. Apparently, the effects of ectopic MIS exposure were titratable, with the Mullerian ducts being the most sensitive to MIS action. Recently, the ontogeny of reproductive abnormalities in these transgenic mice were examined (Lyet et al., 1995). Mullerian duct regression in these transgenic females initiated at the same time as control and transgenic males. A reduction in the number of germ cells in the ovaries of the transgenic females was maximal between E l 6 and birth. This reduction occurred when the transgenic oocytes were still in the leptotene stage of meiotic prophase, whereas normal germ cells had already reached the pachytene stage. Interestingly, abnormal phenotypes were also observed in a proportion of the males (512 1) from the highest-expressing transgenic lines (Behringer et al., 1990). Externally, these transgenic males were feminized, exhibiting mammary gland development, and internally, Wolffian duct
10
YUll MlSHlNA and RICHARD R. BEHRINGER
differentiation was arrested and the testes were undescended. It was postulated that this feminization phenotype was most likely due to a defect in androgen biosynthesis, suggesting that high levels of MIS could influence Leydig cell function. In support of this idea, testosterone levels in the transgenic males were significantly reduced (Lyet et al., 1995). A number of conclusions can be drawn from the results of these transgenic mouse studies. First, MIS can act in vivo as the Mullerian inhibitor. Second, the virilization of the postnatal ovaries and the alteration in Leydig cell function suggest a role for MIS in gonadal differentiation and germ cell development. Finally, the observation that altered levels of MIS resulted in abnormal testicular descent is consistent with the idea that regulated levels of MIS are required for proper descent of the testes (Hutson and Donahoe, 1986).
B. MIS Knock-Out Mice Gene targeting in mouse ES cells provides a method for generating mice carrying mutations in specific genes (Capecchi, 1989). Therefore, to understand the required functions of MIS during embryogenesis and germ cell development, the MIS gene was mutated by homologous recombination in ES cells to generate MIS mutant mice (Behringer et at., 1994). The mutation, an insertion of a selectable marker expression cassette was engineered to simultaneously delete a portion of the first exon, the first intron, and the second exon. Correctly targeted ES clones were obtained and injected into blastocysts to produce mouse chimeras to regenerate mice heterozygous for the MIS mutation. Males and females homozygous for the MIS mutations were recovered from heterozygous intercrosses at the predicted mendelian frequencies. All of the female homozygous mutants possessed a uterus with oviducts and ovaries that were morphologically normal. In addition, all of the females were fertile. Therefore, although MIS is expressed in a regulated manner in the ovary after birth, there is apparently no requirement for MIS expression for normal ovarian function. Perhaps related molecules that are also expressed in granulosa cells might provide redundant or compensatory functions in the absence of MIS. Morphological abnormalities of the reproductive tract were only found in male homozygous mutants. Testes were morphologically normal and completely descended in these males which are in contrast with the results obtained from the transgenic mice expressing high levels of human MIS (Behringer et al., 1990). Moreover, the Wolffian duct system
11
MIS Function During Sexual Development
indifferent gonads
Mullerian duct Wolffian duct
Fetus
vao deferens seminal vesicl
body of the uterus
XY MIS-/-
Tf mY MIS-/-
Figure 5. Schematic representation of male pseudohermaphroditism in XY individuals that lack MIS. XY individuals that only lack MIS differentiate both the Mullerian and Wolffian duct systems and are male in appearance. The presence of both types of reproductive organs severely hinders fertility. The physical association of the resulting oviductal tissue with the Wolffian duct derivatives blocks oviduct coiling. XY individuals that lack MIS and are insensitive to androgens because of the Tim mutation differentiate the Mullerian duct system but the insensitivity to androgens also results in the passive regression of the Wolffian duct system. While these mice have testes, they are female in appearance. Since the Wolffian duct system has been eliminated in these mice the oviductal tissue assumes its normal coiled morphology.
was fully differentiated. However, these mutant males also had a uterus (Figure 5). The uterine horns were physically attached to the vas deferens by connective tissue. While no coiled oviducts were found in these animals, oviductal tissue was present at the distal regions of the uterine horns. Since these MIS-deficient males have testes and both Wolffian and Mullerian duct-derived tissues, they are male pseudohermaphrodites. Nearly all (900/) of the MIS-deficient males were infertile. These males were able to mate with females, but sperm were rarely detected in the uteri of the recipient females. Normal numbers of motile sperm were
12
YUJIMlSHlNA a n d RICHARD R. BEHRINGER
detected in the vas deferens and epididymides of the mutant males, and these were shown to be capable of fertilizing oocytes in viti-o. Thus, the MIS-deficient males were able to produce functional germ cells but the simultaneous development of the Miillerian and Wolffian duct systems structurally interfered with the transfer ofthe sperm into the reproductive tract of females. Histological examination of the testes of the MIS-deficient mice revealed Leydig cell hyperplasia and in one case a testicular tumor of Leydig cell origin. In contrast, no histological abnormalities or tumors were detected in MIS-deficient females. Leydig cell hyperplasia and the development of a tumor in MIS-deficient mice was intriguing because a targeted deletion of the related a-inhibin gene in mice leads to the development of testicular and ovarian tumors (Matzuk et al., 1992). Thus, like inhibin, MIS also appears to function as a gonadal tumor suppressor, though this activity is relatively weak. These loss of function studies demonstrated that MIS is the Mullerian inhibitor and that regression of the Miillerian duct system during fetal male development is important ultimately for male fertility. In addition, MIS is not required for male or female gametogenesis. Furthermore, the Leydig cell hyperplasia and neoplasia suggested that MIS functions in the male gonad to influence Leydig cell proliferation. The effect on Leydig cells was particularly interesting because in the MIS gain-offunction experiments Leydig cell function was also altered in transgenic males (Behringer et al., 1990; Lyet et al., 1995). Finally, the viability of the MIS-deficient mice and the fertility of the homozygous mutant females facilitated subsequent crosses with other relevant mouse mutations (see below).
C. MIS/Tfm Double Mutant Mice After formation of the testis, male sexual differentiation is primarily controlled by two hormones, testosterone and MIS. With the availability of MIS mutant mice, it became possible to generate mice through genetic crosses that lack both MIS and androgen function (Behringer et al., 1994). This was accomplished by exploiting the classic mouse mutation known as testicuEnrfemiizizatiorz (?fin; Green, I990), an X-chromosomelinked mutation that results in feminization of XY mutant individuals due to a mutation in the androgen receptor gene (He et al., 1991; Charest et al., 1991). Thus, TfmlY males are insensitive to androgens and become feminized; they lack Wolffian duct differentiation, and have small and
MIS Function During Sexcia1 Devdopn~ent
13
not fully descended testes in which spermatogenesis is blocked at meiotic prophase. T f X Y males do however produce MIS as evidenced by the regression of the Mullerian duct system. TfmlY MIS-deficient mice were generated by interbreeding the Tfiiz and MIS mutants (Figure 5 ) . These animals were overtly feminized with improperly descended testes and a vaginal opening. Also, Wolffian duct differentiation was eliminated and a uterus had developed. Interestingly, coiled oviducts were present, whereas no coiled oviducts were found in the MIS-deficient male pseudohermaphrodites. These results suggested that the elimination of the Wolffian duct during female development may be required for oviductal morphogenesis.
V. REGULATORY FACTORS F O R MIS A. Does SRY Regulate the Expression of MIS? As discussed above, MIS expression is strictly regulated spatially and temporally in a sexual dimorphic manner. Experiments in transgenic mice demonstrated that 2.0 kb of the 5’ flanking region of the human MIS gene is sufficient for cell-specific gonadal expression of a reporter gene (Peschon et al., 1992). One of the possible regulatory factors for MIS is SRY/Siy because it is expressed before MIS in the fetal gonad. Since SRY contains a motif called the HMG box, SRY is believed to be a sequence-specific DNA-binding protein (Alexander-Bridges et al., 1992). There are two in vitro binding sites for SRY in the human MIS promoter, around - 150 bp (5’-T/C-T/C-TTTGAGA-3’) and -45 to -70 bp (5’-GGGGTCTGTCCTGCACAAACACCCC-3’) upstream of the transcription initiation site (Alexsander-Bridges et al., 1992; Harley et al., 1992; Denny et al., 1992; Haqq et al., 1993). It was demonstrated that SRY can bind 5’-AACAAT-3’ with higher affinity than an HMG box consensus sequence (5’-AIT-AIT-CAAAG-3‘)(Haqq et al., 1994; Harley et al., 1994), however, the closest related sequence within the MIS promoter was 5‘-CACAAA-3‘ in the second binding site (around -50 bp). The binding site around - 1 50 bp is not necessary because 114 bp of the human MIS promoter was sufficient for transcriptional activation by SRY in a cell line derived from the differentiating gonadal ridge of male rat embryos (Haqq et al., 1994). Although a mutant SRY which had one amino acid substitution in the HMG box could not activate MIS expression in the gonadal cell line, mutations in the second SRY/Srry binding site (-45 to -75, including 5’-CACAAA-3’) in the 114 bp MIS promoter
14
YUll MlSHlNA dnd RICHARD R. BEHKINGER
did not diminish transcriptional activation (Haqq et al., 1994). Thus, it is still unclear whether SRY/Siy regulates the expression of M I S in a direct manner. B. SF-1 as a Trans-Regulatory Factor for MIS
In primary Sertoli cells, it has been demonstrated that the first -180 bp of the mouse MIS gene was sufficient for expression (Shen et al., 1994). Within this 180 bp, thcre is a highly conserved element around -90 bp designated MIS regulatory element 1 (MIS-RE- l ) , 5’-CCAAGGTCA-3’ (Shen et al., 1994; Hatano et al., 1994),exclusive ofthe binding sites for SRY. SRY itselfcannot bind the MIS-RE- 1, however, a sequence specific binding protein for the MIS-RE- 1 has becn identified in anuclear extract of rat Sertoli cells. The temporal and spatial patterns of this binding activity coincide with that ofMIS, that is, high in rat embryonic testis, E20, and PO testis, tow in E20 ovary, P20, and older testis, no activity in heart or kidney (Shen et al., 1994; Hatano et al., 1994). Since an antibody against an orphan nuclear receptor known as steroidogenic factor- 1 (SF- 1), or alternatively known as adrenal four-binding protein (Ad4BP) (Rice et al., 199 I ; Lala et al., 1992; Honda et al., 1993) is able to abolish the sequence specific binding activity, this binding protein is believed to be SF-I/Ad4BP(Shen et al. 1994). Similarity ofthe temporal and spatial expression patterns of SF-I with the binding activity also supports this idea (Shen et al. 1994). A cotransfection assay of SF-I and the -1 80 bp MIS construct has been performed using the heterologous HeLa cell line to examine the transactivation ability of SF-1 for MIS. The finding that intact SF-I cannot activate M f S gene expression whereas a deletion mutant of a putative ligand-binding domain of SF- 1 can activate the expression suggests a ligand and/or cofactor is necessary for SF-1 function (Shen et al., 1994). C. Loss of Function of SF-1
SF-l/Ad4BP is encoded by the jtishi tnix~zufactor 1 (Ftz-Ff)gene (Lala et al.. 1992). This gene also encodes embryonal long terminal repeat-binding protein (ELP) by alternative promoter usage and splicing (Ikeda et al., 1993). SF-1 is expressed not only in all primary steroidogenic tissues including the adrenal cortex, testicular Leydig cells, and ovarian granulosa cells, but also in the embryonic urogenital ridge at E9-E9.5 and fetal Sertoli cells (Ikeda et al., 1993; Ikeda et al., 1994). Since this expression pattern suggested that SF- 1 plays multiple roles in
MIS Functroii During Sexifdl Developnient
15
gonadal differentiation, a targeted disruption of Ftz-FI was undertaken to determine its function itz vivo (Luo et a]., 1994, see below). Homozygous mutants ofFfz-Fl were born but began dying at 12 hours after birth and all of them were dead within eight days of birth. Lower corticosterone levels in the mutants and subcutaneous injection of a glucocorticoid/mineralocorticoidcocktail prolonged their survival time, suggesting that the cause of death was adrenocortical insufficiency. The absence of the adrenal glands in the Ftz-FI mutant mice supported this idea. Both male and female Ftz-FI mutant mice also lacked gonads that was consistent with the expression pattern of SF-I and oviducts, uteri, and vagina formed in a normal female pattern in both sexes. These findings suggest that SF-1 is essential for the differentiation of the adrenal glands and gonads. The absence of the MIS-expressing gonads in mutant males would explain the persistence of Miillerian duct derived tissues. Developmental analysis of the gonads of Ftz-FI mutant mice showed that abnormalities were first detected at E12.0. By E12.5 the gonads were almost coinpletely degenerated. Alkaline phosphatase staining of El 1.5 gonads from Ftz-FI mutant mice suggested that SF-I was not necessary for primordial germ cell migration (Luo et al., 1994). Thus, like SRY, SF-1 is an important regulator of gonadal development. The precise relationship of SF-I to SRY and MIS will likely be an actively pursued area of future research.
VI. RECEPTOR FOR MIS A. Identification of a MIS Receptor
The phenotypes of the MIS-overexpressing transgenic mice and the MIS-deficient mice were similar yet different. Some of the explanations that might reconcile these differences probably relate to cross talk phenomena between ligands and receptors (Figure 6). Since the MIS overexpressing transgenic mice are exposed to pharmacological levels of MIS during development, it is possible that this may lead to productive interactions with other related receptors (Figure 6). In the case of MIS-deficient mice, some of the possible phenotypes might be rescued by signals through MIS receptors caused by the binding of other ligands (Figure 6). Therefore, characterization of the receptors for MIS is important to understand the function of MIS during sexual development. The establishment of a binding assay for MIS has been lacking for many years. Recently, a binding assay using FITC-labeled human MIS
normal
MIS-deficient
MIS
mice
E 0.0
MIS
MIS
mice
receptor-deficienl
mice
MIS-dericienl
receptor-deficient
v
I
M I S
0 TGF-P
family
mice
mice
X
X
MIS
mice
F
0
overexpressing
overexpressing
Y
MIS
mice
MIS
receptor
T(;F-P f a n i i l j r e c e p t o r
receptor-deficient
+r e g u l a r
mice
lalk
-------, cross
talk
Figure 6. Schematic expression of possible cross talk phenomena. In normal mice, there may be cross talk between MIS and receptors for other members of the TGF-P family and/or between other members of the TGF-P family and the receptor for MIS (A, dashed line). In MIS expressing transgenic mice, some of the phenotypes may be the result of cross talk between overexpressed MIS and receptors for other members of the TGF-P family (B, bold dashed line). In MIS-deficient mice, someof the phenotypes which might occur could be rescued by the activation of the MIS receptor by cross talk with other members of the TGF-P family (C, dashed line). Thus, novel phenotypes appearing by crossing with MIS overexpressing mice would be the result of cross talk between overexpressed MIS and other types of MIS receptors or receptors for other members of the TGF-P family (D and E, dashed line). The appearance of novel phenotypes by crossing with MIS-deficient mice would suggest that there has been a cross talk between MIS and receptors for other members of the TGF-P family in the MIS receptor-deficient mice (D and F, dashed line). What kinds of phenotypes have been rescued in MIS-deficient mice by cross talk between other members of the TGF-P family and MIS receptor would also be found (C and F, dashed line). 16
MIS Function During Sexual Development
17
was developed (Catlin et al., 1993). According to this method, specific binding sites for MIS were identified in a MIS responsive human vulvar carcinoma cell line (A43 1 ) as 15,000-1 8,000 sites per cell and Kd = 5.8 nM. The molecular weight of the binding protein was determined to be 88 kDa (Catlin et al., 1993). It is not clear if this protein acts as a signal transduct ion molecule. B. Molecular Cloning of a MIS Receptor
In the case of other TGF-P family ligands, two types of membrane bound serine/threonine kinases form heterodimers to transduce a signal (type I and type I1 receptors, molecular weights 55 kd and 7 5 - 8 5 kd, respectively) subsequent to their ligand binding (Kingsley, 1993). Type I1 receptors were cloned by their binding ability for activin (Mathews and Vale, 1991) or TGF-P (Lin et al., 1992). Type I receptors require type II receptors for their ligand binding activity; they have been cloned by sequence similarity of the kinase domain with type I1 receptors (Attisano et al., 1993; Ebner et al., 1993; Franzen et al., 1993). Type I receptors have a characteristic serine/glycine box at the juxtamembrane region which provides phosphorylation sites by type I1 receptor kinases and type I receptors are believed to phosphorylate downstream proteins during signal transduction (Wrana et al., 1994). Recently, candidate genes for the MIS type I1 receptor have been isolated from a rat Sertoli cell cDNA library as an testosterone induced gene (Baarends et al., 1994) and a rabbit fetal ovary cDNA library using a degenerate oligo probe for a conserved kinase domain (di Clemente et al., 1994). Sequence comparisons suggest that these are probably homologs. Although the affinity of the receptor when tranfected into COS cells is very low, specific binding for MIS has been demonstrated (di Clemente et al., 1994). The expression of the putative MIS receptor gene was localized by in situ hybridization to the mesenchymal cells adjacent to the Miillerian ducts, suggesting that MIS most likely alters the surrounding mesenchyme to elicit Mullerian duct epithelium regression. A type I receptor known as R1, alternatively named SKR-1 or ALK-2 (Matsuzaki et al., 1993; ten Dijke et al., 1993), has been suggested to be a receptor for MIS based upon its expression pattern. However, since it is expressed in many other regions besides the mesenchyme surrounding the Mullerian ducts, it is not clear based upon this information that this is indeed a type I receptor for MIS (He et al., 1993). Ligand binding studies for the R1 receptor and the putative MIS type I1 receptor complex
-
18
YUjl MISHINA and RICHARD R. BEHRINGER
should be informative. Isolation of the human MIS receptor gene will facilitate the identification of human PMDS patients with normal levels of MIS that have mutations in the MIS receptor gene. Finally, studies of the MIS receptor gene will open up avenues for the molecular characterization of signal transduction pathways that mediate Mullerian duct regression and Leydig cell proliferation control.
VII. CONCLUSIONS AND PERSPECTIVES In mammalian embryos, there is only a single pair of primordial gonads that can differentiate into either sexual type depending upon the SRYISry genotype. In males, a single-copy gene ofSRY/Sry initiates a cascade of testicular differentiation (Figure 7). In this cascade, Ftz-Fl in Sertoli cell is activated to produce SF- 1/Ad4BP protein to activate MIS expression. A role for Ftz-Fl earlier in the development of the embryonic urogenital ridge is also likely. Unlike the gonads, embryos have two genital duct systems. SRY/Sty expression leads to both induction of male-specific tissues by testosterone produced in Leydig cells, and the regression of female-specific anlagen by MIS produced in Sertoli cells. The absence of these two hormones during fetal development in the female (the hormonal equivalent of no testes) permits Mullerian duct differentiation and does not induce Wolffian duct development. The in vivo outcomes of ectopic MIS exposure or MIS deficiency illustrate the balance required to coordinately differentiate and cause the regression of the respective male and female genital ducts. The observations made in the MIS-deficient mice demonstrate that codevelopment of both genital duct systems interferes with normal development of both systems and ultimately interferes with reproduction and fertility. Thus, it is important to produce individuals that possess only one sexual type of reproductive organs for efficient reproduction and fertility. MIS produced by the Sertoli cells of the fetal testis would interact with its receptors which are located on the mesenchyme cells surrounding the Mullerian duct to transduce a signal that alters those mesenchyme cells to ultimately mediate the regression of the Mullerian duct. This signaling pathway is still unclear, however, experiments are underway to isolate the mouse MIS receptor gene to generate MIS receptor-deficient mice and compare their phenotype with the MIS gain-of-function and loss-offunction animals. These mutant animals will also provide a suitable tool to examine cross talk phenomena (Figure 6).
MIS Function During Sexual Development
19
14- SRMSry
Leydig cells
Sertoli cells
steroid+ genesis
SF-l/Ad4BP
4
testosterone
Figure 7. Summary of the mechanism of sexual dimorphic development. The expression of SRY/Sry initiates testicular differentiation that leads to male steroidogenesis in Leydig cells and Ftz-FI expression in Sertoli cells. With unknown cofactor(s), SF-l/Ad4BP, a protein encoded by Ftz-F7 activates MIS. MIS binds a MIS receptor located in surrounding mesenchyme of the Mullerian ducts to cause regression of Mullerian ducts. Testosterone generated by Leydig cells induces the differentiation of the Wolffian ducts toward male specific reproductive tissues. W, the Wolffian duct; M, the Mullerian duct; MIS-R, MIS receptor.
Exploring the downstream proteins which interact with MIS receptors also should provide important insights. Recently, it was demonstrated that FK506 binding protein 12 (FKBP12) could interact with type I receptors of the TGF-P family (Wang et al., 1994). Since FKE3P12 is
20
YUll MlSHlNA arid RICHARD R. HEHRINGER
expressed in many tissues, i t is unlikely to contribute to a MIS-specific signal transduction pathway, howevcr, this type of approach may delineate what kind of biochemical reactions occur for regression of the Mullerian duct after MISheceptor interactions.
ACKNOWLEDGMENTS W e thank Dr. Martin Matzuk for critical reading o f t h i s manuscript, and Yoshiko
and Kanade Mishina for encouragement. Aided by a grant from the National Institutes of Health HD30284 and N C I CA16672.
REFERENCES Alexander-Bridges, M., Ercolani. L.. Kong, X.F., & Nasrin, N. (1992). Identification of a core motif that is recognized by three members ofthe HMG class of transcriptional regulators: IRE-ABP, SRY and TCF-la. J. Cell. Biochem. 48. 129- 135. Attisano, L., Carcamo, J., Ventura, F., Wcis, F.M.B., MassaguC, J., & Wrana, J.L. (1993). Identification of human activin and TGF-13 type 1 receptors that form heterodirneric kinase complexes with type I I receptors. Cell 7.5, 671-680. Baarends, W.M., van Helmond, M.J.L., Post, M., van dcr Schoot, J.C.M., Hoogerbrugghe, J.W., de Winter, J.P., Uilenbroek, J.T.J.. Karels, B., Wilniing, L.G., Meijers, J.H.C., Themmen, A.P.N., & Grootegoed, J.A. (1994). A novel member of the transmembrane serinehhrconinc kinase receptor family is specifically expressed in the gonads and in mescnchyinal cells adjacent to the Mullerian duct. Development 120, 189%197. Behringer, R.R., Cate, R.L., Froelick, G.J., Palmiter, R.D., & Brinster, R.L. (1990). Abnormal sexual development in transgenic inice chronically expressing Mullerian inhibiting substance. Nature 345. I67 -1 70. Behringer, R.R., Finegold, M.J., & Catc, R.L. ( 1994). Mullerian-inhibiting substance function during mammalian sexual development. Cell 79, 415 426. Bezard, J., Vigier, B., Tran, D., Mauleon, P., & Josso, N. (1987). Ininiunocytochemical study of anti-Miillerian hormone in sheep ovarian follicles during fetal and post-natal development. J. Reprod. Fertil. 80, 50'+516. Budzik, G.P., Powell, S.M., Kamagata, S., & Donahue, P.K. (1983). Miillerian inhibiting substance fractionation by dye affinity chromatography. Cell 34, 307-3 14. Capecchi, M.R. (1989). The new niousc genetics: Altering the genome by gene targeting. Trends Genet. 5,70-76. Carre-Eusebe, D., Imbeaud, S.. Harbison, M., New, M.I., Josso, N., & Picard, J.-Y. (1992). Variants of the anti-Mullerian hormone gene in a compound heterozygotes with the persistent Mullerian duct syndrome and his family. Iliun. Genet. 90, 3 8S-394. Cate, R.L., Mattaliano, R.J., Hession, C., Tizard, R., Farber, N.M., Cheung, A,, Ninfa, E.G., Frey, A.Z., Gash, D.J., Chow, E.P., Fisher, R.A., Bertonis, J.M., Torres, G., Wallner, B.P., Ramachandran, K.L., Ragin, R.C., Manganaro, T.F., MacLaughlin, D.T., & Donahoe, P.K. (1986). Isolation ofthe bovine and human genes for Mullerian
MIS Func-fionDuring SexiiaI Devdopinent
21
inhibiting substance and expression of the human gene in animal cells. Cell 45. 68S698. Cate, R.L., Donahoe, P.K., & MacLaughlin, D.T. ( 1990). Mullerian inhibiting substance. In: Peptidr Growth Fuctor.? arid Their. Recepors I1 (Sporn, M.B., & Roberts, A.B., Eds.). Springer-Verlag, Berlin, pp. 179-2 10. Cate, R.L., & Wilson, C.A. ( 1993). Mullcriaii-inhibiting substance. In: Genes in Mummaliarz Repiduction (Gwatkin, R.B.L., Ed.). Wiley-Liss, New York, pp. 185-205. Catline, E.H., Ezzell. R.M., Donahoe, P.K., Gustafson. M.L., Son, E.V., & MacLaughlin, D.T. (1994). Identification o f a receptor for huinan Miillerian inhibiting substance. Endocrinology 133, 3007- 30 13. Charest, N.J., Zhou, Z.-X., Lubahn, D.B.. Olsen, K.L., Wilson, E M . , & French, F.S. (1991). A fraineshift inutation destabilizes androgen receptor messenger RNA in the Tfk mouse. Mol. Endo. 5, 573-58 1 . Cohen-I-laguenauer, O., Picard, J.-Y., Mattei, M.G., Serero, S., Nguyen, V.C., de Tand, M.F., Guerrier, D., Hors-Cayla, M.C., Josso, N., & Frezal, J. (1987). Mapping of the gene for anti-Miillerian horinone to the short ann of huinan chromosome 19. Cytogenet. Cell. Genet. 44, 2 4 . Denny, P., Swift, S., Connor, F., & Ashworth, A . ( 1992). An SRY-related gene expressed during spermatogenesis in the mouse encodes a sequence-specific DNA-binding protein. EMBO J. 11, 370S3712. di Clemente, N., Ghaffari, S., Pepinsky, R.R., Picau, C., Josso, N., Cate, R.L.. & Vigier. B. (1992). A quantitative and interspecific test for biological activity of antiMullerian hormone: The fetal ovary aroinatase assay. Development 114, 72 1-727. di Clemente, N., Wilson, C., Faure, E., Boussin. L., Carmillo, P., Tizard, R., Picard, J.-Y., Vigier, B., Josso, N., & Cate, R.L. (1994). Cloning, expression and alternative splicing ofthe receptor for anti-Mullerian hormone. Mol. Endocrinol. 8, I 0 0 6 1020. Donahoe, P.K., Cate, R.L., MacLaughlin, D.T., Epstein, J.. Fuller, A.F., Takahashi, M., Coughlin, J.P.. Ninfa, E.G., & Taylor, L.A. (1987). Mullerian inhibiting substance: Gene structure and inechanisin of action of a fetal regressor. Recent Prog. Horm. Res. 43, 43 1 - 4 7 , Ebner, R., Chen, R.J., Shum, L., Zioneheck, T.F., Lee, A.R.. & Derynck, R. (1993). Cloning of a type I TGF-P receptor and its effect on TGF-p binding to the type 11 receptor. Science 260, 1144- 1348. Franzen, P., ten Dijkc, P., Ichijo, H., Yainashita, H., Schulz, P., Heldin,C.H., & Miyazono, K. (1993). Cloning o f a TGF beta type I receptor that forms a heteroineric complex with the TGF beta type I1 reccptor. Cell 75, 6x1 692. Green, M.C. (1990).Catalog of mutant genes and polymorphic loci. In: Gerreficbhrinnts and Strains oj'the Lahorutor:v A4oir.w (Lyon, M.F., & Searle, A.G., Eds.). Oxford University Press, London, pp. 302-363. Gubbay, J., Collignon, J., Koopinan, P., Capel, B., Econoinou, A., Miinsterberg, A., Vivian, N., Goodfellow, P.N., & Lovell-Badge, R. (1990). A gene mapping to the sex-determining region of the mouse Y chromosome is a member of a novel family of embryonically expressed genes. Nature 3415,245-250. Guerrier, D., Tran, D., Vanderwinden, J.M., Flideux, S.. Van Outryvc, L., Legeai, L., Bouchard, M., Van, V.G.. Picard. J.Y., Kahn, A,, & Josso, N. (1989). The persistent Miillerian duct syndrome: A molecular approach. J. Clin. Endocrinol. Metab. 68, 46-52.
22
YUJIMlSHlNA and RICHARD R. REHRINGER
Haqq, C.M., Lee, M.M., Tizard, R., Wysk, M., DeMarinis, J., Donahoe, P.K., & Cate, R.L. ( I 992). Isolation ofthe rat gene for- Miillerian inhibiting substance. Genoinics 12, 665.469. Haqq, C.M., King, C.Y., Donahoe, P.K., & Weiss, M.A. (1993). S K Y recognizes conserved DNA sites in sex-specific promoters. Proc. Nail. Acad. Sci. USA YO, 1097.1101. Ilaqq, C.M., King, C.Y., Ukiyama, E., Falsati, S., Haqq, T.N., Donahoe, P.K., & Weiss, M.A. (1994). Molecular basis of maimnalian sexual determination: Activation of Mullerian inhibiting substance gene expression by S K Y . Science 266. 1494-1 500. Harley, V.R., Jackson, D.I., Hextall, P.J., Hawkins, J.R., Rcrkovit7, C.D., Sockanathan. S., Lovell-Badge, R., & Goodfellow, P.N. (1992). DNA binding activity ofrccombinant S R Y from normal males and X Y females. Scicnce 255,453~456. Harley, V.R., Lovell-Badge, R., & Goodfellow, P.N. (1994). Definition of a consensus DNA binding site for SKY. Nucl. Acids. Res. 22, 1500-1501. IIatano, O., Fdkayama,' K., Imai, T., Watemian, M.R., Takakusu, A.. Ointira, 'F.. & Morohashi, L I . ( 1994). Sex-dependcnt expression ofa transcription factor, Ad4BP. resulting steroidogenic P-450 genes in the gonads during prenatal and postnatal rat development. Development 120, 2787- 2797. He, W.W., Kumar, M.V., & Tindall, D.J. (1991). A frame-shift mutation in the androgen receptor gene causes complete testosterone insensitivity in the testicular-feminized mouse. NLIC~. Acids. Res. l Y , 2373--2378. He, W.W., Gustafson, M.L., Hirobe, S., & Donahoe. P.K. (1993). Developinental expression offour novel serineithreonine kinase receptors homologous to the activin/transforming growth factor-beta type I1 receptor family. Dev. Dyn. 196, 133-142. Honda, %-I., Morohashi, K.I., Nomura, M., Takeya, H., Kitajima, M., & Omtira, T. ( 1993). Ad4BP-regulating steroidogenic P-450 gene is a member of steroid hormone receptor superfamily. J . Biol. Chem. ?6K, 74947502. Hutson, J.M., & Donahoe, P.K. (1986). The hormonal control of testicular descent. Endocr. Rev. 7,27&283. Ikeda, Y., Lala, D.S., Luo, X., Kim, E., Moison, M.-P., & Parker, K.L. (1993). Characterization of the mouse FTZ-(;I gene, which encodes a key rcgulator of steroid hydroxylases. Mol. Endocrinol. 8, 292--295. Ikeda, Y., Shen, W.-H., Ingraham, I I.A., & Parkcr. K.L. ( 1994). Developmental expression of mouse steroidogenic factor 1, an essential regulator of the steroid hydroxylases. Mol. Endocrinol. 8,654-662. Imbeaud, S., Carre-Eusebe, D., Rey, R., Belvillc, C., Josso, N . , & Picard, .I.-Y. (1994). Molecular geneticsofthe persistent Miillerian duct syndrome: Astudy of 19 families. Hum. Mol. Genet. 3, 125--131. Josso, N., Picard, J.Y., & Tran, D. ( 1 977). Thc anti-Mullerian hormone. Rec. Prog. Honn. Res. 33, 117-160. Josso, N. ( 1986). Anti-Mullerian hormonc: New perspectives for a sexist molecule. Endocr. Rev. 7,42 1-433. Josso, N., Cate, R.L., Picard, J.-Y., Vigier, B., di Clemente, N., Wilson, C., Imbeaud, S., Pepinsky, R.B., Guerrier, D., Boussin. L., Legeai, L., & Carre-Eusebe, D. (1993). Anti-Mullerian hormone: The Jost factor. Rec. Prog. Ilorm. Res. 48, 1-59. Jost, A. (1947). Recherches sur la differenciation sexuelle de I'embryon de lapin. Arch. Anat. Micro. Morph. Exp. 36, 27 1 - 3 IS.
MIS Function During S r x u c ~Drvelopnient /
23
Jost, A. ( 1953). Problems of fetal endocrinology: The gonadal and hypophyseal hormones. Rec. Prog. I-lorin. Res. 8, 3 7 9 4 18. Jost, A., Vigier, 8..& Prepin. J. (1972). Freemartins in cattle: The first steps of sexual organogenesis. J. Reprod. Fertil. 29,349--379. King, T.R., Lee, B.K.. Bchringcr, R.R., & Eicher, E.M. (1991). Mapping anti-Mullerian hormone ( A d z ) and related sequences in the mouse: Identification of a new region of homology between MMIJlO and HSAI9p. Genomics 11,273-283. Kingsley, D.M. ( 1 994). The TGF-p superfamily: New members, new receptors, and new genetic tests of function i n different organisms. Genes Dev 8. 133-146. Koopman, P., Gubbay, J., Vivian, N., Goodfellow, P., & Lovell-Badge, R. ( 1 991). Male development of chromosomally fcmale mice transgenic for St?>.Nature 351, 1 17121. Knebelmann, B., Boussin, L., Guerrier, D., Legeai, L., Kahn, A., Josso, N.. & Picard, J-Y. (199 I).Anti-Mullerian horinone Rruxelles: A non-sense mutation in the last exon of the anti-Mullerian hormone gene associated with the persistent Mullerian duct syndrome in three brothers. Proc. Natl. Acad. Sci. USA 88, 3767-377 I . Lala, D.S., Rice, D.A., & Parker, K.L. ( 1 992). Steroidogenic factor 1, a key regulator of steroidogenic enzyme cxpression, is thc mouse homolog of fushi tarazu factor I . Mol. Endocrinol. 6, 1249- 1258. Lin, H.Y., Wang, X., Ng-Eaton, E., Weinberg, R.A., & Lodish, I1.F. (1992). Expression cloning of the 'TGF-P type I1 receptor, a functioiial transmembrane serineithreonine kinase. Cell 68, 775-785. Luo, X., Ikeda, Y., & Parker, K.L. ( I 994). A cell-specific nuclear receptor is essential for adrenal and gonadal development and sexual dif'ferentiation. Cell 77,48 1 4 9 0 . Lyet, L., Louis, F., Forest, M.G.. Josso, N., Behringer. R.R., & Vigier, B. (1995). Ontogeny of reproductive abnormalities induced by deregulation of anti-Mullerian hormone expression in transgenic mice. Biol. Reprod. 5 2 , 4 4 4 4 5 4 . Massague, .I. (1990). The transforming growth factor-p family. Ann. Rev. Cell Bid. 6, 59744I. Mathews, L., & Vale, W.W. ( 1 99 1). Expression cloning of an activin receptor, a predicted transmembrane seriiic kinase. C'ell 65, 973-982. Matsuzaki. K., Xu, J., Wang, F., McKeehari, W.L.. Kruniiiien, L., & Kan, M. (1993). A widely expressed transmembrane serine/threoninc kinase that does not bind activin. inhibin, transforming growth factor fl, or bone morphogenic factor. J. Bid. Chein. 268, 127 18G12723. Matauk, M., Finegold, M.J., Su, J.-G.J., Hseueh, A.J.W.,& Bradley, A. (1992). a-Inhibin is a tumour-suppressor gene with gonadal specificity in mice. Nahire 360.3 13-3 19. Meyers-Wallen, V.N., Donahoe, P.K., Ueno. S . , Manganaro, T.F., & Patterson, D.F. (1989). Mullerian inhibiting substance is present in testes of dogs with persistent Mullerian ducr syndrome. Biol. Reprod. 41, 881-888. Munsterberg, A,, & Lovell-Badge, R. ( I99 I ). Expression of the mouse anti-Mullerian hormone gene suggests a role in both male and female sexual differentiation. Development 113, 6 1 3 4 2 4 . Palmiter, R.D., & Brinster, R.L. (1986). C;ern-line transformation of mice. Ann. Rev. Genet. 20, 46-99, Peschon, J.J., Behringer, R.R.. Cate, R.L., Harwood, K.A., Idzerda, R.L., Brinster, R.L., & Palmiter, R.D. (1992). Direct expression of an oncogene to Sertoli cells in
24
YUJl MlSHlNA and RICHARD R . KEHRINCER
transgenic mice using Mullerian inhibiting substance regulatory sequences. Mol. Endocrinol. 6, 1403- 141 1. Pepinsky, R.B., Sinclair, L.K., Chow, E.P., Maltaliano, R.J., Manganaro, T.F., Donahoe. P.K., & Cate, R.L. (1988).Proteolytic processing of Miillerian inhibiting substancc produces a transforming growth factor-p-like fragment. J. Biol. Chein. 263, 1896 1-
18964. Picard, J.-Y., Benarous, R.. Guerricr, D., Josso. N.. & Kahn, A. ( 1986).Cloning and expression of cDNA for anti-Mullerian hormone. Proc. Natl. Acad. Sci. USA 83,
5464--5468. Pircher, R., Jullien, P., & Lawrence, D.A. (19x6). (3-transforming growth factor is stored in human blood platelets as a latent high molecular weight complex. Biocheni. Biophys. Res. Comm. 136, 30-37. Picon, R.(1969).Action du testicule foetal stir le devcloppeincnt rri vi/ro dcs canaux dc Muller chez Ie rat. Arch. Anat. Microsc. Morphol. Exp. 58. 1-19. Picon, R. (1970).Modification, cheL le rat. au cours du developpeinent du iesticule. de son action inhibitrice sur les canaux de Muller iri vitrn. C.R. Acad. Sci. SCr. D. Paris
271,237&2372. Rice, D.A., Mouw, A.R., Bogerd, A.M., & Parker, K.L. (1991).A shared promoter element regulates the expression of three steroidogenic enzymes. Mol. Endocrinol. 5, 1552-1561. Shen, W.-H., Moore,C.C.D., Ikeda, Y.,Parker, K.L.,& Ingrahani, H.A. (1994).Nuclear receptor steroidogenic factor 1 regulates the Mullerian inhibiting substance gene: A link to the sex determination cascade. Cell 77, 1L20. Sinclair, A.H., Berta, P., Palnier, M.S., Hawkins, J.R., Griffiths, B.L., Smith, M.J., Foster, J.W., Frischauf, A-M., Lovell-Badge. R., & Goodfcllow, P.N. (1990).Agene from the human sex-determining region encodes a protein with homology to a conserved DNA-binding motif. Nature 346, 24@-244. Takahashi, M., Koide, S.S., & Donahoe, P.K. (1986).Mullerian inhibiting substance as oocyte meiosis inhibitor. Mol. Cell. Endocrinol. 47,225-234. ten Dijke, P., Ichijo, H., Franzen, P., Schulz, P., Saras. J., Toyoshima, H., Heldin, C.H., & Miyazono, K. (1993).Activin receptor-like kinases: A novel subclass of cell-surface receptors with predicted serincithreonine kinase activity. Oncogene 8,2879-87. Tran, R., Meusy-Dessole, N., & Josso, N. (1977).Anti-Mullerian honnone is a functional marker of foetal Sertoli cells. Nature 26Y. 41 1 4 1 2 . Ueno, S.,Takahashi, M., Manganaro, T.F., Ragin, R.C., & Donahoe, P.K. (1989).Cellular localization of Mullerian inhibiting substance i n the developing rat ovary. Endocrinology 125, 1060-1066. Vigier, B., Tran, D., du Mesnil du Buisson, F., Ileyman, Y., & Josso, N. (1983).Use of monoclonal antibody techniques to study the ontogeny of bovine anti-Mullerian hormone. J. Reprod. Fertil. 6Y, 207-2 14. Vigier, B.,Watrin, F., Magre, S., Tran, D., & Josso, N. (1987).Purified bovine AMH induces a characteristic freeinartin effect in fetal rat prospective ovaries exposed to it in vitro. Development lOO,43-55. Vigier, B., Forest, M.G., Eychenne, B., Rezard, J.. Garrigou, O., Robel, P., & Josso, N. ( 1989). Anti-Miillerian hormone produces endocrine sex-reversal of fetal ovaries. Proc. Natl. Acad. Sci. USA 86,3684-3688.
MIS Function During Sexual Developnwnt
25
Wang, T., Donahoe, P.K.. & Zervos, A.S. (1404). Specific interaction of type I receptors of the TGF-B family with the immunophilin FKBP-12. Science 265, 6 7 4 4 7 6 . Wrana, J.L., Attisano, L., Wieser, R., Ventura, F., & Massague, J. (1994). Mechanism of activation of the TGF-!3 receptor. Nature 370, 341 347. Wilson, C.A., di Cleinente, N.. Ehrenfels, C., Pepinsky, R.B., Josso, N., Vigier, B., & Cate, R.L. ( I 993). Mullerian inhibiting substance requires its N-terminal domain for maintenance of biological activity, a novel finding within the transforining growth factor-b superfainily. Mol. Endo. 7, 247-257.
This Page Intentionally Left Blank
THE ROLE OF THE dpp-GROUP GENES IN DORSOVENTRAL PATTERNING OF THE DROSOPHlLA EMBRYO
Christine Rushlow and Siegfried Roth
Introduction . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . , . . . . . , . . . . . . . . . . . . . . . . . . . . . . B. dl as a Transcriptional Repressor . . . . . . . . , . . . . . V. The dpp-Group Genes . . . . . . . . . . . . . . . . . . . . . . A. dpp is a Signaling Molecule . . . . . . . . . . . . . . . . B. dpp is a Component of a Graded Patterning Process . . . C. dpp is a Morphogen . . . . . . . . . . . . . . . . . . . . D. tolloid Encodes a Potential Protease . . . . . . . . . . . . E. screw Encodes Another Member of the TGFP Superfamily F. Additional Genes Required to Enhance dpp Activity . . .
I.
11. The Dorsoventral Pattern . . . . . . 111. The doixrl Morphogen Gradicnt . . IV. Targets of the cll Gradient . . . . . . A. dl as a Transcriptional Activator
Advances in Developmental Binlogy Volume 4, pages 27-82. Copyright 0 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-969-9
27
. , . . . . . . . . . . . . . . . . . . . . . . . .
28 29 34 37 38 40 42 43 45 48 49 51 53
28
CHRISTINE RUSHLOW and SIEGFRIED ROTH
G. short gastrulation Inhibits dpp Activity . . . . . . . . . . . . 54 13. twisted gustrulation and zen Act Downstream or in Parallel to dpp . . . . . . . . . . . . . . . . . . . . . . 58 VI. Receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 VII. Formation and Function of the dpp Gradient: Summary and Prospectives . . . . . . . . . . . . . . . . . . . . . 65 A. The dpp Gradient is Dependent on, but not Completely Determined by, the d/ Gradient . . . . . . . . . . 65 B. The dl and dpp Morphogen Gradients Establish Different Parts of the DV Pattern . . . . . . . . . . 67 VIII. A Comparison with Vertebrates . . . . . . . . . . . . . . . . . . . 70 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 73 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
1.
INTRODUCTION
Morphogens are regulatory molecules distributed in monotonic gradients. They specify cell fate in a concentration-dependent manner so that cells acquire different fates based on the amount of morphogen to which they are exposed. The existence of morphogens had long been suggested, but only recently have genetic and molecular techniques allowed the visualization of morphogen gradients and the study of their formation and function. The most detailed picture of the role of morphogen gradients in development emerged from work carried out on Drosophilu embryogenesis. Genetic saturation screens have led to the identification of the vast majority of the genes requircd for patterning of the early embryo (Jurgens et al., 1984; Nusslein-Volhard et al., 1984; Wieschaus et al., 1984; Schupbach and Wieschaus, 1986, 1989). The comprehensiveness of this information allowed a detailed mechanistic analysis of the principles that govern the fonnation of the body axes during early embryogenesis. This analysis revealed the importance of morphogen gradients for both the large-scale organization and the pattern refinement along the two body axes. The following general principles seem to apply: ( I ) Pattern fonnation along the anteroposterior (AP) and dorsoventral (DV) body axes occurs largely independent from each other (reviewed in St. Johnston and Niisslein-Volhard, 1992). Mutations affecting one axis normally do not perturb the pattern along the other axis. Therefore, each axis can be studied separately. (2) Maternal and zygotic contributions have to be distinguished. Under the control of maternally supplied products, morphogen gradients are generated with long-range effects on
The dpp-group GeIJeS
29
polarity and pattern of the body axes. In the case of the anterior and DV systems, the gradients consist of transcription factors which regulate the expression of zygotic genes in a concentration-dependent manner (reviewed in St. Johnston and Nusslein-Volhard, 1992). This process converts the information contained in the maternal morphogen gradient into a more refined pattern of gene expression domains. ( 3 ) In some instances, further pattern refinement again employs a morphogen gradient (reviewed in Anderson et al., 1992). The purpose of this chapter is to investigate such a transition from a maternally generated morphogen gradient to a zygotic gradient which controls further pattern refinement. In Drosoplzila, the DV pattern depends on a single, maternally generated gradient, the nuclear concentration gradient of the dorsal ( d l ) gene product (reviewed in Anderson et al., 1992). This gradient has peak levels ventrally and i t directly organizes the pattern on the ventral side ofthe embryo. However, the patterning of the dorsal half of the embryo, though also dependent on the dl protein gradient, occurs more indirectly via the formation of the dpp morphogen gradient with the highest activity levels i n the dorsalmost regions (Ferguson and Anderson, 1992a; Wharton et al., 1993). In the following, we first describe the differences between patterning on the ventral and the dorsal sides of the Drosophila embryo, then the genes involved in the formation and interpretation of the dpp activity gradient.
If. THE DORSOVENTRAL PATTERN The dorsoventral (DV) pattern of the Drosophila embryo can be subdivided into four major regions, the anlagen of which can be traced back to the cellular blastoderm stage (schcniatized in Figure 1). The developmental fate of these regions can be distinguished by virtue of region-specific morphogenetic movements during gastrulation (Figure 1b), by the stereotypic pattern of mitotic divisions during cell cycle 14 (Figure la), and by the various differentiated structures they produce (Figure lc; Campos-Ortega and Hartenstein, 1985; Wieschaus and NussleinVolhard, 1986; Foe, 1989). The ventralmost part of the pattern gives rise to the mesoderm. It is derived from an 18 cell-wide strip of ventrally located cells of the blastoderm embryo (mitotic domain 10, 610) which invaginates in a morphogenetic movement called ventral furrow formation. The mesoderm gives rise to somatic and visceral musculature, fat body, and gonads.
A
A
/
.
-
dorsal ectoderm
- - .- - - - - - - -
-w
ventral ectoderm
\jP
mesectoderm
6
C dorsal folds
dorsal hairs
pole cells
\ I
Fk
-.cephalic furrow
ventral furrow
ventral denticle bands
Figure 7. Fate map of the D V axis. Embryos are oriented so that anterior is to the left and dorsal is up. (A) Lateral (left) and transverse (right) views of the cellular blastoderm embryo. The main subdivisions of the DV axis (left) and the mitotic domains (6)at the beginning of gastrulation (right) are indicated. The dorsalmost region gives rise to the amnioserosa (domain A). Dorsolateral cells give rise to dorsal ectoderm including dorsal epidermis, the peripheral nervous system, and trachea. The dorsolateral regions comprises mitotic domains 19 and 1 1 . Ventrolateral cells give rise to the ventral ectoderm including the central nervous system and ventral epidermis, and comprises domains N and M. The ventralmost region will become mesoderm (domain 10). Abutting the mesoderm, the rnesectoderm (domain 14) forms as a single-cell-wide stripe that becomes the ventral midline after ventral furrow formation. (B) An embryo undergoing gastrulation reveals region-specific morphogenetic movements. The cephalic furrow appears laterally. The ventral furrow forms as a longitudinal cleft along the ventral midline between 20% and 80% egg length. At the posterior end a plate forms which carries the pole cells towards the dorsal side of the embryo when the germ band starts to extend. As the pole cells shift dorsally, the anterior and posterior dorsal folds begin to form. (C) Cuticle preparation of a first instar larva shows the characteristic array of ventral denticle bands (or belts) and the finer hairs of the dorsal epidermis. Also visible are the head skeleton and the Filzkiirper (Fk) which derives from a defined dorsolateral position within the dorsal ectoderni. 30
The dpp-group Genes
31
The ventrolateral region is called neuroectoderm or ventral neurogenic region or ventral ectodenn. It gives rise to both ventral epidermis and central nervous system (CNS), and is derived from an approximately 20 cell-wide strip of cells on either side of the mesoderm. The pattern of mitotic divisions reveals an early subdivision of this region. Abutting the mesoderm, the mesectoderni (domain 14) forms as a single-cell-wide strip that becomes the ventral midline after ventral furrow formation, and generates the specialized glial cells and neurons of the midline (Crews el al., 1988). Lateral to the mesectoderm mitotic domains M and N can be distinguished which contain CNS and ventral epidermis precursors. The embryonic cuticle produced from these regions contains characteristic heavily pigmented ventral denticles (Figure I c). The dorsolateral region (approximately 16 cells wide at cellular blastoderm) gives rise to dorsal epidermis, the peripheral nervous system, and the tracheal system. The cephalic fold which forms during early gastrulation is a morphogenetic marker for both the dorsolateral and ventrolateral regions (Figure 1 b). The dorsolateral region harbors two mitotic domains: domains 1 1 and 19. The latter is a single cell-wide strip which demarcates the region dorsally. The cuticle produced by the dorsal epidermis is characterized by a lawn of fine hairs (Figure lc). In the terminal regions of the embryo, distinct cuticular structures are derived from specific positions within the dorsolateral region. For example, the labrum (the median tooth) and the spiracle hairs (a structure protecting the tracheal openings posteriorly) are derived from a more dorsal position than the antenna1 sense organs and the Filzkoi-per (a tracheal specialization shown in Figure 1c). The dorsalmost region is a five cell-wide domain which gives rise to the amnioserosa, an extraembryonic tissue. As the germband begins to elongate along the dorsal side of the embryo, two dorsal folds (the anterior and posterior transverse folds) appear (Figure 1b). The cells in this region do not divide after cellularization (domain A). They undergo a change in shape from the blastoderin cuboidal to a flat squamous shape, and their nuclei become polyploid and enlarged. The differentiated amnioserosa cells become compressed and displaced laterally during germband elongation. During germband shortening they spread out and provide a covering over the dorsal side. Later, as the embryo undergoes dorsal closure, the amnioserosa is internalized and undergoes apoptosis (Abrams et al., 1993). The morphological subdivisions of the DV pattern are partially reflected in the requirement for specific zygotic gene activities (summarized
Table 7. Dorsoventral Genes’ ~ _ _ _ ~
Embryonic dorsoventral tissues affected
Gene w
h\
MS MS
snail (sria)”* twist (iwi)’?* rhomboid ( r h ~ )also ~ ’ called ~ veinlet (ve) star ( S p 6 spitz pi)^'^ pointed @rif)5,6 single-minded ( ~ i m ) ’ , ~
‘
decapentaplegic (dpp) tolloid (tld) screw ( s ~ w ) ~ ” ~ shrew (srw)I4 short gastrulation (sog)I8
MES, VE MES, VE MES, VE MES, VE MES, VE AS, DE, VE AS, DE AS, DE AS AS, VE
Expression
-
_ _ _ _ _ _ ~ ~ ~ _ ___
Tvpe ofprotein function
zinc finger3 transcription factor basic helix-loop-helix (bHLH) transcription factor4 transmembrane protein7 transmembrane protein’ TGFn-like protein’ ETS-like protein, transcription factor” bHLH transcription factor’ I BMP-214. TGFP-like secreted signaling r n o l e c ~ l e ’ ~ BMP-1, metalloprotea~e’~ TGFP-like secreted signaling molecule1’ not known potential secreted signaling rnole~ule’~
twisted gastrulation (tsg)” zerkniillt (zen)’* Mothers against dpp (Mad)’4 Medea (Med)24 thick veins (tl~l,)~*-~’ saxophone (sax)”
AS AS
z)lg
AS, DE AS, ? AS, DE, VE AS
mat. zyg mat, zyg mat, zyg mat. zyg
w!z
potential connective tissue growth factor (C-lFG)-likeprotein2’ homeodomain transcription factorz3 nor:el protein25
not known 3 M P type I receptor2G3o BMP type I 32
: Key: MS: mesoderm; MES: Inesectudmn; VE: ventral epidennis; DE: dorsal epidcnnis; A: amnioserosa; mat: rriaternal; ~ y g zygotic
’& w
I . Siiiipson (1983).’2.Nusslein-Vblhard et al. (1984). 3. Boulay c t al. (19S7). 4. Thisse ct al. (1988). 5 Mayer and Nus~lciI1-Vnlhard(,1988). 6 . K i l n and Crews (1993). 7. Bier el al. (1990). 8. Kolodkin et al. (1994). 9. Rutlcdgc ct a!. (1992). 10. Kl6rnbt (,1993). 11, Namhu al. (1991) 12. Irish :ind Gclbari (1987). 1 3 . Padgettctal. (1987). 14. Jurgcns e t a ) .(1984). IS. Shimell etal. (1991). 16. Aroraand Niisslein-Volhard (1992). 17. A w n et al (1994). 18. Zusman et al. (1988). 19. FranGois et al. (1994). 2,O. Zusman and Wieschaus (1985). 21. Mason et al. (1994) 22. Wakimoto et a1 (1984). 23. Rushlow et al. (1987a). 21. Raftery et al. (1995). 25. Scketsky et al. (1995). 26. Lindsley and Zimm (1992). 27. Schupbach and Wieschaus (1989).28. Terracol and Lengyel (1994). 29, Nellen etal. (1994). 30. Penton et al. (1994). 3!. Bmmmef et al. (1994). 32. Xie et al. (1994).
Notes: *Included are the mesoderm-forming genes, the spi-group genes, the dpp-genes. and genes that interact with dpp.
34
CHRISTINE RUSHLOW And SIEGFRIED ROTH
in Table 1 ;reviewed in Rushlow and Arora, 1990; Ferguson and Anderson, 1991). The establishment of the mesodermal anlagen requires the two genes, twist ( M i ) and snail (sna; Simpson, 1983). In the absence of twi orstza, the ventral furrow does not form and mesoderm differentiation does not occur. The dorsal and dorsolateral regions ofthe embryo require the activities of seven genes: decapentnplegic (dpp),tolloid (tld),screw (scw),shrew (srw),twisted gastrulation (tsg),short gastrulation (sog), and zerkizullt (zen; Jurgens et al., 1984; Nusslein-Volhard et al., 1984; Wieschaus et al., 1984; Wakimoto et al., 1984). Since the phenotypes caused by mutations in these genes are similar to those produced by dpp alleles of different strength (Wharton et al., 1993), and since, as will be discussed later, they either act to modulate dpp activity, or they function downstream of dpp (Ferguson and Anderson, 1992a), we call these genes collectively, the dpp group (Arora and Niisslein-Volhard, 1992). The genes described so far are either required in the mesoderm or in the dorsal half of the embryo. No zygotic mutations have been isolated which delete the entire ventrolateral region. However, the spitz-group of genes which includes rhomboid (rho; Bier et al., 1990) and singleminded (sim;Crews et al., 1988), for example, are involved in the specification of the mesectoderm and subregions of the ventral epidermis (Mayer and Nusslein-Volhard, 1988; Kim and Crews, 1993). Interestingly, not all cells along the DV axis are committed simultaneously to execute a specific differentiation program. While cells in the ventral half of the embryo are already committed at early gastrulation, the cells of the dorsal half retain a considerable regulative capacity for several hours after gastrulation (Technau and Campos-Ortega, 1987). This might be a reflection of the fact that the pattern i n both halves of the embryo are established in different ways. While the pattern in the ventral half of the embryo is directly dependent on the dl morphogen gradient, the dorsal half of the embryo is organized by a dpp activity gradient (Anderson et al., 1992).
111. THE dorsal MORPHOGEN GRADIENT While the AP axis of the Drosophila embryo is patterned by the independent action of three groups of maternal genes, the DV axis is organized by a single group of maternal genes which functions to establish a single morphogen gradient (reviewed by St. Johnston and NussleinVolhard, 1992; Anderson et al., 1992). These genes are the 11 dorsal-
The dpp-group Gene5
35
group genes and cacfzi.9(cacf).In the absence of the activity of any one of the dorsal-group genes, the resulting mutant embryos lack all DV polarity. They differentiate dorsal epidermis at all positions of the embryonic circumference and entirely lack lateral and ventral structures. This and the analysis of partial loss-of-function alleles demonstrates that the highest requirement for dorsal-group gene activity is at the ventral side and that lower amounts of activity correspond to ventrolateral and dorsolateral positions (Nusslein-Volhard, 1979; Anderson et al., 1985a, 1985b). The dorsal-group genes mediate a signal transduction process (reviewed by Anderson et al., 1992). Spatial information present at the ventral side of the inner egg shell, the vitelline membrane, initiates the formation of a proteolytic cascade which leads to the localized production of an extracellular ligand molecule (Stein et al., 1991). This ligand is a proteolytic fragment of the spiitzle (sp) protein which is released into the fluid-filled space surrounding the embryo, the perivitelline space (Stein and Nusslein-Volhard. 1992; Morisato and Anderson, 1994; Schneider et al., 1994). The proteolytic spz fragment binds to the transmembrane protein Toll (77) which is uniformly present in the plasma membrane surrounding the embryo and functions as a receptor molecule (Hashimoto et ai., 1988; Schneider et al., I99 1 ). Thereby, the TI receptor becomes activated only at the ventral side of the embryo. Activated TI receptor initiates an intracellular signaling cascade which stimulates the releaseofdlfromthe inhibitorcact(R0thet al., 1991; Geisleretal., 1992; Kidd, 1992) and results in the transport of the dl protein from the cytoplasm to the nucleus. Although the quantity of dl protein remains uniform throughout the embryo, a gradient in the subcellular localization is established such that dl protein is mostly nuclear in ventral regions, mostly cytoplasmic in dorsal regions, and partitioned between the nucleus and cytoplasm in lateral regions (Roth et al., 1989; Rushlow et al., 1989; Steward, 1989). A high magnification view of the gradient is shown in Figure 2. It is in the ventrolateral region where the gradient appears steepest. The shape of the nuclear concentration gradient of dl protein is deteimined, at lcast in part, by a diffusion gradient ofthe active spz fragment in the perivitelline space (Roth, 1993; Morisato and Anderson, 1994; Schneider et al., 1994). Nuclear dl protein acts as a morphogen to specify pattern and position along the DV axis (Roth et al., 1989; Jiang and Levine, 1993). Since all the morphogentic effects of the dl protein depend on its nuclear concentration we use in the following the tern1 “dl morphogen” instead of
D
/"
I tld
n sim
[Ill rho
twi sna
scw
Figure 2. The dorsal morphogen gradient regulates D V zygotic gene expression. A high magnification view of the dl gradient is shown on the left. It was derived from a region of one side of a cellular blastoderm embryo (top) that was labeled with anti-dl and rhodamine-conjugated secondary antibodies, Thus, staining appears white. Notice that dlstaining in the ventral region i s predominantly nuclear, while that in the dorsal region is mostly cytoplasmic. In the ventrolateral region staining i s partitioned between the cytoplasm and nuclei; close inspection of this region will reveal the steepest part of the nuclear gradient. The subdivisions of the DV axis are indicated with horizontal lines (MS, mesoderm; MES, mesectoderm; VE, ventral ectoderm; DE, dorsal ectoderm; AS, amnioserosa).The limits, but not the relative levels, of zygotic D V gene expression are summarized as dark rectangles to the right of the dl gradient (see text for references). Expression patterns that refine during gastrulation are summarized as lighter colored rectangles. dl protein differentially activates the expression of twit m a , rho, and possibly sog; dl represses the expression of dpp, tld, Zen, and possibly tsg. 36
The dpp-group Genes
37
“nuclear dl protein.” A correlation between pattern elements and dl morphogen concentrations can be established through the analysis of various maternal-effect mutations which change both the dl moi-phogen distribution and the DV pattern. Especially informative are mutations which cause a uniform distribution of the dl morphogen. They give rise to embryos which are radially symmetric, that is, they differentiate the same cell type(s) at all DV positions (Anderson et al., 1985b). Such apolar embryos can be generated with: ( I ) a unifoimly high dl moiphogen concentration which will differentiate only mesoderm, (2) an intermediate dl concentration differentiating only neuroectoderm, ( 3 ) low dl concentrations differentiating only certain dorsolateral structures, but no dorsal epidermis, or (4) undetectable amounts of dl differentiating dorsal epidermis and amnioserosa (Roth et al., 1989). Thus, the generation of at least these four differentiation states is not dependent on the interaction between distinct parts ofthe DV pattern. Rather, they are determined by specific thresholds of dl morphogen concentrations. Interestingly, these four DV differentiation states do not completely match the subdivisions of the DV pattern based on morphological criteria. The absence of the n’l morphogen elicits cell fates which are derived from two different regions of the wildtype embryo, the dorsolaterally-derived dorsal epidermis and the dorsally-derived amnioserosa (Konrad et al., 1988). In these embiyos, amnioserosa cells are found randomly dispersed among dorsal epidermal cells around the entire embryonic circumference. This observation demonstrates that the pattern elements of the dorsal side ofthe embryo are not entirely determined by distinct dorsal morphogen concentrations. However, the polarization of the dl region which brings the pattern elements into a specific array so that the amnioserosa occupies the dorsalmost position flanked by dorsal epidermis on both sides, is still dependent on the establishment of the d/ gradient on the ventral side. The four differentiation states identified by genetic and morphological means correspond to three dl protein thresholds. However the molecular analysis of dl target genes (discussed later) indicates that additional thresholds exist within the mesoderm and within the neuroectoderm.
IV. TARGETS OF THE dl GRADIENT The dl protein belongs to the rel/NFKB-family of transcription factors (Steward, 1987; reviewed in Rushlow and Warrior, 1992). It regulates the transcription of several DV zygotic genes, that is, genes required to
38
CHRISTINE KUSHLOW and SIEGFRIED ROTH
specify different parts of the DV pattern (reviewed in Ip and Levine, 1992). At a given DV position the transcriptional activity o f a specitic target gene depends on the quantity o f dl protein present in the nuclei at this position, and on the quantity and quality of df binding sites located in its regulatory sequences. Since the d I protein can act as both a transcriptional activator or repressor, its graded nuclear distribution elicits a variety of different target gene expression patterns (schematized in Figure 2). For example, dl protein activates the expression of mi,sna, and rho, and it represses the expression o f d p p , tld, and zen (reviewed in Ip and Levine, 1992). The dl binding sites in the promoter regions ofthesc target genes have been identified, and mutations that disrupt dl binding in vitro have been shown to affect the expression of corresponding promoterlacZ fusion genes in vivo (Ip et al., 199I , 1992a, 1992b; Jiang et al., 1991; Pan et al., 1991;Thisse et al., 1991 ; Rushlow and Warrior, 1992; Huang et al., 1993; Kirov et al., 1993). These studies have revealed why the pattern on the ventral side of the embryo is directly determined by the shape of the dorsal morphogen gradient while the pattern on the dorsal side emerges as a secondary consequence of the gradient. We first give a brief description of the patterning on the ventral side to highlight the differences between ventral and dorsal patterning. A. dl as a Transcriptional Activator
Several dl protein thresholds can be distinguished on the ventral side of the embryo which correspond to specific domains ofgenc expression. The m i and sna genes which are responsible for mesoderm differentiation, are activated by high levels o f dl in a ventral region which IS approximately 18 cells wide (Jiang and Levine, 1993). Both encode putative transcription factors, mi IS a bHLH protein (Thisse et al., 19SS), and sna is a zinc finger protein (Boulay et a]., I987j, and thus, it is likely that twi and sna regulate the expression o f downstrcam target genes. I t appears that m a preferentially acts as a repressor o f gencs involved in the specification of veiitrolateral fates (mesectoderm, neuroectodermj while twi is responsible for the activation of genes required for mesodermal differentiation (Kosinan et al., 1991; Leptin, 1991). Within the mesodermal region, dl has peak levels only in the ventralmost 12-14 cells. Studies by Jiang and Levine (1993) demonstrated that low affinity dl-binding sites present in one part of the twi promoter are
The dpp-group Genes
39
only sufficient for twi expression in the ventralmost 12-14 cells where there are peak levels of dl protein. Expression in the entire mesodermal region requires a combination of low affinity dl-binding sites with bHLH protein-binding sites. The ubiquitously expressed (maternally derived) bHLH proteins, daughterless and T4, a member of the achaete-scute (AC-S) complex, have been shown to be involved in mesoderm specification (Gonzalez-Crespo and Levine, 1993). Furthermore, mi,being itself a bHLH protein, seems to be autoregulative and probably binds to its own promoter. The lateral borders ofsiza and twi expression differ markedly (Kosman et ai., 1991). twi expression diminishes at its borders much like the dl gradient though with a steeper slope. sna however, shows a high level of expression throughout the presumptive mesoderm and diminishes abruptly at the mesoderm-neuroectoderm boundary. A study of the sna promoter has shown that it contains both dl and m i binding sites (Ip et al., 1992a). It seems that dland twi are both activators ofsna transcription and that they function multiplicatively to ensure strong, uniform expression of sna throughout the ventral domain. The sharp sna border may be formed by multiplying the shallow dl gradient with the steeper m i gradient. rho is initially expressed in ventrolateral stripes which encompass the ventral half of the presumptive neuroectoderm (Bier et al., 1990). The ventral limit of the rho stripe abuts the siza border precisely (see Figure 2). The analysis of the rho promoter shows that it contains a small element which harbors both dl and sna binding sites (Ip et al., 1992b). sna binding leads to repression of rho transcription. Although dl is capable of activating rho in the ventral region, sna acts to keep rho transcription off. rho expression extends considerably further toward the lateral side than twi. Thus, it is activated by lower levels of dl protein than those required for twi expression. This response to low levels of dl protein requires not only high affinity dl binding sites, but in addition, as in the case of the mesoderm, bHLH protein binding sites. Thus, like the mesoderm, the specification of the mesectoderm and parts of the neuroectoderm occurs via a cooperation of ubiquitously present bHLH proteins and spatially restricted nuclear dl protein. It is conceivable that similar mechanisms will govern the expression of sog whose transcripts also accumulate in a ventrolateral stripe that extends even further to the dorsal side than that of rho (see Figure 2;
40
CHRISTINE RUSHLOW and SIEGFRIED ROTH
FranCois et al., 1994). In this case, promoter elements might exist which respond to even lower dl protein concentrations. In summary, the patterning of the ventral side of the embryo employs two principles: (1) dl protein acts, in part, in concert with uniformly distributed bHLH proteins as a concentration-dependent transcriptional activator, and (2) some of the dl target genes are themselves transcriptional regulators that enhance or repress the transcriptional activation by dl protein. Thus, the pattern on the ventral side of the embryo appears to be a direct consequence of a small number of gene regulatory interactions. B. dl as a Transcriptional Repressor
As mentioned above, dl also acts as a transcriptional repressor of dpp, tld, and Zen whose activities are required to specify dorsal and dorsolatera1 pattern elements (reviewed in Ip and Levine, 1992). In the wildtype embryo, these genes are initially expressed in a dorsal-on, ventral-off pattern (Figures 2 and 3; Doyle et al., 1986; St. Johnston and Gelbart, 1987; Shimell et al., 1991). In embryos from dl mutant females, their expression domains extend along the entire DV axis (Rushlow et al., 1987; Ray et al., 1991; Shimell eta]., 1991). The uniform expression of dpp, tld, and zen in these mutant embryos corresponds to the apolar, dorsalized phenotype described above. The repression of dpp, tld, and Zen occurs in the ventral and ventrolateral regions. Thus, very low dl protein concentrations can mediate repression. In fact, it has been shown that the strong dl binding sites in the zen promoter have at least a five-fold higher affinity for dl protein than those of the tud promoter (Jiang et al., 1991; Thisse et al., 1991). How can dl protein act as a transcriptional activator and as a transcriptional repressor in the same nucleus, for example, in a ventral position? Regulatory elements in the dpp, tld, and Zen genes that mediate ventral repression have been identified and are called ventral repression elements (VRE; Doyle et al., 1989; Huang et al., 1993; Jackson and Hoffmann, 1994; Kirov et a]., 1994). They were shown to contain dl binding sites, and mutations in these sites abolish ventral repression. However, the dl binding sites themselves are not sufficient to explain repression. dl binding sites alone, present in an artificial promoter, induce transcriptional activation (Jiang et a]., 1992). Thus, dl protein mediates activation through these binding sites when they are taken out of context
The dpp-group Genes
41
of the Zen promoter. Therefore, sequences other than dl binding sites in the VRE must be important for repression. Kirov et al. (1 993) and Jiang et al. (1993) showed that pyrimidine-rich sequences that lie adjacent to the dl binding sites in the zen VRE are required for repression. When these sequences are mutated, but the dl binding sites left intact, dldependent repression is abolished. Moreover, the mutated VRE now confers dl-dependent activation. Thus, by mutating the pyrimidine (T)rich sequences, the silencer element is converted to an enhancer element, and dl mediates activation rather than repression. It was proposed that another molecule, a co-repressor, binds next to dl and together they mediate repression. Recently, Lehming et al. (1994) described a candidate for the corepressor. Using a yeast system to isolate Drosophilu cDNAs encoding inhibitors of dl protein activity, they found an HMG (high mobility group) protein that converts dl, as well as its vertebrate counterpart, NF-KB,from an activator to a repressor. It is not yet known ifthis protein, called DSPl (Drosophila switch protein), binds to the VRE and/or interacts with the dl protein to confer repression. HMG proteins were first isolated as components of chromatin and it is possible that formation of a protein-DNA complex could interfere with the assembly of a stable initiation complex at the transcription start site. In contrast to the genes activated by dl, the genes so far known to be repressed by dl display similar expression domains (Doyle et al., 1986; St. Johnston and Gelbart, 1987; Shimell et al., 1991). dpp, tld, and zen are confined to a domain comprising about 40% of the embryonic circumference, and although there is some gradation of their expression patterns at the dorsolateral border, presumably due to the dl protein gradient, the borders are approximately the same (Figure 3). Maternaleffect mutations which lead to partial dorsalizations of the DV pattern affect the expression domains of all three genes in a similar manner (Figure 3). It seems, therefore, that the patterning of the dorsal side of the embryo is not initiated by a mechanism of differential repression analogous to the differential activation important for ventral patterning. However, as mentioned above, the dl gradient still has some organizing influence on the patterning of the dorsal side. The remainder of this review considers how members of the dpp group-genes establish a zygotic DV morphogen gradient, presumably of dpp activity, that organizes the pattern in the dorsal half of the embryo.
42
CHRISTINE RUSHLOW and SIEGFRIED ROTH
tld
dPP
zen
D1
Figure 3. dl represses tid, dpp, and Zen. wildtype (wt; top) and D1 (bottom) stage 5 embryos were hybridized with RNA probes synthesized from cDNA templates of tld (left), dpp (center), or Zen (right), embedded, and sectioned. In wildtype dlseems to repress tld, dpp, and zen in a similar manner. To enhance potential differences between the expression domains of tld, dpp, and Zen, embryos derived from Tr632/T15BRE females were used for in situ hybridizations. These embryos show a strong expansion of dorsal fates (D1 phenotype according to Anderson et al., 1985), and have therefore enlarged expressiondomains of tld, dpp, and Zen. The tlddomain seems to be slightly larger than the dppdomain, and the dppdomain seems to be slightly larger than that of zen. But even this strong dorsalization does not reveal a qualitative difference in the expression of these three genes, such that, for example, one of them would show a complete derepression (uniform expression).
V. THE dppGROUP GENES
As mentioned above seven zygotic genes, the dpp group, are required for the specification of pattern in the dorsal region of the embryo (Arora and Nusslein-Volhard, 1992; Ray, 1993). Loss of hnction mutants of six genes-dpp, tld, sew, srw, tsg, and zen-cause a ventralized phenotype in which a deletion of dorsal parts of the pattern is compensated by an expansion of lateral and/or ventrolateral regions (summarized in Figure 4). Mutations in the seventh gene, sog, also cause a partial loss of dorsal fates, but in addition, the extent of the ventral ectoderm is reduced. All
The dpp-group Genes
43
WI
VB
M
Figure 4. Fate map changes in the dpp-group mutants. Adapted from Ray, 1993. Circles represent cross sections through the middle of cellular blastoderm embryos. Each embryo is labeled wildtype (wt) or mutant for each of the dpp-group genes. Regions along the DV axis are differentially shaded and represent mitotic domains (top, left; Foe, 1989). The fate map shifts were determined from the analysis of mitotic domains and regionspecific markers (Arora et al., 1992, Ray, 1993; Mason et al., 1994). The dorsalmost region, domain A, i s affected in all dpp-group mutants. Successive dorsal pattern elements are deleted (domains 19, 11) and ventral elements expanded as the mutant phenotype becomes more ventralized; dpp null embryos (bottom, right) lack all dorsal structures.
the members of the dpp group are early-acting zygotic genes so that their effect is already visible shortly after cellularization when the first morphogenetic movements occur. Mutants of dpp-group genes lead to abnormal gastrulation in which the laterally derived cephalic fold is displaced to a dorsal position and germband extension does not occur normally. The following sections are centered around the dpp gene and how other genes affect or respond to its activity. Many of the genes have been characterized at the molecular level, and in some cases vertebrate counterparts have been identified (summarized in Table 1). A. dpp is a Signaling Molecule
The decapentaplegic (dpp) gene is necessary for several different morphogenetic processes during Drosophilu development (Spencer et
44
CHRISTINE RUSHLOW and SIEGFRIED ROTH
al., 1982). Besides its role in DV patterning (Irish and Gelbart, 1987), it is also involved in midgut morphogenesis (Panganiban et al., 1990), eye development (Heberlein et al., 1993; Kaphingst and Kunes, 1994), and in the establishment of the proximodistal axis of adult appendages (Campbell et al., 1993; Basler and Struhl, 1994). In fact, the name dpp derives from its requirement for the normal development of all 15 major imaginal discs. For the purpose of this chapter we focus only on the role of dpp in DV patterning of the early embryo. The dpp gene product is homologous to members of the transforming growth factor-p (TGFP) superfamily of secreted growth and differentiation factors (Padgett et al., 1987; reviewed in Lyons et al., 1992). Thus, dpp does not instruct cell fate based on direct transcriptional regulation of downstream genes as is the case for the dl morphogen. Rather, dpp acts at the level of intercellular communication. In accordance with these molecular findings, genetic mosaic experiments have shown that dpp can act nonautonomously in imaginal discs (Posakony et al., 1990). Furthermore, characterization of dpp activity and protein localization in the embryonic gut has shown that the dpp protein can undergo limited diffusion (Panganiban et al., 1990). Thus, dpp has properties of an extracellular signaling molecule. More recently, cell surface receptors have been identified that appear to mediate the dpp signal (Table 1 and see below; Childs et al., 1993; Brummel et al., 1994; Nellen et al., 1994; Penton et al., 1994; Xie et al., 1994). The members of the TGFP-superfamily are synthesized as large precursor molecules which are cleaved to release a mature carboxyterminal segment of 110-140 amino acids (reviewed in Massague et al., 1994). The active forms are disulfide-linked homodimers or heterodimers of this carboxy-terminal segment. All members show sequence similarity to the prototype, TGF P l , particularly in the active domain where the most conserved feature is the spacing of seven cysteine residues. The superfamily, which currently includes at least 24 members, has several subgroups. The main groups are the TGFPs, the decapentaplegic-Vg-related (DVR) proteins (including the bone morphogenetic proteins, or BMPs), and the activins (reviewed in Kingsley, 1994a). dpp is most related to the vertebrate bone morphogenetic proteins BMP-2 and BMP-4 (about 75% identical in the mature signaling portion of the molecule). In fact BMP-4 is capable of rescuing a dpp mutant embryo (Padgett et al., 1993) and conversely, dpp protein when implanted in rats engenders the same response of ectopic bone formation as shown by all other BMPs (Sampath et al., 1993). This demonstrates
The dpp-group Genes
45
the great degree of conservation between the vertebrate and fly molecules.
B. dpp is a Component of a Graded Patterning Process The characteristic feature of the dpp-group genes is that mutations in any of these genes cause a loss of dorsal pattern elements (reviewed in Ferguson and Anderson, 1991 ). For most of the dpp-group genes multiple alleles exist which can be ordered in an allelic series. The phenotypic effects caused by these alleles have been analyzed on the basis of changes in the cuticle pattern, the patterns of mitotic domains or the expression of molecular markers (Arora and Niisslein-Volhard, 1992; Ray, 1993; Arora et al., 1994; Franqois et al., 1994; Mason et al., 1994). The most complete phenotypic analysis was derived from the large number of embryonic lethal dpp alleles (Wharton et al., 1993). Partial loss-of-function alleles of dpp were ordered in a series in which the weak alleles affect only the dorsalmost structure, the amnioserosa. These weak alleles could be tightly ordered among themselves to show progressive deletion of the amnioserosa by counting the number of amnioserosa cells. Stronger alleles delete, in addition to the amnioserosa, dorsolateral structures. The progressive loss of these structures is compensated by an expansion and a dorsal shift in position of more ventral pattern elements such as the ventral denticle bands (see Figure 5 ) . Finally, the complete lack of dpp causes not only a loss of all dorsal and dorsolateral structures, but also the loss of dorsal parts of the ventral epidermis. The remaining ventral epidermis replaces the deleted structures so that the mutant embryos consist only of ventral epidermis and mesoderm (Figures 4 and 5).
The correlation between allelic strength and the progressive loss of dorsal pattern elements reveals a graded requirement for dpp activity such that high levels of dpp activity specify the amnioserosa while progressively lower levels specify dorsal epidermis and some parts of the ventral epidermis. Furthermore, the phenotypic series comprises a continuum of fate shifts suggesting that the requirement for dpp activity is continuous, and that dpp is part of a system which generates a gradient of positional information. Allelic series exhibiting a continuum of ventralized phenotypes have also been described for other dpp-group genes, such as tld and scw (Ferguson and Anderson, 1992a; Ray, 1993; Arora et al., 1994). The phenotypes are similar if not identical to those caused by dpp alleles of
46
CHRISTINE RUSHLOW and SIEGFRIED ROTH
The dpp-group Genes
47
corresponding strength, indicating that these genes are part of a system that generates positional information in a graded manner. However, even the strongest alleles of any of the other dpp-group genes do not cause phenotypes as severe as dpp null alleles (Figures 4 and 5). tld and sew are required for the production of the amnioserosa and the dorsalmost parts of the dorsal epidermis. Zen, tsg, and SYW, and to some degree sog, are only required for the differentiation of the amnioserosa (Figures 4 and 5). Double mutants between dpp null alleles and alleles of any of the other dpp-group genes do not lead to phenotypes stronger than the dpp null phenotype (Arora and Niisslein-Volhard, 1992). Thus, none of the other dpp-group genes seem to contribute to the DV pattern in a dppindependent way. Another feature which distinguishes dpp from the other dpp-group genes is its dosage-sensitivity (Spencer et al., 1982). dpp is haplo-lethal. Embryos heterozygous for a deficiency of the dpp region are weakly ventralized (Irish and Gelbart, 1987). This suggests that the level of dpp gene product is critical for proper development. Furthermore, increasing the wildtype gene dosage of dpp can shift cell fates along the DV axis, so that more amnioserosa is present with increased dpp copy number (Wharton et al., 1993). A normal diploid embryo has an average of 130 amnioserosa cells; increasing the dpp copy number to four, increases the number to 325.1100. A transformation of epidermal cells into amnioserosa cells can also be observed if the dpp copy number is changed in the background of apolar dorsalized embryos (dl mutant embryos). As
Figure 5. The larval cuticle produced by mutations of dpp-group genes. Dark-field micrographs of wildtype (wt) and mutant embryos. Embryos are oriented so that anterior is up and ventral is front. A. The cuticular phenotype of a wt embryo demonstrating the normal array of ventral denticle bands and the Filzkorper (Fk). B. t/dB4 mutant embryo showing a partially ventralized embryo. The ventral denticle bands are expanded at the expense of the dorsal epidermis and amnioserosa. C. zenw36mutant embryo showing a weakly ventralized embryo. The Filzkorper are abnormal and internalized, the head is not involuted, and the anterior portion of the gut is externalized. D. dppH1n46 mutant embryo showing a strongly ventralized phenotype. Denticle bands encircle the entire DV circumference at the causes a expense of dorsal epidermis and amnioserosa. E. dpph1n-r27/dpph1n-r92 less severe phenotype resemblingthat of strong tldalleles (compare E and B). F n-r4/dpph In-r2 7 causes a weak ventralization, similar to that of zen mutations (compare F and C).
48
CHRISTINE RUSHLOW and SIEGFRIED ROTH
mentioned earlier these embryos differentiate both dorsal epidermal cells and amnioserosa cells if they have a wildtype dosage of dpp. They differentiate exclusively dorsal epidermis if they have only one copy of dpp, and exclusively amnioserosa (in the segmented body region) if they have four copies. This demonstrates that at least the distinction between amnioserosa and dorsal epidermis depends critically on the level of dpp activity. The experiments mentioned so far reveal that dpp is a crucial component of a system which generates an activity gradient with peak levels in the dorsalmost region. C. dpp is a Morphogen
Ferguson and Anderson ( I 992b) showed that the injection ofdpp RNA into blastoderm embryos can rescue the dpp mutant phenotype. They also demonstrated that dpp RNA injections into wildtype embryos can cause a dorsalization of the pattern. A dorsalization can even be achieved if apolar embryos are used as recipients which normally differentiate only ventral epidermis. These apolar embryos do not express some of the dpp-group genes (see below) indicating that dpp can become active even in the absence of other dpp-group genes. In these experiments, the degree of dorsalization depends on the amount of injected RNA. A steep dose response relationship was observed: two to fourfold increases in the concentration of injected RNA elicit progressively more dorsal cell fates. High levels of dpp drive amnioserosa development, intermediate dpp levels drive dorsal epidermal development, and low levels of dpp permit development of ventral ectoderm. Moreover, local dpp activity can also direct a surprising degree of patterning within the entire ectoderm. Local injections of dpp RNA into apolar embryos polarize the gastrulation movements and produce the array of DV structures typical for the wildtype pattern, including dorsal hairs, lateral denticles, and trapezoidal denticle bands. Thus, in these experiments dpp can define both embryonic polarity and organize detailed pattern. The fact that dpp can function at least partially independently from other dpp-group genes, and that local RNA injections induce DV patterning in a dose-dependent manner strongly suggests that dpp itself acts as a morphogen. More support for this idea comes from further genetic experiments. The similarity of the ventralized phenotypes caused by mutations in dpp and other members of the dpp-group seems to preclude the unambiguous hierarchic ordering of these genes. However, a relationship among the
The dpp-group Genes
49
dpp-group genes could be established on the basis of gene dosage studies (Ferguson and Anderson, 1992a). For example, doubling the dpp gene dosage completely suppresses the phenotypes of weak tld mutants and partially suppresses tld null mutants. This result is significant because it demonstrates that tld function is not absolutely required for dpp activity. In contrast doubling the tld gene dosage does not change the dpp null phenotype. Because an increase in the dosage of dpp can partially bypass the requirement for tld, but an increase of tld gene dosage does not suppress the dpp mutant phenotype, it appears that tld acts to elevate the activity of dpp. Similar results were found for srw and scw, indicating that they are also upstream of dpp and function to increase dpp activity (Ferguson and Anderson, 1992a; Arora et al., 1994). In contrast, thezen and tsg phenotypes cannot be suppressed by additional dpp copies which places them downstream or in parallel to dpp (Ferguson and Anderson, 1992a). Interestingly, increasing the dpp gene dosage in a sog mutant background leads to a new phenotype (Ferguson and Anderson, 1992a). As mentioned above, among the dpp-group genes, sog is unusual since sog mutants show a reduction of dorsalmost fates and ventrolateral fates at the same time (see discussion below). If the dpp gene dosage is increased in a sog mutant background, the ventrolateral region (the neuroectoderm) is even further reduced while the region producing dorsal epidermis is expanded. This indicates that sog might normally inhibit dpp activity in ventral regions. In summary, these genetic and injection experiments strongly suggest that an activated form of dpp forms a morphogen gradient in the dorsal half of the embryo. This gradient is responsible not only for determining cell fate and spatial pattern in dorsal regions, but it also has an organizing influence on the neuroectoderm (Ferguson and Anderson, 1992a). The other dpp-group genes are either upstream of dpp and modulate its activity in a positive (tld, SCW, and s w ) or negative (sog) way, or they are potential targets of dpp function (zen, tsg). Since all but one of the dpp-group genes have been cloned, and more recently, dpp receptors have been identified, a molecular picture emerges of how the dpp activity gradient is established. D. tolloid Encodes a Potential Protease
The genetic data which suggest that tld acts to enhance dpp activity find support from the molecular analysis of the tld gene (Shimell et al.,
50
CHRISTINE RUSHLOW and SIEGFRIED ROTH
1991). tld transcripts accumulate like dpp transcripts in the dorsal 4050% of the blastoderm embryo (Figure 3). Thus, tld and dpp products colocalize in the dorsal half of the embryo. tld encodes a protein of an approximate molecular weight of I16 kD which possesses an N-terminal signal sequence. It is therefore, as dpp, most likely a secreted protein. The tld protein has three sequence motifs: an N-terminal region with similarity to the astacin family of metalloproteases, two EGF-like repeats, and five copies of the CUB repeat which was first found in the human complement proteins C 1r and C 1s. However, most significantly, the overall structure of the tld protein is similar to the human bone morphogenetic protein 1 (BMP-I) (Wozney et al., 1988). tldand BMP-1 are 4 1% identical. The BMPs were originally isolated as components of a protein complex that can induce ectopic bone morphogenesis in rats (Wang et al., 1988; Wozney et a]., 1988). Besides BMP-1, all other so far identified BMPs are TGFP family members (Wozney et al., 1988; Celeste et al., 1990), and as mentioned above two of these, BMP-2 and BMP-4, are 75% identical in their active carboxy terminus to the dpp protein. Since the mammalian homologues of tld and dpp (BMP-1 and BMP-2/4) are present in a complex, it is likely that tld and dpp proteins also physically associate. The biochemical function of the tld/BMP- 1 protease is currently unknown. It is not clear whether BMP- 1 is the elusive maturation enzyme responsible for cleaving the pro-forms of TGFP-type molecules. Most TGFP-type molecules are cleaved at a site which suggests that a subtilisin-like activity is responsible for the processing event (Ozkaynak et a]., 1992; Barr, 1991). It could be that additional processing sites exist which modify the activity of TGFP-type molecules. Alternatively, the tldiBMP- 1 protease could function in a later step such as the liberation of the mature peptide from the latent complex. Due to the difficulty to obtain biochemical data on BMP-I in mammalian systems, the genetic and molecular analysis of the tld locus b m m e s especially important. The tld alleles show a complex complementation behavior (Ferguson and Anderson, 1992b). Some of the alleles are antimorphic. Embryos heterozygous for weak dpp alleles and antimorphic tld alleles die with a weakly ventralized phenotype. In contrast, embryos carrying a deficiency of the tld locus and the same dpp allele are hlly viable. If tld acts as a protease, then the antimorphic alleles could encode products that are still able to form a complex with dpp, although they are proteolytically inactive and consequently render the complex inactive. Childs and O’Connor (1993) have sequenced the antimorphic
The dpp-group Genes
51
tld alleles and find, in fact, that the molecular lesions cluster in the protease domain. While the early expression pattern of tld RNA closely resembles that of dpp RNA, the post cellularization pattern of tld transcripts diverges from that of dpp (Shimell et al., 1991). Maybe, tld has additional dpp-independent functions and conversely, some of the later dpp f h c tions may not require tld. This is especially clear for dpp’s function in imaginal disc development during late larval and pupal stages, since the temperature sensitive period for tld function does not extend past the cellular blastoderm stage (Ferguson and Anderson, 1992a). Interestingly, two other genes have been discovered in Drosophila which are homologous to tld. They have been named tolloid-related-] (tlr-I) and tolloidrelated-2 (tlr-2; Nguyen et al., 1994). tlr-1 is located immediately proximal to the tld gene and is 62% identical to tld. It is required during larval and pupal stages of development and thus may augment some of the later dpp functions.
E. screw Encodes Another Member of the TGFV Superfamily
Like tld, scw enhances dpp function, but is not absolutely required for dpp activity (Arora et al., 1994). This is evident from the fact that the scw mutant phenotype can be rescued by injection of dpp RNA. Thus, scw does not act downstream of dpp, and the dpp signal transduction pathway is functional in scw mutant embryos. scw encodes a novel BMP-like member of the TGFP superfamily (Arora et al., 1994). It shares 40% identity with dpp in the carboxy-terminal active region. The greatest sequence conservation between scw and a vertebrate family member is seen with BMP-6 (49%). This is considerably less than the conservation between dpp and BMP-2 and BMP-4, suggesting that scw is not an ortholog of a known vertebrate member of the TGFP superfamily. Unlike tld which is expressed only on the dorsal side of blastoderm embryos, the scw gene is ubiquitously expressed during early stages of embryogenesis (Arora et al., 1994). Moderate levels of scw transcripts are first detected in a stage 4 embryo toward the end of nuclear cycle 10. The level of scw RNAs rapidly increases during cycles 11-1 2 until the cellular blastoderm stage (stage 5 ) when they decline very rapidly to below detection. Although ubiquitous, Arora et al. demonstrated that scw expression on the dorsal side is suficient for normal development by testing a tld-promoter-scw fusion gene in a scw mutant background. The
52
CHRISTINE RUSHLOW and SIEGFRIED ROTH
lack of scw transcripts in ventral regions had no developmental consequences. This suggests that scw activity is restricted to dorsal regions. It is possible that the activity of scw is restricted by translation of scw message only in dorsal cells or by postranslational modification of scw protein in a limited region of the embryo. For example, the dorsally localized tld protein could interact with either scw or dpp, or with both proteins. scw and tld have identical loss-of-function phenotypes. Thus, scw activation, though not dpp activation, could be entirely meditated by tld. Given the structural similarity of scw and dpp, it is possible that scw functions as a signaling molecule by forming heterodimers with dpp (Arora et al., 1994). Protein-protein interactions between dpp and scw are consistent with the observation of genetic interactions. Specific gain-of-function scw alleles have been identified that fail to complement a recessive, partial loss-of-function dpp allele (Raftery et al., 1995). Embryos carrying a single copy of both mutations die with a partially ventralized phenotype. In contrast, embryos carrying a deficiency for the scw locus and the same allele of dpp are completely viable, indicating that the defective product encoded by the gain-of-function scw allele can block the activity of the remaining functional copy of dpp. Thus, these alleles behave like the tld alleles described above as antimorphs. The antimorphic scw proteins may be incapable of signal transduction, but still sequester active dpp molecules and thus reduce the effective levels of dpp in the embryo. On the basis of the putative physical association of dpp and scw proteins, Arora et al. (1994) propose a model which explains why scw is only required to specify amnioserosa and parts of the dorsal epidermis while dpp has a broader requirement. They suggest that the activity gradient that specifies dorsal pattern may be composed of both scwldpp heterodimers and dpp homodimers. scw and dpp could act combinatorily such that scwldpp heterodimers elicit a stronger response than dpp homodimers. scw homodimers would be inactive since they should be formed in a dpp null background where scw appears not to be active. The assumption that,dpp homodimers elicit a response explains why the scw mutant phenotype is less severe than that of dpp. The stronger signaling response elicited by heterodimers may result from a higher affinity of the heterodimer for a common set of receptors. Besides scw, one other TGFP-like molecule has been discovered in Drosophila and named by its chromosomal localization, 60A (Wharton et al., 1991). dpp and 60A are 55% identical to each other. While dpp is
The dpp-group Genes
53
more closely related to BMP-2 and -4 (75% identity), 60A seems to be related to BMP-5, -6, and -7 (70% identity). Like dpp and scw, 60A is expressed during embryogenesis, however, its pattern of expression is broader than that of dpp (Wharton et al., 1991). The function of 60A is not known since mutations in the locus have not yet been reported. A first hint for a role in development comes from ectopic expression studies. Although both dpp and 60A can induce ectopic bone formation in rats, they elicit different responses if they are ectopically expressed during Drosophila development (Staehling-Hampton et al., 1994). While ectopic expression of dpp induces the formation of dorsal epidermis in ventrolateral positions, ectopic 60A expression has no detectable effects on embryogenesis. Thus, 60A seems not to be involved in DV patterning, nor can its ectopic expression interfere with the function of endogenous dPP. F. Additional Genes Required to Enhance dpp Activity One of the dpp-group genes, shrew (sw) has not yet been cloned. In contrast to tld and SCW, it is required only for the highest levels of dpp activity, since null mutations of s w affect only the amnioserosa (Figure 4;Ferguson and Anderson, 1992a). An increase in the dose of dpp can bypass the requirement for the sw gene. Thus, like tld and scw, snv acts upstream of dpp to increase dpp activity. Genetic screens have been performed to isolate other posi tive-acting components of the dpp signaling pathway (Ferguson and Anderson, 1992a). The assumption was made that such components might be rate-limiting, and should, therefore, enhance the dpp mutant phenotype. Interestingly, in a screen for further zygotic mutations, only alleles of the already known dpp-group genes, tld and scw, were isolated. This reinforces the notion that tld and scw are pivotal elements of the dpp activity gradient. Screens were also performed looking for maternal enhancers of the dpp phenotype, which led to the discovery of two new loci, Mothers against dpp (Mud) and Medea (Med; Raftery et al., 1995). Mutations in these genes cause pupal lethality and exhibit phenotypes which resemble those caused by mutations affecting late dpp functions (gut defects, imaginal disc defects). This indicates that Mud and Medencode rate-limiting components integral to dpp pathways throughout development. Although in the maternal screens multiple alleles of Mud and Med have been isolated, saturation was not achieved since two other maternal genes
54
CHRISTINE RUSHLOW and SIEGFRIED ROTH
have been found which enhance the dpp phenotype. These are thick veins ( t h ) and saxophone (sax), both encoding potential dpp receptors (see below). It has not been determined so far whether Mad and Med act upstream or downstream of dpp. MAD has been cloned and characterized and does not contain informative sequence motifs (Sekelsky et al., 1995). However, given that the MAD protein lacks a secretion signal sequence or transmembrane domain, it seems likely that MAD acts within the cells that produce dpp or within the cells that receive the dpp signal. G. Short Gastrulation Inhibits dpp Activity
Among the dpp-group genes, sog is unusual since mutations cause amnioserosa defects and, at the same time, reduce the extent of the ventral epidermis (Figure 4; Ferguson and Anderson, 1992a). Mutations in other dpp-group genes cause either an expansion of the ventral epidermis or do not affect the ventral epidermis. Surprisingly, the dorsally-located defects caused by sog mutations are due to a requirement of sog activity in ventral regions of the embryo as defined by genetic mosaic studies (Zusman et al., 1988). A better understanding of the sog phenotype resulted from a manipulation of the dpp gene dosage in a sog mutant background. Increasing the dpp gene dosage in embryos with normal sog function affects the amount of amnioserosa, but leaves ventral fates largely unchanged (Wharton et al., 1993). However, sog mutant embryos with three copies of dpp already have a dramatically reduced ventral epidermis (Ferguson and Anderson, 1992a), and with four copies of dpp, they virtually lack all ventral epidermis (shown in Figure 6d). In these embryos the dorsal epidermis is expanded at the expense of the ventral epidermis. This suggests that sog acts to inhibit dpp activity in the ventrolateral region where the genetic mosaic experiments previously located its requirement. An inhibitory action of sog on dpp also explains why sog mutations, unlike any other dpp-group mutation, can rescue the dpp hap1o:insufficiency to a significant extent (Frangois et al., 1994). Although the defects on the dorsal side of sog mutant embryos result ultimately in a loss of dorsalmost cell fates (amnioserosa), initially, the dorsalmost region seems to be expanded (Rushlow and Levine, 1990; Ray et al., 1991; Frangois et al., 1994). This can be demonstrated by changes in the dorsal rho expression domain. Besides the ventrolateral expression domains of rho described above, rho is also expressed in a
The dpp-group Genes
55
dorsal domain which is L T l O cells wide in wildtype embryos. In sog mutant embryos, the dorsal rho domain is 20 cells wide indicating an expansion of dorsalmost fates (Bier et al., 1990). However, normal pattern refinement of dpp and zen (see below) does not occur in this expanded region (Rushlow and Levine, 1990; Ray et al., 1991). Thus, sog seems to be required for the normal subdivision of the dorsal region which is a prerequisite for the establishment of the amnioserosa. Interestingly, while the ventral defects of sog are enhanced by extra copies of dpp, the dorsal defects are suppressed (Ferguson and Anderson, 1992a). Thus, in contrast to sog’s inhibitory action on dpp at the ventral side, sog seems to enhance dpp activity at the dorsal side. These observations suggest that sog is responsible for both the dorsal maximum and the lateral extent of the dpp activity gradient. Thus, sog may redistribute dpp activity to keep it high in dorsal regions and low in ventrolateral regions (Ferguson and Anderson, 1991). Recently the sog gene has been cloned (Franqois et al., 1994). As predicted from genetic mosaic data, sog transcripts accumulate in a broad ventrolateral stripe, 14-1 6 cells wide, during cycle 14. The sog stripe is broader than the ventrolateral rho stripe; it extends one to two cells beyond rho ventrally, and 4-6 cells beyond rho dorsally (see Figure 2). The dorsalmost cells expressing sog abut the ventralmost cells expressing dpp, although the boundary between the sog and dpp expression domains is not absolute, as a few cells express low levels of both transcripts. Expression of sog is progressively lost dorsally during the late blastoderm stage and during early gastrulation. By germband extension sog transcripts are confined to the mesectoderm. sog encodes a protein with a single long hydrophobic domain beginning 56 amino acids from the amino terminus (Franqois et al., 1994). This domain could either serve as a transmembrane domain or as an internal signal sequence. Thus, it is not clear if sog remains membranebound or whether it is secreted. The putative extracellular portion of the sog protein contains four repeats of a novel motif defined by the spacing of 10 cysteine residues (CR 1-CR4). These repeats are distantly related to domains present in thrombospondin, procollagen, van Willebrand factor, laminins and genes of the CEF-1O/CTGF/PlG-M2 family. Some of these cysteine-rich repeats present in sog are separated by dibasic pairs of amino acids which could serve as cleavage sites for tyrosine-like serine proteases. Thus, they might be released as diffusible peptides following proteolytic processing. This would explain the non-autonomous genetic behavior of sog mutations. The peptides could diffuse
56
CHRISTINE RUSHLOW and SIEGFRIED ROTH
4 x dpp'
A
-
sog 4 x dpp'
The dpp-group Genes
57
toward the dorsal side of the embryo and possibly interact with dpp directly. Support for this model comes from studies on procollagen; a region of the procollagen domain in type IV collagen and thrombospondin has been shown to bind TGFP (Paralkaret al., 1991; Murphy-Ullrich et al., 1992). Moreover, the procollagen domain is biologically active as a soluble factor. To explain the sog mutant phenotype one may assume that sog binds dpp and thereby limits its diffusion toward the ventral side. Preventing dpp diffusion may also be necessary to maintain the highest dpp levels dorsally. Alternatively, .Tog may interact with other dpp-group genes, with the dpp receptors or activate its own receptors. Interestingly, although sog does not encode a transcription factor, sog exerts an influence on dpp transcription during the late cellular blastoderm stage (Figure 6; Roth, unpublished results). The dpp expression domain is slightly expanded to comprise 50% of the egg circumference in sog mutant embryos rather than 40% as in wildtype (Roth, data not shown). If in asog mutant background, the dpp gene dosage is increased, an additional expansion of the dpp expression domain occurs (Figures 6b and 6c) which is not observed if the dpp dosage is increased in a wildtype background (Figure 6a). sog mutant embryos with four copies of dpp show high levels of dpp transcripts along the entire embryonic circumference except in the presumptive mesodermal region. This corresponds to the complete loss of the ventral epidermis (Figure 6d). Taking into account the molecular data, sog’s influence on dpp transcription must be indirect. If sog has its own receptor, a consequence of sog signaling could be the transcriptional repression of dpp. However, it is also possible that dpp activates its own transcription and that sog
figure 6. sog inhibits dpp in ventral regions. Embryos with four copies of the dpp gene that were otherwise wt (A) or mutant for sog (B-D) were hybridized with a dpp antisense R N A probe (A-C) or aged for cuticle preparations (D).Lateral views (A,B,D) or ventral views (C) are shown. The dpp expression pattern is slightly expanded (to 50%) in an embryo with four copies of dpp (A), compared to an embryo with the normal two copies (40%; see Figure 3). However, in the absence of sog, this pattern expands greatly to encompass 80% of the circumference, excluding only the presumptive mesodermal (B,C). Thus, the ventrolateral region of the embryo now expresses dpp. This results in a fate change of ventral ectoderm to dorsal ectoderm, and the embryo is entirely covered with Thus sog acts to inhibit dpp transcription in ventrodorsal epidermis (D). lateral regions.
58
CHRISTINE RUSHLOW and SIEGFRIED ROTH
inhibits this auto-regulation of dpp, for example by restricting the diffusion of activated dpp protein. Recently, a Xenopus gene, chordin, was described which seems to be an ortholog of sog (Sasai et al., 1994). sog and chordin have the same overall protein organization. The four cysteine repeats are present in the same relative positions in the two proteins. In addition, the cysteine repeats are the regions with highest similarity between the proteins (e.g., 47% identity for the first repeat; FranGois and Bier, 1995). Interestingly, chordin seems to play a similar role in the DV patterning of Xenopus as sog does in Drosophila. This observation yields a new perspective for a comparison of DV patterning in insects and vertebrates (see below). H. fwisfedgasfrubfion and zen Act Downstream or in Parallel to dpp
Extra copies of the dpp gene are unable to suppress tsg or zen mutations, suggesting that they act downstream or in parallel to dpp, rather than upstream like tld, scw, and S M (Ferguson and Anderson, 1992a). There are several differences in the effects that tsg and zen mutations have on dorsal cell fate determination when compared with mutations in the other dpp-group genes. They affect cell fate of the dorsal midline cells (presumptive amnioserosa and dorsalmost regions of the head) but not of the dorsal ectoderm (Figure 4; Wakimoto et al., 1984; Zusman and Wieschaus, 1985). Using several criteria to determine the fate of cells in mutant embryos (cuticular phenotypes, mitotic domains, and specific cell marker expression patterns), there appears to be a shift in fate of dorsal midline cells to dorsolateral fates (Arora and NiissleinVolhard, 1992; Mason et al., 1994). There does not appear to be an accompanying expansion of the more ventrolateral regions, as is the case in tld, scw, and dpp mutants (see Figure 5 ) . Thus, it seems that zen and tsg are dorsal midline specific genes which act in concert with, or in response to, the highest levels of dpp activity. The tsg gene has recently been cloned and characterized (Mason et al., 1994). Initially tsg transcripts are expressed in the dorsal half of the embryo (see Figure 2), but not uniformly along the AP axis. tsg transcripts appear in two domains along the AP axis, a broad domain in the middle of the embryo that looks like a saddle over the dorsal midline, and an anterior cap. This pattern quickly refines during cellularization; transcripts disappear from the dorsal midline from both domains to give a bilaterally symmetrical pattern which refines into a series of four
The dpp-group Genes
59
diffuse stripes along the AP axis. Transcripts fade during the rapid phase of germband elongation. Curiously, the expression pattern at the end of cellularization does not include all regions of the fate map that are affected by tsg mutations. Some of the dorsal midline cells whose behavior is affected in tsg mutants are more than I0 cells removed from the tsg-expressing domains after cellularization. It is possible that the early low level tsg gene products present at the beginning of cellularization are functional. Alternatively, the protein product may act as a signal to surrounding cells or be part of a signal transduction process. The lack of cell autonomy in mosaic studies (Zusman and Wieschaus, 1985) supports the latter view. The predicted tsg protein appears to be secreted since a potential signal sequence is present at the amino-terminus (Mason et al., 1994). A homology search identified human connective tissue growth factor (CTGF; Bradham et al., 1991), and other related proteins, as sharing a limited region of similarity with tsg in the amino half of the protein. The distinguishing features of this family of proteins is an overall structure exhibiting cysteine-rich amino and carboxy-terminal domains flanking a cysteine-free nonconserved central core. The carboxy-terminal region of one of the cysteine-rich domains in the CTGF family of proteins is related to the cysteine-rich repeats in sog. Two facts about CTGF make the homology with tsg interesting. First, CTGF is a powerful chemoattractant (Bradham et al., 1991). The cytoskeletal rearrangements induced by CTGF during chemoattraction may be similar to those that occur to dorsal cells during amnioserosa differentiation. Second, CTGF is active in mammalian cells that are becoming committed to bone formation, as are the BMP proteins (Mason et al., 1994). Thus dpp, tld, SCW, and tsg act in concert as do BMP-I, BMP2/4, CTGF, and CYR61, a relative of CTGF, in their respective developmental processes. As mentioned before, zen is initially expressed in the dorsal half of the embryo (Doyle et al., 1986; see Figure 2). The factors responsible for its activation (as well as that of dpp and tld) are unknown, but assumed to be general transcriptional activators. Zen is not expressed in ventral regions due to ventral repression by dl proteins (Rushlow et al., 1987). During cellularization, zen RNA and protein refine into a stripe along the dorsal midline about 5-6 cells wide. The length of the stripe extends from the optic lobe region to the presumptive posterior midgut. It is not clear if Zen protein functions early in cellularization or only after refinement into the dorsal stripe. zen is the only dpp-group gene
60
CHRISTINE RUSHLOW and SIEGFRIED ROTH
that encodes a transcription factor. The Zen gene is located in the Antennapedia complex on the third chromosome, and like the other genes in the complex, it contains a homeobox (Doyle et af., 1986; Rushlow et al., 1987). The zen protein is a sequence specific DNA binding protein and was demonstrated to function as a transcriptional activator in transient cotransfection assays (Hoey and Levine, 1988; Han et al., 1989). Thus, Zen was speculated to be a direct regulator of downstream genes involved in the differentiation of the amnioserosa. Zen expression is affected by all other dpp-group genes (Rushlow and Levine, 1990; Ray et af., 1991 ). dpp activity is necessary for maintaining Zen transcription in the dorsal half of the embryo. In dpp, tld, and scw mutants, zen transcripts fade from dorsal regions early in cell cycle 14 and protein is never detected. Experiments where dpp transcripts were injected into young embryos showed that high levels of dpp are sufficient to induce zen transcription (Ferguson and Anderson, 1992b). tsg and sog are necessary for zen refinement into the dorsal stripe (Rushlow and Levine, 1990; Ray et al., 1991). In tsg mutants, zen RNA remains broad and fades earlier than normal. In sog mutants, the zen pattern also remains broad but does not fade as early; Zen protein levels do not, however, reach maximal levels as they do in the wildtype dorsal stripe. Thus, tsg and sog have a similar effect on the refinement of zen RNAs. If the dpp activity gradient is responsible for inducing high levels of Zen in the dorsalmost region at the late blastoderm stage, and this high level of zen is required for those cells to undergo differentiation into amnioserosa, then high levels of zen should be sufficient for amnioserosa formation. Ectopic expression studies support this idea. Expression of Zen driven by the hsp70 promoter in early gastrulating embryos resulted in an expansion of the amnioserosa (Figure 7; Rushlow, unpublished results). Figures 7a-7c show embryos stained with anti-Kriippel (Kr) antibodies which provides a marker for differentiated amnioserosa cells (Hoch et al., 1990; Jacob et al., 199'1). The amnioserosa of a wildtype embryo contains about 130 Kr-staining cells (Figure 7a; Ray et al., 1991). When Zen is expressed in all regions of a gastrulating embryo, the number of Kr-staining cells increases to 2-3 times the wildtype number (Figures 7b and 7c), reminiscent of the situation where extra copies of dpp are present in an otherwise normal embryo (Wharton et al., 1993). Note that even in the more extreme case (Figure 7c), the amnioserosa does not expand into
The dpp-group Genes
61
cuticle
anti-Kr
wt A
hszen
B
C Figure 7. Ectopic expression of zen causes an expansion of the amnioserosa. Embryos undergoinggerm band extension (stage 7) were stained with anti-Kr and peroxidase-conjugated secondary antibodies (A-C), or aged for cuticle preparations (D-F). Heat shock experiments were performed on transformant embryos carrying an hsp70 promoter-zencDNA fusion gene: 3-4 h old embryos were heat shocked at 38°C for 45 minutes. Note the expansion of the amnioserosa in heat-shocked embryos (B and C; C has more amnioserosa than B) when compared to wt (A). Heat-shocked embryos display defects in dorsal closure and germband retraction which might be caused by the enlarged amnioserosa (E,F). Some embryos have reduced denticle belts (F).
the entire ventrolateral region, that is, ventral denticles are still present (Figure 70. This could be a reflection of the timing of Zen ectopic expression; gastrulation may be too late to change the fate of ventral cells, but not of dorsal cells. Heat shock induction performed earlier than gastrulation had detrimental effects on the embryos. Alternatively, perhaps another factor (tsg?) must also be ectopically expressed in ventral regions in order to induce proper amnioserosa development.
62
CHRISTINE RUSHLOW and SIEGFRIED ROTH
VI. RECEPTORS In order to understand how different concentrations ofdpp elicit different cellular responses, the downstream components of the dpp signaling pathway must be analyzed. A first step in this direction was the molecular cloning of dpp receptors which became possible after receptors for members of the TGFP superfamily had been identified in vertebrates (Childs et al., 1993; Brummel et al., 1994; Nellen et al.. 1994; Penton et al., 1994; Wrana et al., 1994b; Xie et al., 1994). Biochemical studies using mammalian tissue culture cells indicated that members of the TGFP superfamily bind to receptors consisting of multiple components (reviewed in Kingsley, 1994a). Originally three molecular size groups, called receptor type I (53 kD), type I1 (70-85 kD), and type 111 (200-400 kD) were identified. Receptor type 111 turned out to be a membrane proteoglycan that binds the TGFP isoforms and facilitates TGFP binding to the receptor proper. Receptor types I and 11 proved to be the components actually involved in signal transduction. They both encode transmembrane serinehhreonine kinases which are only distantly related to each other (40% amino acid identity). In order to constitute a hnctional receptor, both type I and type I1 receptors have to be present. Different sets oftype I and 11 receptors exist for the different subgroups of the TGFP superfamily. For example, type I and I1 activin receptors are distinct from the corresponding pairs of receptors which bind the TGFPs or the BMPs. Recently a mechanism of activation of the TGFP receptor has been described which shows that type I and I1 receptors fulfill different functions during signal transduction (reviewed in Wrana et a]., 1994a). The type I1 transmembrane serine/threonine kinase shows a constitutive autophosphorylation and is able to bind ligand in the absence of type I receptor. However, no signal transduction occurs if the receptor type I is not present. Only ligand bound by receptor type I1 can be recognized by r w p t o r type I. As a result, a stable ternary complex forms between the ligand and the two receptors. This complex formation enables receptor type I1 to phosphorylate receptor type I. Phosphorylated receptor type I mediates the signal to intracellular targets. The knowledge of the mammalian receptor sequences led to the identification of related DrosophiZu genes via homology cloning. First, the two Drosophila receptors with the closest similarity to vertebrate activin type I and I1 receptors were identified (Childs et al., 1993; Wrana et al., 1994b). These Drosophila activin receptor homologs are expressed
The dpp-group Genes
63
maternally, during embryonic and imaginal disc development. If they are co-expressed in tissue culture cells they bind mammalian activin efficiently. Due to the lack of mutations in the corresponding Drosophilu genes, the developmental function of these receptors has not been elucidated. We currently do not know the corresponding Drosophila ligand(s). When the Drosophila activin receptor type I was co-expressed with a BMP-2 type I1 receptor (C. elegans daf-4) it was unable to bind BMP-2 (Wrana et al., I994b). This and the close relationship of BMP-2 to dpp and 60A suggests that the Drosophila activin receptor type I does not mediate the biological effects of dpp or 60A. Homology screening led to the identification of two further type I receptors which corresponded to the Drosophila genes sax and tkv (Brummel et al., 1994; Nellen et al., 1994; Penton et al., 1994; Xie et al., 1994). A comparative sequence analysis revealed that they do not appear to be orthologous to any currently known type I receptor, nor are they closely related to each other or the Drosophila activin receptor type I. Both receptors, if co-expressed with a BMP-2 type I1 receptor (C. elegans daf-4) bind BMP-2 with high affinity (Brummel et al., 1994; Penton et al., 1994). For tkv, the binding of dpp was also demonstrated (Penton et al., 1994). Thus, both may function as dpp receptors in vivo. Both the sax and tkv receptors are maternally expressed and uniformly present in the early embryo (Affolter et al., 1994; Brummel et al, 1994; Penton et a]., 1994). In later stages of development only sax shows ubiquitous expression. tkv, in contrast, displays a complex and dynamic pattern of expression. During cellularization, tkv transcripts accumulate at the dorsal side of the embryo. Shortly before gastrulation transcripts start to accumulate ventraily in the region which gives rise to the mesoderm. Later, the invaginated mesodermal cells continue to express high levels oftranscripts and, in addition, specific regions in the neuroectoderm, the tracheal placodes and visceral mesoderm accumulate tkv transcripts. The phenotypes caused by mutations in tkv and sax genetically support the assumption that both receptors are involved in dpp signaling (Affolter et al., 1994; Brummel et al., 1994; Nellenet al., 1994; Pentonet al., 1994; Terracol and Lengyel, 1994; Xie et al., 1994). Loss of function alleles of tkv are embryonic lethal and cause cuticular defects. They had been first identified in a screen for mutations affecting the larval cuticle pattern (the gene had been called sluter [slu];Niisslein-Volhard et al., 1984). However, this cuticular phenotype results from a late requirement of tkv during dorsal closure while its earlier requirement during DV patterning
64
CHRISTINE RUSHLOW and SIEGFRIED ROTH
is masked by maternally supplied tkv transcripts (Affolter et al., 1994). Using germ-line mosaics, embryos could be produced which lacked both maternal and zygotic tkv gene products (Nellen et al., 1994). These embryos show a severe ventralization which is indistinguishable from the ventralized phenotype caused by a complete loss of dpp function. Severely ventralized embryos are also produced by females which are transheterozygous for certain strong tkv alleles. Using tkv alleles of different strength the same range of phenotypes can be produced which is displayed by the allelic series of hypomorphic dpp alleles (Terracol and Lengyel, 1994). In contrast to tkv, sax mutations do not appear to be zygotically lethal (Nellen et al., 1994). sax was discovered in a screen for maternal-effect mutations (Schupbach and Wieschaus, 1989). The maternal phenotype of the strongest known sax alleles is a weak ventralization (loss of amnioserosa) which resembles the weakest embryonic phenotypes caused by partial loss of dpp function. Thus, although sax is present uniformly in the early embryo, it appears to be required only to interpret peak levels of the dpp gradient. Genetic interactions between dpp and saxltkv further support the proposed molecular interactions. For example, sax and tkv alleles that encode full-length proteins with point mutations in their kinase domain enhance the phenotype of weak dpp mutations in a dominant negative manner. It seems that these mutant proteins are able to interact with and sequester limiting amounts of dpp protein on the outside of the cells without being able to activate the kinase domain within the cell (Nellen et al., 1994). In addition to defects in DV pattern formation of the early embryo, both tkv and sax mutations display defects in later patterning events (Nellen et al., 1994; Brummel et al., 1994; Penton et al., 1994). Like dpp, tkv and sax are required for normal midgut morphogenesis and for patterning in the imaginal discs. Genetic interactions between dpp and saxltkv have also been reported for these late functions. Thus, tkv and sax might be generic dpp receptors which are required throughout development for the multiple aspects of dpp function. The identification of dpp and activin receptors in Drosophila together with the possibility to perform large-scale genetic screens (Raftery et al., 1995) makes it conceivable that work with Drosophila will help to elucidate the intracellular signaling pathway of TGFj3-like molecules. This pathway has remained largely elusive in studies with mammalian cells. In mammalian tissue culture the expression of several genes is known to be modulated as a consequence of TGFP signaling (reviewed
The dpp-group Genes
65
in Attisano et al., 1994). TGFP can control transcription both positively as in the case of collagen and TGFP itself, as well as negatively in the case of transin/stromlysin and c-myc. The growth inhibitory action of TGFP on human keratinocytes might involve a downregulation of cyclin E expression. The nature of these transcriptional responses appear to be quite complex and although both positive and negative TGFP-response elements have been identified, the chain of events between the reception ofthe signal and changes in transcriptional activity remains to be defined. Only recently, a protein was characterized (FKBP- 12) that seems to specifically interact with the type I TGFP receptor (Wang et al., 1994).
VII. FORMATION AND FUNCTION OF THE dpp GRADIENT SUMMARY AND PROSPECTIVES A. The dpp Gradient is Dependent on, but not Completely Determined by, the dl Gradient
The dpp activity gradient is related to the dl gradient via two distinct levels of negative interactions. First, the dpp gradient is confined to the dorsal side of the embryo by the dl-dependent transcriptional repression of dpp itself and of at least one of its activators, tld (St. Johnston and Gelbart, 1987; Shimell et al., 1991). However, the repression of transcription of dpp and other dpp-group genes is not sufficient to explain the formation of the dpp activity gradient. Although there might be slight differences in the expression domains of dpp and tld (Figure 3), these differences appear to be too small to confer the necessary spatial information for the patterning of the dorsal half of the embryo. A second mechanism has to be postulated by which the dl gradient influences the dpp activity gradient. Since there are no detectable differences in the nuclear dl protein concentrations in the dorsal half of the embryonic circumference, it is likely that such a mechanism is initiated in the ventrolateral region and spreads from there towards the dorsal side (Ferguson and Anderson, 1992a, 1992b; Wharton et a]., 1993). df might activate genes in the ventrolateral region whose products diffuse into the dpp expression domain and antagonize dpp activity. The antagonistic influence on dpp might be achieved in several distinct ways. 1. It could occur at the transcriptional level. However, the dpp transcription seems to be uniform in its expression domain (Figure 3).
66
CHRISTINE RUSHLOW and SIEGFRIED ROTH
2. dpp translation could be controlled such that a dpp protein gradient forms with a high point at the dorsal side. 3. The processing and modification of dpp could be controlled either by proteins which interact with dpp itself or with activators of dpp (tld,scw, or snu). 4. The antagonistic mechanism might even act downstream of dpp. A signaling protein (sog) emanating from the ventrolateral side might bind dpp directly or have its own receptors on the dorsal side. The activation of these receptors could lead to a down-regulation of the dpp signaling pathway. If the latter possibility would apply, a dpp activity gradient would form without an actual gradient of (activated) dpp protein. Although this possibility cannot be excluded, the injection of dpp RNA into Drosophila embryos (Ferguson and Anderson, 1992b) and studies with activin in vertebrates (Green and Smith, 1990; Green et al., 1992; Gurdon et al., 1994) suggest that physical gradients of TGFP-like molecules can exist, and that they can exert morphogenetic functions. Genetic and molecular studies reveal that sog is a candidate for a gene which is activated by dl in the ventrolateral region and which antagonizes dpp activity in a non-autonomous way (Zusman et al., 1988; Ferguson and Anderson, 1992a; FranGois et al., 1994). It remains to be shown whether sog protein is in fact secreted and diffuses into the dpp expression domain. As mentioned above, although sog is not a transcription factor, it represses dpp transcription in the ventrolateral region (Figure 6). This observation might indicate that dpp is auto-regulative so that dpp signaling in the embryo induces dpp transcription. sog could interfere with such an auto-regulative mechanism thereby precluding the spreading of dpp transcription toward the ventral side. Although the dpp activity gradient forms in response to the dl morphogen gradient, its shape seems not to be completely determined by the shape of the maternal gradient. For example, the five cell-wide domain giving rise to the amnioserosa is not determined by a specific dl morphogen threshold. Instead it depends on a specific activity level of dpp (Ferguson and Anderson, 1992b; Wharton et a\., 1993). How can the dpp activity gradient on the one hand depend on the dl gradient, and on the other hand contain more spatial information than the maternal gradient does? Mechanisms involved in pattern-sharpening might include positive and negative autoregulation (Meinhardt, 1982). The analysis of dpp null alleles which initially produce normal amounts of dpp RNA reveals
The dpp-group Genes
67
that dpp is required to maintain its own transcription (Ray et al., 199 1). In addition, the spreading of dpp transcription in sog mutant embryos can be explained by an auto-inductive mechanism. Thus, these two observations are in agreement with a positive autoregulation of dpp. Furthermore, the analysis of the refinement of dpp expression also indicates negative feedback loops (Ray et al., 1991). dpp expression is down-regulated in the presumptive amnioserosa region where its initial requirements are highest. This down-regulation depends on genes which themselves are activated by dpp. For example, in a Zen mutant embryo, dpp refinement does not occur. A more systematic analysis of these regulatory interactions might reveal that a combination of positive and negative autoregulation is required to generate refined spatial information. B. The dl and dpp Morphogen Gradients Establish Different Parts of the D V Pattern
The DV pattern of the Drosophilu embryo seems to be specified by at least two gradients whose functions are already required before gastrulation, the maternal dl gradient which acts ventrally and the zygotic dpp gradient which organizes the dorsal pattern. Ventrally, the activation of twi and SNU is crucial for both the establishment of the mesoderm and the mesectoderm. Dorsally, the dpp activity gradient establishes the domains of amnioserosa and dorsal epidermis. It seems that the ventrolateral region, the neurogenic ectodenn, is the default state of the DV pattern which forms in the absence of either mesoderm or dorsal epidermis specification. This can be inferred form the phenotypic analysis of zygotic DV mutants. Both mesodermal genes and dpp seem to cause a repression ofneuroectodermal cell fates. Thus, in asna twi double mutant the lack of the mesoderm is compensated by a shift of the neuroectoderm toward the ventral side (Leptin and Grunewald, 1990; Kosman et al., 1991; Leptin, 1991; Raoetal., 1991), whileinadppmutant theneuroectoderm expands toward the dorsal side (Irish and Gelbart, 1987; Wharton et al., 1993). It also seems that at least parts of the neuroectoderm do not require dl-dependent gene activation. As described above, the complete lack of dl function leads to embryos which develop only dorsal cell fates since dpp is uniformly expressed along the DV axis. If, in such a dl mutant background, dpp activity is removed, embryos result which develop ventral epidermis along their entire circumference (Irish and Gelbart, 1987; Wharton et al., 1993). Thus, the dorsalized phenotype of
68
CHRISTINE RUSHLOW and SIEGFRIED ROTH
dl mutant embryos is due to the repression of neuroectodermal development by dpp, and neuroectoderm can form even in the absence of nuclear dl protein. Therefore, the neuroectoderm is the ground state from which the dl and dpp gradients define ventral and dorsal pattern elements, respectively. Although both dpp and the mesodermal genes repress neuroectoderm development, they also exert an organizing and inductive influence on some parts of the neuroectoderm. Mesoderm specification leads to the establishment of the mesectoderm which induces ventral parts of the ventral epidermis (Kim and Crews, 1993). In sna twi double mutants both mesoderm and mesectoderm are completely lacking and although the mesoderm is replaced by the ventral epidermis, the amount of ventral epidermis seems to be reduced (Arora and Niisslein-Volhard, 1992). Similarly, dpp is not only repressing ventral ectodermal fates, but it seems to be required, as discussed above, for the formation and organization of dorsal parts of the ventral ectoderm (Wharton et al., 1993). Since an organizing influence on the neuroectoderm emanates from both the ventral and the dorsal sides, it is interesting to study embryos which at the same time lack the mesoderm and the dorsal ectoderm. Such embryos result if dpp, sna, and twi gene activities are simultaneously missing (Figure 8; Roth, unpublished results). The triple mutant embryos show an expansion of the neuroectoderm along the entire DV axis (Figure 8a). In this respect they resemble embryos derived from a dl, dpp double mutant (Irish and Gelbart, 1987). However, in contrast to those, they have polarity during gastrulation. The triple mutant embryos (Figure 8c) exhibit a polarized germband extension. Since these embryos have normal dl gradients, they show localized gene expression before gastrulation. For example, rho is expressed in a broad domain which is continuous on the ventral side since sna-dependent repression of rho does not occur (Figure 8b). However, in later stages rho transcripts completely disappear from the ventral side (Figure 8c), since later rho is expressed in the mesectoderm which is not established in the absence of mesoderm formation (Nambu et al., 1990). It is not known whether the early expression of rho has any function, or whether rho.plays any role in the absence of mesectoderm formation. Similarly, several dpp-group genes (sog, zen, tld) might be locally transcribed in the triple mutant, since their early transcription pattern depends on dl. But analogous to the case of rho, it is not known whether they have any function if dpp is not present. Thus, it is not clear where the residual polarity in the dpp sna twi triple mutant embryo originates. dl might play
The dpp-group Genes
69
dpp' sna- twi'
Figure 8. dpp sna twi triple mutant embryos lack mesoderm, mesectoderm; and dorsal ectoderm. A. Cuticular phenotype of the triple mutant, dPPHin46,* sna"'; Df(2R)twF''. Ventral denticles encircle the embryo. B and C. Stage 5 (B) and 6 (C) embryos were hybridized with rho probes. Mesoderm formation does not occur, rho expression diminishes in the ventral region and, is completely absent ventrally in embryos slightly older than (C). Despite this, the triple mutant embryos show a polarized germ band extension.
70
CHRISTINE RUSHLOW and SIEGFRIED ROTH
a not yet identified role in the patterning of the neuroectoderm, some spi-group genes (e.g., rho) might have early functions independent from mesectoderm specification, and finally, not all functions of the dpp-group genes might be mediated by dpp. Whatever the final answer will be, the residual polarity in the triple mutant indicates that a simple model which invokes only mesoderdmesectoderm specification ventrally and dorsal ectoderdamnioserosa specification dorsally is not sufficient to explain all aspects ofpatterning along the embryonic DV axis. A further analysis of the triple mutant promises new insights in the function of either the spi-group or the dpp-group genes.
VIII. A COMPARISON WITH VERTEBRATES Most of the zygotic DV genes in Drosophila have homologs in vertebrates. For example, twi and sna homologs exist in vertebrates which are expressed in the mesoderm and could play a role in mesoderm specification (Hopwood et al., 1989; Sargent and Bennett, 1990; Wolf et al., 1991; Smith et a!., 1992; Hammerschmidt and Nusslein-Volhard, 1993). The close relationship of dpp and BMP2/4 has been confirmed even with functional assays (Padgett et a]., 1993; Sampath et a]., 1993), and the finding that tld is homologous to BMPl suggests that details of the regulative mechanisms which activate TGFP-like molecules are conserved between insects and vertebrates (Shimell et al., 1991). Originally it seemed that the dpp-group genes and the BMPs were involved in different developmental processes, DV patterning and bone formation, respectively (reviewed in Kingsley, 1994b). Other TGFP-like molecules, only distantly related to the BMPs and dpp, have been implicated in DV patterning in vertebrates. More specifically, they have been shown to be involved in mesoderm induction (reviewed in Kessler and Melton, 1994). In Xenopus, during cleavage stages, cells of the vegetal pole, which will later form the endoderm, induce the overlaying marginal zone cells to form mesoderm. The two best-studied mesoderm-inducing TGFP-like molecules are Vgl and activin. Vgl. is a maternally expressed TGFP-like molecule whose mRNA is localized to the vegetal pole of Xenopus oocytes and cleavage stage embryos (Weeks and Melton, 1987). Although the Vgl precursor protein is abundantly expressed, the cleaved mature form has yet to be detected, suggesting that the processing of Vgl is tightly regulated (Dale et al., 1993; Thomsen and Melton, 1993). A stringent regulation of Vgl processing is also suggested by the observation that Vgl RNA injections into
The dpp-group Genes
71
Xenopus embryos have little or no effects. In comparison, although the full activation of dpp is a highly regulated process during normal development (see above), injection of dpp RNAs into Drosophila embryos always caused a dose-dependent dorsalization (Ferguson and Anderson, 1992b). The phenotypic effects of Vgl could only be investigated using hybrid Vgl molecules consisting of an N-terminal proregion from a BMP and the C-terminal region from Vgl. Such hybrid molecules appeared to be processed after injection, so that mature Vgl was released. These injection studies showed that mature Vgl is a potent inducer of dorsal mesoderm (Dale et al., 1993; Thomsen and Melton, 1993). Dose-response studies have been performed for another mesoderminducing TGFP-like molecule, activin, which showed that different activin concentrations elicit different developmental fates (Green and Smith, 1990). Recently, it was shown using combinations of Xenopus vegetal and animal cap tissues that activin can form a morphogentic gradient (Gurdon et al., 1994). The selection of genes expressed by a cell was found to be determined by its distance from the activin source, indicating that activin regulates gene expression in a concentrationdependent manner. Therefore, the existence of an activin concentration gradient was proposed which comprises at least 10 cell diameters (300um in Xenopus animal pole tissue) and requires a few hours for its formation. This gradient can form by passive diffusion since activin can bypass cells that do not themselves respond to the signal nor synthesize protein. These experiments strongly suggest that morphogen gradients of TGFP-like molecules can form in embryonic tissues of a vertebrate. The same principles should apply for Drosophila tissues. There are no apparent physiochemical reasons to exclude the idea that diffusion can contribute to the formation of a gradient of activated dpp. A closer comparison between DV patterning in insects and vertebrates is implied by the recent discovery that BMP-4 is also a mesoderm indwer, and that it is required for the DV patterning of the marginal zone (Dale et al., 1992; Jones et al., 1992; Fainsod et al., 1994). BMP-4 is both maternally and zygotically expressed in Xenopus. In mesoderm induction assays, BMP-4 induces only ventral mesodermal tissues, in contrast to Vgl and activin which can also induce dorsal mesoderm. In experimental situations, BMP-4 expression modifies the induction by activin, resulting in suppression of dorsal mesoderm formation. Thus, it appears that the ventralizing BMP-4 signal is able to override dorsalizing signals. Studies with a truncated BMP-4 receptor show that in the absence of
72
CHRISTINE RUSHLOW and SIEGFRIED ROTH
BMP-4 signaling, dorsal mesoderm forms even in ventral regions of the embryo (Graff et al., 1994; Maeno et al., 1994). This suggests that ventral is not the ground state for all mesoderm as had been inferred, for example, from the completely ventralized phenotype of UV-treated embryos. Interestingly, Xenopus embryos express, in addition to the close dpp homolog, BMP-4, also the sog homolog, chordin (Franqois et al., 1994; Sasai et al., 1994; Franqois and Bier, 1995). While BMP-4 is expressed in the ventral marginal zone, chordin expression is detected in the dorsal marginal zone. Thus, BMP-4 and chordin are expressed in different and potentially abutting domains. The relationship of their expression domains might be similar to those of dpp and sog in Drosophila. Injection of chordin RNA into Xenopus embryos demonstrates that chordin is a potent dorsalizing factor (Sasai et al., 1994). chordin 5 biological effects seem to be opposite to BMP-4 and thus, chordin might antagonize BMP-4 function, like sog antagonizes dpp function. Corresponding molecules involved in DV patterning in both insects and vertebrates seem to fulfill opposing functions. BMP-4 is ventralizing in vertebrates while dpp is dorsalizing in insects; chordin is dorsalizing in vertebrates while sog is ventralizing in insects. These findings have led to the revival of an old discussion. In 1822, Geoffroy St. Hilaire proposed that the arthropod body plan is like the vertebrate body plan turned upside down (Niibler-Jung and Arendt, 1994). According to this idea the longitudinal nerve cords of insects and vertebrates derive from the same centralized nervous system in their common ancestor. Because the nerve cord is located ventrally in insects and dorsally in vertebrates, the ventral side of insects would then correspond to the dorsal side of vertebrates. To explain how a common ancestor can give rise to both vertebrates and insects (arthropods), an inversion of the DV axis during early chordate evolution was suggested (Arendt and Niibler-Jung, 1994). This inversion must result from a change in the gastrulation behavior. The vertebrate mode of gastrulation can be conceptually derived from that efpolychaete annelids in a way which implies an inversion of the DV axis. However, this inversiun hypothesis is challenged by the existence of a group of hemichordates (enteropneusta) which possess both a ventral and a dorsal nerve cord (Peterson, 1995). According to an alternative, and less radical proposal, vertebrates are derived from ciliated larvae resembling those of modern echinoderms (Lacalli, 1995). This idea accounts for a partial inversion of the vertebrate DV axis relative to that of insects. A solution to these interesting evolutionary questions requires that molecular studies are carried out not only in
The dpp-group Genes
73
Drosophilu and vertebrates, but that t h e y are also extended to primitive invertebrates. However, even w i t h o u t a knowIedge of t h e precise evolutionary relationship, it is already a p p a r e n t that the molecules and mechanisms discovered in Drosophila not only have counterparts i n vertebrates, but that they govern similar developmental processes (for other examples see Scott, 1994).
ACKNOWLEDGMENTS We would like to thank several of our colleagues for valuable and stimulating discussions that enabled us to write this manuscript: Laurel Raftery, Rob Ray, Manfred Frasch, Chip Ferguson, Ethan Bier, Kristi Wharton, Kavita Arora, Mike O’Connor, Liz Mason, Larry Marsh, and Rick Padgett. Manfred Frasch first observed the dpp expression changes in sog mutant embryos with multiple copies of dpp. We thank Rob Ray for help in generating the dpp,sna,twi triple mutant, and Ethan Bier for the hsCasper plasmid. We are grateful to Trudi Schiipbach and Paul Wassarman for their patience. We acknowledge Rob Ray and Kavita Arora for providing parts of Figures 1,4, and 5 . We thank Andreas Bergmann, Stefan Schulte-Merker, and Mary Mullins for critical reading of the manuscript.
REFERENCES Abrams, J.M., White, K., Fessler, L.I., & Steller, H. (1993). Programmed cell death during Drosophilu embryogenesis. Development I1 7.29-43. Affolter, M., Nellen, D., Nussbaumer, U., & Basler, K. (1994). Multiple requirements for the receptor serindthreonine kinase thick veins reveal novel functions of TGFP homologs during Drosophila embryogenesis. Development f20,3 105-3 I 17. Anderson, K.V., Jurgens, G., & Nusslein-Volhard. C. ( 1985a). Establishment of dorsalventral polarity in the Drosophilu embryo: Genetic studies on the role of the Toll gene product. Cell 42,779-789. Anderson, K.V., Bokla, L., & Nusslein-Volhard, C. (1985b). Establishment of dorsalventral polarity in the Drosophila embryo: The induction ofpolarity by the Toll gene product. Cell 42, 79 1-798. . Anderson, K.V., Schneider, D.S., Morisato, D., Jin, Y., & Ferguson, E.L. (1992). Extiacellular morphogens in Drosophila embryonic dorsal-ventral patterning. Cold Spring Harbor Symp. Quant. Biol. 57,40%417. Arendt, D., & Nubler-Jung, K. (1994). Inversion of the dorsoventral axis. Nature 371, 26. Arora, K., & Nusslein-Volhard, C. (1992). Altered mitotic domains reveal fate map changes in Drosophila embryos mutant for zygotic dorsoventral patterning genes. Development 114, 1003-1024.
74
CHRISTINE RUSHLOW and SIEGFRIED ROTH
Arora, K., Levine, M., & O’Connor, M.B. ( 1994). The screw geneencodes a ubiquitously expressed ineinber of the TGF-P family required for specification of dorsal cell fates in the Drosophila embryo. Genes & Dev. 8,2588-2601. Attisano, L., Wrana, J.L., Lopez-Casillas, F., & Massague, J. (1994). TGF-P receptors and actions. Biochem. Biophys. Acta 1222,7140. Barr, P.J. ( 199 I). Mammalian subtilisins: The long-sought dibasic processing endoproteases. Cell 66, 1-3. Basler, K., & Struhl., G. ( 1994). Compartment boundaries and the control of Drosophila limb pattern by hedgehog protein. Nature 368,208-2 14. Bier, E., Jan, L.Y., & Jan, Y.N. (1990). rhomboid, a gene required for dorsoventral axis establishment and peripheral nervous system development in Drosophila melanogaster. Genes & Dev. 4 , 190-203. Boulay, J.-L., Dennefeld, C., & Alberga, A. (1987). The Drosophila developmental gene snail encodes a protein with nucleic acid binding fingers. Nature 330, 395-398. Bradham, D.M., Igarashi, A., Potter, R.L., & Grotendorst, G. (1991). Connective tissue growth factor: A cysteine-rich secreted by human vascular endothelial cells is related to the SRC-induced immediate early gene product CEF-10. J. Cell Biol. 114, 1285-1294. Brummel, T.J., Twombly, V., Marques, G., Wrana, J.L., Newfeld, S.J., Attisano, L., Massague, J., O’Connor, M.B., & Gelbart, W.M. (1994). Characterization and relationship of Dpp receptors encoded by the saxophone and thick veins genes in Drosophila. Cell 7&25 1-261. Campbell, G., Weaver, T., & Tomlinson, A. (1993). Axis specification in the developing Drosophila appendage: The role of wingless,decapentaplegic and the homeobox gene aristaless. Cell 74, 1 1 13-1 123. Campos-Ortega, J.A., & Hartenstein, V. ( 1985). The Embryonic Development OfDrosophila melunogaster. Springer Verlag, Berlin. Childs, S.R., & O’Connor, M.B. (1993). Two domains of the tolloid protein contribute to its unusual genetic interaction with decapentaplegic. Dev. Biol. 162,209-220. Childs, S.R., Wrana, J.L., Arora, K., Attisano, L., O’Connor, M.B., & Massague, J. (1993). Identification of a Drosophila activin receptor. Proc. Natl. Acad. Sci. USA 90,9475-9479. Crews, S.T.,Thomas, J.B.,&Goodman, C.S. (1988). The Drosophilasiiigle-mindedgene encodes a nuclear protein with sequence similarity to theper gene product. Cell 52, 143-15 1. Dale, L., Howes, G., Price, B.M.J., & Smith, J.C. (1992). Bone morphogenetic protein - . Q: ‘A ventralizing factor in Xenopus development. Development I f 5,573-585. Dale, L., Matthews, G., & Colman, A:( 1993). Secretion and mesoderm-inducing activity of the TGF-beta-related domain of Xenopus Vgl . EMBO J. 12,447 1-4480. Doyle! H.J., Harding, K., Hoey, T., & Levine, M. (1986). Transcripts encoded by a hoineobox gene are restricted to dorsal tissues of Diasophila embryos. Nature 323, 76-79. Doyle, H.J., Kraut, R., & Levine, M. (1 989). Spatial regulation of zerkniillt: A dorsalventral patterning gene in Drosophila. Genes & Dev. 3, 15 18-1533. Fainsod, A., Steinbeisser, H., & De Robertis, E.M. (1994). On the function of BMP-4 in patterning the marginal zone of theXenopus embryo. EMBO J. 13,5015-5025.
The dpp-group Genes
75
Ferguson, E.L., & Anderson, K.V. (1991 1. Dorsal-ventral pattern formation in thc Drosophilu embryo: The role of zygotically active genes. Current Topics Dev. Biol. 25, 1 7 4 3 . Ferguson, E.L., & Anderson, K.V. (1992a). Localized enhancement and repression of the activity of the TGF-P family member, decapentaplegic, is necessary for dorsal-ventral pattern formation in the Drosophila embryo. Development 114, 583-597. Ferguson, E.L., & Anderson, K.V. (1992b). Decupeiztuplegic acts as a morphogen to organize dorsal-ventral pattern in the Drosophilu embryo. Cell 71,451461. Foe, V. ( 1989). Mitotic domains rcveal early commitment of cells in Drosophila embryos. Development 107, 1-22. Franqois, V., & Bier, E. ( 1 995). Xenopus chordin and Drosoplzilu short gastrulation genes encode homologous proteins functioning in dorsal-ventral axis formation. Cell 80, 19-20. Franqois, V., Solloway, M., O’Neill. J.W., Emery, J., & Bier, E. (1994). Dorsal-ventral patterning of the Di-mophila embryo depends on a putative negative growth factor encoded by the slzor-t gastrulution gene. Genes & Dev. 8,2602-261 6. Geisler, R., Bergmann, A., Hiromi, Y., & Nusslein-Volhard, C. (1992). cactus, a gene involved in dorsoventral pattern formation of Drosophilu, is related to the I kappa B gene family of vertebrates. Cell 71,6 13-62 I . Gonzalez-Crespo, S., & Levine, M. ( 1993). interactions between dorm/ and helix-loophelix proteins initiate the differentiation of the embryonic mesoderm and neuroectoderm in Drosophilu. Genes & Dev. 7, 1703-1 7 13. Graff, J.M., Thies, R.S., Song, J.J., Celeste, A.J., & Melton, D.A. (1994). Studies with a Xenopus BMP receptor suggest that ventral mesoderm-inducing signals override dorsal signals in vivo. Cell 79, 169-1 79. Green, J.B.A., & Smith, J.C. (1990). Graded changes in dose of a Xenopus activin A homologue elicit stepwise transitions in embryonic cell fate. Nature 347,391-394. Green, J.B.A., New, H.V., & Smith, J.C. (1992). Responses of embryonicXenopus cells to activin and FGF are separated by multiple dose thresholds and correspond to distinct axes of the mesoderm. Cell 71, 731-739. Gurdon. J.B., Harger, P., Mitchell, A,, & Lemaire, P. (1994). Activin signalling and response to a morphogen gradient. Nature 371,487492. Green, J.B:A., New, H.V., & Smith, J.C. (1992). Responses of embryonicXenopu.7 cells to activin and FGF are separated by multiple dose thresholds and correspond to distinct axes of the mesoderm. Cell 7 1 , 73 1-739. Hammerschmidt, M., & Nusslein-Volhard, C. (1 993). The expression of a zebrafish gene homologous to Drosophilu snail suggests a conserved function in invertebrate and vertebrate gastrulation. Development 119, 1107-1 11 8. Han, K.,.Levine, M . S . , & Manley, J.L. (1989). Synergistic activation and repression of transcription by Dvosophilu homeobox proteins. Cell 56, 573-583. Hashimoto, C., Hudson, K.L., & Anderson, K.V. (1988). The Toll gene of Drosophila, required for dorsal-ventral embryonic polarity, appears to encode a transmembrane protein. Cell 52,269-279. Heberlein, U., Wolff, T.. & Rubin, G.M. (1993). The TGF beta homolog dpp and the segment polarity gene hedgehog are required for propagation of a morphogenetic wave in the Drosophilu retina. Cell 75,913-926.
76
CHRISTINE RUSHLOW and SIEGFRIED ROTH
Hoch, M., Schroder, C., Seifert, E., & Jackle, H. (1990). Cis-acting control elements for Kruppel expression in the Di-osophilu embryo. EMBO J. 9,2587-2595. Hoey, T., & Levine, M. (1988). Divergent homeobox proteins recognize similar DNA sequences in Di-osophilu. Nature 332, 858-86 I . Hopwood, N.D., Pluck, A., & Gurdon, J.B. (1989). A Xenopus mRNA related to Drosophilu rwist is expressed in response to induction in the mesoderm and the neural crest. Cell 59, 893-903. Huang, J.D., Schwyter, D.H., Shirokawa, J.M., & Courey, A.J. (1993). The interplay between multiple enhancer and silencer elements defines the pattern of decapentuplegic expression. Genes Dev. 7,696704. Ip, Y.T., Kraut, R., Levine, M., & Rushlow, C.A. (1991). The dorsal morphogen is a sequence-specific DNA-binding protein that interacts with a long-range repression element in Drosophila. Cell 64,439-446. Ip, Y.T., & Levine, M. ( 1992). The role of the dorsul morphogen gradient in Drosophilu embryogenesis. Sem. Dev. Biol. 3, 15-23. Ip, Y.T., Park, R.E., Kosman, D., Bier, E., & Levine, M. (1992a). The dorsal gradient morphogen regulates stripes of rhomboid expression in the presumptive neuroectoderm of the Drosophilu embryo. Genes & Dev. 6. 1728-1739. Ip, Y.T., Park, R.E., Kosman, D., Yazdanbakhsh, K., & Levine, M. (1992b). Dorsal-twist interactions establish snail expression in the presumptive mesoderm of the Drosophila embryo. Genes & Dev. 6, 1518-1530. Irish, V.F., & Gelbart, W.M. (I 987). The decupentaplegic gene is required for dorsal-ventral patterning of the Drosophilu embryo. Genes & Dev. I , 868479. Jackson, P.D., & Hoffmann, EM. (1994). Embryonic expression patterns of the Dtvsophila decupentaplegic gene: Separate regulatory elements control blastoderm expression and lateral ectodermal expression. Dev. Dyn. 199,2844. Jacob, Y., Sather, S., Martin, J.R., & 0110,R. (1991). Analysis ofKruppel control elements reveals that localized expression results from the interaction of multiple subelements. Proc. Natl. Acad. Sci. USAR8,5912-5916. Jiang, J., & Levine, M. ( 1 993). Binding affinities and cooperative interactions with bHLH activators delimit threshold responses to the dorsal gradient morphogen. Cell 72, 741-752. Jiang, J., Kosman, D., Ip, Y.T., & Levine, M. (1991). The dorsal morphogen gradient regulates the mesoderm determinant twist in early Drosophilu embryos. Genes & Dev. 5, 1881-1891. Jiang, J., Rushlow, C.A., Zhou, Q., Small, S., & Levine, M. (1992). Individual dorsal binding sites mediate activation and repression in the Drosophilu embryo. EMBO J. 11,3147-3154. Jiang, J., Cai, H., Zhou, Q., & Levine, M. (1993). Conversion of a dorsal-dependent silencer into an enhancer: Evidence for dorsal corepressors. EMBO J. 12, 3201. 3209. Johes, C.M., Lyons, K.M., Lapan, P.M., Wright, C.V.E., & Hogan, B.J.M. ( 1992). DVR-4 (bone morphogenetic protein-4) as a postero-ventralizing factor in Xenopus mesoderm induction. Development 115,639447. Jurgens, G., Wieschaus, E., & Nusslein-Volhard, C. (1984). Mutations affecting the pattern of the larval cuticle in Drosophilu melunoguster.11: Zygotic loci on the third chromosome. Roux’s Archives of Developmental Biology 193,283-295.
The dpp-group Genes
77
Kaphingst, K., & Kunes, S. (1994). Pattern formation in the visual centers of the Drosophila brain: wingless acts via decapentaglegic to specify the dorsoventral axis. Cell 78,437-448. Kessler, D.S., & Melton, D.A. ( I 994). Vertebrate embryonic induction: Mesodermal and neural patterning. Science 266, 59-04, Kidd, S . (1992). Characterization of the Drosophila cactus locus and analysis of interactions between cactus and dorsal proteins. Cell 71, 623435. Kim, S.H., & Crews, S.T. (1993). Influence of Drosophila ventral epidermal development by the CNS midline cells and spitz class genes. Development 118, 893-90 !. Kingsley, D.M. (1994a). The TGF-beta superfamily: New members, new receptors, and new genetic tests of function in different organisms. Genes & Dev. 8, 133-146. Kingsley, D.M. (1994b). What do BMPs do in mammals? Clues from the mouse short ear mutation. Trends Genet. 10, 1 6 2 1 . Kirov, N., Zhelnin, L., Shah, J., & Rushlow, C. (1993). Conversion o f a silencer into an enhancer: Evidence for a co-repressor in dorsal-mediated repression in Drosophila. EMBO J. 12,3193-3199. Kirov, N., Chiids, S., O’Connor, M., & Rushlow, C. (1994). The Duosophila dorsal morphogen represses the tolloid gene by interacting with a silencer element. Mol. Cell. Biol. 14,713-722. Konrad, K.D., Goralski, T.J.. & Mahowald, A.P. (1988). Developmental genetics of the gastrulation defective locus in Drosophila melanogaster. Developmental Biology 127, 133-142. Kosman, D., Ip, Y.T., Levine. M., & Arora, K. (1 99 I). Establishment of the mesodermneuroectoderm boundary in the Drosophila embryo. Science 254, I 1 8-1 22. Lacalli, T.C. (1995). Dorsoventral axis inversion. Nature 373, 110-1 1 1 . Lehming, N., Thanos, D., Brickman, J.M., Ma, J., Maniatis, T., & Ptashne. M. (1994). An HMG-like protein that can switch a transcriptional activator to a repressor. Nature 371, 175-179. Leptin, M. (199 I). twist and srzail as positive and negative regulators during Dmvophila mesoderm development. Genes & Dev. 5, 1.568-1576. Leptin, M., & Grunewald, B. (1990). Cell shape changes during gastrulation in Drosophi/a. Development 110, 73-84. Lyons, K.M., Jones, C.M., & Hogan, B.L. (1992). The TGF-beta-related DVR gene family in mammalian development. Ciba Found. Symp. 16.5, 219-230; 230-234. Maeno, M., Ong, C.R., Suzuki, A,, Ueno, N., & Kung, H.-F. ( 1994). A truncated bone morphogenetic protein 4 receptor alters the fate of ventral mesoderm to dorsal mesoderm: Roles of animal pole tissue in the development of the ventral mesoderm. Proc. Natl. Acad. Sci. USA 91, 10260-10264. Mason, E.D., Konrad, K.D., Webb, C.D., & Marsh, L.J. (1994). Dorsal midline fate in Dro%Yophilaembryos requires twisted gastrulation, a gene encoding a secreted protein related to human connective tissue growth factor. Genes & Dev. 8, 1489IiOl. Massague, J., Attisano, L., & Wrana, J.L. (1994). The TGF-P family and its composite receptors. Trends Cell Biol. 4, 172-178. Mayer, U., & Nusslein-Volhard, C. (1988). A group of genes required for pattern formation in the ventral ectoderm of the Dimophila embryo. Genes and Development 2, 1 4 9 6 15 11.
78
CHRISTINE RUSHLOW and SIEGFRIED ROTH
Meinhardt, H. (1982). Models of Biological Pattern Formation.Academic Press, London. Morisato, D., & Anderson, K.V. (1994). The spatzle gene encodes a component of the extracellular signaling pathway establishing the dorsal-ventral pattern of the Drosophila embryo. Cell 76,677488. Murphy-Ullrich, J.E., Schultz-Cherry, S., & Hook, M. (1992). Transforming growth factor$ complexes with thrombospondin. Mol. Biol. Cell. 3. 181-188. Nambu, J.R., Frank, R.G., Hu, S., & Crews, S.T. (1990). The single-minded gene of Drosophila is required for the expression of genes important for the development of CNS midline cells. Cell 63, 63-75. Nellen, D., Affolter, M., & Basler, K. (1994). Receptor serinekhreonine kinases implicated in the control of Dro.sophila body pattern by decapentaplegic. Cell 78, 225-237. Nguyen, T., Jamal, J., Shimell, M.J., Arora, K., & O’Connor, M.B. (1994). Characterization of tolloid related-I: A BMP- I -like product that is required during larval and pupal states of Drosophila development. Dev. Biol. 166, 569-586. Nubler, K., & Arendt, D. (1994). Is ventral in insects dorsal in vertebrates. Roux’s Arch. Dev. Biol. 203,357-366. Nusslein-Volhard, C. (1979). Maternal effect mutations that alter the spatial coordinates ofthe embryo ofDrosophilamelanogaster. In: Determinants ofspatial Organization (Subtelny, S . , & Koenigsberg, I.R., Eds.). Academic Press, New York, pp. 185-2 11. Nusslein-Volhard, C., Wieschaus, E., & Kluding, J. (1984). Mutations affecting the pattern of the larval cuticle in Drosophila nielanogaster I: Zygotic loci on the second chromosome. Roux’s Archives of Developmental Biology 193.267-282. Ozkaynak, E., Rueger, D.C., Drier, E.A., Corbett, C., Ridge, R.J., Sampath, T.K., & Oppermann, H. (1990). Op-l cDNA encodes an osteogenic protein, in the TGF-P family. EMBO J. 9,2085-2093. Padgett, R.W., Johnson, R.D.S., & Gelbart, W.M. (1987). A transcript from aDrosophila pattern gene predicts a protein homologous to the transforming growth factor-p family. Nature 325, 81-84. Padgett, R.W., Wozney, J.M., & Gelbart, W.M. (1993). Human BMP sequences can confer normal dorsal-ventral patterning in the Drosophila embryo. Proc. Natl. Acad. Sci. USA 942905-2909. Pan, D., Huang, J-D., & Courey, A.J. (1991). Functional analysisofthe Drosopfiila twist promoter reveals a dorsal-binding ventral activator region. Genes & Dev. 5 , 18921901. Panganiban, G.E., Rashka, K.E., Neitzel, M.D., & Hoffmann, EM. (1990). Biochemical characterization of the Drosophila DPP protein, a member of the transforming growth factor p family of growth factors. Molecular and Cellular Biology 10, 2669-2677. Paralkar, V.M., Hamkonds, R.G., & Reddi, A.H. (1991). Transforming growth factor beta type 1 binds to collagen IV of basement membrane matrix: Implications for development. Dev. Biol. 143, 303-308. Penton, A,, Chen, Y., Staehling-Hampton, K., Wrana, J.L., Attisano, L., Szidonya, J., Cassill, J.A., Massague, J., & Hoffmann, F.M. (1994). Identification of two bone rnorphogenetic protein Type I receptors in Drosophila and evidence that Brk25D is a decapentaplegic receptor. Cell 78,239-250. Peterson, K.J. (1995). Dorsoventral axis inversion. Nature 373, I 11-1 12.
The dpp-group Genes
79
Posakony, L.G., Raftery, L.A., & Gelbart, W.M. (1990). Wing formation in Drosophilu rnelunoguster requires decupei~ruplegicgene function along the anterior-posterior compartment boundary. Mech. Dev. 33,6%82. Raftery, L.A., Twombly, V., Wharton, K., & Gelbart, W.M. (1995). Genetics screens to identify elements of the decapentuplegic signaling pathway in Drosophila. Genetics 139,241-254. Rao, Y., Vaessin, H., Jan, L.Y., & Jan, Y.-N. (1991). Neuroectoderm in Drosophilu embryos is dependent on the mesoderm for positioning but not for formation. Genes and Development 5, 1577-1588. Ray, R.B. (1993). Dorsal-ventral patterning in the Drosophila embryo. Ph.D. thesis, Harvard University. Ray, R.P., Arora, K., Niisslein-Volhard, C., & Gelbart, W.M. (1991). The control of cell fate along the dorsal-ventral axis of the Drosophila embryo. Development 113, 35-54. Roth, S. ( 1 993). Mechanisms of dorsal-ventral axis determination in Drosophila embryos revealed by cytoplasmic transplantations. Development I 1 7, 1385-1396. Roth, S., Stein, D., & Niisslein-Volhard. C. (1989). Agradient ofnuclear localization of the Dorsal protein determines dorsoventral pattern in the Drosophila embryo. Cell 59, I 1 89-1 202. Roth, S., Hiromi, Y., Godt, D., & Nusslein-Volhard, C. (1991). Cactus. a maternal gene required for proper formation of the dorsoventral morphogen gradient in Drosophih embryos. Development 112, 37 1-388. Rushlow, C.A., & Arora, K. (1990). Dorsal-ventral polarity and pattern formation in Drosophila. Pattern formation in Drosophilu. Sem. Cell Biol. I , 173-184. Rushlow, C., & Levine, M. (1990). Role of the zer.kiziillt gene in dorsal-ventral pattern formation in Drosophila. Adv. Genet. 27. 277-307. Rushlow, C., & Warrior, R. ( 1 992). The re1 family of proteins. BioEssays 14, 8%95. Rushlow, C.A., Frasch, M., Doyle, H., & Levine, M. (1987). Maternal regulation of zerkniillt: A homeobox gene controlling differentiation of dorsal tissues in Dvosophilu. Nature 330,583-586. Rushlow, C.A.. Han, K., Manley, J.L., & Levine, M. (1989). The graded distribution of the dorsuf morphogen is initiated by selective nuclear transport in Dmsophilu. Cell 59, 1165-1177. Rutledge, B.J., Zhang, K., Bier, E., Jan, Y.N., & Perrimon, N. (1992). The Drosophila spitz gene encodes a putative EGF-like growth factor involved in dorsal-ventral axis formation and neurogenesis. Genes & Dev. 6, 1503-1 5 17. Sampath, T.K., Rashka, K.E., Doctor, J.S., Tucker, R.F., & Hoffman, F.M. (1993). Dtvsophila transforming growth .factor j3 superfamily proteins induce endochondral bone formation in mammals. Proc. Natl. Acad. Sci. USA 90,60044008. Sargent,'M.G., & Bennett, M.F. (1990). Identification in Xenopus of a structural homoIogue of the Drosophila gene snail. Development 109,967-973. Sasai, Y., Lu, B., Steinbeisser, H., Geissert, D., Gont, L.K., & De Robertis, E.M. (1994). Xenopus chordin: A novel dorsalizing factor activated by organizer-specific homeobox genes. Cell 79,779-790. Schneider, D.S., Hudson, K.L., Lin, T.Y., & Anderson, K.V. (1991). Dominant and recessive mutations define functional domains of Toft, a transmembrane protein
80
CHRISTINE RUSHLOW and SIEGFRIED ROTH
required for dorsal-ventral polarity in the Drosophila embryo. Genes & Dev. 5 , 797-807. Schneider, D.S., Jin, Y., Morisato, D., & Anderson, K.V. (1994). Aprocessed form ofthe Spatzle protein defines dorsal-ventral polarity in the Drosophila embryo. Development 120, 1243-1250. Schupbach, T., & Wieschaus, E. (1989). Female sterile mutations on the second chromosome of Drosophifa melanogaster I. Maternal effect mutations. Genetics 121, 101-1 17. Schupbach, T., & Wieschaus, E. ( I 986). Maternal-effect mutations altering the anteriorposterior pattern of the Drosophila embryo. Roux’s Archives of Developmental Biology 195, 302-317. Scott, M.P. (1994). Intimations of a creature. Cell 79, 11 21-1 124. Sekelsky, J.J., Newfeld, S.J., Raftery, L.A., Chartoff, E.H., & Gelbart, W.M. (1995). Genetic characterization and cloning of Mothers against dpp, a gene required for decapentaplegic function in Drosophila melanogaster. Genetics 139, 1347-1 358. Shimell, M.J., Ferguson, E.L., Childs, S.R., & O’Connor, M.B. (1991). The Drosophila dorsal-ventral patterning gene tolloid is related to human bone morphogenetic protein I.Cell 67,469-48 1. Simpson, P. (1983). Maternal-zygotic gene interactions during formation of the dorsoventrai pattern in Drosophila embryos. Genetics 105, 6 15-632. Smith, D.E., Franco del Amo, F., & Gridley, T. (1992). Isolation of Sna, a mouse gene homologous to the Dvosophila genes snail and escargot: Its expression pattern suggests multiple roles during postimplantation development. Development 116, 1033-1039. Spencer, F.A., Hoffmann, F.M., & Gelbart, W.M. (1982). Decapentaplegic: A gene complex affecting morphogenesis in Drosophila melanogaster. Cell 28,45 1-46 1. St. Johnston, R.D., & Gelbart, W.M. (1987). Decapentaplegic transcripts are localized along the dorsal-ventral axis of the Drosophila embryo. EMBO J. 6,2785-2791. St. Johnston, R.D., & Nusslein-Volhard, C. ( 1 992). The origin of pattern andpolarity in the Drosophila embryo. Cell 68,201-2 19. Staehling-Hampton, K., Jackson, P.D., Clark, M.J., Brand, A.H., & Hoffmann, F.M. ( 1994). Specificity ofbone morphogenetic protein-related factors: Cell fate and gene expression changes in Drosophila embryos induced by decapentaplegic but not 6OA. Cell Growth & Diff. 5,585-593. Stein, D., & Nusslein-Volhard, C. ( 1992). Multiple extracellular activities in Drosophila egg perivitelline fluid are required for establishment of embryonic dorsal-ventral polarity. Cell 68,429-440. Stein, D., Roth, S., Vogelsang, E., & Nusslein-Volhard, C. (1991). The polarity of the dorsoventral axis in the Drosophila embryo is defined by an extracellular signal. Cell 65, 725-735. Steward, R. (1987). dorsal, an embryonic polarity gene in Drosophila is homologous to the vertebrate proto-oncogene, c-re/.Science 238,692694. Steward, R. (1989). Relocalization of the dorsal protein from the cytoplasm to the nucleus correlates with its function. Cell 59, 117%-1188. Technau, G.M., & Campos-Ortega, J.A. (1987). Cell autonomy of expression of neurogenic genes ofDrosophilu melanogaster. Proc. Natl. Acad. Sci. USA84,4500-4504.
The dpp-group Genes
81
Terracol, R., & Lengyel. J.A. ( 1994). The thick veiris gene of Drosophila is required for dorsoventral polarity of the embryo. Genetics 138, 165-178. Thisse, B., Stoetzel, C., Gorostiza-Thisse, C., & Perrin-Schmitt, F. (1988). Sequence of the twist gene and nuclear localization of its protein in endomesodermal cells of early Drosophila embryos. EMBO J. 7,2 175-2 183. Thisse, C., Perrin-Schmitt, F., Stoetzel, C., & Thisse, B. (1991). Sequence-specific transactivation of the Di*o.sophilatwist gene by the dorsal gene product. Cell 65, 1191-1201. Thomsen, G.H., & Melton, D.A. (1993). Processed Vgl protein is an axial mesoderm inducer in Xenopus. Cell 74,433441. Wakimoto, B.T., Turner, F.R., & Kaufinan, T.C. (1984). Defects in embryogenesis in mutants associated with the Antennapedia gene complex of Drosophila rnelanogaster. Dev. Biol. 102, 147-1 72. Wang, E.A., Rosen, V., Cordes, P., Hewick, R.M., Kriz, M.J., Luxenberg, D.P., Sibley, B.S., & Wozney, J.M. (1988). Purification and characterization of other distinct bone-inducing factors. Proc. Natl. Acad. Sci. USA 85, 9484-9488. Wang, T., Donahoe, P.K., & Zervos, A S . (1994). Specific interaction oftype 1 receptors of the TGF-P family with the iinmunophilin FKBP- 12. Science 265,674-676. Weeks, D.L., & Melton, D.A. (1987). A maternal mRNA localized to the vegetal hemisphere in Xenopirs eggs codes for a growth factor related to TGF-P. Cell 51, 86 1-867. Wharton, K.A., Thomsen, G.H., & Gelbart, W.M. ( I99 I). Drosophila 60A gene, a new TGF-P family member is closely related to human bone morphogenetic proteins. Proc. Natl. Acad. Sci. USA 88,9214-9218. Wharton, K.A., Ray, R.P., & Gelbart, W.M. (1993). An activity gradient of decapentaplegic is necessary for the specification of dorsal pattern elements in the Drosophila embryo. Development 117, 807-422. Wieschaus, E., & Nusslein-Volhard, C. ( 1986). Looking at embryos. In: Dro.sopliila:A Practical Approach (Roberts, D:B.. Ed.). IRL Press, Washington DC, pp. 199-227. Wieschaus, E., Nusslein-Volhard, C., & Jurgens, G. (1984). Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster: 11. Zygotic loci on the X-chromosome and fourth chromosome. Wilhelm Roux’s Arch. Dev. Biol. 193, 296-307. Wolf, C., Thisse, C., Stoetzel, C., Thisse, B., Gerlinger, P., & Perrin-Schmitt, F. (1991). The M-twist ofMus is expressed in subsets of mesodermal cells and is closely related to the Xenopus X-twi and the Drosophila twist genes. Dev. Biol. 143,363-373. Wozney, J.M., Rosen, V., Celeste, A.J., Mitsock, L.M., Whitters, M.J., Kriz, R.W., Hewick, R.M., & Wang, E.A. ( 1988).Novel regulators ofbone formation: Molecular clones and activities. Science 242, 1528-1 534. Wrana, J.L., Attisano, L., Wieser, R., Ventura, F., & Massague, J. (1994a). Mechanisms of activation of the TGF-P receptor. Nature 370,341-347. Wrana, J.L., Tran, H., Attisano, L., Arora, K., Childs, S.R., Massague, J., & O’Connor, M.B. (1 994b). Two distinct transmembrane serinehhreonine kinases from Drosophila melanogaster form an activin receptor complex. Mol. Cell. Biol. 14,944-950. Xie, T., Finelli, A.L., & Padgett, R.W. (1994). The Drosophila saxophone gene: Aserine threonine kinase receptor ofthe TGF-beta superfamily. Science 263, 1756-1 759.
82
C H R I S T I N E R U S H L O W and SIEGFRIED R O T H
Zusman, S.B., & Weischaus, E. (1985). Requirement for zygotic gene activity during gastrulation in Drosophila melanogaster. Dev. Biol. 111,359-371. Zusman, S.B., Sweeton, D., & Wieschaus, E.F. (1988). Short gastrulation, a mutation causing delays in stage specific cell shape changes duringgastrulation in Drosophila melanogaster. Dev. Biol. 129,417427.
THE TERMINAL GENE HIERARCHY OF DROSOPHlLA A N D THE GENETIC CONTROL OF TISSUE SPECIFICATION AND MORPHOGENESIS
Man Lun R. Yip and Howard D. Lipshitz
Introduction , . . . . . . , . . . . . . . . . . . . . . . . . Maternally Encoded Components of the Terminal Pathway A. The Transmembrane Receptor . . . . . . . . . . . . . B. The Extra-embryonic Ligand . . . . . . . . . . . . . . C. Cytoplasmic Signal Transduction . . . . . . . . . . . D. ConservationofSignalTransductionPathways . . . . 111. The Zygotic Effectors of Terminal Cell Fates . . . . . . . . A. Subdivision of the Termini into Distinct Tissues . . . . B. Subdivision into Endoderm versus Ectoderin . . . . . C. Genes that Program Ectodermal Development . . . . D. Genes that ProgramEndodermal Development . . . . E. Genesthatcontrol Morphogenetic Movement . . . . 1. 11.
Advances in Developmental Biology Volume 4, pages 83-146. Copyright 0 1996 by JAl Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-969-9
83
. . . 84 . . . 1 00 . . , 100 . . . 102 . . . 104 . . . 108 . . . 109 . . . 109 . . . 111 . . . 114 . . . 120 . . . 121
84
M A N LUN R. YIP and H O W A R D D. LlPSHlTZ
1V. Mechanisms for Establishing Distinct Cell Fates Within the Termini . . . . . . . . . . . . . . . . . . . . . . . . . 124 A. More Central versus More Terminal Fates . . . . . . . . . . 124 B. Dorsal versus Ventral Fates . . . . . . . . . . . . . . . . . . 125 C. Anterior versus Posterior Fates . . . . . . . . . . . . . . . . 129 V. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Note Added in Proof . . . . . . . . . . . . . . . . . . . . . . . . 132 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 132 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
1. INTRODUCTION In Drosophilu the establishment of cell fates along the dorso-ventral and antero-posterior axes of the embryo is accomplished by genetic hierarchies that act both maternally and zygotically (for reviews see Lipshitz, 1991; St. Johnston and Nusslein-Volhard, 1992). While only one pathway is required for the specification of dorso-ventral fates, three interacting pathways confer fates along the antero-posterior axis: these have been referred to as the “anterior” (head and thorax), “posterior” (abdomen), and “terminal” (asegmental termini) pathways (Niisslein-Volhard et al., 1987). Axis specification is initiated during oogenesis and involves interactions between the developing oocyte and the surrounding, somaticallyderived follicle cells. Two main consequences are (1) the asymmetric deposition of localized determinants within the oocyte: bicoid RNA for the anterior pathway and nunos RNA for the posterior pathway (St. Johnston and Nusslein-Volhard, 1992; Ding and Lipshitz, 1993), and (2) local deposition (or activation) in the extracellular matrix that surrounds the oocyte, of ligands for uniformly distributed receptors that are synthesized in the early embryo: TOLL for the dorso-ventral system and the TORSO receptor tyrosine kinase for the terminal pathway (Lipshitz, 1991;St. Johnston and Niisslein-Volhard, 1992). In this review we focus on the terminal pathway. More than 40 genes have been defined genetically and/or molecularly as residing in the terminal pathway and more than 20 additional identified genes are likely to function in this hierarchy (Table 1). Since neither the genetic nor the molecular screens have been saturated, this list represents only a subset of the loci that function in the terminal pathway. The initial components of the pathway, from ligand production and activation through receptor activation and cytoplasmic signal transduction, are encoded by maternally synthesized molecules. This signal results in the control of differ-
Table 1. Terminal Hierarchy Members-Genes or Likelv to Reside in the Pathwav
Known
~
Gene (abbreviation)
Molecular Details
A bdominal-B (Abd-B) 1. Homeodomain transcription factors
armadillo (arm)
co VI
(ABD-BI, ABD-BII). 2. Zygotic. 3. ABD-BII in parasegments 14 and 15; ABD-BI in parasegments 13-15. 1. Homolog of human plakoglobin. 2. Zygotic. 3. Uniformly distributed in embryos.
cactus (cact)
I . Homolog of IKB. 2. Maternal. 3. Uniformly distributed in embryos.
cap 'n 'collar (cnc)
1. Leucine zipper transcription factor. 2. Zygotic.
Mutant Phenotppes (alleles)
References
lof: Deletion of posterior spiracles and Sanchez-Herrero et al. (1985), filzkorper. Homeotic transformation of Casanova ( 1990), Celniker et al. (1990) A8 toward A4lA5.
lof: Head and tail defects.
Riggleman et al. (1989, 1990), Peifer andwieschaus (1 990)
lof: Dependent on the allelic combination, variable degree of ventralization including deletion of anal plates and filzkorper. Enhance torsogofphenotypes. gof: Dependent on the allelic combination, variable degree of dorsalization. In most extreme case anal plates and filzkorper are deleted. lof: Missing labrum and dorsal pouch.
Schupbach and Wieschaus (1989), Roth et al. (1991), Strecker et al. ( 1991), Geisler et al. (1992), Kidd ( 1992)
Mohler et al. (1991, 1995), Mohler (1993) (Continued)
Table 7. (Continued) Gene (abbreviation)
caudaI (cad) 0,
0
corkscrew (csw)
cubitus interruptus/Cell (ci/Ce)
Molecular Details 3. Terminal expression in an anterior cap activated by TORSO and BICOID at nuclear cycle thirteen. Subsequent repression by GIANT anteriorly, TAILLESS posteriorly and the dorsoventral genes ventrally, restrict expression to part of the labral primordium. 1. Homeodomain transcription factor. 2. Maternal and zygotic, 3. Maternal product distributed in a posterior-to-anterior gradient in early embryo. Zygotic expression includes the primordia of midgut, hindgut and A9, A 10, telson primordia. 1. Putative tyrosine phosphatase with two SH2 domains. 2. Maternal and zygotic. 3. Uniformly distributed in embryo.
1.Zinc-finger protein. 2. Zygotic.
Mutant Phenohtues (alleles)
References
lof: No anal tuft, abnormal anal pads, Macdonald and Struhl(1986), terminal sense organs lost or abnormal. Mlodzik and Gehring (1987a), No maternal and no zygotic: Liu and Jack (1 992) Deletion of A9, A I0 and telson; abnormal Malphigian tubules
No maternal: Deletion of acron and Perrimon et al. (1989), labrum in the anterior; posterior Perkins et al. (1992) midgut and Malpighian tubules posteriorly. Defects in morphogenetic movements. Suppress toug"l phenotypes. Head and tail defects. Orenic et al. (1 987, 1990)
3. At cellular blastoderm stage, uniformly expressed from 20-90% EL; later resolves into fifteen broad stripes (the last two stripes in posterior terminal domain). I . Homeodomain transcription factor. cut (ct) 2 . Zygotic. 3. Expressed in Malpighian tubule anlagen. decapentaplegic (dpp) 1. TGF-P family growth factor. 2 . Zygotic. 3. Initial expression occurs in the dorsal 40% of the central region of the embryo but extends around the two poles to include more ventral cells within the termini. Expression at the termini is dependent on torso function. I. Novel protein. dbhevelled (dsh) 2. Maternal and zygotic. 3. Uniformly distributed in early embryos. dorsal (dl)
1. Homolog of REL/NFKB transcription factor. 2. Maternal. 3. Protein is translocated into the nuclei on the ventral side and in the termini of the embryo.
lof: Abnormal Malphigian tubules.
Blochlinger et al. (1988), Liu et al. (1991), Liu and Jack (1992)
Null: Loss of acron, labral structures, posterior spiraclular hair, filzkoper, anal tuft and pads. Fail to complete germband extension. lof: Rudimentary and uneverted posterior spiracles and filzkorper. Enhance tordofphenotypes.
Irish and Gelbart (l987), Padgett et al. (1987), St. Johnston and Gelbart (l987), St. Johnston et al. (1990), Casanova (1991), Strecker et al. (1991), Ferguson and Anderson (1 992), Wharton et al. (1993) Perrimon and Mahowald (l987), Perrimon et al. (1989), Klingensmith et al. (1994), Noordermeer et al. (1994), Siegfried et al. (1994) Nusslein-Volhard et al. (l980), Steward (1 987, 1989), Steward et al. (l988), Roth et al. (1989), Rushlow et al. (1989a), Strecker et al. (1991)
No maternal: Lack posterior spiracles and filzkorper.
No maternal: Loss of endodermal gut and Malpighian tubules. Suppress tors&0’ phenotypes.
-
~ - _ _ _ -
(Continued)
Table 7. (Continued) Gene (abbreviation)
cc
m
Molecular Details Mutant Phenotypes (alleles) 1. Protein with two SH2 domains and downstream of lof: Suppress tors&’’phenotypes. receptor kinases (dt-k) one SH3 domain. 2. Maternal (?) and zygotic. 3. ? I . MAP kinase activator (MEK homolog). No maternal: Typical terminal Dsorl 2. Maternal and zygotic. phenotypes. 3. Expressed throughout development. No maternal and zygotic: Poor cuticular development. Doininant maternal suppressor of tor: trk, f,i(l)ph, tsl, l(1)ph loss of function mutations and dominant maternal enhancer oftoFofphenotypes. 1. Homologous to C-SRC. Dsrc Ectopic expression of wildtype or mutant 2. Maternal and zygotic. SRC at high level causes germband 3. Uniformly distributed at retraction defects. cellularization and gastrulation. empty spiracles (ems) 1. Homeodomain transcription factor. lof: Optic lobes and filzkorper absent. 2. Zygotic. Suppress HS-tailless phenotypes. 3. Expressed in procephalic lobe anteriorly, filzkorper anlagen posteriorly.
References Doyle and Bishop (1993), Olivier et al. (1993), Simon et al. (1993) Tsuda et a]. ( 1 993)
Simon et al. (1983, 1985), Kussick et al. (1993)
Dalton et al. (1989), Strecker et al. (1992), Walldorf and Gehring (1992)
engrailed (en)
even-skipped (eve)
folded gastrulation flog) forkhead (frk)
fs(1)Nasrat fs(1)N)
\of: Reduced posterior spiracles. I , Homeodomain transcription factor. 2. Zygotic. 3. Expression is seen in parasegments 14 and 15. Later, complex expression in the head including spots in the brain and clypeolabrum. lof: Reduced posterior spiracles, tuft and 1. Homeodomain transcription factor. fi Iz korper. 2. Zygotic 3. Expression in the seventh stripe is dependent on tailless. I . Putative secreted, novel protein. 2. Zygotic. 3. Expressed in a posterior polar cap (@-lo% EL). 1. Transcription factor withforkhend domain. 2. Zygotic. 3. Midgut, foregut and hindgut. l.? 2. Maternal. 3. ?
DiNardo et al. (1985), Komberg et al. (1985), Poole et al. (1985), Jurgens (1987), Diederich et al. (199 I), Schmidt-On and Technau (1992) Nusslein-Volhard et al. (1985), Macdonald et al. (1986), Frasch et al. (1987), Frasch and Levine ( 1987), Goto et al. (1989) lof: Failure in anterior and posterior Wieschaus et al. ( 1984), midgut invagination. Suppress tomF" Zusman and Wieschaus ( 1985), phenotypes. Strecker et?il. (1991), Costa et al. ( 1 994) lof: Anterior and posterior midgut Jurgens and Weigel ( 1988), primordia fail to undergo migration in Weigel et al. (1989a, 1989) order to fuse into a unit structure, instead they disintegrate. lof: Collapsed egg and deletion of acron Degelmann et al. (1986, 1990) and labrum in the anterior and tail structures posterior to the A7 segment. 2 11: Deletion of acron and labrum in the anterior and tail structures posterior to the A7 segment. DH 1 : Temperature-sensitive, collapsed egg plus posterior terminal defect ( 18°C) or central deletion (25°C). ______~-(Continued)
Table 7. (Continued) Gene (abbreviation)
Molecular Details
Mutant Phenowpes (alleles)
fs(1)polehole Ifs(1)ph)
1. ? 2. Maternal. 3. ?
.fused 0"ir)
1. Serineheonine kinase. 2. Maternal and zygotic. 3. Maternal transcript uniform in embryo through germband extension. 1. Homeodomain transcription factor. lof: Defects in structures derived from 2. Zygotic. several posterior terminal domain 3. The seventh stripe is dependent on segments: reduced filzkorper and torso and tailless. spiracular hair (A8), posterior lateral sense organs (A9), anal sense organs (A-10) and telson. The posterior spiracles remain on the dorso-lateral surface. 1. Homologous to GTPase activating lot Enhances to$ofphenotypes. protein. 2. Maternal (?) and zygotic. 3. ? 1. Transcription factor with leucine lof: Missing the labrum, epistomal zipper. sclerite and dorsal bridge. 2. Zygotic. Suppress torsgo'
fuslii tarazu (ftz) a 0
Gap I
giant (go
lof: Collapsed egg. 190 1 : Deletion of acron and labrum in the anterior and tail structures posterior to the A7 segment. No or reduced maternal: Defects in A8
References Degelmann et al. (1990)
Perrimon and Mahowald (1987), Preat et al. (1990), Therond et al. (1993) Laughon and Scott (1 984), Wakimoto et al. (1984), Jiirgens (1987)
Buckles et al. (1992), Gaul et al. ( 1 992), Rogge et al. (l992), Doyle and Bishop, (1993) Mohler et al. (1 989), Eldon and Pirrotta (1991), Strecker et al. (199l), Capovilla et al. (1992)
gooseberry (gsb)
hairy (h)
2 hedgehog (hh)
hindsight (hnt)
HNF-4
3. At cellular blastoderm, horseshoeshaped domain (stripe 1) at.8&95% EL. 1. Two paired domain- and homeodomain-containing proteins. 2. Zygotic. 3. Striped expression in parasegments 14 and 15; later, additional expression occurs more posteriorly. 1. b-HLH transcription factor. 2. Zygotic. 3. Dorsal head patch at 85-95% EL and the seventh stripe fall within the terminal domain. Expression in the seventh stripe is dependent on tailless. 1. Transmembrane protein. 2. Zygotic. 3. Early gastrula: expressed in dorsal anterior spot at 97% EL; posterior 3 stripes in terminal domain. 1. Putative zinc-finger transcription factor. 2. Zygotic. 3. Midgut. 1.Zinc-finger transcription factor. 2. Maternal and zygotic.
No details reported
Baumgartner et al. (1987), Perrimon and Mahowald (1 987)
lngham et al. ( 1 9 8 9 , lof: Defects in the anterior lateral and dorso-medial sense organs, filzkorper Ish-Horowicz et al. (1985). and fell posteriorly, and median tooth Jurgens (1 987), Mahoney and Lengyel (l987), of the labrum anteriorly. Hooper et al. (1989), Rushlow et al. (1989b) Head skeleton and tail abnormal. Mohler (1988), Lee et al. (l992), Mohler and Vani ( 1992)
lof: Failure of germband retraction. Suppress tors#of phenotypes.
lof: Midgut and Malpighian tubules fail to form.
Wieschaus et al. (1984), Strecker et al. (1991), Yip and Lipshitz (in preparation) Zhong et al. (1993) .
~
_
_
-
(Continued)
Table 1. (Continued) Gene (abbreviation)
Molecular Details
3. Maternal transcripts distributed uniformly. Zygotic transcripts appear in anterior and posterior midgut. 1. JAK family tyrosine kinase. hopscotch (hop) 2. Maternal and zygotic. 3. Uniformly distributed in embryos. huckebein (hkb) 1. SP-I EGR-like zinc-finger transcription factor. 2. Zygotic. 3. Two polar caps (&12% EL and 9(r 100% EL). hunchback (hb) 1.Zinc-finger transcription factor. 2. Maternal and zygotic. 3. At cellular blastoderm, posterior stripe from 10-25% EL. Kruppel (Kr) 1.Zinc-finger transcription factor. 2. Zygotic. 3. Malphigian tubule anlagen. I(1)pole hole (I(1)ph) 1. RAF serinekhreonine kinase homolog. 2. Maternal and zygotic. 3. Uniformly distributed in embryo.
Mutant Phenotypes (alleles)
No maternal and zygotic: Deleted or reduced A8 segment and defects in posterior spiracles. lof: Deletion of endodermal midgut. Abnormal morphogenetic movements. Suppress t o r . d o fphenotypes.
Cot Deletion of the A8 segment.
References
Perrimon and Mahowald ( 1 986), Binari and Perrimon (1 994) Weigel et al. (l990), Bronner and Jackle (1991), Bronner et al. ( 1994)
Lehmann and Nusslein-Volhard (1987, Tautz et al. ( 1 987), Strecker et al. (1991) lof: Malphigian tubules missing. Gaul et al. (1987), Harbecke and Janning (1 989), Liu and Jack (1992) No maternal: Deletion of acron and Nishida et ai. ( I 988), labrum in the anterior and tail Ambrosio et al. (1989b), structures posterior to the A7 segment. Melnick et al. (1993) Defects in morphogenetic movements. Suppress torso@” phenotypes.
lines (lin)
1. ?
2. Zygotic. 3. ?
naked (nkd)
odd-paired (opa) co W
odd-skipped (odd)
paired (prd)
1. '? 2. Zygotic. 3.? 1. Zinc-finger protein. 2. Zygotic. 3. Single broad domain at 2 W O % EL.
1. Zinc-finger protein. 2. Zygotic. 3. At early gastrulation stages, anterior pole and stripes fourteen and fifteen fall within the terminal domain. 1. Paired domain and homeodomaincontaining transcription factor. 2. Zygotic.
No maternal and zygotic: Embryos degenerate 7 hr after fertilization. Suppress tors0S"f phenotypes. lof: Head defects; missing A8, posterior spiracles and anal pads. Synergistic interactions with tailless alleles. Suppress tors$"f and HS-tailless phenotypes. lof: Defective head and abnormal posterior spiracles.
Niisslein-Volhard et al. (1984), Strecker et al. (1991, 1992)
Jurgens et al. ( 1 984), Martinez Arias et al. (1988), Perrimon and Smouse (1989) Jurgens (1 987), Benedyket al. (1994)
lof: Defects in multiple posterior structures: the anterior lateral sense organs, dorso-medial sense organs, anal sense organs, spiracular hair, fell and filzkorper. lof: Ectopically located median tooth that Coulter and Wieschaus (1988), is frequently misshapen; defective anal Coulter et al. (1990) tuft and pads.
lof: Lack and telson and posterior sense organs.
Nusslem-Volhard et al. (1983, Bopp et al. (1986), Frigerio et al. (1986), Kilchherr et al. (1986) - ____-
(Continued)
Table 7. (Continued) Gene (abbreviation)
patched (ptc)
a porcupine (Port)
ras I
rolled (ro
runt (run)
Molecular Details
3. At cellular blastoderm, anterior head patch (87-93% EL) and thirteenth and fourteenth stripes fall within the terminal domains. 1. Putative transmembrane protein. 2. Zygotic. 3. Expression includes parasegments 13 and 14, the labrum, stomodeum and hindgut anlagen (where it is eventually restricted to the Malpighian tubules). I.? 2. Maternal and zygotic. 3. ? I . U S , GTPase protein 2. Maternal and zygotic. 3. Expressed throughout development. 1. MAP kinase. 2. Maternal and zygotic. 3. Expressed throughout development. 1. runt domain-containing transcription factor. 2. Zygotic.
Mutant Phenowpes (alleles)
References
lof: Enlarged anal tuft. Defects in all tail sense organs.
Jurgens (1 987), Hooper and Scott (1 989), Nakano et al. (1 989)
No maternal: Head and tail defects.
Perrimon et al. (1 989), Siegfried et al. (1994)
Lev et al. (1 985), Simon et al. (1991), Doyle and Bishop ( 1993), Lu et al. (1 993) lof: Suppress tors&"/ phenotypes. Biggs and Zipursky ( 1992), gof (Sem): Deletion of central structures Biggs et al. (l994), Brunner et al. ( 1994b) similar to tors&'/ phenotypes. Gergen and Wieschaus (1985), Phenotypes not examined in termini. Gergen and Butler (l988), Kania et al. ( 1990)
lof: Suppress tors&"/ phenotypes. g o t Phenocopy tors$" phenotypes.
serpent (srp)
shaggyheste white 3 (sgg/zw3)
W
vI
short gastrulation (so&
sloppy paired (slp)
3. At cellular blastoderm, anterior head patch ( 7 5 4 5 % EL) and the seventh stripe fall within the terminal domain. Later, an additional stripe and proctodeal expression occur posterior to the original seventh stripe. l.? 2. Zygotic. 3. ? 1. Serinehhreonine kinase homologous to glucose synthetase kinase 3. 2. Maternal and zygotic. 3. Uniformly distributed in early embryos. 1. Putative secreted protein homologous to Xenopus CHORDIN, with four cysteine-rich domains. 2. Zygotic. 3. At cellular blastoderm, present ventrally in central region; largely absent from termini. 1. Twoforkhead domain-containing transcription factors. 2. Zygotic. 3. At syncytial blastoderm, anterior polar cap (70-1 00% EL) that undergoes rapid changes. Later, complex expression occurs in the procephalon. At germband extended stage, the last two stripes (in A8 and A9) fall within the terminal domain.
,
lof: Homeotic transformation of endodermal midgut to ectodermal foreguvhindgut. No maternal: Defective head and abnormal posterior spiracles.
Jurgens et al. (1984), Reuter (1994)
lof: Delayed formation and closure of anterior and posterior midgut invaginations.Germband extension incomplete. Suppress tor.&‘’ phenotypes.
Zusman and Wieschaus (1985), Zusman et al. (1988), Strecker et al. (1991), Francois et al. (1994), Francois and Bier (1995)
Phenotypes not examined in termini.
Grossniklaus et al. ( 1992), Cadigan et al. (1994)
Perrimon et al. (l989), Perrimon and Smouse (1989), Bourouis et al. (1990)
(Continued)
Table 1. (Continued) Gene (ahhreviation)
rD 0
Molecular Details
Son of sevenless (Sos) 1. homologous to RAS guanine nucleotide exchange factor. 2. Maternal and zygotic. 3. ? spalt (sal) 1.Zinc-finger transcription factor. 2. Zygotic. 3. Anterior horseshoe-shaped domain at 8 W 6 % EL. Posterior stripe at 1220% EL is dependent on tailless. I . Homologous to mammalian Brackyuty T-relatedgene (Trg) ( r ) gene. 2. Zygotic. 3. Hindgut and anal pad primordia. 1. Homologous to steroid hormone tailless (tlr') receptor superfamily. 2. Zygotic. 3. Initially two polar caps (&20% EL and 80-100%0 EL). The anterior cap is refined into a dorsal horseshoe-shaped domain (76-89% EL). 1.Zinc-finger transcription factor. tailup (tup) 2. Zygotic. 3. ?
Mutant Phenowpes (alleles)
References
No maternal: Deletion of acron and labrum in the anterior and tail structures posterior to the A7 segment. lof: Suppress tors@'/ phenotypes. lof: Partial homeotic transformation of A9 and A10 toward A8 segment.
Rogge et al. (1991), Simon et al. (1991), Bonfini et al. (1992), Doyle and Bishop (1993) Jurgens (1988), Kuhnlein et al. (1994), Schuh and Jaeckle cited in Jurgens and Hartenstein (1993)
l o t Deletion of hindgut and anal pads.
Kispert et al. (1994)
lof: Abnormal clypeolabrum, optic lobes Strecker et al. (1986, 1988, 1989), and procephalic lobe. Missing Klingler et al. (1988), segments A8, A9, AIO, hindgut and Pignoni et al. (1990) Malpighian tubules. Suppress tors#"/ phenotypes.
l o t Failure of germband retraction. Suppress to@ phenotypes.
Nusslein-Volhard et al. (1984), Strecker et al. (1991)
tenaxinmiodd 0 2 (ten "Iodz)
tolloid (tld)
torso (tor)
1. Protein homologous to TENASCIN with 8 EGF repeats and 10 fibronectin Ill repeats. 2. Zygotic. 3. At cellular blastoderm stage, protein present in an antero-dorsal head patch and in the posterior midgut primodium. Later, protein also accumulates in stomodeum. 1. Homolog of bone morphogenetic protein- 1 (BMPL), a putative metalloendopeptidase. 2. Zygotic. 3. Initial expression occurs in the dorsal 40% of the central region of the embryo but extends around the two poles to include more ventral cells within the termini. 1. PDGF family receptor tyrosine kinase . 2. Maternal.
lof: Abnormal telson.
Baumgartner et al. (1994), Levine et al. (1994)
lof: Missing pharyngeal skeleton anteriorly and filzkorper posteriorly. Enhance torso@"phenotypes.
Shimell et al. (1991), Strecker et al. (1991), Ferguson and Anderson (1992), Childs and O'Connor, (1994)
lof: Deletion of acron and labrum in the anterior and tail structures posterior to the A7 segment. No midgut or hindgut formed; reduced foregut. Defects in morphogenetic movements. Expansion of central, segmented fate map into termini. gof: Expansion of terminal fate map into central, segmented region. Deletion of central s e g m e n t a w structures.
Schiipbach and Wieschaus (1986a, 1986b 1989), Klingler et al. (1988), Casanova and Struhl(1989), Sprenger et al. (1989), Strecker et al. (1989)
(Continued)
Table 1. (Continued) Gene (abbreviation)
torso-like (tsl)
Mutant Phenowpes (alleles)
Molecular Details
3. Germline expression. Translated in embryo after fertilization. RNA and protein uniformly distributed in embryo. lof Deletion of acron and labrum in the 1. Putative secreted novel protein. anterior and tail structures posterior to 2. Maternal and zygotic. the A7 segment. No midgut or hindgut 3. Maternally expressed in the somatic formed; reduced foregut. Defects in follicle cells adjacent to the two poles morphogenetic movements. Expansion of the oocyte. of central, segmented fate map into termini. ee: Deletion of central trunk structures. l o t Abnormal head and tail (filzkorper ].Zinc-finger transcription factor. absent). 2. Maternal and zygotic. 3 . Maternal transcripts distributed uniformly. Zygotic transcripts appear in anterior and posterior midgut. 1. Probable ligand for TORSO receptor lof: Deletion of acron and labruin in the anterior and tail structures posterior to tyrosine kinase, with similarity to the A7 segment. No midgut or hindgut spatzle. formed; reduced foregut. Defects in 2. Maternal. morphogenetic movements. Expansion 3. Uniformly distributed in early emof central, segmented fate map into bryos. termini. lof: Abnormal morphogenetic 1. Limited homology to human movements, head defects and connective tissue growth factor. condensed, retracted spiracles. 2. Zygotic. Suppress torso@' phenotypes. 3. Anterior dorsal cap at syncytial blastoderm stage.
References
Stevens et al. (l990), Savant-Bhonsale and Montell (1 993), Martin et al. (1994)
'
tramtrack (ttk)
trunk (trk)
twisted gastrulation (tsd
Read and Manley (l992), Brown and Wu (1993), Xiong and Montell (1993)
Schupbach and Wieschaus (1986a, 1986b 1989), Casanova and Struhl (l993), Casanova et al. (1995)
Zusman and Wieschaus (1985), Zusman et al. (1988), Streckeret al. (1991), Mason et al. (1994)
u-shaped (ush)
unpaired (upd)
wingless (wg)
zerkniillt (Zen)
I. ?
2. Zygotic. 3. ? l.? 2 . Zygotic.
lof Failure of germband retraction. Suppress tors#ofphenotypes.
Nusslein-Volhard et al. (1984), Strecker et al. ( 1 991)
lof Abnormal posterior spiracles; filzkorper altered in appearance or absent. lof Lack head structures and filzkorper.
Wieschaus et al. (1984), Gergen and Wieschaus (1986)
3. ? 1. Vertebrate INT-1 proto-oncogene homolog. 2. Zygotic. 3. At early blastodenn, initial expression at the stomodeum and proctodeum. Subsequently expression includes parasegments 13 and 14. Later, complex expression in the head including a spot in the labral region. 1. Homeodomain transcription factor. lof Deletion of optic lobes in the anterior 2. Zygotic. terminus.
3. Initial expression occurs in the dorsal 40% of the central region of the
Baker (1987, 1988a, 1988b), Perrimon and Mahowald ( 1987), Rijsewijk et al. (1987), Schmidt-Ott and Technau (1992)
Wakimoto et al. (1984), Doyle et al. (1 986), Rushlow et al. (1 987a), Strecker et al. ( 1 992)
Suppress torsgoi phenotypes.
embryo but extends around the two poles to include more ventral cells within the termini. Expression at the termini is dependent on torso function. Notes: Molecular details: 1. = nature of the gene product;2. = maternal and/or zygotic expression; 3. = expression patterns.Abbreviations: A# =abdominal segment #; ee = ectopic expression; EL = egg length, with 0% representing the posterior tip and 100% representing the anterior tip; gof = gain of function; lof = loss of function or hypomorphic phenotype; ? = unknowdunreported.
100
MAN LUN R. YIP and HOWARD D. LlPSHlTZ
ential zygotic gene expression. We begin our description of the pathway at the level of the transmembrane receptor, TORSO, that receives the terminal signal (Section 11). We describe what is known about ligand production and activation, and how the signal is transduced by a cytoplasmic phosphorylation cascade. We then focus (Section 111) on the zygotic genes that implement detailed aspects of terminal cell fate specification. Finally (Section IV), we consider how the' outcome of reading the same signal is modulated within the termini along their antero-posterior and dorso-ventral axes, as well as how cells in the anterior terminal region are programmed to undergo distinct developmental fates from those in the posterior terminal region.
II. MATERNALLY ENCODED COMPONENTS OF THE TERMINAL PATHWAY A. The Transmembrane Receptor The torso gene encodes a transmembrane receptor tyrosine kinase related to those of the PDGF family (see Table 1 and Figure 1; Casanova and Struhl, 1989; Sprenger et al., 1989) that receives the extra-embryonic terminal signal and transduces it into the embryo. Since expression of the torso gene is restricted to the germline of the mother, torso mutations have a strictly maternal effect. Loss of function torso (torso'O') alleles result in an inability of cells at both the anterior and posterior termini of the embryo to adopt terminal cell fates. Instead the cells in the termini adopt central fates: the segmented region expands into the termini (Figure 1;Schupbach and Wieschaus, 1986b). The torso'"'a1leles encode proteins with no functional kinase activity. Gain of function torso alleles (tors&"') result in the reciprocal phenotype: the terminal fate map expands into the central region with consequent loss of segments (Figure 1; Klingler et al., 1988; Schiipbach and Wieschaus, 1989; Strecker et al., 1989). The tors& alleles behave genetically as hypermorphs (Strecker et al., 1989) and have alterations in the extracellular domain that result in constitutive, ligand-independent activation of their kinase activity (Figure 1; Sprenger and Nusslein-Volhard, 1992; Casanova and Struhl, 1993). The reciprocal phenotypes of torso gain of function versus torso loss of function alleles indicate that the torso gene has dual functions during embryogenesis: (1) to promote terminal cell fates in the polar regions, and (2) to repress central, segmented cell fates in the termini (Strecker et al., 1989; Strecker and Lipshitz, 1990).
Drosophila Terminal Gene Hierarchy
101
A. Perivitelline Space
+RL3:H242L
RI:Q846Stop+
gain of function alleles
U U
B.
Figure 1. The TORSO receptor tyrosine kinase: (A) structure; (B) mutant phenotypes. (A) All gain of function torso alleles map to the extra-cellular domain, which is shown cross-hatched; while all loss of function torso alleles map to the cytoplasmic kinase domain (stippled) (Sprenger and Nusslein-Volhard, 1992). The transmembrane domain is shaded black. (B) Schematic representation of the effects of torso mutations on the fate map of the embryo. The terminal domains of the fate map are shaded; the central, segmented region is unshaded; anterior is to the left and dorsal toward the top of the page. In embryos from torso loss of function females, the terminal portions of the fate map are lost and the cells in the termini adopt central, segmented fates (Schupbach and Wieschaus, 1986b). In embryos from torso gain of function females, the central portions of the fate map are lost and the cells in the central region adopt terminal fates (Klingler et al., 1988; Schupbach and Wieschaus, 1989; Strecker et at., 1989).
102
MAN LUN R. YIP and HOWARD D. LIPSHITZ
The torso gene is expressed maternally and its transcripts are uniformly distributed in the egg and early embryo (Casanova and Struhl, 1989; Sprenger et al., 1989). These are translated after fertilization and TORSO protein is found on the surface of the entire early embryo (Casanova and Struhl, 1989).How then is the terminal signal transduced only in the termini? All experimental data so far support the model that TORSO’Sligand (or its activity) is spatially restricted to the perivitelline space at the termini (see the next section). Since the TORSO receptor is present in excess, it will sequester the limited amount of ligand and prevent it from diffusing away from the termini (Casanova and Struhl, 1993) . As a result the terminal signal is only transduced at the poles. It is assumed, by analogy to what is known about vertebrate receptor tyrosine kinases, that ligand binding results in receptor oligomerization with consequent transient phosphorylation of the TORSO receptor’s cytoplasmic domain on tyrosine residues, probably through autophosphorylation (Yarden and Ullrich, 1988; Schlessinger and Ullrich, 1992). This activation of the kinase activity of the receptor initiates a cytoplasmic signal transduction cascade. Before detailing this cascade, we discuss what is known about the TORSO receptor’s ligand and its activation.
6. The Extra-embryonic Ligand Mutations in four maternal genes [fs(l)Nasrat,fs(l)polehole, torsolike, and trunk] produce phenotypes closely resembling those of torso mutations (Table 1 and Figure 2) and behave genetically as if they act upstream of the TORSO receptor based on their inability to suppress torsOgof phenotypes. Three of these Ifs(l)Nasrat, fs(l)polehole, and trunk] are expressed in the germline (Perrimon and Gans, 1983; Schupbach and Wieschaus, 1986a) while the fourth, torso-like, is expressed in the follicle cells that surround the oocyte (Stevens et al., 1990). A clone of between six and 30 mutant torso-like follicle cells at the posterior end of the developing oocyte is sufficient to cause deletion of the telson in the posterior terminal region (Stevens et al., 1990). Consistent with this spatially restricted requirement, expression of the torso-like gene is restricted to the anterior border cells and posterior follicle cells, and ectopic expression of TORSO-LIKE during oogenesis generates phenotypes similar to those produced by tors0gOf mutations (Savant-Bhonsale and Montell, 1993; Martinet al., 1994).While the TORSO-LIKE protein has an amino-terminal signal sequence and is secreted, it bears no obvious similarity to other proteins. Thus, the nature of the localized
Drosophila Terminal Gene Hierarchy
103
figure 2. The TORSO-mediated signal transduction pathway. The T.ORS0 receptor (TOR) is shown interacting with its ligand, TRUNK (TRK), which is converted into its active form by the TORSO-LIKE protein (TSL). Two cytoplasmic signal transduction pathways are shown, one to the right and one to the left of the TOR receptor. To the right, the signal is transduced through DRK, SOS, RAS, GAP, RAF, DSORl, RL, and affects the phosphoryfation state of transcription factors X and Y that then control nuclear transcription of the downstream effector genes tailless (tll) and huckebein (hkb). To the left, possible additional participants, SRC and CSW, are shown. For details of the functions and phenotypes of each component, see the text. Symbols: Mickey Mouse ear = SH2/SH3 domains; rectangle = phosphatase domain; ellipsoids in the cytoplasm = kinase domain; ? = uncertain step or player in the pathway. information provided by TORSO-LIKE is unclear; it may fknction to activate an otherwise uniformly-distributed inactive ligand for the TORSO receptor (Figure 2 ) . The latter possibility appears most likely since trunk has recently been shown to encode a secreted protein with homology to SPATZLE, the likely ligand for TOLL (Casanova et al., 1995). Thus, it is likely that TORSO-LIKE activates TRUNK which binds to TORSO (Figure 2).
104
MAN LUN R. YIP and HOWARD D. LlPSHlTZ
While the torso-like and trunk loss of function mutant phenotypes are identical to those of torso loss of function phenotypes, those produced byfs(l)Nusrut andfs(1)pole hole are more pleiotropic. Females homozygous for most alleles offs(l)Nusrat orfs(l)pole hole lay normal-looking eggs that collapse soon after deposition, are permeable to neutral red, and burst upon removal of the chorion (Degelmann et al., 1990). These alleles are assumed to be genetically amorphic. One mutant allele at each locus Ifs(l)Nusru$" and fi(1)pole hole'9o'], results in the typical terminal class mutant phenotype, while two alleles offs(1)Nasrut Ifs(l)Narut' andfi(I)NusruPH'] result in both the collapsed egg and the embryonic terminal phenotypes. Interestingly,fs(l)NusrupH', which is a temperature sensitive allele, results in phenotypes that range from ones similar to those produced by torso loss of function alleles to ones similar to weak torso gain of function phenotypes. It has been speculated that fs(1)Nusrut and fs(1)pole hole encode vitelline membrane proteins or extracellular matrix-associated proteins that interact with the TORSO protein or its ligand (Figure 2) and/or provide structural support for the oocyte (Degelmann et al., 1990; Lipshitz, 1991). C. Cytoplasmic Signal Transduction
The cytoplasmic signal transduction pathway initiated by the TORSO receptor shares many components with other receptor tyrosine kinase (RTK)-mediated pathways, such as those initiated by SEVENLESS in the developing eye (Hafen et al., 1994) and by the Drosophila homolog of the epidermal growth factor receptor (DER) in various tissues (Shilo and Raz, 1991; Shilo, 1992). The known components are listed in Table 1 and Figure 2 and are summarized here. The downstream ojreceptor kinuses (drk) gene encodes a protein containing SRC homology 2 (SH2) and SRC homology 3 (SH3) domains (Olivier et al., 1993; Simon et al., 1993). drk mutations were initially identified as extragenic modifiers of SEVENLESS receptor tyrosine kinase activity in the eye. SH2 domains are capable of binding receptor tyrosine kinases in vivo and in vitro in a phosphorylation-dependent manner (Pawson and Schlessinger, 1993). Consistent with such a role, DRK protein has been shown to bind SEVENLESS protein in vitro and in vivo (Olivier et al., 1993; Simon et al., 1993). A role for DRK in the embryonic terminal gene hierarchy is suggested by the fact that torsOgof allele; such a genetic interphenotypes are suppressed by the dr@fsev)2B action would place DRK downstreamof the TORSO receptor. Consistent
Drosophila Terminal Gene Hierarchy
105
with this interpretation, the drkE(JeidZ5 allele has a single amino acid substitution in the SH2 domain (Doyle and Bishop, 1993; Olivier et al., 1993), thus it is likely to interfere with binding of DRK to the activated TORSO receptor, hence affecting the efficiency of coupling to downI stream signaling molecules. RAS, a protein with GTPase activity, is involved in many intracellular signal transduction pathways (Lowy and Willumsen, 1993). Its activity is regulated by bound guanine nucleotide (Bourne et al., 1990, 1991): RAS is active when bound to GTP and inactive when GDP is bound. Cycling between the two states largely depends on two opposite activities: ( I ) an RAS guanine nucleotide exchange factor that catalyzes the exchange of GDP for GTP, and (2) an RAS GTPase activating protein (GAP) that stimulates the intrinsic GTPase activity of RAS protein to hydrolyze bound GTP. A Drosophila homolog of RAS, RASI, was first implicated in the SEVENLESS signaling pathway (Simon et al., 1991). Loss of fhction mutations in the Rasl gene cause embryonic lethality and function as dominant suppressors of the tors&"' phenotype (Doyle and Bishop, 1993), while maternal expression of a dominant activated RAS 1 protein (RAS 1G'"'3) phenocopies the tors@"fphenotype (Lu et al., 1993). Further, direct injection of variants of mammalian p2 1ras protein into Drosophila embryos have indicated a role in the TORSO signaling pathway: activated p21ra5(p2 1 "-,)' partially rescues the torso'"' phenotype while dominant-negative forms of p2 Iras (p2 lrasN17) block the terminal signal in wildtype embryos and generate a torso'"' phenotype (Lu et al., 1993). Mutations in the Drosophila Gapl gene, which encodes a protein with homology to mammalian GTPase activating protein (GAP), were identified as producing extra R7 cells in a sensitized sevenless genetic background (Gaul et al., 1992; Rogge et al., 1992) and as causing hyperinnervation of the R7/R8 retinotopic map (Buckles et al., 1992). The homology to mammalian GAP suggested a function in the SEVENLESS pathway through regulation of RAS 1 activity. GAP 1 may also function as a negative regulator in the TORSO-mediated pathway since Gap/ mutations enhance tors&"' phenotypes (Doyle and Bishop, 1993). The Drosophila homolog of RAS guanine nucleotide exchange factor is encoded by the Son of sevenless (Sos) gene. Gain of function alleles of Sos were isolated originally as dominant suppressors of sevenless mutations (Rogge et al., 1991; Simon et al., 1991; Bonfini et al., 1992). Most loss of function alleles of Sos are lethal and embryos derived from germline clones of Sos'"' show either a terminal class phenotype (zygotic
106
MAN L U N R. YIP and HOWARD D. LIPSHITZ
genotype: Sod+) or very little differentiation (zygotic genotype: Sos/Sos; Lu et al., 1993). Thus, the torso''/ phenotype produced by direct injection of a dominant negative form of p2lraS(see above) is probably due to competition with endogenous RAS for binding to the SOS protein (Lu et al., 1993). So~'~~alleles act as dominant suppressors of tors&'/phenotypes (Doyle and Bishop, 1993). l(1)pole hole [distinct from.fs(l)pole hole, discussed above] encodes the Drosophila homolog of the RAF serinelthreonine kinase (D-RAF; Nishida et al., 1988; Ambrosio et al., 1989b). D-RAF is expressed and required both maternally and zygotically, but it is maternally expressed D-RAF that fhctions specifically in the terminal signal transduction pathway. Embryos derived from homozygous l(1)pole hole germline clones show a typical terminal class phenotype (Ambrosio et al., 1989b) and embryos that are deprived of both maternal and zygotic l(1)pole hole function show incomplete development and massive cell death (Ambrosio et al., 1989a; Melnick et al., 1993). While suppression of the tors&"'phenotype by l(1)pole hole mutations places D-RAF downstream of the TORSO receptor (Ambrosio et al., 1989b), the failure of injection of activated ~ 2 1 to " ~rescue embryos that lack maternal D-RAF activity suggests that D-RAF functions downstream of RAS in the terminal pathway (Lu et al., 1993). Similar conclusions have been reached from molecular genetic analyses of RAS/RAF-mediated vulva1 induction in Cuenorhubditis eleguns (Sternberg, 1993). Recent data have demonstrated a GTP-dependent direct interaction between the amino-terminal domain of RAF and the effector domain of RAS (Vojtek et al., 1993; Warne et al., 1993; Zhang et al., 1993). The Drosophilu CORKSCREW protein bears homology to cytoplasmic tyrosine phosphatases that also contain two SH2 domains (Perkins et al., 1992). As is the case for D-RAF and SOS, CORKSCREW activity is required both maternally and zygotically. Embryos derived from eggs produced by homozygous corkscrew germline clones show characteristic terminal class pattern defects, as well as twisted gastrulation and germband retraction defects reminiscent of those exhibited by embryos derived from torso"f mothers (see also Section 111 below). Mutations in the corkscrew gene suppress torsog" phenotypes, suggesting that corkscrew lies downstream of TORSO in the terminal pathway, however, the actual position of CORKSCREW in the cytoplasmic signal transduction pathway is uncertain. As for null Z(1)pole hole alleles, null corkscrew alleles do not completely abolish expression of tailless (see Section 111). Embryos from which maternal contributions of functional D-RAF and
Drosophila Terminal Gene Hierarchy
107
RAF and CORKSCREW have been eliminated lack posterior tailless expression; this was initially taken to suggest a possible interaction of CORKSCREW with D-RAF (Perkins et al., 1992). However, injection of p2 1v-m protein into embryos lacking maternal corkscrew can rescue the phenotype, positioning CORKSCREWupstream of RAS1 (Lu et al., 1993), whereas similar experiments (see above) place D-RAF downstream of RASI . Recent biochemical data from mammalian cells have shown that GlU32 (the DRK homolog), can either directly bind activated PDGF receptor (the mammalian TORSO homolog; Arvidsson et a]., 1994) or form a complex with SHPTP2 (the mammalian homolog of CORKSCREW; also known as SYP/PTPlD/PTP2C) and then bind activated PDGF receptor (Bennett et al., 1994; Li et al., 1994). That the same is likely to hold true for the Drosophilu terminal pathway is supported by the data that positions CORKSCREW upstream of RAS 1. The Dsorl gene encodes the Drosophilu homolog of the mammalian MAP kinase activator (MEK) and yeast PBS2, STE7, and BYRl (Tsuda et al., 1993).Mutations in Dsori were identified as dominant suppressors of l(1)pole hole and were subsequently found to suppress other terminal class mutants Vs(1)pole hole, torso-like, trunk, and corkscrew]. Elimination of functional maternal DSORl from embryos produces phenotypes similar to those in embryos produced by l(i)pole hole germline clones. These results suggest that DSORl acts downstream of D-RAF. ADrosophilu homolog of MAP kinase (Biggs and Zipursky, 1992) is encoded by the rolled locus. Rolled mutants exhibit defects in multiple receptor tyrosine kinase signaling pathways (Biggs et al., 1994; Brunner et al., 1994b). Gain of function mutations in rolled cause similar phenotypes to those produced by torsPfrnutations. Loss of function mutations in rolled can suppress terminal class phenotypes caused by tors8”falleles (Brunner et al., 1994b). The position of D-SRC, the Drosophilu homolog of mammalian C-SRC, in the terminal signal transduction pathway remains unclear. Ectopic expression of wildtype and mutant forms of D-SRC interfere with eye development and with germband retraction in the embryo (Kussick et al., 1993). This phenotype is dependent on functional U S 1 activity. In mice, the oncogenic form of C-SRC, V-SRC, can constitutively phosphorylate and enhance SYP/PTP1D (CORKSCREW)activities (Feng et al., 1993; Vogel et al., 1993). Moreover, RAF and SRC coimmunoprecipitatewhen expressed in vitro, indicating possible direct interactions between RAF and SRC (Cleghon and Morrison, 1994). By
108
MAN LUN R. YIP and HOWARD D. LIPSHITZ
analogy, D-SRC may function downstream of the TORSO receptor and interact with CORKSCREW and D-RAE Four of the segment polarity genes (see Section III[C]) are expressed maternally and mutations result in terminal defects in embryos derived from germline clones. Two of these (shaggy/zeste-white3 and fused) encode serinehhreonine kinases, one (dishevelled) a novel protein, and one (porcupine) has not yet been studied molecularly. For logistical reasons these are considered in our discussion of pair-rule genes, however, it should be noted here that their gene products may reside in the maternally encoded signal transduction pathway. D. Conservation of Signal Transduction Pathways
As we have emphasized, the terminal signal transduction pathway in the Drosophila embryo shares many components with related receptor tyrosine kinase-mediated pathways both in Drosophila and in other metazoa. The best studied examples include the SEVENLESS-mediated pathway in Drosophila eye development (Hafen et al., 1994); various growth factor responses in mammalian cells (Yarden and Ullrich, 1988; Schlessinger and Ullrich, 1992); vulva1 induction in Caenorhabditis elegans (Stemberg, 1993); and the pheromone response in both Saccharomyces cerevisiae and Schizosaccharomycespombe (Kurjan, 1993). Typically, ligands activate the pathway by binding to transmembrane receptors, inducing oligomerization, and conformational changes that are transmitted across the cell membrane and catalyze the phosphorylation and activation of the receptor’s kinase activity. Phosphorylation of cytoplasmic signal transduction molecules follows. The participation of small GTP-binding proteins such as RAS, or of G proteins, in these pathways is important. Downstream of the GTP-binding proteins are RAF and the MAP kinase cascade which eventually modulate the activities of transcriptional regulators through phosphorylation. Studies in other systems have identified transcriptional regulators that are modulated by the MAP kinase cascade. In mammalian systems, the activity of transcription factors-such as ELK-1 (which belongs to the ETS-family), C-JUN, and C-MYC-is affected by their phosphorylation state, and this state is regulated by the MAP kinase cascade (Hunter and Karin, 1992). In the Drosophila SEVENLESS-mediatedpathway, it has been shown that a Drosophila homolog of C-JUN as well as two ETS-family proteins (YAN and POINTED) are required for R7 development and that the latter can be phosphorylated in vitro by the MAP
Drosophila Terminal Gene Hierarchy
109
kinase cascade (Brunner et al., 1994a; O’Neill et al., 1994). In the Drosophilu tracheal system-development of which is regulated by the Drosophilu homolog of the FGF receptor, BREATHLESS-POINTED functions to program cell growth and migration (Klambt, 1993). At this point, little is known abodt transcription factors in the Drosophilu embryo that might be the mediators of the terminal signal into the nucleus, however, it is likely that one or more of the above proteins (or related proteins yet to be identified) will be involved. One unanswered question involves the nature of the control exerted by such activated (or repressed) transcriptional control proteins. Clearly not all of the genes that they turn on and off can be the same in all tissues and at all times. In other words, an outstanding question-and one of our major areas of ignorance-is how conserved cytoplasmic signal transduction pathways and their target transcription factors result in differential gene regulation. The next section will focus on the specific readout of the terminal signal in terms of zygotic genes that respond to it. We then go on (Section IV) to consider how cells are instructed as to their positions along the dorso-ventral and anterior-posterior axes within the termini, as well as how, the anterior terminal region is elevated from the “ground state” represented by the posterior terminus through interaction of the terminal and anterior genetic pathways.
111. T H E Z Y G O T I C EFFECTORS O F T E R M I N A L CELL FATES A. Subdivision of the Termini into Distinct Tissues
Before describing the assorted zygotic terminal pathway genes and their roles, it is first useful to define some of the details of terminal development. Much of the emphasis in discussions of antero-posterior axis specification in Drosophila has been on the central, segmented part of the embryo. The maternal hierarchies involved in these regions are the so-called “anterior” and “posterior” systems, and the developmental problem in the central region is largely one of subdividing a domain into repeating units-parasegments/segments-and then elaborating differences among those repeated units (St. Johnston and Nusslein-Volhard, 1992; Martinez Arias, 1993; Pankratz and Jackle, 1993). (Parasegments are segment-sized embryonic developmental units that are out of phase with the actual morphologically defined segments by half of a segment [Martinez Arias and Lawrence, 19851.) This is a problem specific to
110
MAN LUN R. YIP and HOWARD D. LlPSHlTZ
insects that use the long germband mode of development in which the entire segmented region is specified from a single field of cells at the blastoderm stage. In contrast, short germband insects and other metazoa do not face this problem; segmental repeats either are not present or they develop sequentially with consequent differences in the nature of positional specification and the mechanisms used to implement it. in contrast to the antero-posterior axis in the segmented region of Drosophilu, cell fate specification mechanisms along the dorso-ventral axis and in the embryonic termini more closely resemble those implemented generally in other metazoa. That is, cells must be instructed as to their position and then must develop into different tissues (rather than metameric units) dependent on their location. Along the dorso-ventral axis, this largely involves subdivision into (from ventral to dorsal): mesoderm, neuro-ectoderm, epidermal ectoderm, and extra-embryonic tissues. In the termini, this process involves subdivision into (from terminus toward center) endoderm, intestinal ectodenn, epidermal ectoderm, and (in the anterior) brain. Since these processes more closely resemble those in other metazoa, it is probably no accident that, in terms of molecular pathways, the maternally encoded components of both the dorso-ventral and the terminal systems use cell-cell interaction and signal transduction mechanisms to instruct cells as to their fates, rather than the more specialized system of localized cytoplasmic determinants used by the maternal anterior and posterior systems (Lipshitz, 1991;Ding and Lipshitz, 1993). Given this view, we shall focus on the effectors of terminal fates, not in terms of drawing arbitrary boundaries on a field of cells, but rather by focusing on how cells are directed into distinct tissue fates with consequent morphogenetic outcomes that include coordinated invagination and cell movement. It should be emphasized at this point that the terminal gene pathway does not act in isolation: the anterior terminal region gives rise to structures and tissues (e.g., the brain) distinct from those that form posteriorly. This is accomplished through an interaction of the terminal, the anterior, and the dorso-ventral pathways in the anterior terminal domain (see IV[C] below). Similarly, the dorsal domains of the anterior, and posterior termini have different fates from the ventral domains and this is accomplished through interaction of the terminal and dorso-ventral pathways at the termini (see IV[B] below). The zygotic genes that implement terminal cell fates have been defined using two related strategies. First, genetic screens and genetic-
Drosophila Terminal Gene Hierarchy
111
interaction tests have been used to define zygotically active genes with phenotypes that are similar to (or subsets of) those produced by the maternal terminal class genes (reviewed in Lipshitz, 1991). Where possible, it has been useful to complement such analyses with tests for possible genetic interaction with defined maternal components (e.g., by tests for interaction with tors0g'f mutations; Strecker et al., 1989, I99 1, 1992). Second, the definition of cloned genes that are expressed in the embryonic termini and examination of their expression patterns in terminal pathway mutants, have also made it possible to position many of these genes in the terminal regulatory hierarchy. Taken together, these approaches have led to the definition of more than 50 zygotic genes that are involved (or are likely to be involved) in specifying aspects of cell fate in the anterior and posterior termini (Table 1; for a recent review see Jiirgens and Hartenstein, 1993). B. Subdivision into Endoderm versus Ectoderm
Two genes function to subdivide the termini into distinct tissues: huckebein and tailless (Table 1 and Figure 3). Tailless is required to program the development of much of the ectoderm and its derivatives at the termini, including the brain at the anterior (Jiirgens et al., 1984; Strecker et al., 1986, 1988), while huckebein is required for the establishnient of the endoderm at both poles and also the intestinal ectoderm and labhm anteriorly (Weigel et al., 1990; Bronner and Jackle, 1991; Bronner et al., 1994; Reuter and Leptin, 1994). Tailless was the first of the zygotic genes shown by genetic interaction tests to reside in the terminal hierarchy: torsOgofphenotypes are suppressed by loss of function tailless mutations (Klingler et al., 1988; Strecker et al., 1989). The tailless gene encodes an "orphan" receptor belonging to the steroid-receptor superfamily to which it shows homology in both its putative ligand-binding and its DNA-binding domains (Pignoni et al., 1990). Transcription of the tailless gene is activated directly by the TORSO-mediated terminal signal and occurs in those regions ofthe termini in which it is known to function (Table 1 and Figure 3). The anterior terminal expression of TAILLESS undergoes more dramatic changes than does its posterior expression; these are dependent on input from the BICOID-mediated anterior system and the dorsoventral pathways (see Section IV for more detailed discussion). Genetic analyses (Streeker et al., 1989; Strecker and Lipshitz, 1990) led to the conclusion, later supported by molecular studies (Steingrimsson et al.,
112
MAN LUN R. YIP and HOWARD D . LlPSHlTZ
1
I
I
100
90
80
i %EL 70
I
I
1
30
20
10
, %EL 0
(continued)
Figure 3. Fate maps of the anterior and posterior termini (A) and expression/functional domains of selected terminal pathway genes (B). (A) The anterior terminal domain extends from roughly 70-1 00% egg length (EL), while the posterior terminal domain extends from 0-25% EL. Abbreviations of structures on the fate map: A# = abdominal segment #; Ac = acron; Amg = anterior midgut; Ap = anal pads; f = fell; fk = filzkorper; Hg = hindgut; Lr = labrum (clypeolabrum); Mp = Malpighian tubules; Pmg = posterior midgut; Te = telson; At = tuft. Small ovals in the anterior represent regions that will contribute to the head skeleton: dbr = dorsal bridge; da = dorsal arms; Ir = labrum; vp = vertical plates. Sensory organs are symbolized by filled squares or asterisks: anteriorly, the labrai sense organ (lrso), and posteriorly the dorsal medial sense organ hair (asterisk in A8), anterolateral sense organ hair (asterisk in A9, anterior), posterior-lateral sense organ hair (asterisk in A9, posterior), anal sense organ hair (asterisk in A1 01, antero-lateral sense organ peg sensillum (square in A8), posterior-lateral sense organ peg sensillum (square in A9), anal sense organ peg sensillum (square in AIO), dorsal medial sense organ peg sensillum (square in f-fk region). (Figures based on those in Jurgens et at., 1986, 1987; Kuhn et at., 1992; Jurgensand Hartenstein, 1993.)
1991), that TAILLESS has dual functions in the termini: to promote terminal cell identities and to suppress central cell identities in the terminal region. Ectopic expression of TAILLESS in embryos using a heat-inducible promoter can phenocopy tors&'' phenotypes. This is achieved through activation of genes such as hunchback and repression of genes such as Kriippel and knirps (Steingrimsson et al., 1991). Due to the absence of the posterior midgut anlagen, huckebein mutant embryos fail to undergo posterior midgut invagination and they also
-
Drosophila Terminal Gene Hierarchy I
Amg
Lr
Ac
Fg
I
1
I
I
=
m
V W -
113 I
I
1
1
1
1
1
torso
huckebein tailless lines
m
A10 TeApHgMp Pmg
A9
AR
,
empty spiracles
=
hunchback giant caudal spalt Abd-BI
D
m
Abd-BII Kruppel
R?
=
fushi tarazu hairy even-skipped forkhead HNF-4 serpent
folded gastrulation
D
l?zZm y/////7/11
hindsight decapentaplegic zerknullt
figure 3. (Continued) Fate maps of the anterior and posterior termini (A) and expression/functionaI domains of selected terminal pathway genes (B). (B) Bars represent expression and/or functional domains of selected terminal hierarchy genes. Cross-hatching represents uncertainty in the extent of the expression or functional domain. (Figure based on Jurgensand Harten-
stein, 1993.) show defects in germband extension. Similarly, the anterior midgut and stomodeal invaginations are abnormal. The formation of the anterior gut structures requires synergistic interactions between huckebein and the dorso-ventral genes, snail and twist (Reuter and Leptin, 1994). Huckebein encodes a protein with similarity to the SPl/EGR family of zincfinger transcription factors. As in the case of tailless, mutations in huckebein suppress phenotypes (Weigel et al., 1990). Ectopic expression of HUCKEBEIN perturbs segmentation in the ectoderm and
114
MAN LUN R. YIP and HOWARD D. LIPSHITZ
suppresses the segregation of mesoderm (Bronner and Jackle, 1991 ; Bronner et al., 1994). Huckebein RNA and HUCKEBEIN protein are expressed at the syncytial blastoderm stage in polar caps that overlap with the zones of TAILLESS expression. Unlike the gap genes expressed in the central, segmented region, the products of the huckebein and tailless genes do not cross-regulate each other's expression (Bronner and Jackle, 199 1). Embryos that are homozygous for both huckebein and tailless mutations show posterior terminal region phenotypes that are almost identical to those in embryos derived from homozygous torsolofmothers (Weigel et al., 1990), suggesting that much of terminal development-articularly at the posterior-is programmed through the combined action of these two zygotic effector genes. C. Genes that Program Ectodermal Development
Numerous genes function downstream of tailless and huckebein to specify the details of ectodermal development in the termini (Table I and Figure 3), and these are summarized here. Mutant lines embryos die with terminal and central region defects (Nusslein-Volhard et al., 1984). The terminal defects are similar to those of tailless mutants; indeed, lines and tailless mutations show synergistic interactions (Strecker et al., 1991). In addition lines mutations suppress the phenotypes caused by both torsdofmutationsand ectopic expression of TAILLESS (Strecker et al., 1991, 1992). These results suggest that lines acts downstream of tailless in the terminal pathway. Lines has not yet been analyzed molecularly. Hunchback mutations are also able to suppress torsdofphenotypes (Strecker et al., 1991). The hunchback gene encodes a zinc-finger transcription factor that is expressed both maternally and zygotically, and it functions in the termini as well as in the central region of the embryo. Only the terminal functions of hunchback are considered here. Specifically, zygotic expression of HUNCHBACK in the posterior region is required for the proper formation of the seventh and eighth abdominal segments, the most centrally located part of the posterior terminal epidermal ectoderm (Lehmann and Nusslein-Volhard, 1987). This expression is dependent on the terminal signal. It is activated by TAILLESS and repressed by HUCKEBEIN (Casanova, 1990; Bronner and Jackle, 1991).
Drosophila Terminal Gene Hierarchy
115
The giant gene encodes a transcriptional regulatory protein with characteristics of the b-ZIP family of DNA-binding proteins (Capovilla et al., 1992) and functions to specify fates in the termini and the central region of the embryo. Mutations in giant suppress tors0g'f phenotypes (Strecker et al., 1991). The major terminal pathway function of giant is in the anterior where it is required for the development of the labrum (Petschek et al., 1987; Mohler et al., 1989; Petschek and Mahowald, 1990). In torsoLDfembryos, the most anterior stripe of GIANT expression (stripe 1) is missing while in torsdo' embryos this stripe expands centrally. The combined activities of TAILLESS and HUCKEBEIN are required to activate stripe 1 GIANT expression (Eldon and Pirrotta, 1991). The empty spiracles gene encodes a homeodomain-containing transcription factor that functions both inside and outside the terminal developmental domain. Within the posterior domain it programs the formation of the filzkiirper that reside inside the posterior spiracles (Dalton et al., 1989)and anteriorly it functions in optic lobe development (Walldorf and Gehring, 1992). Empty spiracles mutations suppress phenotypes induced by ectopic expression of TAILLESS, suggesting that it acts downstream of TAILLESS in the terminal pathway (Strecker et al., 1992). The tramtrack gene also functions in both terminal and central embryonic .development, as well as postembryonically. It was originally identified based on TRAMTRACK protein binding cis-regulatory sequences of two pair-rule genes, fushi tarazu, and even-skipped. Tramtrack encodes two related proteins, p69 and p88, that differ in their carboxy-terminal domains, and possess different pairs of C,H, zincfingers (Harrison and Travers, 1990; Brown et al., 1991; Read and Manley, 1992). Null alleles of tramtrack are embryonic lethal; mutant embryos show severe cuticular defects and fail to form filzkorper posteriorly and certain head structures anteriorly. These terminal defects suggest that tramtrack may reside in the terminal pathway (Xiong and Montell, 1993). The spalt mutation causes incomplete transformation of the labial segment into a prothorax-like segment, and the tail into structures that resemble the eighth abdominal segment. These have been interpreted as homeotic transformations toward more central identities, indicating that spalt functions to confer posterior terminal identities in parasegments 14 and 15 (Jurgens, 1988). Spalt encodes a zinc-finger-containing protein and is likely to be a transcriptional regulator (Kuhnlein et al., 1994). The
116
MAN LUN R. YIP and HOWARD D. LIPSHITZ
posterior expression domain of SPALT coincides with the TAILLESS expression domain; spalt expression is absent in tailless mutant embryos (Schuh and Jaeckle, cited in Jiirgens and Hartenstein, 1993). The Abdominal-B gene of the bithorax complex encodes two homeodomain transcription factors referred to as ABDOMINAL-BI and ABDOMINAL-BII (also called M-ABDB and R-ABDB). Within the termini, ABDOMINAL-BI specifies cell identity in the epidermal ectodermal derivatives of parasegment 13 while ABDOMINAL-BII specifies cell identity in the epidermal ectodermalderivatives of parasegments 14 and 15 of the embryo (Celniker et al., 1990). TAILLESS activates ABDOMINAL-BI expression in parasegment 13 (Reinitz and Levine, 1990). Activation of ABDOMINAL-BII expression in parasegments 14 and 15 is absent in embryos from torso”f females and in tailless mutant embryos, and ABDOMINAL-BII is ectopically expressed in embryos from tors&’’ females (Casanova, 1990). These data indicate that both ABDOMINAL-BI and ABDOMINAL-BII hnction in the terminal hierarchy. The forkhead gene is under the dual control of HUCKEBEIN and TAILLESS and functions in both midgut (endodermal) and hindgut (ectodermal) development, including the Malpighian tubules (Jurgens and Weigel, 1988; Weigel et al., 1989a). Forkhead encodes a transcription factor (Weigel et al., 1989b).With regard to its ectodermal functions, inforkhead mutants hindgut and foregut are replaced by more centrally derived structures, suggesting that FORKHEAD is required to confer terminal versus central cell fates in these regions. The Drosophila homolog of the mouse Brachyury (T) gene, T-related gene (Trg;),is required for the formation of the hindgut and the anal pads (Kispert et al., 1994). The T gene product of mice has been shown to bind DNA and possibly to function as a transcriptional regulator (Kispert and Herrmann, 1993; Herrmann and Kispert, 1994). Expression of the Drosophila Trg gene is dependent on TAILLESS but not on FORKHEAD. Repression of Trg expression in the posterior midgut primordium is dependent on HUCKEBEIN. Antibodies directed against the mouse T protein cross-react with the Drosophilu TRG protein as well as putative T homologs in the gut primordia of embryos of the short germband insects, Locusta and Tribolium, suggesting possible evolutionary conservation of the’mechanisms used to establish hindgut identity (Kispert et al., 1994). The Malpighian tubules derive from the hindgut and form at the junction of the hindgut and the midgut. They are specified by TAILLESS
Drosophila Terminal Gene Hierarchy
117
and HUCKEBEIN, which overlap in expression in this region and function jointly there. Four genes are required for the proper formation of the Malpighian tubules (Harbecke and Janning, 1989; Gaul and Weigel, 1991; Liu and Jack, 1992): cut, which encodes a homeodomaincontaining protein (Blochlinger et dl., 1988); Kriippel, which encodes a Zn-finger transcription factor (Rosenberg et al., 1986); caudal, which encodes a homeodomain-containing transcription factor (Macdonald and Struhl, 1986); and HNF-4, a Drosophila homolog of hepatocyte nuclear factor-4 (Zhong et al., 1993). In cut, Kruppel and caudal mutants the cells that would normally give rise to the Malpighian tubules take on hindgut characteristics, indicating that the wildtype functions of these genes is to convert the identity of these progenitor cells from a “ground” hindgut state to that of Malpighian tubule. Forkhead expression is activated by TAILLESS and HUCKEBEIN. FORKHEAD then activates Kruppel expression, and KRUPPEL activates expression of caudal and cut. The caudal and cut genes are expressed independently of each other. HNF-4 is expressed in the primordia of the Malpighian tubules and mutant embryos lacking zygotic HNF-4 activity fail to form the Malpighian tubules. Detailed phenotypic analyses of HNF-4 mutants have not yet been reported. Besides its function in Malpighian tubule development, zygotic expression of CAUDAL is also required for the formation of the anal pads, struckres derived from the most terminal region of the posterior cuticular ectoderm (Macdonald and Struhl, 1986). Zygotic expression of the caudal gene in this region is dependent on TAILLESS (Mlodzik and Gehring, 1987b). The cnc (cap ’n ’collar) gene encodes a potential leucine zipper transcription factor and is expressed in the anterior of the embryo, including the labral primordium in the anterior terminal domain (Mohler et al., 199I). An anterior cap ofcnc gene expression is activated by the BICOID and TORSO-mediated pathways at nuclear cycle thirteen (see also IV[C] below). Later its posterior extent is refined by TAILLESS and SPALT, its anterior extent by GIANT, and its ventral extent by the dorso-ventral hierarchy; all of these interactions repress cnc and so restrict its domain of expression to part of the labral primordium (Mohler, 1993). Mutations in cnc have recently been reported and affect labral development (Mohler et al., 1995). The “pair-rule” and “segment polarity” genes were initially identified on the basis of their role in the development of the central, segmented region of the embryo; the pair-rule genes function to subdivide this
118
MAN LUN R. YIP and HOWARD D. LIPSHITZ
region into segmentally repeated units known as parasegments while the segment polarity genes are required for maintenance of the parasegmental borders. Most of the pair rule and segment polarity genes are also expressed in the embryonic termini. Their phenotypes and h c t i o n s there are less well defined (Table l), and their positions in the terminal gene hierarchy remain to be studied in detail. We consider these genes briefly below. There are 11 pair-rule genes: hairy, even-skipped,fushi tarazu, oddpaired, odd-skipped,ten"lodd Oz,paired, hopscotch,runt, sloppypaired, and unpaired. Hairy encodes a helix-loop-helix family transcription factor (IshHorowicz et al., 1985; Rushlow et al., 1989b) that is expressed in a patch of cells in the dorsal head region in the anterior terminal domain while, posteriorly, the seventh stripe of hairy expression falls within the terminal domain (Ingham et al., 1985). Expression of this posterior stripe is strongly reduced or absent in tailless mutant embryos, placing hairy in the terminal gene hierarchy downstream of tailless (Mahoney and Lengyel, 1987; Hooper et al., 1989). In the termini, hairy mutants shows defects in structures that derive from these regions (Jiirgens, 1987). The even-skippedgene encodes a homeodomain-containingtranscription factor (Nusslein-Volhard et al., 1985; Macdonald and Struhl, 1986; Frasch et al., 1987). Even-skipped mutants exhibit posterior terminal defects (Niisslein-Volhardet al., 1985) coincident with the seventh stripe of EVENSKIPPED expression. This stripe is absent in tailless mutant embryos (Frasch and Levine, 1987; Goto et al., 1989). Thefushi tarazu gene encodes a homeodomain-containingtranscription factor (Laughon and Scott, 1984). Fushi tarazu mutants exhibit defects in structures derived from the posterior terminal region (Wakimot0 et al., 1984; Jiirgens, 1987) and coinciding with its seventh stripe of expression. This stripe is absent in embryos derived from torso'Of females and in tailless mutant embryos, and is expanded in embryos produced by tors0g"fmothers (Mahoney and Lengyel, 1987; Strecker et al., 1989, 1991; Strecker and Lipshitz, 1990). Several additional pair-rule genes are expressed in the termini and mutants exhibit defects in the development of structures derived from these regions. Presumably these reside in the terminal gene hierarchy, but analyses of their expression in terminal pathway mutants have not been reported. These genes are odd-paired,which encodes a zinc-finger containing protein (Jiirgens, 1987; Benedyk et al., 1994); odd-skipped, which also encodes a zinc finger protein (Coulter and Wieschaus, 1988;
Drosophila Terminal Gene Hierarchy
119
Coulter et al., 1990); ten"/odz, which encodes a protein homologous to TENASCIN, with eight EGF and 11 fibronectin I11repeats (Baumgartner et al., 1994; Levine et al., 1994);paired,which encodes a transcription factor with two highly conserved domains, a paired box and a homeodomain (Nusslein-Volhard et al., 1985; Bopp et al., 1986; Frigerio et al., 1986; Kilchherr et al., 1986); and hopscotch, which encodes a JAKfamily tyrosine kinase (Perrimon and Mahowald, 1986; Binari and Perrimon, 1994). No analyses of terminal mutant phenotypes have been reported for the remaining pair-rule genes: runt, sloppy paired, and unpaiped. Two of these have been analyzed molecularly: runt encodes a runt domaincontaining transcription factor (Gergen and Butler, 1988) and sloppy paired encodes twoforkhead domain-containing transcriptional regulatory proteins (Grossniklaus et al., 1992). Both of these are expressed within the terminal domains and so are likely to function in the terminal gene hierarchy. The segment polarity genes are involved in the establishment and maintenance of the parasegments through control of cell-cell interactions (reviewed in Martinez Arias, 1993), and 12 of these are known to be expressed in the termini: patched, engrailed, wingless, dishevelled, shaggy/zeste-white 3, armadillo, naked, porcupine, hedgehog, cubitus interruptus,gooseberry, andfused. While expression within the terminal domain as well as terminal mutant phenotypes have been reported for several of the segment polarity genes, no detailed analyses have been conducted that would position them within the terminal hierarchy. Consequently they are only considered briefly here (Table 1): patched encodes a putative transmembrane protein and mutants have defects in the tail (Jurgens, 1987); wingless encodes a protein homologous to vertebrate INT- 1 and mutant embryos lack structures derived from both termini (Perrimon and Mahowald, 1987); engrailed encodes a homeodomain-containing transcription factor and mutant embryos have posterior terminal defects (Jurgens, 1987); dishevelled encodes a novel maternally synthesized protein that is uniformly distributed in the embryo (Klingensmith et al., 1994), and embryos derived from germline clones lack certain posterior terminal derivatives (Perrimon and Mahowald, 1987); shaggy/zeste-white 3 also encodes a maternally synthesized, uniformly distributed protein-in this case a serinelthreonine kinase homologous to glucose synthetase kinase 3 (Bourouis et al., 199OFand embryos derived from germline clones have defects in the anterior and posterior termini (Perrimon and Smouse, 1989); armadillo
120
MAN LUN R. YIP and HOWARD D. LlPSHlTZ
encodes a protein homologous to vertebrate plakoglobin and mutant embryos exhibit terminal defects (Peifer and Wieschaus, 1990); hedgehog encodes a transmembrane protein that is expressed in the termini which exhibit defects (Mohler, 1988; Lee et al., 1992; Mohler and Vani, 1992); cubitis interruptus encodes a zinc-finger protein that is expressed in the termini, and mutant embryos exhibit terminal defects (Orenic et al., 1987, 1990); gooseberry encodes a paired domain- and homeodomain-containing protein that is expressed in the termini (Baumgartner et a]., 1987; Perrimon and Mahowald, 1987);fusedencodes a serine/threonine kinase that is maternally expressed, and embryos from mutant females or germline clones exhibit posterior terminal defects (Preat et al., 1990). Finally naked mutant embryos (Perrimon and Smouse, 1989), andporcupine embryos derived from germline clones have defects in the head and tail (Perrimon et al., 1989; Siegfried et al., 1994).As discussed above (Section II[C]), the four maternally expressed segment polarity genes may reside in the maternally encoded signal transduction pathway. D. Genes that Program Endodermal Development
The list of genes that fimction in endodermal development is much shorter than the list of those that function in ectodermal development (Table 1). This is likely to be an artefact with two origins. First, most mutant screens have focused on the cuticle (which is an ectodennal derivative) and thus have biased the identification of loci toward the ectoderm. Second, cuticle phenotypes are easier to identify and study than those in the gut, thus biasing phenotypic analyses away from the endoderm. It is likely that many of the genes that function in ectodermal development also function in endodermal development but phenotypes in the endoderm have not yet been defined. The serpent gene functions to specify midgut as distinct from foreguthindgut identity: in serpent mutants endodermal midgut is transformed into ectodermal foregutkindgut (Reuter, 1994). Serpent mutant embryos also fail to undergo germband retraction; this may be an indirect consequence of incorrect specification of midgut cell identity. There is as yet no information on the nature of the serpent gene product. As mentioned above, the FORKHEAD transcription factor functions in programming both midgut (endodermal) and hindgut (ectodermal) development. The anterior and posterior midgut primordia of forkheud mutant embryos invaginate, but fail to undergo migration and disintegrate. This defect has been interpreted as evidence that the midgut
Drosophila Terminal Gene HierJrchy
121
primordia cells fail to adopt appropriate fates in the absence of FORKHEAD function and thus die, indicating that continued expression of FORKHEAD is necessary for the cells of the midgut primordia to differentiate (Weigel et al., 1989a). Apart from its expression and fiinction in Malpighian tubule development (III[C] above) HNF-4 is expressed in the primordia of the midgut. Mutant embryos lacking zygotic HNF-4 activity fail to form the midgut (Zhong et al., 1993). E. Genes that Control Morphogenetic Movement
During normal D I . O . S Oembryogenesis, ~~~~~I the posterior midgut and hindgut primordia are brought internally by the combined processes of midgut and hindgut invagination. and germband extension. Anteriorly, invaginations move the antcl-ior midgut and thc foregut priniordia internally. All of these processes are accomplished by a combination of' local cell rearrangement and cell shape alterations (Campos-Ortega and Hartenstein. 1985; Sweeron et al.. 199 1 : Costa et al., 1993. 1994). The germband subsequently retracts bringing the body parts into their final locations along the antero-posterior axis: this process is largely dependent on cell shape changes with ccll rearrangements playing a minor role (Canipos-Ortcga and Hartenstein, 19x5; Martinez-Arias, 1995). Subscquintly, the anterior and posterior midgut continue growing toward cacti other and finally fuse to tomi one continuous intestinal structure. In addition to exhibiting defects in the specification of positional and tissue identity in the termini, embryos derived from honiozygous r o t . . d f mutant mothers show dcfects in morphogenesis (Schupbach and Wieschaus, 1986b; Strecker et al.. 1989. 199 I , 1992). Embryos derived from homozygous tot:ro'"' mothers lack cclls with midgut and hindgut identities. They do not undergo midgut invagination, and they also have defects in getmband extension: instead of moving dorsally around the posterior pole and then anteriorly toward the head, the tip of the germband remains at the posterior end of the embryo (Schupbach and Wieschaus, 1986b). The germband is eventually thrown into deep folds or forms spirals as it extends. These spiralled embryos are reminiscent of those derived from c'm w w mothers (see Il[C] above). It is difficult to distinguish whether these defects in morphogenetic movement ate a secondary consequence of misspecification of gut identity, or whether the genes that program these movements are regulated independently of the gut identity genes by the TORSO-signaling pathway.
122
M A N LUN R. YIP and H O W A R D D . LlPSHlTZ
Genetic analyses indicate that the zygotic gastrulation mutant,folded gastrulation, can suppress tors&'?' mutant phenotypes (Strecker et al., 1991, 1992). This may be taken as circumstantial evidence for direct regulation of folded gastrulation by the TORSO-mediated pathway (Table 1 and Figure 3). Hemizygousfoldedgastrulation mutant embryos are defective in ventral furrow formation in the central region and in posterior midgut invagination in the posterior terminal domain (Zusman and Wieschaus, 1985; Costa et al., 1994). Only the latter derives from the posterior terminal region and will be discussed here. During normal gastrulation, at the site of posterior midgut invagination, somatic cells immediately dorsal to the pole cell cluster initiate apical constriction. Subsequently, such constrictions commence in cells located further dorsally, then in those positioned laterally and finally in cells on the ventral side of the pole cell cluster. Infblded gastrulation mutant embryos, initiation of apical constriction in the dorsal cells is normal but the subsequent propagation of apical constrictions to other cells is defective, and as a result, posterior midgut invagination does not occur (Costa et al., 1994). Genetic analysis indicates that the fblded gusti-ulation gene product acts locally and that over-expression of FOLDED GASTRULATION protein can induce ectopic cell shape changes (Costa et al., 1994). The FOLDED GASTRULATION protein exhibits no obvious homology to any known protein (Costa et al., 1994), however, since it contains a potential amino-terminal signal sequence, i t may be secreted. Expression of FOLDED GASTRULATION commences in the ventral furrow and the posterior midgut primordium about thirty minutes before the first apical constrictions of cells in these regions (Costa et al., 1994). It has been speculated that FOLDED GASTRULATION functions as a local signal that coordinates cell shape change (Costa et al., 1994). Posterior expression of FOLDED GASTRULATION is dependent on the terminal pathway; it is reduced by huckebein, tailless, and forkhead mutations individually and is completely abolished in huckebein tailless double mutant embryos (Costa et al., 1994).These data suggest that the folded gastrulation gene resides downstream of the tailless and huckebein genes in the gut development hierarchy, but they do not resolve the issue of whether the morphogenetic defects are a primary or a secondary effect. It was shown some time ago that, if embryos from torso mutant females do manage to undergo germband extension, there are later defects in the process of germband retraction which is delayed and/or
Drosophila Terminal Gene Hierarchy
123
incomplete (Strecker et al., 1989, 1991; Strecker and Lipshitz, 1990). Several zygotic mutants exhibit a similar failure of germband retraction, including hindsight, tailup, and u-shaped (Table 1 and Figure 3 ; Niisslein-Volhard et al., 1984; Wieschaus et al., 1984). These act as suppressors of tors&”’ phenotypes, ‘suggesting that they might reside in the terminal gene hierarchy (Strecker et al., 1991). Among the three, hindsight mutations are the strongest suppressors of tors&” phenotypes. Studies of hindsight mutant embryos using time-lapse video microscopy have indicated that germband extension is normal (Yip and Lipshitz, in preparation), thus focusing attention on the specificity of the germband retraction defect. This is further supported by the fact that the midgut and hindgut form normally in hindsight mutant embryos (Yip and Lipshitz, in preparation), excluding a possible secondary effect of misspecification of the identity of these tissues (in contrast to serpent; see III[D] above), and supporting the possibility of primary control of the germband retraction process by the TORSO-mediated signal acting through hindsight. The hind.fight gene encodes a putative transcription factor with 14 widely spaced C,H, Zn-fingers, that is transcribed in the posterior terminal region in the presutnptive posterior midgut primordium (Yip and Lipshitz, in preparation). Activation of hindsight gene transcription in this region fails in embryos from torso mutant mothers, suggesting that it is controlled in response to the terminal signaling pathway. Hindsight transcription occurs normally in tailless mutant embryos but does not occur in huckeheirr mutant embryos (Yip and Lipshitz, in preparation), consistent with hindsight residing downstream of huckehein in the midgut cell fate specification hierarchy. Expression of HINDSIGHT persists in the posterior midgut through the end of germband extension (Yip and Lipshitz, in preparation). Since HINDSIGHT is expressed in the midgut but not in the mesoderm or the epidermal ectoderm which undergo cell shape changes during germband retraction (Yip and Lipshitz. in preparation), it must program a cell-cell signaling pathway that initiates or coordinates the process of germband retraction. The u-shaped gene has been cloned and shown to encode a zinc-finger protein (Simpson and Gelbart, personal communication), while the tailup gene has not yet been analyzed molecularly (Table 1). It is not yet known whether tailup a n d o r u-shaped are transcriptionally regulated by HINDS I GHT.
124
MAN LUN
R. YIP and HOWARD D. LlPSHlTZ
IV. M E C H A N I S M S F O R ESTABLISHING D I S T I N C T CELL FATES WITHIN T H E T E R M I N I While we have outlined many of the downstream effector genes that function in the termini, we have yet to address issues relating to how distinct cell fates are specified within the termini with respect to ( 1 ) more central versus more terminal, (2) dorsal versus ventral, and (3) anterior versus posterior. Such differences are likely to derive from several mechanisms. First, there is evidence that the TORSO-mediated signal or the response to this signal is graded within the termini (Casanova and Struhl, 1989, 1993), and this is likely to be relayed into the initial differential regulation of effector genes. Second, while many ofthe same zygotic effector genes are turned on at both termini (see Figure 3 and Section Ill), their domains of expression are modulated by the dorsoventral hierarchy in both termini as well as by the anterior gene hierarchy at the anterior, thus leading to distinct outcomes in cell fate specification. Such a mechanism relies on spatial control of zygotic effector gene transcription by the maternal axis specification pathways. Third, while several of the same zygotic effector genes are turned on at both termini, their regulatory capacities are modulated by the dorso-ventral hierarchy and also at the anterior by the anterior gene hierarchy, resulting in the control of at least partially distinct subsets of target genes in different regions (see below). Fourth, there are differences in the battery of effector genes turned on at the anterior versus the posterior, and along the dorso-ventral axis (see Sections H I and IV[B]), again brought about by the combined action of the terminal and the other axis-specifying hierarchies. This last mechanism differs from the third in that the integration of terminal and other axial information occurs at the level of the transcriptional control of the zygotic effector genes, rather than further downstream at the level of effector gene action on their target genes. A. More Central versus More Terminal Fates
In addition to differences between the anterior and posterior termini (see Section IV[C]), both termini must acquire differences along their antero-posterior and their dorso-ventral axes. We have considered at length (Section 111) the various effector genes that are differentially expressed along the antero-posterior axis of the termini. It is not clear how the more central cells in the termini are first specified to become distinct from the more terminal ones. Experiments have suggested that
Drosophila Terminal Gene Hierarchy
125
the TORSO receptor tyrosine kinase is activated to different extents, with the highest levels of activation at the poles and the lower levels in the more central regions of the termini (Casanova and Struhl, 1989).There is additional evidence suggesting that this differential activation of the receptor might be a consequence of a non-uniform distribution (or activation) of the ligand for the TORSO receptor which, again, is likely to be highest at the poles (Casanova and Struhl, 1993). Just what this differential activation means at the molecular level is unclear: presumably either more receptors bind ligand and thus transduce more signal per unit area of membrane in the more terminal region than more centrally; or the occupancy rate for the ligand-receptor complex is higher in one region than the other. Either way, since the cytoplasmic signal transduction pathway appears to be conserved, it is plausible that the read-out of differential receptor activity is differential phosphorylation of transcription factors such as those in the JUN- and/or ETS-families (see Section II[ D]). Presumably this results in differential activity of these transcriptional regulators and thus differential activation of downstream effector genes in distinct regions of the termini. B. Dorsal versus Ventral Fates
Next we consider the dorso-ventral axis specification genes and their possible, role in specifying dorso-ventral differences within the termini. All of the so-called dorso-ventral pathway genes were identified based on genetic screens that examined the central, segmented region of the embryo. Thus, it does not follow apriwi that they are also expressed in the termini or, if expressed there, that they are necessarily involved in dorso-ventral axis specification in the termini. That dorso-ventral axis specification in the termini is controlled, at least in part, by the terminal pathway was revealed by the fact that both torso'"'and torso@"mutations result in abnormalities in the dorso-ventral axis: torso"' mutations result in ventralization of terminal region cells while tors@'/ mutations result in dorsalization of central region cells (Strecker et al., 1991, 1992). Several of the dorso-ventral pathway genes (both maternal and zygotic) are expressed in the termini, and transcriptional control of zygotic dorso-ventral genes in the termini is regulated by the terminal pathway (see below and Table 1). Using the torsd'fphenotypes as the starting point, six zygotic and two maternal dorso-ventral mutations were identified as either suppressors or enhancers (zygotic: decapentaplegic, tolloid, short gastrulation, twisted gastrulation, zerkniillt, and pointed;
126
MAN L U N R. YIP and HOWARD D. LlPSHlTZ
maternal: dorsal and c a m s ; [Strecker et al., 19911).Many of these genes have been shown to be required for normal terminal development, although it is not always clear that the mutant phenotypes can most readily be explained in terms ofa dorso-ventral axis defect in the termini rather than a more general requirement. Below we outline what is known about several of these genes, their expression and hnction with particular reference to the termini. It should be remembered that they also function in the central region of the embryo but those functions, as well as their expression there, are not considered in detail here. The decapentaplegic gene encodes a protein with homology to a family of mammalian secreted proteins that includes transforming growth factor+, inhibin, Miillerian inhibiting substance, and bone morphogenetic proteins (BMPs; Derynk et al., 1985; Mason et al., 1985; Cate et al., 1986; Padgett et al., 1987; Wozney et al., 1988). It is believed that these proteins exert their influence on cells that express the corresponding receptors. Null mutations ofdecuprntaplegic (the decapentaplegic""' alleles) are zygotic lethal and result in the dorsal and dorsolateral cells of mutant embryos adopting more ventral cell fates (Irish and Gelbart, 1987). At the anterior end, a hole is present because little or no cuticle is formed and head structures are missing; posteriorly, there are defects in the development of dorsally-derived terminal structures. lnitial expression of DECAPENTAPLEGIC in the termini extends around the two poles to include more ventral cells; by the end of germband extension, this terminal expression refines into patches (St. Johnston and Gelbart, 1987; Ray et al., 1991). The tolloid gene encodes a product homologous to mammalian bone morphogenetic protein-I (BMP- 1; Shimell et al., 1991). BMPs were initially identified in mammalian cells as critical components of protein extracts that can direct cartilage and bone formation (Wozney et al., 1988). Among the seven characterized BMPs, all except BMP-I show sequence similarity to the TGF-13 superfamily (see DECAPENTAPLEGIC above). It has been postulated that BMP-1 acts as metalloprotease and is involved in activating the latent forms ofthe other BMPs (Wozney etal., 1988; Dumermuthetal., 1991; Shimelletal., 1991). Tolloidmutant embryos show defects in the dorsal 40% of the segmented region and are slightly ventralized there (Jurgens et al., 1984; Shimell et al., 1991). In the head and tail, dorsally derived structures are missing (Jurgens et al., 1984; Shimell et al., 1991). Null phenotypes of tolloid can be suppressed by increasing the dosage of the wildtype decapentaplegic gene, suggesting that TOLLOID acts upstream of DECAPENTAPLEGIC (Ferguson
Drosophila Terniinal Gene Hierarchy
127
and Anderson, 1992). This has since been confirmed at the molecular level: the protease domain of TOLLOID is required for the activation of DECAPENTAPLEGIC (Riddihough and Ish-Horowicz, 1991 ; Finelli et al., 1994). Since decapentaplegic and tolloid encode products homologous to different components of the'BMP complex in mammals. the fact that decapentuplegicand rolloid interact genetically to specify the dorsoventral axis of the Drosophilu embryo suggests evolutionary conservation of this cell-cell interaction pathway. Initial expression ofTOLLOlD at the syncytial blastoderm stage i s very similar to that of DECAPENTAPLEGIC, however, in contrast to DECAPENTAPLEGIC, during cellularization TOLLOID expression disappears from the poles (Shimell et al., 1991). The zerknullt gene encodes a homeodomain transcriptional regulatory protein. It is required for the tormation of dorsal tissues, including the amnioserosa in the central region and the optic lobes in the anterior terminal domain (Wakimoto et al., 1984). Initial expression of ZERKNULLT in the termini at the syncytial blastoderm stage is very similar to that of DECAPENTAPLEGIC and TOLLOID. By the cellular blastoderm stage, however, terminal ZERKN ULLT expression becomes restricted to two dorsal patches in the head (Doyle et al., 1986: Rushlow et al., 1987b). In the central region of the embiyo, initial expression of thesc zygotic genes is under the control of the maternal dorso-ventral genes (e.g., Rushlow et al., 19874. Specifically, nuclear localization of DORSAL, a REL/NFKB-related transcription factor, on the ventral side of the embryo represses exprcssion of these zygotic genes thus restricting their initial expression to the dorsal side (Ray et al., 1991). Subsequent refinements of decupentqhgic.. rolloid. and zerkniillt expression are probably a result of regulatory interactions within the zygotic component of the dorso-ventral hierarchy. In contrast to the central region. the expression of the decupentuplegic, rolloid, and zer-knullt genes in the two termini is controlled by the TORSO-mediated terminal pathway, which also overrides repression by DORSAL on the ventral side of the termini (Ray et al., 1991 : Rusch and Levine, 1994). Control by the TORSO-mediated terminal pathway is consistent with the observed genetic interactions between mutations in these three loci and tor:&'"' alleles (Strecker et al., 1991, 1992). For example, mutations in decupentaplegic, tolloid, and cactus (a maternally encoded negative regulator o f DORSAL, which is homologous to I K B :
128
M A N LUN R . YIP and HOWARD D. LlPSHlTZ
Geisler et al., 1992; Kidd, 1992)are enhancers of tors@'/ alleles, while dorsal mutations function as suppressors (Strecker et al., 1991, 1992). Mutations in three additional zygotic dorso-ventral loci interact genetically with torso@"alleles: tu~i.ctedgusrrirl~itiorz and shor(~astru1ation mutations behave as suppressors while pointed mutations act as enhancers (Strecker et al., 199I , 1992). Embryos hemizygous for twisted gustrulation or short gasrnrlation exhibit defects during gastrulation as a result of the dorsal amnioserosa cells undergoing abnormal cell shape changes (Zusman and Wieschaus, 1985; Zusman et al., 1988). In the termini, misted gastrulation mutant embryos show head and tail defects (Lindsley and Zimm, 1992). Mosaic analysis indicates that, while twisted gastrtrlation activity is required on the dorsal side o f the embryo, short gastrulation functions in the termini and ventrally (Zusman and Wieschaus, 1985; Zusman et al., 1988). Genetically, twisted gastr-trlotion probably lies downstream of, or parallel to, deccrpentaplegic while short ga.c.rt.ulutiot~is required to repress decupentaplegic activity ventrally in the central region of thc embryo (Ferguson and Anderson, 1992). The TWISTED GASTRULATION protein shows limited homology to human connective tissue growth factor (Mason ct al., 1994) while SHORT GASTRULATION is homologous to Xeizopt1.7 CHORDIN (Francois et al., 1994; Francois and Bier, 1995). foinled belongs to thc so called .spitz-groupof genes that was initially identified based on their functions i n the ventral ectodcrm ofthe embryo (Mayer and Nusslein-Volhard. 1988). Pointed mutant embryos have defects in the anterior and posterior tennini (Mayer and NussleinVolhard, 1988; Klambt, 1993).The pointed gcne encodes two ovcrlapping transcripts, P I and P2, that share 3'-sequence (Klambt, 1993).This common 3'-sequence encodes an ETS domain, which has been shown to be important for DNA-binding ofother ETS-family transcription factors (Karim et al., 1990). The 5'-region of the longer transcript, P2, encodes an additional domain of homology to a subset of ETS-like proteins. PI and P2 show differential expression patterns and activities during embryogenesis (Klambt, 1993; Scholz et al., 1993). Recent data indicate that the POINTEDp2protein is a target of M A P kinase in the SEVENLESS-mediated signaling pathway in the eye (Brunner et al., 1994a; O'Neill et al., 1994). It is possible that the expression and/or function of one (or both) of the POINTED proteins is regulated similarly by the TORSO-mediated pathway in the embryonic termini.
Drosophila Terrnmal Gene H i e r a r c h
129
C. Anterior versus Posterior Fates
While both the anterior and posterior termini give rise to endodermal and ectodermal tissues, there are also distinct differences in the tissues formed at the two termini. SpecifEally, the brain is derived from the acron within the anterior domain of TORSO function and the development of epidermal and intestinal ectodermal derivatives is quite different in the two termini. Several effector genes function at both termini: these include tailless, huckebein, lines, emphi spiracles, tramtrack,forkhead, spalt, sevpent, NNF-4, hairy, and tvingless (see Section 111). However, the details of their spatial expression patterns differ between the two termini. In addition, several effector genes function in only one of the two termini: for example, hunchback. Abdominal-B, T-related gene, Kriippel, cut, caudal,,fOlde~gustriilation, and hindsight function only in the posterior terminal region and gianr functions only in the anterior terminal region (see Section I l l ) . How zygotic effector genes might be dif'ferentially controlled at the two termini is exemplified by analysis of transcriptional regulation of the tailless gene (Figure 4).In the posterior terminal domain, TAILLESS is expressed as a symmetrical cap at both the syncytial and the cellular blastoderm stages (Pignoni et a]., 1990). In contrast, while TAILLESS is also initially expressed as a symmetrical cap in the anterior terminal region at the syncytial blastoderm stage, TAILLESS expression subsequently retracts from the most anterior and ventral regions and becomes restricted to the acron by the cellular blastoderm stage (Pignoni et al., 1990). The initial, symmetrical expression of TAILLESS at the anterior is programmed largely by the TORSO-mediated pathway, while the subsequent restriction of TAILLESS expression to the acron is accomplished through combined action of the anterior pathway (probably direct regulation of tuilless gene transcription by the BICOID homeodomain protein) and the dorso-ventral pathway (repression of tailless transcription ventrally; Liaw and Lengyel, 1992; Pignoni et al., 1992). In addition to control of zygotic terminal pathway effector gene transcription by the anterior and dorso-ventral hierarchies, recent experiments have demonstrated a reciprocal action of the TORSO-mediated phosphorylation cascade upon the BICOID homeodomain protein (Ronchi et al., 1994).Specifically, it has been shown that certain zygotic target genes that reside in the anterior gene hierarchy are initially transcriptionally activated by BICOID, but later become repressed in the most anterior
130
M A N LUN R. YIP and H O W A R D D. LlPSHlTZ
Figure 4. Mechanisms by which the anterior and the dorso-ventral pathways restrict expression of the tailless terminal pathway effector gene to the acronal region of the anterior terminal domain. Initially (left diagram), tai//ess expression (shaded) i s activated throughout the anterior terminus through the combined action of BlCOlD at the anterior tip and the TORSO-mediated pathway (Pignoni et al., 1992). Subsequently (right diagram), the TORSO-mediated pathway results in the phosphorylation of BICOID, which now represses the tailless gene at the anterior tip; while DORSAL represses the tailless gene ventrally. Together, these result in restriction of tailless expression to the acron (shaded). The postulated repressive action of phosphorylated BlCOlD on tailless transcription is highly speculative and is an attempt to resolve published results (Pignoni et al., 1992; Ronchi et al., 1994) in a simple manner. To date, no in vitro evidence for such repression has been reported. Positive regulation i s indicated by arrows; negative regulation by the Ts. In the right diagram, although in principle the TORSO-mediated pathway could positively regulate taillessexpression at the anterior tip and in the ventral region, only the final outcome-negative control of tailless expression by BlCOlD and DORSAL-is shown.
Drosophila Terminal Gene Hierarchy
131
terminal region of the embryo (Figure 4). This repression is dependent on the function of the TORSO receptor tyrosine kinase and its downstream genes in the cytoplasmic signal transduction pathway such as D-RAF, which were shown to result in phosphorylation of the BICOID protein (Ronchi et al., 1994).Repression did not, however, require either TAILLESS or HUCKEBEIN. Thus, cross-regulation among the axis-specifying hierarchies can occur both among the matemally-encoded proteins and at the level of transcriptional control of zygotic effector genes. Further, these analyses indicate that the cross-regulation is bidirectional; the above examples demonstrate that the anterior pathway can regulate the terminal pathway and vice versa. Clearly this lends power and flexibility to the differential cell fate specification machinery.
V. CONCLUSIONS Analysis of the terminal pathway in Drosophilu has begun to provide us with insights into how localized activation of a generally expressed receptor can be used to provide spatial cues during development. Many components of the cytoplasmic signal transduction cascade are used at other times and in other places during Drosophilu development. There is increasing evidence that these signal transduction cascades have been conserved among metazoa. Numerous zygotically expressed terminal pathway effector genes have been identified. These control the specification of terminal positional identity, the development ofterminal tissues and terminal morphogenetic cell shape changes and movements. Several mechanisms appear to be used to modulate the general cytoplasmic signal into varied patterns of effector gene expression. These include differential activation of the transmembrane receptor as well as combinatorial action of the terminal, anterior, and dorso-ventral genetic pathways in distinct regions within the termini. This likely results in activation of overlapping batteries of effector genes in spatially distinct patterns within the termini, resulting in subdivision of the terminal domains into groups of cells with distinct fates. Understanding of the link between the maternally encoded cytoplasmic signal transduction cascade and the control of zygotic effector genes remains poor. In addition, the details of how the effector genes are differentially regulated and how they in turn specify the details of cell and tissue fates, remain key areas for future analysis.
M A N LUN R. YIP m d H O W A R D 0 . LIPSHI-lZ
I32
NOTE ADDED IN PROOF Since this review was written in 1994 a n additional pair-rule gene with expression and function in the termini has been reported (Hou ct al., 1996; Yan et al., 1996).
ACKNOWLEDGMENTS We thank numerous investigators in the L>/'O.\.ophi/0terminal development arid morphogenesis research fields for providing u s with reprints arid preprints; and the following for critical comments on the inanuscript: A . Bashirullah, P. Becker, S. Celniker, T. Clandinin, M. Lmnka, P. Sternberg. and S. Ward. Man Lun K. Yip was supported by a predoctoral fellowship froin the Howard Hughes Medical Institute. Our research on the terminal gene hierarchy and the genetic control of niorphogenesis at C'altech \vas supported by thc Anierican Cancer Society (LIB- 14), the Gustavus and Louise Pfeiffer Rcsearch Foundation, and a gift for the support of genetics research li-urn Millard and Muriel Jacobs. Our research on thc genetic control of niorphogeiiesis at thc liospital for Sick C'hildrcn is supported by the National C'aiicer Institutc of C'anacia.
REFERENCES Ainbroaio. L.. Mahouald. A P.. & Pcrrinion. N ( I 9 X C ) a ) . / ( / ) / i o / c h o / ~is rcquircd maternally for pattern ti)riiiation i n [lie tcriniiial region5 ofthe cinbrvo. Decelopmcnt 106. 1 4 5 1 5 8 .
Ambrosio. L.. M;ihowald, A.P.. & Perrimon. N . ( 1089h).Kequircment ofthe D/~i).\o/dii/0 ruf'homolog for l w s o function. Nature .i1-7.288 2 0 I . Arvidsson, A.K.. Rupp, E.. Nhnberg, k... Dowiwnrd, J.. Ronnstrand. L . , Wennstrom. S.. Schlessinger, J.. Heldin. C'.t I.. & C'laesson-Wclsh. L. ( 1994) 'lyr-7 I6 in the platcletderived growth factor beta-receptor kinasc insert i s involvcd i n GRHZ binding and Ras activation. Mol. Cell. Hiol. /1.67 l S 4 ~ 7 2 0 , Baker, N.E. ( 19x7). Molecular cloning oi sequcnccs from ~i,rng/es\.a segment polarity The spatial distribution o f a transcript i n embryos. EMRO J . 6. gene in D~.o.ci)phi/o: I705 ~I 774. Baker. N , E ( 198Xa). Einbryonic and iinaginal requirements t o i - ii,r~,y/es.\.a segi11cnIpolarity gene in Dru.sophilu. I h . Biol. 1-75, 9 G I 9 X . Baker, N.E. (1988b). Localization of transcripts from the \titig/e,\.s gene in whole Dtwophilo embryos. Development 103. 289 -299 Baumgartner. S., Bopp, D.. Burr!. M.. & Noll, M . (1987). Structure oftwo genes at thc goosebeny locus related to the paired gene and their spatial expression during Drusophila embryogenesis. Gcncs and Dev. 1. 1247-1 267. Bauingartner. S., Martin. D.. Hagios, C.. & C'hiyuct-Ehrisinann, R. ( 1994). leii-m, a Dru.wphi/a gene related to tenascin. is a new pair-rule gene. EMBO J . 13. 37283740.
Drosophila Terminal Gene H ierxch y
13.3
Renedyk, M.J., Mullen. J.R., & DiNardo, S. (1994). odd-paired: a zinc finger pair-rule protein required for the timely activation o f cv7gruiled and ~ir7glc.r.rin Dro.rophi1a embryos. Genes and Dev. X , 105-1 17. Bennett, A.M., Tang, T.L., Sugimoto. S.. Walsh. C.T., & N e d . B.G. (1994). Protein-tyrosine-phosphatase SHPTP2 couples platelet-derived growth factor receptor beta to Ras. Proc. Natl. Acad. Sci. lJSA 91. 7335 7339. Diggs, W.11.. Ill. & Zipursky. S.L. (1092). Primary structure, expression, and signaldependent tyrosine phosphorylation ol' a Dr.o.vophi1tr homolog o f extracellular signal-rcgulatcd kinase. Proc. Natl. Acad. Sci. USA 89, 6295-0209. Higgs, W.11.. 111, Zavitt, K.11.. Dickson. D.. win der Straten. A,, Brunner, D., Hafen, E., & Zipursky. S.L. (1994). The D,v.cophi/o 1.011ed locus encodes a M A P kinnse required i n the .seiwr/:cs signal transduction pathway. EMBO J. 13, 1628-1 635. Binari, R., & Perrimon, N. (1994). Stripc-specific regulation o f pair-rule genes by hop.rco/c/r,a putative lak family tyrosine kinasc in Dr~o.rop/ri/u. Gencs and Dev. 8.
30&3 12.
.
Blochlinger, K.. Hodmer. R Jack J.. Jan. L..Y.. & Jan, Y.N. (1988). Primary structure and expression o f a product from c'iri. n locus involved in specifying sensory organ I ~ / u . 3.1.1. 62%635. identity in D / . o . w J ~ /Nature Bonfini. H.R.. Karlovich. ('.A , I h g u p t a . (.., RC Ranerjec, U. ( 1992). Thc Sori o/ .sc~~~eti/e.cs gene product: A putative activator of Ras. Science 255, 003-400. Bopp, D., Burri, M.. Raumgartncr, S., Frigcrio, (i.,6i Noll, M. ( 1986). Conservation o f h large protein domain in the segmentation gcnc ptrir.et/ and i n functionally related genes o f Dro.~op/ri/u. Cell 47. 1033-1040 Bourne, I1.R.. Sanders, D.A., & McC'ormick. F. (1990). The GTPase superfamily: A conserved switch for diverse cell functions. Nature 348. I2S.132. Bourne, H.R.. Sanders, I>.A., & McCormick. F. (1991). The GTPase supert8mily: Conserved structure and molecular mcchanism. Nature 349, I 17-1 27. Bourouis, M., Moore, P., Ruel, L.. (irau, Y., tleitzler, P.. & Simpson, P. ( 1990).An early embryonic product of the gene sh~rggyencodes a serindthreonine protein kinase related to the CDC2Xkdc2+ subfamily. EMBO J Y, 2877- -2884. Bronner, G., & Jackle, H. ( 1991 ). Control and function o f terminal gap gcne activity in the posterior pole region ofthc l ~ ~ 0 . ~ 0 p hembryo. ik7 Mcch. Develop. 35,2 0 5 2 1 I . Bronner. G., Chu-LaGraff, 0..Doc, C.Q.. Cohcn. H., Wcigel. D.,Taubert. H., & Ja'ckle, H. ( 1994). Splicgr-likc zinc-finger prokin required for endoderm specification and Nature 36Y. 664- 668. germ-layer formation in Dru~.wphi/tr. Brown, J.L., Sonoda, S., Ucda. 11.. Scott, M.P.. & Wu, c'. (1991). Rcprcssion o f the Drnsophrla / i ~ d r /i u r a x (f/:) xgnientation gene. b.MBO J. 10. 665 674. Brown, J.L., & wu. C. (1993). Rcprcssion of Dro.vophi/u pair-rule segmentation genes by ectopic expression of /r.crnrtr.uck.Development / I 7.4S58. Brunner. D., Ducker, K., Oellers, N., Ilafen, 17.. ScholL, Ii., & Klambt. C. (1994a). The ETS domain protein Pointed-P2 is a target o f M A P kinase i n the sevenlcxr signal transduction pathway. Nature 3 70, 3 x 6 3 8 9 . Brunner, D., Oellers, N., Szabad, J., Biggs I l l , W.11.. Zipursky, S.L.. & Hafen, E. ( 1994b). A gain-of-function mutation i n Dro.rophilu M A P kinase activates multiple receptor tyrosine kinase signaling pathways. Ccll 76. 875-888.
134
MAN LUN R. YIP and HOWARD 0. LlPSHlTi!
Buckles, G.R., Smith, Z.D.J., & Katz, F.N. (1992). mip causes hyperinnervation of a retinotopic map in Drosophila by excessive recruitment of R7 photoreceptor cells. Neuron 8, 1015-1029. Cadigan, K.M., Grossniklaus, U., & Gehring, W.J. (1994). Localized expression of sloppypuired protein maintains the polarity ofDrosophila parasegments. Genes and Dev. 8 , 8 9 9 1 3 . Campos-Ortega, J.A., & Hartenstein, V. ( 1985). The Embryonic Development qfDrosophila melanogaster: Springer Verlag, Berlin. Capovilla, M., Eldon, E.D., & Pirrotta, V. (1992). The giant gene of Drosophila encodes a b-ZIP DNA-binding protein that regulates the expression of other segmentation gap genes. Development 114,99-112 Casanova, J. (1990). Pattern formation under the control of the terminal system in the Drosophila embryo. Development 110, 62 1428. Casanova, J. (1991). Interactions between torso and dorsal, two elements of different transduction pathways in the Drosophila embryo. Mech. Dev. 36,4145. Casanova, J., & Struhl, G. ( I 989). Localized surface activity of torso, a receptor tyrosine kinase, specifies terminal body pattern in Drosophilu. Genes and Dev. 3,202S2038. Casanova, J., & Struhl, G. (1993). The torso receptor localizes as well as transduces the spatial signal specifying terminal body pattern in Drosophila. Nature 362, 152-1 55. Casanova, J., Furriols, M., McCormick, C.A., & Struhl, G. (1995). Similarities between trunk and spatzle, putative extracellular ligands specifying body pattern in Drosophila. Genes and Dev. 9,253P-2544. Cate, R.L., Mattalliano, R.J., Hession, C., Tizard, R., Farber, N.M., Cheung, A., Ninfa, E.G., Frey, A.Z., Gash, D.J., Chow, E.P., Fisher, R.A., Bertonis, J.M., Torres, G., Wallner, B.P., Ramachandran, K.L., Ragin, R.C., Manganaro, T.F., MacLauglin, D.T., & Donahoe, P.T. ( 1986). Isolation of the bovine and human genes for Mullerian inhibiting substance and expression of the human gene in animal cells. Cell 45, 685-698.
Celniker, S.E., Sharma, S., Keelan, D.J.. & Lewis, E.B. (1990). The molecular genetics of the bithorax complex o f Drosophila: cis-regulation in the Abdominal-B domain. EMBO J. 9,42774286. Childs, S.R., & O’Connor, M.B. (1994). Two domains of the tolloid protein contribute to its unusual genetic interaction with decupentaplegic. Dev. Biol. 162,209-220. Cleghon, V., & Morrison, D.K. (1994). Raf-l interacts with fyn and src in a non-phosphotyrosine-dependent manner, J. Biol. Chem. 269, 1 7749-1 7755. Costa, M., Sweeton, D., & Wieschaus, E. (1993). Gastrulation in Drosophilu: Cellular mechanisms of morphogenetic movements. In: The Development qf’ Drosophila rnelanoguster (Bate, M., & Martinez-Arias, A,, Eds.). Cold Spring Harbor Laboratory Press, New York, pp. 425465. Costa, M., Wilson, E.T., & Wieschaus, E. ( I 994). A putative cell signal encoded by the folded gastrulation gene coordinates cell shape changes during Drosophila gastrulation. Cell 76, 1075-1089. Coulter, D.E., & Wieschaus, E. (1988). Gene activities and segmental patterning in Drosophilu: Analysis of odd-skipped and pair-rule double mutants. Genes and Dev. 2, 1812-1823.
Coulter, D.E., Swaykus, E.A., Beran-Koehn, M.A., Goldberg, D., Wieschaus, E., & Schedl, P. (1990). Molecular analysis of odd-skipped, a zinc finger encoding
Drosophila Terminal Gene Hierarchy
135
segmentation gene with a novel pair-rule expression pattern. EMBO J. 9, 37953804. Dalton, D., Chadwick, R., & McGinnis, W. ( I 989). Expression and embryonic function of empty spiracles: A Drosophila homeo box gene with two patterning functions on the anterior-posterior axis of the embryo. Genes and Dev. 3, 1940-1956. Degelmann, A., Hardy, P.A., & Mahowald, A.P. ( 1990). Genetic analysis of two femalesterile loci affecting eggshell integrity and embryonic pattern formation in Drosophilu rnelunogaster. Genetics 126,427434. Degelmann, A., Hardy, P.A., Perrimon, N., & Mahowald, A.P. (1986). Developmental analysis of the torso-like phenotype in Drosophila produced by a maternal-effect locus. Dev. Biol. f15,47%489. Derynk, R., Jarrett, J.A., Chen, E.Y., Eaton, D.H., Bell, J.R., Assoian, R.K., Roberts, A.B., Sporn, M.B., & Goeddel, D.V. (1985). Human transforming growth factor-f3 complementary DNA sequence and expression in normal and transformed cells. Nature 3/6,701-705. Diederich, R.J., Pattatucci, A.M., & Kaufman, T.C. (1991). Development and evolutionary implication of labial, Deformed, and eiigruiled expression in the Di-osophila head. Development 113,273-281. DiNardo, S., Kuner, J.M., Theis. J., & O’Farrell, P.H. ( 1 985). Development-of embryonic pattern in D.melanoguster as revealed by accumulation of the nuclear engrailed protein. Cell 43, 5-9. Ding, D., & Lipshitz, H.D. (1993). Localized RNAs and their functions. BioEssays 10, 651-658. Doyle, H.J., & Bishop, J.M. ( I 993). torso, a receptor tyrosine kinase required for embryonic pattern formation, shares substrates with the sevenless and EGF-R pathways in Drosophila. Genes and Dev. 7, 633446. Doyle, HJ.; Harding, K., Hoey, T., & Levine, M. (1986). Transcripts encoded by a homoeo box gene are restricted to dorsal tissues of Drosophila embryos. Nature 323, 76-79. Dumermuth, E., Sterchi, E.E., Jiang, W.P., Wolz, R.L., Bond, J.S., Flannery, A.V., & Beynon, R.J. (1991). The astacin family of metalloendopeptidases. J. Biol. Chem. 266,2 138 1-2 1385. Eldon, E.D., & Pirrotta, V. (1991). Interactions of the Dro.vophifa gap gene giant with maternal and zygotic pattern-forming genes. Development 111, 367-378. Feng, G.S., Hui, C.C., & Pawson, T. (1993). SH2-containing phosphotyrosine phosphatase as a target of protein-tyrosine kinases. Science 259, 1607-1 6 I 1. Ferguson, E.L., & Anderson, K.V. (1992). Localized enhancement and repression of the activity o f the TGF-fi family member, deecapentaplegic,is necessary for dorsal-ventral pattern formation in the Drosophila embryo. Development If 3, 583-597. Finelli, A.L., Bossie, C.A., Xie, T., & Padgett, R.W. (1994). Mutational analysis of the Drosophila tolloid gene, a human BMP-I homolog. Development 120, 861470. Francois, V., & Bier, E. (1995). Xei?opuschordin and Drosophila short gastrulation genes encode homologous proteins functioning in dorsal-ventral axis formation. Cell 80, 19-20. Francois, V., Solloway, M., O’Neill, J.W., Emery, J., & Bier, E. (1994). Dorsal-ventral patterning of the Drosophila embryo depends on a putative negative growth factor encoded by the shortgastriilation gene. Genes and Dev. 8, 2602-261 6.
136
M A N LUN R. YIP and HOWARD D. LlPSHlTZ
Frasch, M., & Levine, M. (1987). Complementary patterns of even-skipped and,fitshi larazu expression involve their differential regulation by a common set of segmentation genes in Drosophila. Genes and Dev. I , 981-995. Frasch, M., Hoey, T., Rushlow, C., Doyle, H., & Levine, M. ( 1 987). Characterisation and localisation of the even-skipped protein of Drosophila. EMBO J. 6, 749-759. Frigerio, G., Burri, M., Bopp, D., Baumgartner, S., & Noll, M. ( 1986). Structure of the segmentation gene paired and the Drosophila PRD gene set as a part of a gene network. Cell 47,735-746. Gaul, U., & Weigel, D. (1991). Regulation of Kruppel expression in the anlage of the Malpighian tubules in the Drosophila embryo. Mech. Dev. 33, 57-68. Gaul, U., Seifert, E., Schuh, R., & Jackle, H. (1987). Analysis of Kruppel protein distribution during early Drosophila development reveals post-transcriptional regulation. Cell 50, 63p647. Gaul, U., Mardon, G., & Rubin, G.M. (1992). A putative Ras GTPase acts as a negative regulator of signaling by the sevenless receptor tyrosine kinase. Cell 68, 1007-10 19. Geisler, R., Bergmann, A., Hiromi, Y., & Nusslein-Volhard, C. (1992). cactus, a gene involved in dorsoventral pattern formation of Drosophila, is related to the I-kappa-B gene family of vertebrates. Cell 7/, 613421. Gergen, J.P., & Butler, B.A. (1988). Isolation of the Drosophila segmentation gene runt and analysis of its expression during embryogenesis. Genes and Dev. 2,1179-1 193. Gergen, J.P., & Wieschaus, E. (1985). The localized requirements for a gene affecting segmentation in Drosophila: Analysis of larvae mosaic for runt. Dev. Biol. 109, 32 1-335.
Gergen, J.P., & Wieschaus, E.F. f 1986). Localized requirements for gene activity in segmentation of Drosophila embryo: Analysis of armadillo, ,fused, giant and LO?paired in mosaic embryos. Roux’s Arch. Dev. Biol. 195,4%62. Goto, T., Macdonald, P., & Maniatis, T. (1989). Early and late periodic patterns of even-skipped expression are controlled by distinct regulatory elements that respond to different spatial cues. Cell 57,413422. Grossniklaus, U., Pearson, R.K., & Gehring, W.J. (1992). The Drosophila sloppypaired locus encodes two proteins involved in segmentation that show homology to mammalian transcription factors. Genes and Dev. 6, 1030-105 1. Hafen, E., Dickson, B., Brunner. D., & Raabe, T. (1994). Genetic dissection of signal transduction mediated by the sevenless receptor tyrosine kinase in Drosophila. Prog. Neurobio. 42,287-292. Harbecke, R., & Janning, W. (1989). The segmentation gene Kriippel of Drosophiln melanogasrer has homeotic properties. Genes and Dev. 3, 114-122. Harrison, S.D., & Travers, A.A. ( 1 990). The tramtrack gene encodes a Drosophila finger protein that interacts with the,fiz transcriptional regulatory region and shows a novel embryonic expression pattern. EMBO J. 9,207-2 16. Herrmann, B.G., & Kispert, A. ( 1 994). The T-genes in embryogenesis. Trends Genet. 10,280-286. .
Hooper, J.E., & Scott, M.P. (1989). The Drosophila patched gene encodes a putative membrane protein required for segmental patterning. Cell 59,75 1-765. Hooper, K.L., Parkhurst, S.M., & Ish-Horowicz, D. (1989). Spatial control of hair?, protein expression during embryogenesis. Development 107,489-504.
Drosophila Terminal Gene Hierarchy
137
Hou, X.S., Melnick, M.B., & Perrinon, N. (1996). Muizlle acts downstream of the Drosophilu HOPIJAK kinase and encodes a protein similar to the mammalian STATS. Cell 84,411419. Hunter, T., & Karin, M. (1992). The regulation oftranscription by phosphorylation. Cell I 70,375-387. Ingham, P.W., Howard, K.R., & Ish-Horowicz, D. (1985). Transcription pattern of the Drosophilu segmentation gene hairy. Nature 3 18,43!U45. Irish, V.F., & Gelbart, W.M. (1987). The decupentoplegic gene is required for dorsalventral patterning of the Drosop/zilu embryo. Genes and Dev I, 868479. Ish-Horowicz, D., Howard, K.R.. Pinchin, S.M., & lngham, P.W. (1985). Molecular and genetic analysis of the hairy locus in Drosophilu. Cold Spring Harb. Symp. Quant. Biol. 50, 135-144. Jurgens, G. ( I 987). Segmental organization ofthe tail region in the embryo ofD~-osop/iilu melunoguster. Roux’s Arch. Dev. Biol. 196, 14 1-1 57. Jurgens, G. ( 1988). Head and tail development of the Drosophilu embryo involves spult, a novel homeotic gene. EMBO J. I , 189-1 96. Jurgens, G., & Hartenstein, V. (1993). The terminal regions of the body pattern. In: The Development ojDrosophilu melunoguster (Bate, M., & Martinez Arias, A., Eds.). Cold Spring Harbor Laboratory Press, New York, pp. 687-746. Jiirgens, G., & Weigel, D. (1988). Terminal versus segmental development in the D ~ ~ s o p h iembryo: la The role of the homeotic gene,fbr-khead. Roux’s Archiv. Dev. Biol. 197,345-354. Jurgens, G., Wieschaus, E., Nusslein-Volhard, C., & Kluding, H. (1984). Mutations affecting the pattern of the larval cuticle in Drosophilu melanoguster*.11. Zygotic loci on the third chromosome. Roux’s Arch. Dev. Biol. 193,283-295. Jurgens, G.; Lehmann, R., Schardin, M., & Nusslein-Volhard, C. (1986). Segmental organisation of the head in the embryo of Dr-osophila melunogaster. Roux’s Arch. Dev. Biol I95,359-377. Kania, M.A., Bonner, A S . , Duffy, J.B., & Gergen, J.P. ( 1990). The Drosophilu segmentation gene r-unt encodes a novel nuclear regulatory protein that is also expressed in the developing nervous system. Genes and Dev. 4, 1701-1 7 13. Karim, F.D., Urness, L.D., Thummel, C.S., Klemsz, M.J., McKercher, S.R., Celada, A,, van Beveren, C., Maki, R.A., Gunther, C.V., Nye, J.A., & Graves, B.J. (1990). The ETS-domain: A new DNA binding motif that recognizes a purine rich core DNA sequence. Genes and Dev. 4, 145 1-1453. Kidd, S. (1992). Characterization of the Drosophilu cactus locus and analysis of interactions between cactus and dona1 proteins. Cell 71, 623-635. Kilchherr, F., Baumgartner, S., Bopp, D., Frei, E., & Noll, M. (1986). Isolation of the paired gene of Drosophila and its spatial expression during early embryogenesis. Nature 321,49-99. Kispert, A., & Hemnann. B.G. (1993). The Brachyury gene encodes a novel DNA-binding protein. EMBO J. 12,321 1-3220. Kispert, A., Herrmann, B.G., Leptin, M., & Reuter, R. (1994). Homologs of the mouse Brachytrry gene are involved in the specification of posterior structures in Drosophilu, Tiiholium, and Locrrstu. Genes and Dev. 8,2137-2150.
138
M A N LUN R. YIP and HOWARD D. LlPSHlTZ
Klambt, C. (1993). The Drosophila gene pointed encodes two ETS-like proteins which are involved in the development of the midline glial cells. Development 117, 1 6 3 176. Klingensmith, J., Nusse, R., & Perrimon, N. (1994). The Drosophila segment polarity gene disheveltedencodes a novel protein required for response to the wingless signal. Genes and Dev. 8, 11 8-130. Klingler, M., Erdelyi, M., Szabad, J., & Nusslein-Volhard, C. (1988). Function of torso in determining the terminal anlagen ofthe Drosophila embryo. Nature 335,275-277. Kornberg, T., Siden, I., O’Farrell, P., & Simon, F. (1985). The engruiled locus of Drosophila: In situ localization of transcripts reveals compartment-specific expression. Cell 40,45-53. Kuhnlein, R.P., Frommer, G., Friedrich, M., Gonzalez-Gaitan, M., Weber, A., WagnerBemholz, J.F., Gehring, W.J., Jackle, H., & Schuh, R. (1994). spa/? encodes an evolutionarily conserved zinc finger protein of novel structure which provides homeotic gene function in the head and tail region of the Drosophila embryo. EMBO J. 13, 168-179. Kurjan, J. (1993). The pheromone response pathway in Saccharomyces cerevisiae. Ann. Rev. Genet. 27, 147-179. Kussick, S.J., Basler, K., & Cooper, J.A. (1993). Rasl-dependent signaling by ectopically-expressed Drosophila src gene product in the embryo and developing eye. Oncogene 8,2791-2803. Laughon, A., & Scott, M.P. (1984). Sequence of a Drosophila segmentation gene: Protein structure homology with DNA-binding proteins. Nature 3/0,25-3 I . Lee, J.J., von Kessler, D.P., Parks, S., & Beachy, P.A. (1992). Secretion and localized transcription suggest a role in positional signaling for products of the segmentation gene hedgehog. Cell 71,33-50. Lehmann. R., & Nusslein-Volhard. C. (1987). hunchhack, a gene required for segmentation of an anterior and posterior region of the Drosophila embryo. Dev. Biol. 119, 402-4 17. Lev, Z., Kimchie, Z., Hessel, R., & Segev, 0. (1985). Expression of ras cellular oncogenes during development of Drosophila melanogaster. Mol. Cell. Biol. 5. 1540-1 542. Levine, A., Bashan-Ahrend, A., Baudi-Hadrian, O., Gartenberg, D., Menasherow, S., & Wides, R. (1994). odd Oz, a novel Drosophila pair rule gene. Cell 77, 587-598. Li, W., Nishimura, R., Kashishian, A., Batzer, A.G., Kim, W.J.H., Cooper, J.A., & Schlessinger, J. ( 1994). A new function for a phosphotyrosine phosphatase: Linking GRB2-Sos to a receptor tyrosine-kinase. Mol. Cell. Biol. 14, 509-5 17. Liaw, G.J., & Lengyel, J.A. (1992). Control of faillessexpression by hicoid, dorsal and synergisitically interacting terminal regulatory elements. Mech. Dev. 4 0 , 4 7 4 1 . Lindsley, D.L., & Zimm, G.G. (1992). The Genome oJ’ Drosophila melanogaster. Academic Press, San Diego. Lipshitz, H.D. (1991). Axis specification in the Drosophila embryo. Cum. Op. Cell Biol. 3,966975. Liu, S., & Jack, J. (1992). Regulatory interactions and role in cell type specification of the Malpighian tubules by the cut, Kruppel, and caudal genes of Drosophila. Dev. Biol. 150, 133143.
Drosophila Terminal Gene Hierarchy
139
Liu, S., McLeod, E., & Jack, J. (1991). Four distinct regulatory regions of the cut locus and their effect on cell type specification in Drosophilu. Genetics 127, 151-159. Lowy, D.R., & Willumsen, B.M. (1993). Function and regulation of Ras. Ann. Rev. Biochem. 62,851491. Lu, X., Chou, T.-B., Williams, N.G., Roberts, T., & Perrimon, N. (1993). Control ofcell fate determination by p21‘ “SIRasl,an essential component of torso signaling in Drosophilu. Genes and Dev. 7,621432. Macdonald, P. M., & Struhl, G. ( 1986). Amoleculargradient in early Drosophilu embryos and its role in specifying body pattern. Nature 324,537-545. Macdonald, P.M., Ingham, P., & Struhl, G. (1986). Isolation, structure, and expression ofeven-skipped A second pair-rule gene of Drosophilu containing a homeobox. Cell 47,721-734. Mahoney, P.A., & Lengyel, J.A. (1987). The zygotic segmentation mutant tailless alters the blastoderm fate map of the Drosophila embryo. Dev. Biol 122,464470. Martin, J.R., Ralbaud, A., & 0110, R. ( 1994). Terminal pattern elements in Dr*osophilu embryo induced by the torso-like protein. Nature 367, 741-745. Martinez Arias, A. (1993). Development and patterning of the larval epidermis of Drosophilu. In: The Development of Drosophilu melunogaster (Bate, M., & Martinez Arias, A,, Eds.). Cold Spring Harbor Laboratory Press, New York, pp. 5 17-608. Martinez Arias, A,, & Lawrence, P.A. (1985). Paraseginents and compartments in the Drosophilu embryo. Nature 313,639442. Mardnez Arias, A., Baker, N.E., & Ingham, P.W. (1988). Role of segment polarity genes in the definition and maintenance of cell states in the Drosophilu embryo. Development 103, 157-170. Mason, A.J., Hayflick, J.S., Ling,N., Esch, F.,Ueno,N., Ying, S.-Y., Guillemin, R.,Naill, H., &. Seeburg, R.H. (1985). Complementary DNA sequences of ovarian follicular fluid inhibin show precursor structure and homolog with transfonning growth factor-p. Nature 319,659463. Mason, E.D., Konrad, K.D., Webb, C.D., & Marsh, J.L. (1994). Dorsal midline fate in Drosophilu embryos requires twiyted gastrulation, a gene encoding a secreted protein related to human connective tissue growth factor. Genes and Dev. 8, 14894501. Mayer, U., & Nusslein-Volhard, C. (1988). A group of genes required for pattern formation in the ventral ectodenn of Drosophilu embryo. Genes and Dev. 2, 1496-1 51 I . Melnick, M.B., Perkins, L.A., Lee, M., Ambrosio. L., & Perrimon, N. (1993). Developmental and molecular characterization of mutations in the Diosophilu-ruf’ serinelthreonine protein kinase. Development 118, 127-138. Mlodzik, M., & Gehring, W.J. (1987a). Expression of the cuirdal gene in the germ line ofDrosophilu: Fonnation ofan RNAand protein gradient during early embryogenesis. Cell 48,46S478. Mlodzik, M., & Gehring, W.J. (1987b). Hierarchy of the genetic interactions that specify the anteroposterior segmentation pattern of the Drosophilu embryo as monitored by caudal protein expression. Development 101,421-435. Mohler, J. (1988). Requirements for hegdehog, a segment polarity gene, in patterning larval and adult cuticle of Drosophilu. Genetics 120, 1061-1072.
140
MAN LUN R. YIP and HOWARD D. LlPSHlTZ
Mohler, J. ( I 993). Genetic regulation of CNC expression in the pharnygeal primordia of Drosophilu blastodenn embryos. Roux’s Arch. Dev. Biol. 202,214-223. Mohler, J., & Vani, K. (1992). Molecular organization and embryonic expression of the hedgehog gene involved in cell-cell communication in segmental patterning of Drosophilu. Development 115,957-97 1. Mohler, J., Eldon, E.D., & Pirrotta, V. (1989). A novel spatial transcription pattern associated with the segmentation gene, giant, of Drosophilu. EMBO J. 8, 15391548. Mohler, J., Vani, K., Leung, S., & Epstein, A. (1991). Segmentally restricted cephalic expression of a leucine zipper gene during Drosophilu embryogenesis. Mech. Dev. 34, 3-10. Mohler, J., Mahaffey, J.W., Deutsch, E., & Vani, K. (1995). Control of Drosophilu head segment identity by the bZIP homeotic gene cnc. Development 121,237-247. Nakano, Y., Guerrero, I., Hidalgo, A,, Taylor, A., Whittle, J.R.S., & Ingham, P. W. (1989). Aprotein with several possible membrane-spanning domains encoded by the Drosophila segment polarity gene patched. Nature 341,508-5 13. Nishida, Y., Hata, M., Ayaki, T., Ryo, H., Yamagata, M., Shimizu, K., & Nishizuka, Y. (1988). Proliferation of both somatic and germ cells is affected in the Drosophilu mutants of rufproto-oncogene. EMBO J. 7,775-781. Noordermeer, J., Klingensmith, J., Perrimon, N., & Nusse, R. (1994). dishevelled and armadillo act in the wingless signaling pathway in Drosophilu. Nature 367,80-83. Nusslein-Volhard, C., Lohs-Schardin, M., Sander, K., & Cremer, C. ( 1980). Adorso-ventral shift of embryonic primordia in a new maternal effect mutant of Di-osophilu. Nature 283,474476. Nusslein-Volhard, C., Wieschaus, E., & Kluding, H. (1984). Mutations affecting the pattern ofthe larval cuticle in Diasophilu melunoguster. I. Zygotic loci on the second chromosome. Roux’s Arch. Dev. Biol. 193,267-282. Nusslein-Volhard, C., Kluding, H., & Jurgens, G. (1985). Genes affecting the segmental subdivision of the Drosophilu embryo. Cold Spring Harb. Symp. Quant. Biol. 50, 14S154. Niisslein-Volhard, C., Frohnhofer, H.G., & Lehmann, R. ( 1987). Determination of anteroposterior polarity in Drosophilu. Science 238, 1675168 1. O’Neill, E.M., Rebay, I., Tjian, R., & Rubin, G.M. (1994). The activities of two Ets-related transcription factors required for Drosophilu eye development are modulated by the Ras/MAPK pathway. Cell 78, 137-147. Olivier, J.P., Raabe, T., Henkemeyer, M., Dickson, B., Mbamalu, G., Margolis, B., Schlessinger, J., Hafen, E., & Pawson, T. (1993). A Drosophila SH2-SH3 adaptor protein implicated in coupling the sevenless tyrosine kinase to an activator of rus guanine nucleotide exchange, Sos. Cell 73, 179-1 91. Orenic, T., Chidsey, J., & Holmgren, R. (1987). Cell and cubitus interruptus Dominant: Two segment.polarity genes on the fourth chromosome in Drosophilu. Dev. Biol. 124,5&56. Orenic, T.V., Slusarski, D.C., b o l l , K.L., & Holmgren, R.A. (1990). Cloning and characterization of the segment polarity gene cubitus interruptus Dominant of Drosophila. Genes and Dev. 4, 1053-1067.
Drosophila Terminal Gene Hierarchy
141
Padgett, R.W., St. Johnston, R.D., & Gelbart, W.M. (1987). A transcript from a Drosophila pattern gene predicts a protein homologous to the transforming growth factor-p gene family. Nature 325, 81-84. Pankratz, M.J., & Jackle, H. (1993). Blastoderm segmentation. In: The Development of Drosophila melunogaster (Bate, M., ,& Martinez Arias, A., Eds.). Cold Spring Harbor Laboratory Press, New York. pp. 467-5 16. Pawson, T., & Schlessinger, J. (1993). SH2 and SH3 domains. Cum.Biol. 3,434-442. Peifer, M., & Wieschaus, E. (1990). The segment polarity gene armadillo encodes a functionally modular protein that is the Drosophila homolog of human plakoglobin. Cell 63, 1167-1178. Perkins, L.A., Larsen, I., & Perrimon, N. ( 1992). corkscrew encodes a putative protein tyrosine phosphatase that functions to transduce the terminal signal from the receptor tyrosine kinase torsso.Cell 70, 225-236. Perrimon, N., and Gans, M. ( 1983). Clonal analysis of tissue-specificity of recessive female-sterile mutations of Drosophila melanoguster using a dominant female-sterile mutation Fs(l)K1237. Dev. Biol. 100, 365373. Perrimon, N., & Mahowald, A.P. ( 1 986). I(l)hopscotch, a larval-pupal zygotic lethal with a specific maternal effect on segmentation in Drosophila. Dev. Biol. 1 1 8 , 2 8 4I . Perrimon, N., & Mahowald, A.P. (1987). Multiple functions of segment polarity genes in Drosophila. Dev. Biol. 119,587400. Perrimon, N., & Smouse, D. (1989). Multiple functions o f a Drosophilu homeotic gene, zeste-white 3, during segmentation and neurogenesis. Dev. Biol. 135,287-305. Perrimon, N., Engstrom, L., & Mahowald, A.P. (1989). Zygotic lethals with specific maternal effect phenotypes in Drosophila melanogaster. 1. Loci on the X-chromosome. Genetics 121,333-352. Petschek, J., & Mahowald, A.P. ( 1 990). Different requirements for I(1)giant in two embryonic domains of Drosophilu melunogaster. Dev. Genet. 11, 88-96. Petschek, J., Perrimon, N., & Mahowald, A.P. (1987). Region specific effects in I(1)giunt embryos of Drosophilu. Dev. Biol. 119, 177-189. Pignoni, F., Balderelli, R.M., Steingrimmson, E., Diaz, R.J., Patapoutian, A,, Merriam, J.R., & Lengyel, J.A. (1990). The Drosophila gene tailless is expressed at the embryonic termini and is a member of the steroid receptor superfamily. Cell 62, 151-163. Pignoni, F., Steingrimsson, E., & Lengyel, J.A. (1992). bicoid and the terminal system activate tailless expression in the early Dro.sophilu embryo. Development 115, 2 3 S 2 5 1. Poole, S.J., Kauvar, L.M., Drees, B., & Kornberg. (1985). The engrailed locus of Drosophila: Structural analysis of an embryonic transcript. Cell 40,3743. Preat, T., Therond, P., Lamour-Isnard, C.. Limbourg-Bouchon, B., Tricoire, H., Erk, I., Mariol, M.C., & Busson, D. (1990). A putative serinehhreonine protein kinase encoded by the segment polarity fused gene of Drosophila. Nature 347,87789. Ray, R.P., Arora, K., Niisslein-Volhard, C., & Gelbart, W.M. (1991). The control of cell fate along the dorso-ventral axis of the Drosophila embryo. Development 113, 35-54. Read, D., & Manley, J.L. ( 1992). Alternatively spliced transcripts of the D~-o.sophila tramtrack gene encode zinc finger proteins with distinct DNA binding specificities. EMBO J. 11, 1035-1044.
142
MAN LUN R. YIP and HOWARD D. LlPSHlTZ
Reinitz, J., & Levine, M. (1990). Control of the initiation of homeotic gene expression by the gap genes giaiir and tailless in Drosophila. Dev. Biol. 140.57-72. . Reuter, D. (1994). The gene serpent has homeotic properties and specifies endodenn versus ectoderm within the Drosophila gut. Development 120, 1123-1 135. Reuter, D., & Leptin, M. (1994). Interacting function ofsnail,twist and huckehein during the early development of germ layers in Drosophilu. Development 120; I 137-1 150. Riddihough, G.,& Ish-Horowicz, D. ( 1 99 I). Individual stripe regulatory elements in the Drosophila hairy promoter respond to maternal, gap, and pair-rule genes. Genes and Dev. 5,84-54. Riggleman, B., Wieschaus, E., & Schedl, P. (1989). Molecular analysis of the armadillo locus: Uniformly distributed transcripts and a protein with novel internal repeats are associated with a Drosophila segment polarity gene. Genes and Dev. 3 , 9 6 1 13. Riggleman, B., Schedl, P., & Wieschaus, E. ( I 990). Spatial expression of the Drosophila segment polarity gene armadillo is post-transcriptionally regulated by wingless. Cell 63,549-560. Rijsewijk, F., Schuermann, M., Wagenaar, E., Parren, P., Weigel, D., & Nusse, R. (1987). The Drosophila homolog of the mouse mammary oncogene int- I is identical to the segment polarity gene wingless. Cell 50,649457. Rogge, R.D., Karlovich, C.A., & Banerjee, U. (199 I). Genetic dissection of a neurodevelopmental pathway: Son of sevenless functions downstream of the .seven/es.sand EGF receptor tyrosine kinases. Cell 64,39-48. Rogge, R., Cagan, R., Majumdar, A,, Duianey, T., & Banerjee, U. (1992). Neuronal development in the Di-osophila retina: The sextm gene defines an inhibitory component in the developmental pathway of R7 cells. Proc. Natl. Acad. Sci. USA 89, 5271-5275. Ronchi, E., Treisman, J., Dostatni. N., Struhl, G.,& Desplan, C. ( 1994). Down-regulation of the Drosophila morphogen hicoid by the forsoreceptor-mediated signal transduction cascade. Cell 74,347--355. Rosenberg, U.B., Schroder, C., Preiss, A., Kienlin, A., Cote, S., Riede, I., & Jackle, H. (1986). Structural homology of the product of the Drosophila Kriippel gene with Xenopus transcription factor IIIA. Nature 31Y, 336-339. Roth, S., Stein, D., & Nusslein-Volhard, C. (1989). A gradient of nuclear localization of the dorsal protein determines dorsoventral pattern in the Drosophila embryo. Cell 59, I 189-1 202. Roth, S., Hiromi, Y., Godt, D., & Nusslein-Volhard, C. (1991). cactus, a maternal gene required for proper formation of the dorsoventral morphogen gradient in Drosophilu embryos. Development 112, 37 1-388. Rusch, J., & Levine, M. (1994). Regulation of the dorsal morphogen by the Toll and torso signaling pathways: A receptor tyrosine kinase selectively masks transcriptional repression. Genes and Dev. 8, 1247-1257. Rushlow, C., Frasch, M., Doyle, H., & Levine, M. ( 1 987a). Maternal regulation of zerkniillt: A homeobox gene controlling differentiation of dorsal tissues in Drosophila. Nature 330, 583-586. Rushlow, C.A., Doyle, H., Hoey, T., & Levine, M. (1987b). Molecular characterization of the zerkniilli region of the Antennupedia gene complex in Di-osophila.Genes and Dev. I , 1268-1279.
Drosophila Terminal Gene Hierarchy
143
Rushlow, C.A., Han, K., Manley, J.L., & Levine, M. (1989a). The graded distribution of the dorsal morphogen is initiated by selective nuclear transport in Drosophila. Cell 59, 1165-1 177. Rushlow, C.A., Howe, K.R., Pinchin, S.M., Hogan, A,, Lardelli, M.T., & Ish-Horowicz, D. (l989b). The Drosophila hairy prptein acts in both segmentation and bristle patterning and shows homology to N-myc. EMBO. J. 8.3095-3 103. Sanchez-Herrero, E., Vernos, I., Marco, R., & Morata, G. (1985). Genetic organization of Drosophila bithorax complex. Nature 313, 108-1 13. Savant-Bhonsale, S., & Montell, D.J. (1 993). tovso-like encodes the localized determinant of Drosophila terminal pattern formation. Genes and Dev. 7,254&2555. Schlessinger, J., & Ullrich, A. ( 1992). Growth factor signaling by receptor tyrosine kinase. Neuron 9,383-391. Schmidt-Ott, U., & Technau, G.M. (1992). Expression of en and wg in the embryonic head and brain of Drosoplzilu indicates a refolded band of seven segment remnants. Development 116, 1 1 1-125. Scholz, H., Deastrick, J., Klaes, A,, & Klambt, C. (1993). Genetic dissection ofpointed, a Drosopliila gene encoding two ets-related proteins. Genetics 135,455-468. Schupbach, T., & Wieschaus, E. (1986a). Germline autonomy of maternal-effect mutations altering the embryonic body pattern of Drosophila. Dev. Biol. 113,443-448. Schupbach, T., & Wieschaus, E. ( 1986b). Maternal-effect mutations altering the anteriorposterior pattern of the Drosopkila embryo. Roux's Arch. Dev. Biol. 195,302-3 17. Schupbach, T., & Wieschaus, E. (1989). Female sterile mutations on the second chromosome of Drosophila melanogaster. 1. Maternal effect mutations. Genetics 121, 10 1-1 17. Shilo, B.-Z. (1992). Roles of receptor tyrosine kinases in Drosophila development. FASEB J. 6,2915-2922. Shilo, B.-Z., & Raz, E. (1991). Developmental control of the Drosophila EGF receptor homolog DER. Trends Genet. 7,388-392. Shimell, M.J., Ferguson, E.L., Childs, S.R., & O'Connor, M.B. (1991). The Drosoplda dorsal-ventral patterning gene tolloid is related to human bone morphogenetic protein-I. Cell 67,469-481. Siegfried, E., Wilder, E.L., & Perrimon, N. (1994). Components of wingless signaling in Drosophila. Nature 367, 76-80. Simon, M.A., Kornberg, T.B., & Bishop, J.M. (1983). 3 loci related to the src oncogene and tyrosine-specific protein-kinase activity in Dvosophilu. Nature 302, 837-839. Simon, M.A., Drees, B., Kornberg, T.B., & Bishop, J.M. ( 1985). The nucleotide sequence and the tissue-specific expression of Dvovophila c-src. Cell 42, 83 1-840. Simon, M.A., Bowtell, D.D., Dodson, G.S., Laverty, T.R., & Rubin, G.M. (1991). rasl and a putative guanine nucleotide exchange factor perform crucial steps in signaling by the sevenless protein tyrosine kinase. Cell 67, 701-716. Simon, M.A., Dodson, G.S., & Rubin, G.M. (1993). An SH3-SH2-SH3 protein is required for ~ 2 1 ~ "activation ' and binds to sevenless and Sos proteins in vitro. Cell 73, 169-177. Sprenger, F., & Nusslein-Volhard, C. (1992). tor.70 receptor activity is regulated by a diffusible ligand produced at the extracellular terminal regions of the Drosophila egg. Cell 71,987-1 00 1.
144
M A N LUN R . YIP and HOWARD D. LlPSHlTZ
Sprenger, F., Stevens, L.M., & Niisslein-Volhard,C. (1989). The Drosophilu gene torso encodes a putative receptor tyrosine kinase. Nature 338,478-483. St. Johnston, D., & Gelbart, W.M. ( 1987).decuperrtaplegictranscripts are localized along the dorsal-ventral axis of the Drosophilu embryo. EMBO J. 6,2785-2791. St. Johnston, D., Hoffmann, EM., Blackman, R.K., Segal, D., Grimaila, R., Padgett, R.W., Irick, H.A., & Gelbart, W.M. (1990). Molecular organization of the decapentuplegic gene in Drosophila melanogaster. Genes and Dev. 4, 11 14-1 127. St. Johnston, D., & Niisslein-Volhard,C. ( 1992). The origin ofpattern and polarity in the Drosophilu embryo. Cell 6 4 2 0 1-2 19. Steingrimsson, E., Pignoni, F., Liaw, G.J., & Lengyel, J.A. (1991). Dual role of the Drosophila pattern gene tailless in embryonic termini. Science 254,4 18-42 1. Sternberg, P.W. (1993). Intercellular signaling and signal transduction in C. eleguns. Ann. Rev. Genet. 27,497-52 1. Stevens, L.M., Frohnhofer, H.G., Klingler, M., & Nusslein-Volhard, C. ( 1990). Localized requirement for torso-like expression in follicle cells for development of terminal anlagen of the Drosophila embryo. Nature 346,66(M62. Steward, R. (1987). dorsal, an embryonic polarity gene in Drosophilu, is homologous to the vertebrate proto-oncogene, c-ref.Science 238,692494. Steward, R. (1989). Relocalization of the dorsal protein from the cytoplasm to the nucleus correlates with its function. Cell 59, 1 179-1 188. Steward, R., Zusmann, S.B., Huang, L.H., & Schedl, P. (1988). The dorsal protein is distributed in a gradient in early Drosophilu embryos. Cell 55,487495. Strecker, T., Kongsuwan, K., Lengyel, J., & Merriam, J. (1986). The zygotic mutant tailless affects the anterior and posterior of the Drosophilu embryo. Dev. Biol 113, 64-76. Strecker, T., Merriam, J.R., & Lengyel. J.A. (1988). Graded requirement for the zygotic terminal gene, tailless, in the brain and tail region of the Diasophilu embryo. Development 102,72 1-734. Strecker, T.R., Halsell, S.R., Fisher, W.W., & Lipshitz, H.D. (1989). Reciprocal effects of hyper- and hypoactivity mutations in the DIasophilu pattern gene tor.^^. Science 243, 1062-1066. Strecker, T.R., & Lipshitz, H.D. (1990). Functions of the Drosophila terminal genes in establishing embryonic pattern. In: Developmental Biology, UCLA Symp. Mol. Cell. Biol, New Series, Vol. 125 (Davidson, E.H., Ruderman, J.V.,& Posakony, J.W., Eds.). Wiley-Liss, New York, pp. 85-94. Strecker, T.R., Yip, M.L.R., & Lipshitz, H.D. (1991). Zygotic genes that mediate torso receptor tyrosine kinase functions in the Drosophila embryo. Proc. Natl. Acad. Sci. USA 88,5824-5828. Strecker, T.R., Yip, M.L.R., & Lipshitz, H.D. (1992). Genetic control ofcell fate in the termini of the Drosophilu embryo. Dev. Biol. 150,422426. Sweeton, D., Parks, S., Costa, M., & Wieschaus, E. (1991). Gastrulation in Drosophila: The formation of the ventral furrow and posterior midgut invaginations. Development 112,775-789. Tautz, D., Lehmann, R., Schnurch, H., Schuh, R., Seifert, E., Kienlin, A,, Jones, K., & Jackle, H. (1987). Finger protein of novel structure encoded by hunchback,a second member of the gap class of Drosophilu segmentation genes. Nature 327, 383-389.
Drosophila Terminal Gene Hierarchy
145
Therond, P., Busson, D., Guillemet, E., Limbourg-Bouchon, B., Preat, T., Terracol, R., Tricoire, H., & Lamour-Isnard, C. ( 1993). Molecular organization and expression pattern of the segment polarity gene jiised of Drosophila melanoguster. Mech. Dev. 44,6540. Tsuda, L., Inoue, Y.H., Yoo, M.-A., Mizunq, M., Hata, M., Lim, Y.-M., Adachi-Yamada, T., Ryo, H., Masamune, Y., & Nishida, Y. (1993). A protein kinase similar to MAP kinase activator acts downstream of the Rafkinase in Drosophila. Cell 7 2 , 4 0 7 4 14. Vogel, W., Lammers, R., Huang, J., & Ullrich, A. (1993). Activation ofaphosphotyrosine phosphatase by tyrosine phosphorylation. Science 259, I61 1-1614. Vojtek, A.B., Hollenberg, S.M., &Cooper, J.A. (1993). Mammalian Ras interactsdirectly with the serinekhreonine kinase raf. Cell 74, 205214. Wakimoto, B.T., Turner, F.R., & Kaufman, T.C. (1984). Defects In embryogenesis in mutants associated with the Antennapedia gene complex in Drosoplzilu melunoyuster. Dev. Biol. 102, 147-172. Walldorf, U., & Gehring, W.J. ( 1992).empr?/spiruc/es,a gap gene containing a homeobox involved in Drosophila head development. EMBO J. JJ, 2247-2259. Warne, P.H., Viciana, P.R., & Downard, J. (1993). Direct interaction of Ras and the amino-terminal region of Raf-l ~n vitro. Nature 364, 352-355. Weigel, D., Bellen, H.J., Jiirgens, G., & Jackle, H. (1989a). Primordium specific requirement of the homeotic gene fork head in the developing gut of the Drosophila embryo. Roux's Archiv. Dev. Biol. 198,201-210. Weigel, D., Jurgens, G., Kiittner, F., Seifert, E., & Jackle, H. (1989b). The homeotic gene fork head encodes a nuclear protein and is expressed in the terminal regions of the Drosophila embryo. Cell 57,645-658. Weigel, D., Jurgens, G., Klingler, M., & Jackle, H. (1990). Two gap genes mediate maternal terminal pattern information in Drosophila. Science 248,495498. Wharton, K.A., Ray, R.P., & Gelbart, W.M. (1993). An activity gradient of decapentuplegic is necessary for the specification of dorsal pattern elements in the Drosophilu embryo. Development I 1 7,807-822. Wieschaus, E., Nusslein-Volhard, C., & Jurgens, G. (1984). Mutations affecting the pattern of the larval cuticle in Drosophilu melanoguster. I. Zygotic loci on the X chromosome and the fourth chromosome. Roux's Archiv. Dev. Biol. 193,296-307. Wozney, J.M., Rosen, V., Celeste, A.J., Mitsock, L.M., Whitters, M.J., Kriz, R.K., Hewick, R.M., & Wang, E.A. (1988). Molecular cloning of novel regulators of bone formation. Science 242, 1528-1 534. Xiong, W-C., & Montell, C. (1993). trurntrack is a transcriptional repressor required for cell fate determination in the Drosopkilu eye. Genes and Dev. 7, 10851096. Yan, R., Small, S., Desplan, C., Dearolf, C.R., & Damell, J.E. (1996). Identification of a Stat gene that functions in Drosophila development. Cell 84,42 1-430. Yarden, Y., & Ullrich, A. ( I 988). Growth factor receptor tyrosine kinases. Ann. Rev. Biochem. 57,443-478. Zhang, X.F., Settlemam, J., Kyriakis, J.M., Takeuchi-Suzuki, E., Elledge, S.J., Marshall, M.S., Bruder, J.T., Rapp, U.R., & Avruch, J. (1993). Normal and oncogenic ~ 2 1 " ' proteins bind to the amino-terminal regulatory domain of c-Raf- 1. Nature 364, 308-313.
146
M A N LUN R. YIP and HOWARD D.LlPSHlTZ
Zhong, W., Sladek, EM., & Darnell, J.E. (1993). The expression pattern o f a Drosophilu homolog to the mouse transcription factor HNF-4 suggests a determinative role in gut formation. EMBO J. 12,537-544. Zusman, S.B., & Wieschaus, E.F. ( 1985). Requirements for zygotic gene activity during gastrulation in Drosophilu melunoguster. Dev. Biol. 111,359-371. Zusman, S.B., Sweeton, D., & Wieschaus, E.F. (1988). short gastrulation, a mutation causing delays in stage-specific cell shape changes during gastrulation in Drosophilu melanogaster. Dev. Biol. /29,4 17427.
ANTERIOR-POSTERIOR POLARIZATION AND MESODERM INDUCING FACTORS IN THE PREGASTRULA MOUSE EMBRYO: COMPARISON TO CHICK A N D FROG EMBRYOS
Rosemary F. Bachvarova
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Frog Embryo as a Paradigm Defining Mesoderm Inducing Factors and Inducers of the Organizer . . . 111. Origin of Anterior-posterior Asymmetry in the Chick and Mouse Embryo . . . . . . . . . . . . . . . . . . . . . IV. Comparison of Fate Maps of Chick and Mouse Embryos with that of the Frog: Possible Location of Inducers of Mesoderm and Inducers of the Organizer . . . . . . V. The Hypoblast of the Chick Embryo . . . . . . . . . . . . . . . A. Origin of the Hypoblast . . . . . . . . . . . . . . . . . . . B. Role of the Hypoblast and Related Structures . . . . . . . .
148
11.
Advances in Developmental Biology Volume 4, pages 147-189. Copyright 0 1996 by JAl Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-969-9
147
149 152
156
I60 160 162
148
VI.
VII.
VIII. IX.
ROSEMARY F. BACHVAROVA
C . Goosecoid Expression: High Expression in Cells of the Organizer, Low Expression in Lower Layers before Gastrulation . . . . . . . . . . . . . . . . . . . . . . 164 The Visceral Endoderm of the Mouse Embryo . . . . . . . . . . . 165 A. Origin and Functions of the Visceral Endoderm . . . . . . . 165 B. Expression of Goosecoid and HNF3P . . . . . . . . . . . . 167 Sites of Expression of Potential Inducing Factors 169 and the Effects of Activin . . . . . . . . . . . . . . . . . . . . . A. Expression of Factors in the Mouse Embryo . . . . . . . . . 169 B. Expression of Factors in the Chick Embryo and Possible Role of the Germ Wall . . . . . . . . . . . . . . . . I72 C. Effect of Activin on Chick and Mouse Embryos . . . . . . . 173 Mouse Mutations Affecting the Onset of Gastrulation . . . . . . . 173 Recapitulation and Conclusions . . . . . . . . . . . . . . . . . . 176 A. Maternal Supply of Factors to the Early Vertebrate Embryo . 177 B. Role of the Visceral Endoderm of the Mouse Embryo . . . . 178 C. A Model for Initiation of the Primitive Streak 179 in the Mouse Embryo . . . . . . . . . . . . . . . . . . . . . 18 I Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
1. INTRODUCTION Gastrulation is the process by which future mesoderm and endoderm cells move from the surface of the embryo to the interior, the broad outlines of anterior-posterior and dorsal-ventral cellular identities are established, and tissue layers are brought into the proper associations for development of the organ rudiments. Between fertilization and gastrulation, two crucial events occur; these are the establishment of two axes which provide cues for the polarized activities occurring at gastrulation and the inductive processes that result in the appearance of mesoderm. This review is intended to cover the events leading up to and including early gastrulation of the mouse embryo, from implantation at 4.5 days of development through initiation of the primitive streak at 6.5 days. The origins of anterior-posterior asymmetry and the role of the maternal environment will be considered. The frog embryo is used as the paradigm to define the primary agents which induce mesoderm and confer anterior-posterior asymmetry expressed in the appearance of the organizer. The chick is probably more closely related to mammals than is the frog, and its embryo is examined in detail for clues to the location, source, and nature of the mesoderm inducers and of anterior-posterior positional
Inducing Factors in the Pregastrula Mouse Embryo
149
information. Evidence is scanty on several of the issues raised, but it is hoped that this synthesis of available data from three of the most intensely studied vertebrate embryos will help clarify avenues for future research.
II. THE FROG EMBRYO AS A PARADIGM DEFINING MESODERM INDUCING FACTORS AND INDUCERS OF THE ORGANIZER In the frog embryo, gastrulation is initiated at the dorsal lip of the blastopore which appears within the dorsal equatorial zone. At this site involuting cells move inward and anteriorly to form dorsal mesoderm and endoderm layers. At progressively later times more ventral and posterior mesoderm and endoderm enter at the lateral and ventral lips around the circular blastopore. The early dorsal lip marks the position of the organizer, a structure which organizes itself and surrounding tissues to form dorsal axial structures (anterior head mesendoderm and notochord). The organizer was defined in the well-known experiments of Spemann and Mangold (1 924) by its ability to induce a new axis when transplanted to the ventral side of another embryo. The secondary axis is composed of both the transplanted organizer and cells recruited from the host embryo. Events leading up to establishment of the organizer are now partially understood, The animal-vegetal axis of the frog egg is fixed in the ovary. The event which determines the second axis is called symmetrization, since bilateral symmetry is conferred simultaneously. In amphibian embryos, this second axis is called the dorsal-ventral axis. In Xenopus embryos the orientation of this axis is triggered at fertilization by the sperm entry point (Gerhart et al., 1986, 1989). During the first cell cycle the egg cortex rotates 30" with respect to the inner cytoplasm around the future right-left axis of the embryo. The future dorsal side forms opposite the sperm entry point in the region where the vegetal cortex rotates up over the cytoplasm of the animal hemisphere. If rotation is prevented, development is radially symmetric and follows a ventral pattern on all sides. Cortical rotation results in formation of a Nieuwkoop Center in the dorsal vegetal region (Figure la). This center was identified by its ability to induce a dorsal lip with organizer properties in the equatorial zone above it in various transplantation experiments. For example, if the equatorial region of the embryo is discarded and the vegetal region is
ant
posterior
8.CHICK
A FROG
extraembrvonic ctoderm
posterior
epiblast
distal C. MOUSE
Figure 1. Possible homologies in induction of mesoderm and the organizer by vertebrate embryos. (A) A diagram showing the two early signals generally accepted in current models of mesoderm induction in the frog embryo. Ventral mesoderm inducing signals (small arrows) are produced from the lower vegetal zone symmetrically, while the inducer of the organizer (dorsal anterior mesoderm) (large arrow) is emitted from the dorsal vegetal zone, the NieuwkoopCenter (NC). The location of the future organizer is indicated by asterisks (*). (B) A hypothetical model showing the possible location in the unincubated chick embryo of signals homologous to the two signals of the frog embryo, Ventral and posterior mesoderm inducing signals may come from the germ wall all around the blastoderm. Inducers of anterior mesoderm may come from the posterior germ wall. Location of the future organizer (anterior primitive streak cells) is indicated by asterisks (*). (C)A hypothetical model showing the possible location in the pregastrula mouse embryo of signals homologous to mesoderm inducing signals of lower vertebrates. Inducers of mesoderm (arrows)may come from all sides via the visceral endoderm, especially in the extraembryonic region of the egg cylinder. Inducers of the organizer may come from the posterior visceral endoderm or more likely from all sides, and anterior-posterior polarity may be present only in the embryonic ectoderm before gastrulation. Location of the future organizer cells is indicated by asterisks (*). 150
Inducing Factors in the Pregastrula Mouse Embryo
1 s1
rotated 180" and recombined with the animal region, an axis develops that corresponds to the original dorsal-ventral axis of the vegetal portion with an organizer above the dorsal vegetal zone (Nieuwkoop, 1973; Nieuwkoop and Sutasurya, 1979; Gerhart et al., 1989). Thus, the Nieuwkoop Center is defined as a center in the vegetal region that receives information from the original symmetrization event and induces a Spemann organizer in an adjacent region of the embryo, but does not itself contribute to the embryonic axis. In the current view, mesoderm induction during cleavage involves a minimum of two factors or sets of factors that are emitted from the vegetal zone, one inducing ventral or general mesoderm emitted from all regions of the vegetal zone, and one inducing the organizer (dorsal mesendoderm) emitted from the Nieuwkoop Center in the dorsal vegetal zone (Slack, 1994; see Figure la). These factors interact with the equatorial zone via cell contact or diffusion over a short distance, and effects on the blastocoel roof are prevented by the presence of the blastocoel. The organizer is viewed as emitting a third type of signal whose action is to dorsalize adjacent mesodermal regions, producing intermediate mesodermal cell types. This review is concerned primarily with events leading up to the appearance of the organizer (i.e., with the nature and source of the first two signals). There are several factors that are candidates for the two early inducers (for reviews see Slack, 1994; Dawid, 1994). Potential endogenous inducers have been defined as those factors present at the right time and place, and possessing appropriate functional attributes revealed in several types of assays. One assay tests the ability of factors to induce various kinds of mesoderm in animal caps (a region of future ectoderm dissected out of the blastocoel roof) and to induce rapid (immediate early) expression of transcription factors such as Him-I, forkhead, and goosecoid, genes normally expressed in the early dorsal lip (see Slack, 1994; Dawid, 1994). Another is the ability of the mRNA encoding a factor to induce a secondary axis when injected into the ventral side ofearly embryos. Candidate factors include the TGFP family members Vg-1 (Thomsen and Melton, 1993) and activin, and a Wnt family member such as Wnt-1 1 (Ku and Melton, 1993), and FGFs. Wnts are not active on their own, but are able to potentiate the effects of other factors in inducing dorsal mesoderm (Kimelman and Christian, 1992). A Wnt plus activin or plus Vg- 1 are strong candidates for the inducer of the organizer (i.e., these combinations of factors are highly active in inducing dorsal anterior mesoderm). Vg-1 is present as an abundant
152
ROSEMARY F. BACHVAROVA
maternal mRNA localized in the vegetal hemisphere. Activin is a potent inducer, producing a range of mesodermal cell types from dorsal to ventral in a concentration-dependent manner (Thomsen et al., 1990; Green eta]., 1994; Gurdon et al., 1994). Activin protein is found in the egg and early embryo (Asashima et al., 1991; Fukui et al., 1994), and is most likely taken up by oocytes from material secreted by follicle cells (Dohrmann et al., 1993; Rebagliati et al., 1993), but its location within the early embryo is unknown. A Wnt activity could act as the primary polarizing agent potentiating the activity of homogeneously distributed factors or could act in concert with a gradient of Vg-1 or activin activity. FGFs are important at least in formation of posterior ventral mesoderm (Amaya et al., 1991), and BMPs are also candidates for inducers of ventral mesoderm (Dawid, 1994). Besides localized action of inducing factors within the embryo, receptors may be localized or limiting, since ectopic expression of an activin receptor can produce ectopic mesoderm (Kondo et al., 1991). Formation of all mesoderm can be blocked by the expression of a truncated activin receptor expected to block activin receptor function in a dominant negative fashion (Hemati-Brivanlou and Melton, 1992). This receptor also blocks the action of Vg- 1 (Schulte-Merker et al., 1994) and perhaps other factors in the TGFP family, so the results demonstrate that some factor(s) in this family is crucial in mesoderm formation. A strong case can be made for a crucial role of an activin-like factor in fish embryos (Wittbrodt and Rosa, 1994). In addition to inductive processes, autonomous factors within the equatorial zone of the frog embryo also contribute to early specification of mesoderm, since the early embryonic marginal zone can produce dorsal axial structures when transplanted to ectopic sites (Gallagher et al., 1991). The animal region also possesses a latent polarity. Relatively large animal caps cultured in the presence of activin are able to form structures called embryoids with a well defined cranial-caudal axis (Sokol and Melton, 1991), and differences between dorsal and ventral animal cells can be detected as early as the eight-cell stage (Kinoshita et al., 1993).
111. ORIGIN OF ANTERIOR-POSTERIOR ASYMMETRY IN THE CHICK AND MOUSE EMBRYO In birds and mammals, as in the frog, determination of the second major axis is a key event. In these forms, the second axis (analogous to the
Inducing Factors in the Pregastrula Mouse Embryo
153
dorsal-ventral axis of the frog) is the anterior-posterior axis. There is a partial alignment of the dorsal-ventral and anterior-posterior axes in the pregastrula fate map of both frog and bird (Stern et al., 1992). The dorsal-ventral axis of the chick egg is established before ovulation. The second axis is established b r i n g the symmetrization event that occurs around 18 hours after fertilization and seven hours before laying, as the embryo passes down the uterus (Kochav and Eyal-Giladi, 1971; Eyal-Giladi, 1984). Within the uterus the whole yolk rotates and the blastodisc is tilted as it floats on the rotating yolk; the future posterior end lies at the upper edge. The first sign of anterior-posterior polarity occurs as the multilayered disc is transformed into the one cell thick area pellucida, a process which starts at the future posterior side. When the process is complete, the area pellucida is surrounded by the area opaca or germ wall, a ring of cells covered by a layer of ectoderm and merging with the yolk below. The molecular imprint of the symmetrization event may be carried as a gradient or as a localized region in the area pellucida and/or in the germ wall. Anterior-posterior polarization is definitively established and readily observed when the primitive streak and organizer are formed at the posterior side of the area pellucida 8-10 hours after laying. For an outline of normal mouse development, see Figure 2. The 3.5 day midblastocyst appears radially symmetric around the dorsal-ventral axis with the inner cell mass defining the dorsal side (Figure 2a). The first obvious sign of anterior-posterior asymmetry appears at the time of primitive streak formation on the posterior side of the egg cylinder at 6.5 days. As in the chick, the primitive streak is the site of the first mesoderm formation where epiblast cells move outward (downward with respect to the epithelium) and migrate between the epiblast and the visceral endoderm (Figure 2e). The node appears at the anterior end of the primitive streak at about seven days and has properties similar to that of the frog organizer (Waddington, 1956; Beddington, 1994). Careful histological analyses of 3.5-9.5 day embryos fixed and sectioned in the uterus at known angles has provided evidence that a second axis actually appears at the time of implantation; this axis can be traced forward in time and is found to correspond to the anterior-posterior axis (Smith, 1980, 1985; see for a review Gardner, 1990). Moreover, the blastocyst becomes positioned during early implantation in one of two orientations with respect to the uterus. These lead directly to the two orientations kfiown for the 6.5-8.5 day embryo in which the cranialcaudal axis is positioned horizontally across the uterus approximately
154
ROSEMARY F. BACHVAROVA
Figure 2. Normal development of the mouse embryo. (A) Mid-blastocyst at 3.5 days of development. The inner cell mass is located on the future dorsal side. The polartrophectoderm over the inner cell mass and the mural trophectoderm form a complete sphere. (B) During implantation the inner cell mass attains an asymmetric shape extending further ventrally on the future posterior side (right side in the diagram). The primary endoderm has segregated at the lower surface of the inner cell mass and the parietal endoderm has begun to migrate out under the mural trophectoderm. (C) Embryo at five days. The inner cell mass cells have proliferated to form a mass of embryonic ectoderm and the polar trophectoderm has pro1iferated to form a mass af extraembryonic ectoderm. The embryonic ectoderm i s often called epiblast to indicate that its cells give rise to ectodermal, mesodermal, and endodermal derivatives. The embryonic and extraembryonic regions of the visceral endoderm can be seen covering the embryonic and extraembryonic ectoderm, respectively. (continued)
Inducing Factors in the Pregastrula Mouse Embryo
155
Figure 2. (continued) The parietal endoderm has migrated out under the mural trophectoderm. (D) Embryo at 6.25 days just before appearance of the primitive streak. The cylindrical shape of the embryo has been established by further downward growth of the extraembryonic and embryonic ectoderm and establishment of the proamniotic cavity. (E) Embryo at 6.5 days as the primitive streak is forming on t h e posterior (right) side. The primitive streak is seen as a thickening of the wall of the cylinder, where epiblast cells move outward toward the visceral endoderm to form mesoderm cells that move laterally, proximally, or distally between the endoderm and ectoderm.
parallel to the left-right axis of the mother with the head oriented 50% of the time to the right and 50% to the left. The sequence of events described by Smith ( I 980, 1985) is briefly as follows. Before implantation the blastocyst lies on its side with its dorsal-ventral axis aligned with the cranial-caudal axis of the uterus. The blastocyst contacts the side of the uterine tube, and depending on whether it is the right or left side of the uterus, the uterine reaction pushes the embryo into one of two orientations with the dorsal surface facing anteriorly or posteriorly. The side facing the ventral floor of the uterine lumen will later become the posterior side. By 4.25 days, the posterior side shifts upward and the anterior-posterior axis aligns approximately with the anterior-posterior axis of .the uterus. The embryo displays bilateral symmetry with an asymmetric inner cell mass elongated ventrally on the posterior side (Figure 2b). A blastocyst isolated after delayed implantation has a similar asymmetric appearance (Gardner, 1990). Later, the anterior-posterior axis of the embryo shifts to approximately right-left in the uterus, in one of the two orientations described above. These results are consistent with the possibility that the initial information for posterior is received from the ventral floor of the uterus, and information for the right side from whichever side of the uterine wall is touched first. The results are also consistent with an alternative possibility that a previously existing anterior-posterior asymmetry of the embryo aligns with an asymmetry of the uterus. As far as is known, other vertebrates use an environmentallyderived asymmetric influence to orient their second axis, so that the former possibility seems more likely. Whatever cues are provided by the uterus, no specific factors are required, since normal development can occur through this period in vitro (Hsu, 1979). Thus, the embryo must be able to use some feature of the
156
ROSEMARY F. BACHVAROVA
tissue culture environment, such as the location of the bottom of the dish to orient its axis, or its own inherent axis forming mechanism is able to stabilize and amplify some minor random asymmetric fluctuation. The important point is that from about five days of development onward either the ectoderm or the endoderm or both potentially carry information that is different on the anterior vs the posterior side. This may exist as a center of positive activity on the posterior side, or a gradient of activity extending from one side to the other.
IV. COMPARISON OF FATE MAPS OF CHICK AND MOUSE EMBRYOS WITH THAT OF THE FROG: POSSIBLE LOCATION OF INDUCERS OF MESODERM AND INDUCERS OF THE ORGANIZER Fate maps of the pregastrula and early gastrula stage amphibian, bird, and mouse embryos are shown in Figure 3. During evolution, the largest shifts in the positions of the major regions have apparently occurred between a presumed common ancestor resembling the amphibian with moderate amounts of yolk in the egg, and birds or reptiles with much larger amounts of yolk. The changes can be understood as a large expansion of the yolk mass and a shift of the hture mesoderm toward the posterior side (see Waddington, 1956; compare the side views of
Figure 3. Fate maps of amphibian, chick, and mouse embryos indicating the positions of major mesoderm and endoderm components. Presumptive areas are not actually separated by sharp boundaries due to cell mixing which occurs to different extents in different species (there is more mixing in the chick; see Hatada and Stern, 1994); the areas indicated have the highest probability of becoming a certain cell type if followed during deveiopment. (A) Side view of an amphibian early blastula. The site of the Nieuwkoop Center (NC) is indicated. Asterisks mark the site of the organizer (prechordal plate or head mesendoderm). Based on Slack (1 991), and Nieuwkoop and Sutasurya (1979). (B) View of the vegetal region of the amphibian blastula, distorted so that the equatorial zone is in view. (C) Chick embryo at laying, with the blastoderm, that floats as a flat disc on the surface of the yolk, imagined as bulging upward so that half can be viewed from the side; a large yolk mass extends below. See also the legend of panel (D). (D) Dorsal view of the chick embryo at the same stage as in (C). (continued)
Inducing Factors in the Pregastrula Mouse Embryo
157
A e
C
0
CHICK
F ectoderm -mesoderm
MOUSE
CHICK AND MOUSE
Figure 3. (Continued)The area pellucida consists of the central disc plus the marginal zone ( M a . Most of the marginal zone will form extraembryonic mesoderm. The position of Koller’s Sickle (KS) is indicated. Asterisks mark the position of the organizer. The fate map is not accurately known at this stage and only approximate positions are indicated. The notochord and endoderm regions largely overlap. Based on Vakaet (1984, 1985) and Hatada and Stern (1994). (E) Dorsal view of the mouse early gastrula as it would appear if the embryonic cylinder were flattened out. Adapted from Lawson et al. (1991 and extraembryonic ectoderm added. Asterisks mark the position of the organizer. (F) Dorsal view of the chick or mouse gastrula. The position of the developing notochord is indicated by dashed lines. Note that between (D) and (F), or between (E) and (F), the future mesoderm and endoderm have converged on the midline and moved anteriorly. For the chick, based on Vakaet (1 984, 1985). See also Schoenwolf and Watterson (1989). For the mouse, based on Snow (1 981) and Tam (1 989).
158
ROSEMARY
F. BACHVAROVA
amphibian and chick embryos in Figures 3a and 3c). Although the fate map of the just laid chick embryo is not accurately known, the main mesoderm-forming regions are distributed along with the future extraembryonic mesoderm in a broad rim around the posterior half of the blastoderm near the germ wall (Figures 3c and 3d; Vakaet, 1984, 1985; Hatada and Stern, 1994); the position of the mesoderm is clearly related to its position in the amphibian blastula. The future endoderm and mesoderm of the chick embryo then converge on the caudal midline and extend anteriorly in the midline (see dorsal views in Figures 3d and 30. Most dramatically, the endoderm moves into a midline position extending along the early primitive streak (Vakaet, 1970, 1985). This transition involves extensive movements within the epiblast layer, including an approximately 90' turn as the two sides converge. As a result, the embryo-forming regions of the two sides are now aligned abutting each other (Figure 30. The fate map is better known at this stage and the main regions are aligned in the same order along the primitive streak as they are aligned in the amphibian around the blastopore (compare Figures 3b and 30. The primitive streak does not surround the yolk as does the blastopore of the amphibian, and during gastrulation a three-layered elongated disc is formed that does not engulf the yolk. As in the amphibian, future notochord and paraxial mesoderm converge from lateral positions toward the site of involution and then move forward in the midline. However, a significant component of medial movement toward the midline and anterior extension occurs during the pregastrula phase. From the fate maps, the region of the chick embryo that is most likely to be homologous to the vegetal zone of the amphibian is the yolky endoderm of the germ wall (Figures la, 1b, 3a, and 3c). This region forms extraembryonic endoderm in the chick; in the frog it also may not make any contribution to the embryo, rather it may provide yolk only (see Cooke, 1991). The marginal zone of the area pellucida consists of a ring of epiblast lying just inside of the area opaca; it can be seen more or less clearly from soon after laying to formation of the primitive streak (Figures 3c and 3d; Vakaet, 1970; Eyal-Giladi, 1984,1991; Stem, 1990). The marginal zone is also a candidate for a homolog of the amphibian vegetal zone. Thus, a Nieuwkoop Center may be located either in the posterior germ wall or the posterior marginal zone. In other words, a signal homologous to that coming from the dorsal vegetal region in the frog to induce the organizer may come from the posterior marginal zone and/or posterior germ wall of the chick at an early stage (see Stem, 1992).
lnducing Factors in the Pregasrrula Mouse Embryo
159
Before the convergence movements of the epiblast, the future organizer region is adjacent to the posterior marginal zone, a configuration similar It is at this point that a to that in the frog blastula (Figures 1 and 3 A). chick Nieuwkoop Center could act,in a mode most similar to that of the frog. After the convergence movements, the germ wall is adjacent to the caudal end of the primitive streak and the node is located some distance away (Figure 30. Continuing induction to support the formation and development of the primitive streak as it elongates away from the posterior margin may be supplied by the underlying hypoblast (see below). The six-day prestreak mouse embryo is a cylinder closed off at the lower end and divided roughly into a lower embryonic half and an upper extraembryonic half (Figures Ic and 2d). Both regions of the cylinder consist of two layers, an inner layer of ectoderm and an outer layer of visceral endoderm. The ectoderm is called embryonic ectoderm, primitive ectoderm, or epiblast in the embryonic region, and extraembryonic ectoderm in the extraembryonic region. The visceral endoderm over the embryonic region is called embryonic visceral endoderm, although it does not contribute to the true embryonic endoderm, which is derived solely from the epiblast. The mouse embryo is much smaller than frog and chick embryos. The 6.5 day egg cylinder is about 0.1 mm in diameter, while the frog egg is about 1 mm and the chick blastoderm about 3 rnm in diameter. Nevertheless, the fate map of the epiblast of the mouse embryo can be directly related to that of the chick (compare Figures 3d and 3e). If in one’s imagination the embryonic portion of the cylinder is cut off and the upper edges expanded, the whole may be flattened out into a disc. The fate map as determined by Lawson et al. (1991) can then be applied to the disc and the result (Figure 3e) is similar to that of the chick. If the extraembryonic region of the cylinder is expanded and added to the map, it will form a belt around the outside corresponding approximately to the germ wall or area opaca of the chick. There is no clear intermediate zone of ectoderm comparable to the marginal zone of the chick embryo. The fate maps suggest that mesoderm inducers homologous to those coming from the vegetal zone of the frog or germ wall of the chick could come from the extraembryonic visceral endoderm in the mouse (Figure 1). Medial convergence and alignment of mesoderm-forming areas along the primitive streak occur between the early primitive streak and mid primitive streak stages (Figures 3e and 30. These are comparable to the movements in the pregastrula chick embryo, but apparently take place
160
ROSEMARY F. BACHVAROVA
somewhat later than in the chick. Perhaps the delay is related to the small size of the mouse embryo and short distances of migration involved. The possibility that inducers of the organizer are emitted by homologous regions of the embryo in different vertebrates will be explored. Major differences between the amniotes and the amphibian are likely to be related to convergence to form the primitive streak and extension of the anterior streak away from the primary inducing region.
V. THE HYPOBLAST OF THE CHICK EMBRYO At the onset of gastrulation the hypoblast is a complete cellular layer lying below the epiblast of the chick embryo. It is generally regarded as an important carrier of anterior-posterior information as gastrulation approaches (Nieuwkoop and Sutasurya, 1979; Eyal-Giladi, 1984, 1991 ; Khaner, 1993). The posterior germ wall and posterior marginal zone were identified above as potential sites for the inducer of the organizer. What is the relationship of the hypoblast to the posterior germ wall, and might its role be to carry inducing signals forward from the posterior marginal zone along the length of the primitive streak? A. Origin of the Hypoblast
When the chick egg is laid, the embryo consists of a single-layered circular area pellucida (epiblast) overlying a cavity above the yolk. Outside the area pellucida lies a ring of area opaca (germ wall) cells covered by a layer of ectoderm. The hypoblast is formed as a complete layer in the approximately I0 hours between the start of incubation and appearance of the primitive streak. At the onset of hypoblast formation, the area pellucida is divided into a central disc and a narrow outer ring, the marginal zone. The marginal zone is underlain by a loosely attached medial extension of the germ wall. The hypoblast forms under the central disc. The hypoblast does not contribute to the definitive embryo, since gut endoderm comes entirely from cells ingressing from the epiblast at the primitive streak (Fontaine and Le Douarin, 1977; Vakaet, 1984). It is generally agreed that two different zones of the hypoblast can be distinguished. (1) The hypoblast of anterior and lateral regions arises as scattered cells moving downward from the upper epiblast layer by a process called polyingression. (2) The hypoblast that underlies the future primitive streak in the posterior region of the area pellucida arises at a structure called Koller’s sickle. Koller’s sickle lies at the border between
161
Inducing factors in the Pregastrula Mouse Embryo
the central disc and the posterior marginal zone and consists of a ridge of cells on the ventral surface of the epiblast forming a crescent in the posterior quadrant of the embryo (see Figures 3c and 3d). Here more concentrated clusters of cells appear in the lower layer and migrate anteriorly (Eyal-Giladi and Kocha;, 1976; Vakaet, 1984; Stern, 1990). Two origins for this posterior hypoblast have been proposed (Figure 4). The first is by centrally directed growth of cells from the medial extension of the germ wall (Stem, 1990). In this view, cells of the posterior germ wall expressing the cell surface antigen HNKl are continuous with and give rise to HNKl positive cells of the hypoblast (Canning and Stem, 1988). Alternatively, other workers have concluded that it arises from epiblast cells in the posterior marginal zone (Eyal-Giladi et al., 1992). Earlier workers proposed an epiblast origin for the hypoblast of both bird and reptile embryos (Nelsen, 1953). Another similar view is that the posterior marginal zone epiblast undergoes more concentrated polyingression related to the accumulation of cells in the posterior epiblast (Vakaet, 1970, 1984); once in the lower layer the cells generated move
\ /
germ wall
Figure 4. Possible models of development and functions of the posterior marginal zone, germ wall, and the hypoblast in the chick embryo. A diagram of a mid-sagittal section of the posterior region of an early chick blastoderm is shown with the epiblast above, the forming hypoblast below, and the middle layer in between. The marginal zone (MZ) and Koller’s sickle (KS) are indicated. Dark arrows indicate that the posterior hypoblast (cross hatched) may be derived from the posterior germ wall (dark shading) or from the epiblast of the marginal zone (light shading). Middle layer and posterior hypoblast cells express goosecoid (g). The posterior germ wall, the hypoblast, and the posterior marginal zone are potential sources of inducing factors acting on the epiblast (light arrows).
162
ROSEMARY F. BACHVAROVA
anteriorly. The posterior hypoblast forms and spreads anteriorly shortly before the medial convergence and anterior movement within the epiblast described above. Cell labeling studies on the origin of the hypoblast have yielded contradictory results. Stern ( 1990) states that labeled posterior epiblast cells remain in the epiblast, and regeneration of the hypoblast requires the presence of the medial portion of the germ wall. Eyal-Giladi et al. (1992) state that presence or absence of the medial germ wall has no effect on formation of the hypoblast and that labeled posterior marginal zone epiblast move into the hypoblast. In addition and perhaps partially reconciling these opposing views, more complex events may occur at Koller’s sickle. Stem (1 990) and Izpisua-Belmonte et al. (1 993) note the presence of small mesenchyme-like cells which appear to form a middle layer between the hypoblast and the epiblast, and Vakaet (1985) states that the convergence of the epiblast toward the posterior midline “results in an inward bending of the upper layer” and “part of the pre-endoblast is thus shifted into the middle layer.” This middle layer of cells is not well characterized and even its existence is controversial (see EyalGiladi et al., 1992). A possible conclusion is that the layers below the posterior epiblast may have two different origins; perhaps the middle layer comes from the epiblast, and the lower layer from the germ wall. 8. Role of the Hypoblast and Related Structures
The filly formed hypoblast may play a role in induction and positioning of the primitive streak. When the hypoblast is rotated with respect to the epiblast, the primitive streak that forms aligns at least partially with the original axis of the hypoblast; in some cases the hypoblast may even induce anew primitive streak (see Waddington, 1932,1933; Eyal-Giladi, 1984). It appears relatively late, and its function is most likely to promote the convergence and extension within the epiblast that carries future anterior mesendoderm away from the posterior edge of the area pellucida. It may be regarded as a late Nieuwkoop Center extending forward to provide prolonged induction under the reacting epiblast. Like the frog vegetal zone, it appears to be divided into two functionally different regions, the anterior-lateral and the posterior. It forms in a manner quite different from the frog vegetal zone, partly by delamination and partly by polarized growth from Koller’s sickle, but differences in the way the cells of a Nieuwkoop Center develop are to be expected, since the polarizing events at symmetrization are different.
Inducing Factors in the Pregastrula Mouse Embryo
163
An alternative view relegates the hypoblast to a subsidiary role. In this view the organizer is induced early while its precursors are near the posterior marginal zone, and the organizer itself is responsible for later “inducing” functions of the hypoblast. Cells moving into the middle layer at Koller’s sickle could be early medendoderm cells important in forming the organizer at the node and could account for the inductive properties of the hypoblast (see below). The role of the hypoblast would be reduced to a passive or permissive role. The two main candidates for the site of an earlier Nieuwkoop Center before formation of the hypoblast are the same two structures that are candidates for the precursors of the hypoblast, that is the posterior marginal zone epiblast and the posterior germ wall. One or both of these would receive the original posteriorizing signal and emit anterior mesoderm inducing factors. If the early marginal zone, Koller’s sickle, and area opaca of a freshly laid embryo are rotated with respect to the epiblast, a primitive streak is induced at the position of the original posterior marginal zone (see Eyal-Giladi, 1991). Transplantation of a portion of the posterior marginal zone, including part of Koller’s sickle, to a lateral position in the same embryo results in a primitive streak at the new site (Khaner and Eyal-Giladi, 1989). This and additional experiments suggest that the posterior marginal zone represents the peak of a gradient of activity, which localizes the axis and polarity of the primitive streak (Khaner, 1993; Eyal-Giladi, 1991). However, the epiblast at Koller’s sickle contains precursors of cells of the primitive streak, suggesting that this inducing activity may not be analogous to that of a Nieuwkoop Center, rather it may be the expression of primitive streak organizer activity. Also, the posterior marginal zone epiblast often has some lower layer cells associated with it; these are probably derived from the germ wall. Thus, little is known about the location of a possible Nieuwkoop Center in the chick embryo, that is, inducing cells that do not themselves form part of the axis of the embryo. Whether the germ wall can induce on its own at some earlier stage has not been tested. While it is likely that there is significant overlap in mesoderm inducers between frog and chick, much more work is necessary at the cellular and molecular levels to identify and determine the roles of the different cell layers in the chick. As in the frog, information for anterior-posterior polarity is contained within the epiblast as well as in the hypoblast. The central disc of epiblast from a freshly laid egg (at least some of the hypoblast being contained within it) is capable of forming a primitive streak (Khaner, 1993). Later
164
ROSEMARY F. BACHVAROVA
the central disc isolated just before gastrulation requires posterior marginal zone or hypoblast to express its ability to form a primitive streak, but it can do so even in the presence of a disaggregated and reaggregated hypoblast which must lack anterior-posterior information (Eyal-Giladi, 1984).
C. Goosecuid Expression: High Expression in Cells of the Organizer, l o w Expression in Lower Layers before Gastrulation The goosecoid gene encodes a transcription factor with homology to the bicoid and gooseberry genes in Drosophila (see Cho et al., 1991). High expression of goosecoid marks the cells of the organizer region in all three organisms (Beddington and Smith, 1993). Quite detailed information is available from in situ hybridization studies on the early expression pattern of goosecoid in frog and chick embryos, as well as mouse (see below). The level of sensitivity of detection of goosecoid RNA may be different in different experiments; nevertheless, the comparison provides tentative support for the view that a low level of goosecoid expression defines a region with Nieuwkoop Center inducing activity. Frog
In the frog embryo, high expression of goosecoid is seen in the early dorsal lip and in its progeny, the future anterior mesendoderm migrating forward during gastrulation (Cho et al., 1991). Goosecoid is expressed in animal caps as an early response to inducers of dorsal mesoderm such as activin (Cho et al., 1991; Green et al., 1994). Ventral equatorial zone cells overexpressing goosecoid take on properties of organizer cells, including mobility toward the future head region of the embryo and mobilization of adjacent host cells (Niehrs et al., 1993). Interestingly, goosecoid expression has also been observed in the pregastrula vegetal zone by RNase protection, perhaps corresponding to the autonomous (not dependent on cell interaction) expression described in the same report (Lemaire and Gurdon, 1994). Chick Goosecoid expression in the chick embryo has been described using in situ hybridization by Izpisua-Belmonte et al. (1993), and briefly by Hume and Dodd (1 993). Expression first appears in the prestreak embryo in the posterior sickle-shaped region called Koller’s sickle (see Figures
Inducing Factors in the Pregastrula Mouse Embryo
165
3 and 4). Sections show that expression occurs in some cells of a middle layer between the epiblast and hypoblast and in some cells within the hypoblast. A little later expression extends anteriorly, apparently coincident with the advancing anterior margin of the posterior hypoblast. Then strong expression is seen in a circular region that represents the anterior portion of the newly formed primitive streak; stained cells are found in lower and middle layer cells. Later, it is expressed in all three layers at the center of the definitive node. Cell labeling experiments show that middle layer cells expressing goosecoid below Koller ’s sickle become part of the middle and lower layers of the definitive node (IzpisuaBelmonte et al., 1993). Transplantation of the region of Koller’s sickle with cells expressing goosecoid can induce an ectopic axis (Izpisua-Belmonte et al., 1993). However, some of these cells may have organizer rather than inducer activity. In summary, some or all of the posterior hypoblast expresses goosecoid at least at a low level, and an early middle layer appearing at Koller’s sickle also expresses at a moderate level. Later, high expression is found in newly formed mesendoderm of the anterior primitive streak and then quite high expression in all layers at the node. Strong expression of goosecoid appears to be associated with important functions of the organizer region. The low level of expression seen in the vegetal zone of the frog and in the lower layers of the chick may not represent a major function for goosecoid in these cells. However, it may be useful as a marker either of an early inducing region or possibly of early organizer cells.
VI. THE VISCERAL ENDODERM OF THE MOUSE EMBRYO A. Origin and Functions of the Visceral Endoderrn
At the onset of gastrulation the embryonic visceral endoderm occupies a position similar to that of the hypoblast in the chick, lying under the epiblast that will form the embryo (Figure 2d). However, the two layers are directly apposed in the mouse, while in the chick the two layers are loosely attached. As in the chick, the embryonic visceral endoderm will later form only extraembryonic endoderm (Lawson and Pedersen, 1987). As noted above, the position of the extraembryonic visceral endoderm can be compared to the germ wall endoderm of the chick. Unlike the
166
ROSEMARY F. BACHVAROVA
chick there is no gap between the two regions of endoderm in the mouse embryo. Alternatively, the extraembryonic visceral endoderm could correspond to the anterior-lateral hypoblast, and the embryonic visceral endoderm to the posterior hypoblast of the chick. In the mouse embryo, the primary endoderm arises during implantation at about four days of development when it segregates from the inner cell mass. It then divides into the visceral endoderm remaining as the lower layer of the inner cell mass, and the parietal endoderm which migrates out under the trophectoderm (Figures 2b and 2c). The visceral endoderm soon forms a tight epithelial layer (Enders et al., 1978), and part of its role appears to be to protect the embryonic ectoderm, and to mediate exchange between the environment (maternal plasma and extracellular fluids) and the ectoderm from five days through gastrulation. As the extraembryonic ectoderm develops from the polar trophoblast, two regions of the visceral endoderm can be distinguished, the extraembryonic region adjacent to the extraembryonic ectoderm, and the embryonic region adjacent to the embryonic ectoderm (Figures 2c and 2d). At 5.5 days the visceral endoderm cells are generally cuboidal in shape and active in endocytosis with abundant microvilli and apical vesicles. They take up proteins, most of which are digested and the amino acids released toward the ectoderm. Some proteins may be transported intact across the epithelium by transcytosis; in addition, some may be synthesized and secreted by the visceral endoderm. All these activities have been clearly shown for the visceral endoderm from later stage rodent embryos (Jollie, 1990). By six days the visceral endoderm is columnar in the extraembryonic region and in a belt around the junction between the extraembryonic and embryonic regions; it is squamous over most of the embryonic region indicating lower endocytotic activity. The squamous form of the visceral endoderm over the embryonic region at ECE6.5 may result from specific properties of the adjacent epiblast, that is, the epiblast may supply different signals than the extraembryonic ectoderm. By E7-E7.5 the posterior visceral endoderm underlying the primitive streak is columnar, while the anterior visceral endoderm remains squamous. The important role of the visceral endoderm has been demonstrated by the effects of targeted disruption of the hepatocyte nuclear factor-4 (HNF-4) gene, a transcription factor originally defined for its role in liver-specific gene expression. It is expressed in the visceral endoderm from 4.5 days (Duncan et al., 1994) and homozygous mutant embryos exhibit delayed and defective gastrulation associated with cell death
Inducing Factors in the Pregastrula Mouse Embryo
167
primarily in the distal epiblast (Chen et al., 1994). This result raises the possibility that the visceral endoderm supplies, in addition to mesoderm inducing factors, survival factors specific for the epiblast. By analogy to the chick hypoblast, the embryonic visceral endoderm may be an inducing layer. Moreovdr, unlike the chick hypoblast, it is in a position to provide Nieuwkoop Center factors from an early stage.
6. Expression of Goosecoid and HNF-3P In the prestreak 5.5 and six-day embryo, goosecoid is expressed at a low level in the embryonic visceral endoderm (Figure 5a; Manova et al., unpublished results). In the early streak embryo, abundant transcripts are found in the anterior mesendoderm about halfway down the posterior side of the embryo; this region is homologous to the early dorsal lip of the frog. At this stage goosecoid is also expressed in the ectoderm of the anterior primitive streak and is up-regulated in the posterior visceral endoderm underlying the primitive streak and anterior to it (Blum et al., 1992; Conlon et al., 1994; Manova et al., unpublished results). Expression continues through at least seven days in anterior mesoderm. Thus, the pattern is very similar in chick and mouse embryos from primitive streak stages onward, and in prestreak stages both show expression in the lower (endodermal) layer. However, in the chick embryo goosecoid is expressed only in the posterior hypoblast, while in the pregastrula mouse embryo goosecoid is expressed in a radially symmetric fashion throughout the visceral endoderm. This is consistent with the different modes of formation of the hypoblast and visceral endoderm. In the mouse embryo there are no middle layer cells like those of the chick embryo under Koller’s sickle that express goosecoid before appearance of the primitive streak. Hepatocyte nuclear factor-3P (HNF-3P) is a transcription factor with a winged helix DNA binding domain. Like HNF-4, it is enriched in adult liver. Its homologs are expressed in the frog embryo in the dorsal lip of the blastopore, the future head mesendoderm, and in the notochord as well (Dirksen and Jamrich, 1992; Ruiz-i-Altaba et al., 1993). In mouse embryos HNF-3P transcripts appear at about six days in prestreak embryos in visceral and parietal endoderm around the embryo (Manova et al., unpublished results). In early streak embryos expression is initiated in the anterior region of the forming primitive streak and increases in the visceral endoderm under the primitive streak (Ang et al., 1993; Monaghan et al., 1993; Sasaki and Hogan, 1993; Ruiz-i-Altaba et
168
ROSEMARY F. BACHVAROVA
Figure5 In situ hybridization analysis of the expression of goosecoidand HNF-3P in early postimplantation mouse embryos. (A). High magnification bright field view of a 5.5-6 day mouse embryo section hybridized to an antisense goosecoid RNA probe labeled with P33. The visceral endoderm (arrows)around the embryonic portion of the cylinder contains goosecoid mRNA as shown by the presence of black grains over the cells. Scale bar = 20 pm. (B,C). Bright and dark field views, respectively, of a section of a 6.5 day HNF-4 homozygous mutant embryo hybridized to an HNF-3P antisense probe. The visceral and parietal endoderm around the whole embryo are labeled, and a more intense spot of labeling (arrow)appears in the posterior visceral endoderm on the posterior side of the embryonic region. Scale bar = 100 pm.
al., 1993; Chen et al., 1994). Expression continues in the anterior primitive streak and node, and in the visceral and parietal endoderm. Later, it appears in the head process and notochord as they extend anteriorly in the midline from the node. High expression in the posterior visceral endoderm occurs independent of that in the epiblast in HNF-4 mutant mice in which development of the epiblast is severely delayed (Figures 5b and 5c; Chen et al., 1994, unpublished results). This indicates that high expression at this site is not a secondary result of formation of the primitive streak. The phenotype of homozygous mutant embryos lacking HNF-3P expression is described later (Section VIII).
Inducing Factors in the fregastruta Mouse Embryo
169
In addition, nodal, a factor important in primitive streak formation (see below), is expressed in embryonic visceral endoderm from at least six days of development (Manova et al., unpublished results).
VII. SITES OF EXPRESSION OF POTENTIAL INDUCING FACTORS AND THE EFFECTS OF ACTlVlN A. Expression of Factors in the Mouse Embryo As discussed previously, it is proposed that inducers ofmesodermmay be provided by the extraembryonic and/or embryonic visceral endoderm, and inducers of the organizer possibly from posterior visceral endoderm. The three most likely candidates for the inducer of the organizer in frog embryos are activin and Vg- 1 with a potentiating activity from a Wnt. More is known about the expression of growth factors in the mouse than in chick embryos, so the mouse will be discussed first. Of these, activin is the only one for which clearly homologous genes have been identified so far in mammals. The closest counterparts of Vg-I are BMP-6 and GDF-3 (Kingsley, 1994); a close homolog of Wnt 11 has not been found. Activin has been localized at an appropriate time and place to be involved in mesoderm induction in the mouse embryo (Manova et al., 1992; Albano et al., 1994). Activin is agrowth factor in the TGFP family; it is a homo- or heterodimer of activin PA and PB polypeptide chains (see Kingsley, 1994). Activin PA mRNA is first observed by in situ hybridization at 4.5 days in a few uterine decidual cells near the implantation site. At 5.5 days it is localized in the decidua near the distal tip of the embryo and by 6.5 days in decidual cells around most of the embryonic portion of the egg cylinder. At 7.5 days the mRNA is no longer observed at this location. mRNA for the activin inhibitor follistatin is expressed in the decidua outside the region expressing activin PA at 5.5 to 6.5 days; it also appears in parietal endoderm from about seven days (Albano et al., 1994), suggesting that any in vivo action of activin PA on the embryo is likely to occur between about five and seven days of development. Activin protein can be seen by immunohistochemistry to reach the cavity around the embryo at 5.5-6.5 days (Manova et al., unpublished results). The receptor complex for activin is a heterodimer of type I and type I1 serine-threonine kinase receptors (Attisano et al., 1993; Carcamo et al., 1994). The location of the two main types of activin type I1 receptors has been analyzed (Manova et al., 1994) by in situ hybridization. Type
170
ROSEMARY F. BACHVAROVA
IIB is most abundant and is located in the epiblast from at least as early as 5.5 days and becomes increasingly strong in the epiblast at 6.5 and 7.5 days. The second receptor called Type 11 is also present in the embryo at 6.5 and 7.5 days, probably distributed through all cell types. The presence of type I receptors, which are also required for a functional response, has not yet been determined. Aplausible scenario for the exposure of receptor-bearing epiblast cells to activin is presented in Figure 6. Activin Aprotein (dimers of PAchains) secreted by decidual cel Is diffuses to the extraembryonic visceral endoderm (trophoblast and parietal endoderm do not provide a tight barrier). It is taken up and in part transported by transcytosis (see King, 1982) across the visceral endoderm layer. It is released to the extracellular fluid of the extraembryonic ectoderm, and interacts with cell surface receptors on nearby epiblast cells to influence their fate. Because it is transported more actively by the columnar endoderm of the extraembryonic region, cells in the proximal epiblast receive a higher exposure to the factors. The expression pattern and location of activin protein appears to be radially symmetric. By analogy with the frog, activin receptors may be activated by other factors, including a Vg-1 -like factor yet to be discovered. The role proposed here for activin may in fact be mediated in part by other factors in the TGFP family. Other factors expressed in the uterine decidua before gastrulation that could have effects on the embryo are activin B (Albano et al., 1994), TGF Pl and P2 (Tamada et al., 1990; Manova et al., 1992), TGF a (Tamada et al., 1991), as well as BMP-2a, BMP-4, and BMP-I which are expressed at 6.5 and 7.5 days (Lyons et al., 1990; Jones et al., 1991; Fukagawa et al., 1994). TGF p2 protein has been found within the visceral endoderm (Slager et al., 1991). Maternally provided TGFPl is important for normal development, although this occurs relatively late in embryogenesis (Lettorio et al., 1994). Female mice lacking activin PB can support normal gastrulation of their progeny, so that uterine expression of this factor is not required (Vassalli et al., 1994). FGFs appear to be involved in mesoderm induction in frogs and are necessary at least for development of posterior ventral mesoderm (Amaya et al., 1991). FGFs are expressed in early postimplantation mouse embryos. Most relevant for this discussion, FGF4 appears in the ICM of the blastocyst and continues in the epiblast; it is down-regulated in the anterior epiblast and expression continues only in the posterior epiblast at 6.25 days before the onset of gastrulation (Niswander and Martin, 1992). FGF8 is expressed in the 6 and 6.5 day posterior epiblast,
Inducing Factors in the Pregastrula Mouse Embryo
f
171
extraembryonic visceral endoderm
Figure 6. Expression of activin and activin receptors at the implantation site before gastrulation. Activin mRNA and protein (A) are found in the decidua around the implantation site. Activin protein diffuses through the trophobl&t and parietal endoderm to the yolk sac cavity around the embryo. It i s proposed that activin is transported across the visceral endoderm by transcytosis where it interacts with Type II activin receptors (R) expressed on the epiblast and at a lower level on the extraembryonic ectoderm (Manova et al., 1994). Other factors produced in the maternal decidua may follow the same route. and to a lesser extent in visceral endoderm (Crossley and Martin, 1995). Also, FGF3 is expressed in parietal endoderm (Wilkinson et al., 1988), and FGF5 is expressed in embryonic visceral endoderm and epiblast from six to seven days (Haub and Goldfarb, 1991). Two other factors for which direct homologs have not yet been identified in frogs are expressed in early postimplantation embryos in suggestive patterns. An EGF-like factor called cripto is expressed in the posterior epiblast at six days of development before primitive streak formation (Dono et al., 1993), but little is known about its activity. Perhaps most interesting is nodal, a factor in the TGFP family. mRNA
172
ROSEMARY F. BACHVAROVA
for nodal is found in the epiblast at 5.5 days, and in the posterior epiblast, the proximal anterior epiblast, and posterior visceral endoderm at 6.5 days (Zhou et al., 1993; Conlon et al., 1994). It is possible that like cripto and FGFs 4 and 8, it is expressed asymmetrically in the embryonic ectoderm before primitive streak formation. The phenotype of embryos lacking nodal expression is described below (Section VIII). In conclusion, while it is plausible that the embryonic visceral endoderm supplies mesoderm inducing factors to the epiblast, none have yet been identified that fit this simple pattern. The extraembryonic visceral endoderm may transport factors produced in the decidua. Several factors are expressed in the epiblast and the earliest asymmetric expression of factors is observed there. So far, asymmetric expression of growth factor receptors has not been observed before appearance of the primitive streak. B. Expression of Factors in the Chick Embryo and Possible Role of the Germ Wall There is less information on the location of factors in the chick embryo. bFGF is present before gastrulation and appears to play a role in induction of ventral mesoderm (Mitrani et al., 1990a). Recently, it has been shown that conditioned medium from the hypoblast can substitute for the hypoblast in eliciting a primitive streak in central discs (EyalGiladi et al., 1994). A prominent 28 kD protein in the medium has not yet been hlly characterized. It is interesting to consider the possibility that some factors may be stored in the yolk, and released by the yolk cells of the germ wall. In fact, immunoglobulins are stored in the yolk and transferred to the embryo, although at a stage after gastrulation (Buxton, 1952). Activin may be one factor that is stored in the yolk. It is made in high amounts in follicle cells of the frog and mouse, and activin protein appears in oocytes and remains into early development (see above and Albano et al., 1993). It is plausible that activin is also taken up by oocytes in the chick ovary. This would present an appealing homology between location of maternal factors in the vegetal region of the frog and location of factors in the area opaca of the chick mobilized from the yolk. It would also supply an appealing homology to the provision of factors to the mouse embryo via extraembryonic visceral endoderm as discussed above for activin A. Factors provided by the germ wall may be supplied in a radially symmetric fashion, or could be released preferentially on the posterior side as a
Inducing Factors in the Pregastrula Mouse Embryo
173
consequence of symmetrization, if primary information is received by the germ wall. Evidence that posterior germ wall cells express relatively high levels of the HNK-1 antigen (Canning and Stern, 1988; Cooke et al., 1994) implies that these are in some way different from other germ wall cells. C. Effect of Activin on Chick and Mouse Embryos
Central epiblast discs from pregastrula stage chick embryos cultured in the presence of activin can form a primitive streak with axial mesoderm, while in the absence of activin only ventral mesoderm such as blood islands is formed (Mitrani et al., 1990b). Localized application of activin to early epiblasts results in formation of a primitive streak aligned with the site of activin application, and Wnt activity potentiates the effect of activin (Ziv et al., 1992; Cooke et al., 1994). More widespread application of activin around the marginal zone or scattered across the lower surface of the epiblast resulted in “widespread but patternless streak-like activity” (Cooke et al., 1994). Treatment of 6.4 day mouse embryos with activin results in expression of goosecoid throughout the epiblast rather than in one spot on the posterior side (Blum et al., 1992). Exposure of embryonal carcinoma PI 9 cells to activin induces transient expression of Brachyury, a marker of primitive streak cells (see Beddington and Smith, 1993) and goosecoid (Vidrica‘ireet al., 1994), and inhibits differentiation (van den Eijnden-van Raaij et al., 1991). These results are analogous to effects of activin on the frog embryo in which localized application can induce a new axis, but widespread availability results in an attempt to make dorsal mesoderm throughout the animal half (Cooke, 1991). They are consistent with the fact that receptors for activin are widely distributed before gastrulation in the frog (Hemmati-Brivanlou et al., 1992) and mouse embryos (see preceding). They suggest that activin-like activity in vivo is either localized, or is present everywhere at a low level below that required for induction of anterior mesoderm. Activin activity may also be regulated by the balancing effect of inhibitory influences, such as a BMP (see Dawid, 1994).
VIII. MOUSE MUTATIONS AFFECTING THE ONSET OF GASTRULATION Several classic mutations, insertional mutations, and targeted gene disruptions are recessive lethals known to affect the mouse embryo between
1 74
ROSEMARY F. BACHVAROVA
5.5 and 7.5 days (Green, 1989; Holdener-Kenny et al., 1992 and references below). Of these, histological analyses suggest delay and general affects on the epiblast in sd, velvet, blind, andfugl (encodes a protein homologous to RNAZ in yeast involved in general RNA metabolism; DeGregori et al., 1994). Targeted disruption of the HNF-4 transcription factor expressed in the visceral endoderm results in cell death within the epiblast, delay of gastrulation, and arrest at the stage of early notochord formation, probably due to decreased provision of general and specific factors from the visceral endoderm (Chen et al., 1994). The tW5 mutation has a phenotype quite similar to that of HNF-4 mutants. Hp58 is strongly expressed in visceral endoderm (Lee et al., 1992) and encodes a protein similar to one involved in targeting some proteins to the lysosome in yeast (Bachhawat et al., 1994); its phenotype is less severe than that of HNF-4 mutants and may result from relatively mild effects on the visceral endoderm (Radice et al., 1991). Targeted disruption of the transcription factor evxl, which is expressed at a low level in embryonic visceral endoderm from five to six days, results in arrest at approximately 5.5 days (Spyropoulos and Capecchi, 1994). These mutations have provided clear evidence for an important role of the visceral endoderm before and during gastrulation, but do not indicate whether it produces specific inducing factors. Three mutations are known which affect the formation of mesoderm. Eed mutants form fairly normal extraembryonic mesoderm, but little embryonic mesoderm (Faust et al., 1995). Two mutations lead to the absence of mesoderm within the embryo. Embryos mutant for the TGFP family member nodal develop into folded layers of visceral endoderm and ectoderm, and cell death is obvious within the epiblast (Conlon et al., 1991,1994; Iannaccone et al., 1992; Robertson et al., 1992). As noted above, the nodal factor is expressed in 5.5 and 6.5 day epiblast including regions that will form mesoderm. These results strongly implicate premesoderm cells in synthesis of a factor affecting their own survival and development toward mesoderm. Embryonic stem cells mutant for nodal are nevertheless able to form a variety of mesodermal types upon differentiation in vitro (Robertson et al., l992), indicating that the factor promotes mesoderm formation rather than being required for it. In fact, nodal may be a key molecule in establishing polarity within the embryo. Interestingly, although nodal mutant embryos do not form mesoderm, a few embryos express the primitive streak marker Brachyury in a pattern that suggests more widespread expression than normal (Conlon et al., 1994). Thus, in the absence of nodal, other more widely distributed
inducing Factors in the Pregastrula Mouse Embryo
175
msd mutants
6-6.25days nodal mutants show a variable block at this +stage but ES cells are able to form mesoderm
6.5days Figure 7. Model for the initiation and development of the primitive streak in the mouse embryo. From 5.5-6.5 days, goosecoid(gsc)and HNF-3P are expressed symmetrically in the embryonic visceral endoderm. It is proposed that expression in the posterior epiblast of cripfo, FCF-4, FGF8, and possibly nodal and other factors at about 6.25 days initiates expression of Brachyury, HNF-3P and goosecoid (gsc) in the posterior epiblast at 6.5 days. These factors also act on the posterior embryonic visceral endoderm (dark shading) to upregulate expression of HNF-3P, nodal, and goosecoid. The combined action of genes expressed within the posterior epiblast and of factors expressed in the posterior embryonic visceral endoderm support the development of the primitive streak and appearance of mesoderm. Possible time points of action of two mutations that prevent mesoderm formation in the embryo are indicated.
176
ROSEMARY F. BACHVAROVA
factors like activin impinging on the embryo appear to be present at a level close to the threshold for induction of Bruchyuyexpressing premesoderm. In the presence of nodal, which is expressed primarily in the posterior epiblast, mesoderm formation may be potentiated on the posterior side. This side would then become the focus of mesoderm formation, and in turn would suppress mesoderm formation in other areas. For a discussion of this phenomenon in the chick embryo, see Cooke et al. ( 1994). Another locus called msd is also required for mesoderm formation (Holdener et al., 1994). In mutant msdembryos, genes such as Bruchyury and goosecoid (normally appearing in the primitive streak) are not expressed. At six days FGF4 is apparently down-regulated in both the anterior and posterior epiblast instead of on the anterior side only, suggesting absence of a positive factor on the posterior side. Mutant embryonic stem cells cannot produce mesodermal derivatives in vitro or in vivo, suggesting a defect early in the pathway of mesoderm determination. See Figure 7 for a model of the points at which nodal and msd may act. The phenotype of homozygous mutant embryos Iacking HNF-3P expression has recently been described (Ang and Rossant, 1994; Weinstein et al., 1994).The embryos lack a notochord, but in some cases form a recognizable axis with several pairs of somites, indicating that some organizer activity is present. Goosecoid expression occurs early, but the cells are located in the proximal region of the embryo and appear disorganized. Cells expressing Brachyury initially fail to move downward. From these results it is clear that HNF-3P is an important intermediate in interpreting polarizing and mesoderm inducing signals and conferring node and notochord properties on the cells expressing it.
IX. RECAPITULATION AND CONCLUSIONS Amajor question in comparative vertebrate embryology is: At what point in early development do the different embryonic types converge and start to use homologous genes and mechanisms? Certainly, the earliest stages of the three embryos considered here are quite different; the symmetrization event occurs at different stages and may be registered in different molecular forms. However, by the onset of gastrulation there are clear homologies in the fate maps and in the cellular ingression and gene expression patterns (Beddington and Smith, 1993), and there are functional homologies between the node and dorsal lip (Waddington, 1956;
Inducing Factors in the Pregastrula Mouse Embryo
177
Beddington, 1994). Thus, at some point shortly before gastrulation we would expect a convergence of molecular mechanisms. A. Maternal Supply of Factors to the Early Vertebrate Embryo I
The results on activin expression in frog and mouse presented earlier suggest evolutionary conservation of provision of certain maternal factors to the embryo but with different routes from maternal cells to the embryo. The proposed routes parallel those for nutritional supply to the embryo. In the frog and chick, vitellogenin is made in the liver, taken up by the oocyte, deposited preferentially in the vegetal half, and later digested and the break down products transmitted to the active embryonic cells. In the mammal nutritional proteins are found in the environment of the embryo, are taken up and digested by the visceral endoderm, and the products released immediately to the inner ectoderm cells. For growth factors, in the frog and chick this mode would OCCUT by uptake of factors secreted from somatic ovarian cells followed by storage in the yolky region of the egg. During embryogenesis these factors would undergo regulated release from a stored form. In the mammal, maternal protein factors may also be stored in the egg (see Albano et al., 1993), but the major source of maternal factors for the pregastrula embryo would be the decidua within the uterine wall (see Figure 6 ) .Such proteins would diffuse into the implantation cavity and reach the visceral endoderm. Uptake of the factors would be more active in the columnar visceral endoderm around the junctional region and around the extraembryonic ectoderm than in the more squamous embryonic visceral endoderm. Uptake and release by the vegetal region of the frog eggembryo may be compared to uptake and release by the extraembryonic visceral endoderm. In the mouse embryo, transport of factors across the visceral endoderm by transcytosis in vesicles is proposed; in chick and frog, storage in vesicles or in association with yolk and later release to the extracellular space at a multicellular stage. The targets of these factors would be the adjacent regions of the embryo. Exposure to a higher concentration of factors would occur at a homologous position in the responding region of the embryo, that is, in the equatorial region of the frog, in the marginal zone and outer rim of the central disc of the chick, and in the proximal epiblast of the mouse. A major feature evolved in the mammal would be the elevated expression of factors in some somatic cells of the uterus.
178
ROSEMARY F. BACHVAROVA
In this view the vegetal zone, yolky germ wall, and extraembryonic visceral endodeddecidua are considered to be homologous, and all three embryonic systems would use these tissues to provide similar mesoderm inducers. Such a mechanism might include production of the asymmetric signal of the Nieuwkoop Center via differential release or processing, but we assume here that it produces primarily symmetric signals.
B. Role of the Visceral Endoderm of the Mouse Embryo It should be noted that the extraembryonic visceral endoderm develops quite differently among different mammals. For example, in the human the extraembryonic visceral endoderm forms the yolk sac below the embryonic disc and is not in a position to mediate access of factors to the epiblast, a function largely carried out by trophectoderm. However, it should be noted that, as in the mouse embryo, the periphery of the embryonic disc is more accessible to uterine factors. Also, the embryonic visceral endoderm is in a position to play a similar role in human and mouse embryos. As described above, one role of the extraembryonic visceral endoderm may be transport of maternal factors to the embryo. In addition, although there is little evidence on this point, it may synthesize factors that promote mesoderm formation. Factors coming from this source would be supplied at higher concentration to the proximal epiblast. The mouse embryonic visceral endoderm may be similar to the chick hypoblast in supplying inducing factors, and yet it displays several major differences. The function of the embryonic visceral endoderm is to protect the embryo and to some extent to supply nutrition. The chick epiblast needs little protection and the germ wall supplies sufficient nutrients. Related to these functions, the mouse embryonic visceral endoderm arises one and one-half to two days before gastrulation, but the hypoblast is completed just before gastrulation in the chick. The embryonic visceral endoderm and much of the hypoblast do have a similar origin in polyingression or delamination from the epiblast. However, a major difference between the two types is that the chick hypoblast shows strong asymmetry in its mode of formation. As far as is known, in the mouse there is no process corresponding to the localized formation of the posterior hypoblast at Koller’s sickle. Also, the chick hypoblast shows strong asymmetry in goosecoid expression, while the pregastrula mouse embryonic visceral endoderm expresses goosecoid throughout.
Inducing Factors in the Pregastrula Mouse Embryo
179
There is no suggestion of a functional asymmetry, that is, that rotation of the mouse embryonic visceral endoderm would position the primitive streak in a new place. As noted above, the convergence and anterior migration of epiblast occurs later in the mouse than in tde chick and as far as is known is not preceded or accompanied by movements in the posterior visceral endoderm. In fact, formation of the embryonic visceral endoderm at about five days may be equivalent to formation of the posterior hypoblast, that is, all the embryonic visceral endoderm may be equivalent to posterior hypoblast. This is suggested by the fact that they both expressgoosecoid, albeit at low levels. No further movements would be required to provide a special endoderm under the early primitive streak. These considerations suggest that the embryonic visceral endoderm may produce inducers of dorsal mesoderm, but in a radially symmetric fashion. It would then be predicted that the entire epiblast is exposed to early influences toward mesoderm and would more readily form mesoderm than future ectoderm regions of amphibian and bird. The asymmetry appearing in the embryonic visceral endoderm after the onset of gastrulation suggests that the posterior embryonic visceral endoderm may be important in supporting development of the primitive streak once it has been initiated. It should also be noted that some of the same genes characteristically expressed in the forming and definitive node are also expressed in the embryonic visceral endoderm before gastrulation. In conclusion, the two primary candidates for sources of induction in the mouse are comparable to those in the chick, that is, the extraembryonic visceral endoderm and the embryonic visceral endoderm. It is very likely that each plays a role in supplying factors homogeneously around the embryo, and there is no evidence that either of these is polarized before appearance of the primitive streak, C. A Model for Initiation of the Primitive Streak in the Mouse Embryo
As discussed above in Section 111, the initial polarizing event probably occurs around the time of implantation. Which layer or region of the embryo carries this information and presents factors in an asymmetric fashion? Based on the lack of asymmetry in origin, gene expression patterns, and rhorphology, the extraembryonic and embryonic visceral endoderm have been provisionally ruled out as polarized sources of
180
ROSEMARY F. BACHVAROVA
factors. Thus, other sources of inducers of the organizer or other mechanisms must be considered. The only other possible source of anterior-posterior signals is the ectoderm. The embryonic ectoderm, which diverges early from the extraembryonic ectoderm at the time of formation of the inner cell mass, may have its own independent imprint of the symmetrization event. Alternatively, the embryonic ectoderm may become polarized as a secondary result of polarization of the extraembryonic ectoderm. In other words, the extraembryonic ectoderm may emit asymmetric signals received by the embryonic ectoderm. Interestingly, the ectoplacental cone formed from extraembryonic ectoderm exhibits distinct morphological asymmetry in the period from 5.5 to 6.5 days (Smith, 1985). Whatever the preceding mechanisms, as far as is known at this point, it is the embryonic ectoderm that first displays expression of genes specifically on the posterior side. This includes expression of FGF4, and FGF8, and cripto, an EGF-like factor, in the posterior epiblast at days, 6-12 hours before gastrulation (see Figure 7). The transcription factor evx-I shows a similar expression pattern (Dush and Martin, 1992). The TGFP family member nodal is also a likely candidate for early asymmetric expression between 5.5 and 6.5 days (see Conlon et al., 1994). In fact, nodal may be important in establishing polarity, since some nodal mutant embryos show widespread expression of Bruchyur?, (Conlon et al., 1994). Mutant embryonic stem cells can form mesoderm in vitro (Robertson et al., 1992), again suggesting that nodal is required for organization within the embryo rather than for mesoderm formation per se. Also, in HNF-30 mutant embryos early Bruchyury expression is not localized to the posterior side and elongation of the primitive streak by descent ofBrachyuryexpressing cells is delayed and incomplete (Ang and Rossant, 1994). These results suggest that the early expression of HNF-30 in the posterior epiblast is a key event in reinforcing anteriorposterior axis formation. In the model of Figure 7 it is proposed that the early asymmetric expression of factors in the posterior epiblast sets off events leading to the formation of the primitive streak and node. These factors are envisaged as acting within the context of factors provided symmetrically by the visceral endoderm as discussed above. It is proposed that signals originating in the posterior epiblast act back on the posterior epiblast to turn on expression of several genes (e.g., Bruchyuuy, HNF-3P, and goosecoid).They also act in parallel on the posterior embryonic visceral endoderm where expression ofgoosecoid and HNF-3 p become distinctly
Inducing Factors in the Pregastrula Mouse Embryo
181
higher than in the anterior visceral endoderm, and expression of nodal is initiated. The asymmetry of HNF-3j3 expression in the endoderm is particularly clear in HNF-4 mutant embryos (Figures 5b and 5c). The embryonic visceral endoderm is visibly polarized at this point and the posterior embryonic visceral endoderm would thereafter take on a more active role in supporting development of the primitive streak and node. It is interesting to note that the three genes expressed highly in the posterior embryonic visceral endoderm are also expressed in the posterior epiblast. This suggests that the inducing region and the organizer region share some molecular activities, and that the two layers cooperate in formation of the primitive streak. Similarly, in the frog embryo, the Nieuwkoop Center and the organizer may express shared dorsalizing agents (see Dawid, 1994). Thus, the structure of the mouse embryo that is most similar to aNieuwkoop Center is the posterior embryonic visceral endoderm underlying the forming primitive streak at 6.5-7.5 days. The main difference from the amphibian Nieuwkoop Center is that it arises late, simultaneous with or after the initiation of gastrulation, thus, it apparently does not receive and transmit the effects of the original symmetrization event. We are left then with the most likely possibility that the asymmetric signals inducing the organizer arise in the ectoderm itself and there is no direct equivalent of the Nieuwkoop Center in the endodermal region of the pregastrula mouse embryo. Perhaps this is related to the relatively small size of the embryo, permitting polarity to be established reliably within one cell layer. The inducing mechanism in the posterior epiblast would work in coordination with radially symmetric signals from the embryonic and extraembryonic visceral endoderm with strong homologies at the molecular level to events in lower vertebrates. In the chick and frog, polarity within the ectoderm-epiblast appears to be a second system working in parallel with the dominant information in the vegetal region (hypoblast); the ectoderm system may have become the most important or only source of anterior-posterior information in the mammal. The extent to which homologous genes may act within the ectoderm-epiblast of the different types of embryo remains to be elucidated.
ACKNOWLEDGMENTS We are grateful to Elizabeth Lacy, Stephen Duncan, and William Chen for helpful comments on the manuscript. Our research presented in this review was supported in part by NIH grant HD 069 10.
ROSEMARY F. BACHVAROVA
182
REFERENCES Albano, R.M., Groome, N., & Smith J.C. (1993). Activins are expressed in preimplantation mouse embryos and in ES and EC cells and are regulated on their differentiation. Development I 1 7, 71 1-723. Albano, R.M., Arkell, R., Beddington, R.S.P., & Smith, J.C. (1994). Expressionofinhibin subunits and follistatin during postimplantation mouse development: Decidual expression of activin and expression of follistatin in primitive streak, somites and hindbrain. Development 1 2 0 , 8 0 3 4 13. Amaya, E., Musci, T.J., & Kirschner, M.W. (1991). Expression of a dominant negative mutant of the FGF receptor disrupts mesoderm formation in Xenopus embryos. Cell 66,257-270. Ang, S.-L., Wierda, A., Wong, D., Stevens, K.A., Cascio, S., Rossant, J., & Zaret, K.S. (1993). The formation and maintenance of the definitive endoderm lineage in the mouse: Involvement of HNF3(fbrkhead proteins. Development 119, 130 1-13 15. Ang, S.-L., & Rossant, J. (1994). HNF-3P is essential for node and notochord formation in mouse development. Cell 78.561-574. Asashima, M., Nakano, H., Uchiyama, H., Sugino, H., Nakamura, T., Eto, Y., Ejima, D., Nishimatsu, S., Ueno, N., & Kinoshita, K. (1991). Presence of activin (erythroid differentiation factor) in unfertilized eggs and blastulae ofXenopus luevis. Proc. Nat. Acad. Sci. USA 88,651I 4 5 14. Attisano, L., Carcamo, J., Ventura, F., Weis, F.M.B., Massague, J., & Wrana, J.L. (1993). Identification of human activin and TGFP type I receptors that form heteromeric kinase complexes with type I1 receptors. Cell 75,67 1-680. Bachhawat, A.K., Suhan, J., & Jones, E.W. (1994). The yeast homolog of H
58, a mouse gene essential for embryogenesis, performs a role in the delivery of proteins to the vacuole. Genes Dev. 8, 1379-1387. Beddington, R.S.P. (1994). Induction of a second neural axis by the mouse node. Development 120,613420. Beddington, R.S.P., & Smith, J.C. ( 1993). Control of vertebrate gastrulation: Inducing signals and responding genes. Cum Op. Genet. Dev. 3,655-66 1. Blum, M., Gaunt, S.J., Cho, K.W.Y., Steinbeisser, H., Blumberg, B., Bittner, D., & De Robertis, E.M. (1992). Gastrulation in the mouse: The role of the homeobox gene goosecoid. Cell 69, 1097-106. Buxton, A. (1952). On the transference of bacterial antibodies from the hen to the chick. J. Gen. Microbiol. 7,268-286. Canning, D.R., & Stern, C.D. (1988). Changes in the expression of the carbohydrate epitope HNK-I associated with mesoderm induction in the chick embryo. Development 104,643-656. Carcamo, J., Weis, F.M.B., Ventura, F., Wieser, R., Wrana, J., Attisano, L., & Massague, J. (1 994). Type I receptors specify growth-inhibitory and transcriptional responses to transforming growth factor P and activin. Mol. Cell. Biol. 14,3810-3821. Chen, W.S., Manova, K., Weinstein, D.C., Duncan, S.A., Plump, A.S., Prezioso, V.R., Bachvarova, R.F., & Damell, J.E. (1994). Disruption of the HNF-4 gene, expressed in visceral endoderm, leads to cell death in embryonic ectodenn and impaired gastrulation of mouse embryos. Genes Dev. 8,24662477.
Inducing Factors in the Pregastrula Mouse Embryo
183
Cho, K.W.Y., Blumberg, B., Steinbeisser, H., & De Robertis, E.M. (1991). Molecular nature of Spemann’s organiser: The role of the Xenopus homeobox gene goosecoid. Cell 67, I 1 11-1120. Conlon, F.L., Barth, K.S., & Robertson, E.J. (1991). A novel retrovirally induced embryonic lethal mutation in the moupe: Assessment of the developmental fate of embryonic stem cells homozygous for the 4 13.d proviral integration. Development 111, 969-98 1. Conlon, F.L., Lyons, K.M., Takaesu, N., Barth, K.S., Kispert, A., Hemnann, B., & Robertson, E.J. (1994). A primary requirement for nodal in the formation and maintenance of the primitive streak in the mouse. Development 120, 191%1928. Cooke, J. (1991). The arrangement of early inductive signals in relation to gastrulation; Results from frog and chick. In: Gustrulution (Keller, R., Clark, W.H., & Griffin, F., Eds.). Plenum Press, New York, pp. 79-99. Cooke, J., Takada, S., & McMahon, A. (1994). Experimental control of axial pattern in the chick blastoderm by local expression of Wnt and activin: The role of HNK- I positive cells. Dev. Biol. 164, 5 13-527. Crossley, P.H., & Martin, G.R. (1995). The mouse Fsfx gene encodes a family of polypeptides and is expressed in regions that direct outgrowth and patterning in the developing embryo. Development 121,439-45 I . Dawid, I.B. (1994). Embryonic induction of axis formation in Xenopus laevis: Growth factor action and early response genes. In: Growth Factors and Signal Transduction zn Development (Nilsen-Hamilton, M. Ed.). John Wiley & Sons, New York, pp. 199-226. DeGregori, J., Russ, A., von Melchner, H., Rayburn, H., Priyaranjan, P., Jenkins, N.A., Copeland, N.G., & Ruley, H.E. (1994). A murine homolog of the yeast RNAl gene is’required for postimplantation development. Genes Dev. 8, 265-276. Dirksen, M:L., & Jamrich, M. (1 992). A novel, activin-inducible, blastopore-lip-specific gene of Xeriopus hevis contains a forkhead DNA-binding domain. Genes Dev. 6, 59M08. Dohrmann, C.E., Hemmati-Brevanlou, A., Thomsen, G.H., Fields, A,, Woolf, T.M., & Melton, D.A. (1993). Expression of activin mRNA during early development in Xenopus luevis. Dev. Biol. 157, 47-83. Dono, R., Scalera, L., Pacitico, F., Acampora, D., Persico, M.G., & Simeone, A. (1993). The murine cripto gene: Expression during mesoderm induction and early heart morphogenesis. Development 118, 1157-1 168. Duncan, S.A., Manova, K., Chen, W.S., Hoodless, P., Weinstein, D.C., Bachvarova, R.F., & Darnell, J.E. (1994). Expression of transcription factor HNF-4 in the extraembryonic endoderm, gut, and nephrogenic tissue of the developing mouse embryo: HNF-4 is a marker for primary endoderm in the implanting blastocyst. Proc. Nat. Acad. Sci. 91,7598-7602. Dush, M.K., & Martin, G.R. (I 992). Analysis ofmouse Evx genes: Evx-I displays graded expression in the primitive streak. Dev. Biol. 151,273-287, Enders, A.C., Given, R.L., & Schlafke, S. (1978). Differentiation and migration of endoderm in the rat and mouse at implantation. Anat. Rec. 190,65-78. Eyal-Giladi, H. ( 1984). The gradual establishment of cell commitments during the early stages of chick development. Cell Differ. 14,245-255.
184
ROSEMARY F. BACHVAROVA
Eyal-Giladi, H. (1991). The early embryonic development of the chick, as an epigenetic process. Crit. Rev. Poultry Biol. 3, 143-166. Eyal-Giladi, H., & Kochav, 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. I. General morphology. Dev. Biol. 49, 321-337. Eyal-Giladi, H., Debby, A., & Harel, N. ( 1992). The posterior section of the chick’s area peliucida and its involvement in hypoblast and primitive streak formation. Development 116,819-830. Eyal-Giladi, H., Lotan, T., Levin, T., Avner, O., & Hochman, J. (1994). Avian marginal zone cells function as primitive streak inducers only after their migration into the hypo blast. Development I2 0,2 50 1-2 509. Faust, C., Schumacher, A,, Holdener, B., & Magnuson, T. (1995). The eed mutation disrupts anterior mesoderm production in mice. Development 121,273-285. Fontaine, J., & Le Douarin, N.M. (1977). Analysis of endoderm formation in the avian blastoderm by the use of quail-chick chimeras. J. Embryol. Exp. Morph. 41, 20s222. Fukagawa, M., Suzuki, N., Hogan, B.L.M., & Jones, C.M. ( 1994). Embryonic expression of mouse bone morphogenetic protein- 1 (BMP- 1), which is related to the Dlasophila dorsoventral gene tolloid and encodes a putative astacin metalloendopeptidase. Dev. Biol. 163, 175-183. Fukui, A., Nakamura, T., Uchiyama, H., Sugino, K., Sugino, H., & Asashima, M. (1994). Identification of activins A, AB, and 3 and follistatin proteins in Xenoprrs embryos. Dev. Biol. 163,279-281. Gallagher, B.C., Hainski, A.M., & Moody, S.A. (1991). Autonomous differentiation of dorsal axial structures from an animal cap cleavage stage blastomere in Xenopus. Development 112, l 1 0 ~ l 1 1 4 . Gardner, R.L. ( 1990). Location and orientation of implantation. In: Establishing a Successful Human Pregnancy (Edwards, R.G., Ed.). Raven Press, New York, pp. 22S238. Gerhart, J., Black, S., Scharf, S., Gimlich, R., Vincent, J . 2 , Danilchik, M., Rowning, B., & Roberts, J. (1986). Amphibian early development. BioSci. 36,541-549. Gerhart, J, Danilchik, M., Doniach, T., Roberts, S., Rowning, B., & Stewart, R. (1989). Cortical rotation of the Xenopus egg: Consequences for the anteroposterior pattern of embryonic development. Development (Suppl.), 37-5 1. Green, J.B.A., Smith, J.C., & Gerhart, J.C. (1994). Slow emergence of a multithreshold response to activin requires cell-contact-dependent sharpening but not prepattern. Development 120,2271-2278. Green, M.C. (1989). Catalog ofmutant genes and polymorphic loci. In: Genetic Variants and Strains ofthe Laboratory Mouse (Lyon, M.F. & Searle, A.G., Eds.) pp. 12403. Oxford University Press, New York. Gurdon, J.B., Harger, P., Mitchell, A., & Lemaire, P. (1994). Activin signalling and response to a morphogen gradient. Nature 371,487492. Hatada, Y., & Stem, C.D. (1994). A fate map of the epiblast of the early chick embryo. Development 120,2879-2890. Haub, O., & Goldfarb, M. (1991). Expression of the fibroblast growth factor-5 gene in the mouse embryo. Development 112,397406.
Inducing Factors in the Pregastrula Mouse Embryo
185
Hemmati-Brivanlou, A., & Melton, D.A. (1992). A truncated activin receptor inhibits mesoderm induction and formation of axial structures in Xenopus embryos. Nature 359,609-614. Hemmati-Brivanlou, A,, Wright, D.A., & Melton, D.A. ( 1992). Embryonic expression and functional analysis of a Xenopus qctivin receptor. Devel. Dyn. 194, 1-1 I . Hemmati-Brevanlou, A., Kelly, O.G., & Melton, D.A. ( 1994). Follistatin, an antagonist of activin, is expressed in the Spemann organizer and displays direct neuralizing activity. Cell 77,283-295. Holdener, B.C., Faust, C., Rosenthal, N.S., & Magnuson, T. (1994). msd is required for mesoderm induction in mice. Development 120, 1335-1 346. Holdener-Kenny, B., Sharan, S.K., & Magnuson, T. ( 1 992). Mouse albino-deletions: From genetics to genes to development. BioEss. 14,831439. Hsu, Y.-C. (1979). In vitro development of individually cultured whole mouse embryos from blastocyst to early somite stage. Devel. Biol. 68,453-461. Hume, C.R., & Dodd, J. (1993). Cwnt-8C: A novel Wnt gene with a potential role in primitive streak formation and hindbrain organization. Development 119, 1 1471160. Iannaccone, P.M., Zhou, X., Khokha, M., Boucher, D., & Kuehn, M.R. (1992). Insertional mutation of a gene involved in growth regulation of the early mouse embryo. Devel. Dyn. 194, 198-208. Izpisua-Belmonte, J.C., De Robertis, E.M., Storey, K.G., & Stern, C.D. (1993). The homeobox gene goosecord and the origin of organizer cells in the early chick blastoderm. Cell 74,64=59. Jollie, W.P. (1990). Development, morphology, and function of the yolk sac piacenta of laboratory rodents. Teratol. 41,361-381. Jones,’C.M., Lyons, K.M., & Hogan, B.L.M. (1991). Involvement of bone morphogenetic protein-4 (BMP-4) and Vgr-l in morphogenesis and neurogenesis in the mouse. Development 111,53 1-542. Khaner, 0. (1993). Axis determination in the avian embryo. Curr. Top. Dev. Biol. 28, 155-1 80. Khaner, O., & Eyal-Giladi, H. (1989). The chick’s marginal zone and primitive streak formation. 1. Coordinate effect of induction and inhibition. Dev. Biol. 134,206-2 14. Kimelman, D., & Christian, J.L. (1992). Synergistic principles of development: Overlapping patterning systems inXeriopcrs mesoderm induction. Development 116,l-9. King, B.F. (1982). The role of coated vesicles in selective transfer across yolk sac epithelium. J. Ultrastr. Res. 79,273-284. Kingsley, D.M. ( 1994). The TGF-p superfamily: New members, new receptors, and new genetic tests of function in different organisms. Genes Dev. 8, 133-146. Kinoshita, K., Bessho, T., & Asashima, M. (1993). Competence prepattern in the animal hemisphere of the 8-cell Xenopus embryo. Dev. Biol. 160,27&284. Kochav, S., & Eyal-Giladi, H. ( 197 1). Bilateral symmetry in chickembryo, determination by gravity. Science 171, 1027-1029. Kondo, M., Tashiro, K., Fujii, G., Asano, M., Miyoshi, R., Yamada, R., Muramatsu, M., & Shiokawa, K. (1991). Activin receptor inRNA is expressed early in Xenopus embryogenesis and the level of expression affects body axis formation. Biochem. Biophys. Res. Comm. 181,684690.
186
ROSEMARY F. BACHVAROVA
Ku, M., & Melton, D.A. (1993). Xwnt-11: A maternally expressed Xenopus wnt gene. Development 119, 1161-1173. Lawson, K.A., & Pedersen, R.A. (1987). Cell fate, morphogenetic movement and population kinetics of embryonic endoderm at the time of germ layer formation in the mouse. Development 101,627-652. Lawson, K.A., Meneses, J.J., & Pedersen, R.A. (1991). Clonal analysis ofepiblast fate during germ layer formation in the mouse embryo. Development 113, 89 1-9 I 1. Lee, J.J., Radice, G., Perkins, C.P., & Constantini, F. (1992). Identification and characterization of a novel, evolutionarily conserved gene disrupted by the murine Hp58 embryonic lethal transgene insertion. Development I1 5,277-288. Lemaire, P., & Gurdon, J.B. (1994). A role for cytoplasmic determinants in mesoderm patterning: Cell-autonomous activation of the goosecoid and Xwnt-8 genes along the dorsoventral axis of early Xenopus embryos. Development 120, 1191-1 199. Lettorio, J.J., Geiser, A.G., Kulkarni, A.B., Roche, N.S., Sporn, M.B., & Roberts, A.B. (1994). Maternal rescue of transforming growth factor-pl null mice. Science 264, 1 9 3 6 1938. Lyons, K.M., Pelton, R.W., & Hogan, B.L.M. (1990). Organogenesis and pattern formation in the mouse: RNA distribution patterns suggest a role for bone morphogenetic protein-2a (BMP-2a). Development IOY, 833444. Manova, K., Paynton, B.V., & Bachvarova, R.F. (1992). Expression of activins and TGFPl and TGF f32 RNAs in early postimplantation mouse embryos and uterine decidua. Mech. Dev. 36, 141-152. Manova, K., De Leon, V., Angeles, M., Kalantry, S., Giarre, M., Attisano, L., Wrana, J., & Bachvarova, R.F. (1995). mRNAs for activin receptors I1 and IIB are expressed in mouse oocytes and in the epiblast of pregastrula and gastrula stage mouse embryos. Mech. Dev. 4Y,3-11 Mitrani, E., Gruenbaum, Y., Shohat, H., & Ziv, T. (1990a). Fibroblast growth factor during mesoderm induction in the early chick embryo. Development 109,387-393. Mitrani, E., Ziv, T., Thomsen, G., Shimoni, Y., Melton, D.A., & Bril, A. ( 1990b). Activin can induce the formation of axial structures and is expressed in the hypoblast of the chick. Cell 63,495-50 1. Monaghan, A.P., Kaestner, K.H., Grau, E., & Schutz, G. (1993). Postimplantation expression patterns indicate a role for the mouse forkhead/HNF-3 a, pi and y genes in determination of the definitive endoderm, chordamesoderm and neurectoderm. Development 119, 567-578. Nelsen, O.E. (1953). Comparative Emhyology qfthe Ver-tehr-ate.?.Blakiston Co., New York. Niehrs, C., Keller, R., Cho, K.W.Y., & De Robertis, E.M. (1993). The homeobox gene goosecoid controls cell migration in Xenopus embryos. Cell 72,49 1-503. Nieuwkoop, P.D. (1973). The “organization center” of the amphibian embryo: Its origin, spatial organization, and morphogenetic action.” Adv. Morph. 10, 1-39. Nieuwkoop, P.D., & Sutasurya, L.A. (1979). Primordial Germ Cells in Chordates. Cambridge University Press, New York. Niswander, L., & Martin, G.R. (1992). FGF-4 expression during gastrulation, myogenesis, limb and tooth development in the mouse. Development 114, 755-768.
Inducing Factors in the Pregastrula Mouse Embryo
187
Radice, G., Lee, J.J., & Costsantini, F. (1991). Hp58,an insertional mutation affecting early postimplantation development of the mouse embryo. Development 111, 801811. Rebagliati, M.R., & Dawid, I.B. ( 1993). Expression of activin transcripts in follicle cells and oocytes ofXeriopus 1aevi.s. Dev. Biol. 159, 574-580. Robertson, E.J., Conlon, EL., Barth, K.S., Costantini, F., & Lee, J.J. (1992). Use of embryonic stem cells to study mutations affecting postimplantation development in the mouse. In: Postimplantation Development in the Mouse. CIBA Foundation Symp. 165, pp. 237-250. Ruiz-i-Altaba, A,, Prezioso, V.R., Darnell, J.E., & Jessell, T.M. (1993). Sequential expression of HNF-3P and HNF-3a by embryonic organizing centers: The dorsal liphode, notochord, and floor plate. Mech. Dev. 44,91-108. Sasaki, H., & Hogan, B.L.M. ( 1 993). Differential expression of multiple fork head related genes during gastrulation and axial pattern formation in the mouse embryo. Development 118,47-59. Schoenwolf, G.C., & Watterson, R.L. ( 1 989). Labomtory Studies ofchick, Pig,and Frog Embryos. Macrnillan, New York. Schulte-Merker, S., Smith, J.C., & Dale, L. (1994). Effects of truncated activin and FGF receptors and of follistatin on the inducing activities of BVgl and activin: Does activin play a role in mesoderm induction? EMBO .I.13,3533-3541. Slack, J.M.W. (1991). From Egg to Embryo. Cambridge University Press, New York. Slack, J.M.W. (1994). Inducing factors inXenopus early embryos. Curr. Biol. 4,11&126. Slager, H.G., Lawson, K.A., van den Eijnden-van Raaij, A.J.M., De Laat, S.W., & Mummery, C.L. (1991). Differential localization of TGF-P2 in mouse preirnplantation and early postimplantation development. Dev. Biol. 145,205-2 18. Smith; L.J. (1980). Embryonic axis orientation in the mouse and its correlation with blastocyst relationships to the uterus. I. Relationships between 82 hr and 4 114 days. J. Embryol. Exp. Morph. 55,257-277. Smith, L.J. (1985). Embryonic axis orientation in the mouse and its correlation with blastocyst relationships to the uterus. 11. Relationships from 4I/4 to 9l/2 days. J. Embryol. Exp. Morph. 89, 15-35. Snow, M.H.L. ( 198I). Autonomous development of parts isolated from primitive-streakstage mouse embryos. Is development clonal? J. Embryol. Exp. Morphl. 65 (suppl.), 269-287. Sokol, S., & Melton, D.A. (1991). Pre-existent pattern in Xenopzrs animal pole cells revealed by induction with activin. Nature 351,409-4 11. Spemann, H., & Mangold, H. ( 1 924). Induction of embryonic primordia by implantation of organizers from a different species. In: Foiiridation of Experimental Emhiyology (Willier, B.H., & Oppenheimer, J.M., Eds.) pp. 144-184. Hafner Press, New York. Spyropoulos, D.D., & Capecchi, M.R. ( 1 994). Targeted disruption of the even-skipped gene, evxl, causes early postimplantation lethality of the mouse conceptus. Genes and Dev. 8, 1949-1 96 1. Stem, C.D. ( I 990). The marginal zone and its contribution to the hypoblast and primitive streak of teh chick embryo. Development 109,667-682. Stem, C.D. (1992). Mesoderm induction and development of the embryonic axis in amniotes. Trends Genet. 8, 158-1 63.
188
ROSEMARY F. BACHVAROVA
Stem, C.D., Hatada, Y., Selleck, M.A.J., & Storey, K.G. (1992). Relationships between mesoderm induction and early organogenesis. Development (suppl.), 15 1-156. Tam, P.P.L. (1989). Regionalisation of the mouse embryonic ectoderm: Allocation of prospective ectodermal tissues during gastrulation. Development 107,5547. Tamada, H., McMaster, M.T., Flanders, K.C., Andrews, G.K., & Dey, S.K. (1990). Cell type-specific expression of transforming growth factor+ 1 in the mouse uterus during the periimplantation period. Mol. Endocrinol. 4,965-972. Tamada, H., Das, S.K., Andrews, G.K., & Dey, S.K. (1991). Cell-type-specific expression of transforming growth factor-a in the mouse uterus during the peri-implantation period. Biol. Reprod. 45,365-372. Thomsen, G.H., & Melton, D.A. (1993). Processed Vgl protein is an axial mesoderm inducer in Xenopus. Cell 74,43344 1. Thomsen, G., Woolf, T., Whitman, M., Sokol, S., Vaughan, J., Vale, W., & Melton, D.A. (1990). Activins are expressed early inXenoptrs embryogenesis and can induce axial mesoderm and anterior structures. Cell 63, 485-493. Vakaet, L. ( 1970). Cinephotomicrographic investigations of gastrulation in the chick blastoderm. Arch. Biol. 81,387426. Vakaet, L. ( I 984). Early development of birds. In: Chimeras in Developniental Biology (LeDouarin, N., & McLaren, A,, Eds.) pp. 7 1-88. Academic Press, New York. Vakaet, L. (1985). Morphogenetic movements and fate maps in the avian blastoderm. In: Molecular Determinants qj’Animal Form (Edelnian, G.M., Ed.) pp. 10&109. A.R. Liss, New York. van den Eijnden-van Raaij, A.J.M., van Achterberg, T.A.E., van der Kruijssen, C.M.M., Piersma, A.H., Huylebroeck, D., De Laat, S.W., & Mummery, C.L. ( 1991). Differentiation of aggregated murine PI 9 embryonal carcinoma cells is induced by a novel visceral endoderm-specific FGF-like factor and inhibited by activin. Mech. Dev. 33, 157-166. Vassalli, A., Matzuk, M.M., Gardner, H.A.R., Lee, K-F., & Jaenisch, R. (1994). Activinhnhibin PB subunit gene disruption leads to defects in eyelid development and female reproduction. Genes Dev. 8,414-427. Vidricaire, G., Jardina, K., & McBurney, M.W. ( 1 994). Expression ofthe Brachyuygene during mesoderm development in differentiating embryonal carcinoma cell cultures. Development 120, 1 15-122. Waddington, C.H. (1932). Experiments on the development of chick and duck embryos, cultivated in vitr-0.Phil. Trans. Roy. SOC.London B 221, 179-230. Waddinton, C.H. ( 1 933). Induction by endodenn in birds. Wilhelm Roux Arch. Entw. Mech. Org. 128,502-521. Waddington, C.H. (1956). Principles qf’Emhtyology. The Macmillan Co., New York. Weinstein, D.C., Ruizi Altaba, A.,Chen, W.S., Hoodless, P., Prezioso, V.R., Jessell,T.M., & Darnell, J.E. ( 1994). The winged-helix transcription factor HNF-3P is required for notochord development in the mouse embryo. Cell 78,575-588. Wilkinson, D.G., Peters, G., Dickson, C., & McMahon, A.P. (1988). Expression of the FGF-related proto-oncogene int-2 during gastrulation and neurulation in the mouse. EMBO J. 7,691-695. Wittbrodt, J., & Rosa, F.M. (1994). Disruption of mesoderm and axis formation in fish by ectopic expression of activin variants: The role of maternal activin. Genes Dev. 8, 1448-1462.
inducing Factors in the Pregastrula Mouse Embryo
189
Zhou, X., Sasaki, H., Lowe, L., Hogan, B.L.M.,& Kuehn. M.R. ( 1993). Nodal is a novel TGF-P-like gene expressed in the mouse node during gastrulation. Nature 362, 543-547. Ziv, T., Shimoni, Y.,& Mitrani, E. (1992). Activin can generate ectopic axial structures in chick blastoderm explants. Development 21 5, 689494.
This Page Intentionally Left Blank
HUMAN Y CHROMOSOME FUNCTION IN MALE GERM CELL DEVELOPMENT
Peter H. Vogt
I. Introduction . . . . . . . . . . . . . . , . . . . . . . . . . . . . 11. . Human Gametogenesis . . . . . . . . . . . . . . . . . . . . . . A. Male Germ Cell Development During Embryogenesis . . . B. Male Germ Cell Development Before Puberty . . . . . . . C. Male Germ Cell Development During and After Puberty . . 111. Y-Mutations Interfering with Male Fertility . . . . . . . . . . . . A. Y-Mutations in Yp Interfering with Male Fertility . . . . . . B. Y-Mutations in Yq Interfering with Male Fertility . . . . . IV. Molecular Deletion Mapping of the Azoospermia Factor (AZF) in Yq 1 1 . . . . . . . . . . . . . . . . V. AZF Candidate Genes Expressed In Testis . . . . . . . . . . . . A. The RBMGene Family . . . . . . . . . . . . . . . . . . . . B. The SMCY Gene . . . . . . . . . . . . . . . . . . . . . . . C. The SPGY Locus . . . . . . . . . . . . . . . . . . . . . . . VI. AZF and Y Chromosome Structure in Male Germ Line . . . . .
Advances in Developmental Biology Volume 4, pages 191-257. Copyright 0 1996 by JAI Press h e . All rights of reproduction in any form reserved. ISBN: 1-55938-969-9
191
192 194 197 199 202 . 2 11 212 214 222 229 229 232 234 236
PETER H. VOGT
192
VII. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
241 245 246
1. INTRODUCTION Most important gene activities during embryonic development are those which act as switch signals at specific developmental stages. They initiate specific cascades of molecular events downsrream up to the next switch signal and so forth until the specialized fate of each cell is found and manifested. Switch signal(s) determining the fate of the germ cell to become a male or female germ cell are active during early embryogenesis. After determination, switch signals during the differentiation process of the male and female germ cells seemed to be different. Meiosis is induced in embryonic female germ cells, but repressed in embryonic male germ cells (McLaren, 1994). If one assumes that gene activities acting as switch signals are expressed at the most critical steps during any cell differentiation pathway, they should be expected during male germ cell development (1) at the onset of spermatogonia proliferation and differentiation before and at puberty, (2) at the onset of meiosis, and (3) at the onset of spermatid development starting with the formation of the acrosome and the flagellum as well as the condensation of the nucleus. It was analyzed in different species that there is an initial period during early development during which a stem cell population proliferates before it splits into the somatic and germ cell lineages. In the mouse, this corresponds to 7.2 days after fertilization; in Volvox, the sixth cell division is the decision point. Cell size seemed to be important for cell fate (for a review see McLaren, 1991, 1994; Ciba Foundation Symposium, 1994). It is expected that this decision point is also controlled by a genetic switch signal. Genes for male germ cell development are conserved on the Y chromosome of the Drosophila, mouse, and human. For the Drosophila and mouse, the Y spermatogenesis function was discussed recently in different excellent reviews (Hennig, 1989, 1993; Bonaccorsi and Lohe, 1991; Burgoyne, 1991). Therefore, this review concentrates on the function of the human Y chromosome during spermatogenesis. First, arguments that the human Y chromosome must also include genetic factor(s) important for male germ cell differentiation were drawn
Human Y Chromosome and Spermatogenesis
193
from the observation of terminally deleted Y chromosomes in the karyotype of six sterile males with azoospermia (Tiepolo and Zuffardi, 1976). In all cases, the deletion included the large heterochromatin block in the long Y arm (Yql2) and an undefined amount of the adjacent euchromatic part (Yql 1). Since at that time, nogody believed in any genetic activity of heterochromatin, it was postulated that Y factor(s) important for male germ cell development should be located in Yq 1 I . They were defined as AZoospermia Factor (AZF) as their deletion correlates to the sterile male phenotype “azoospermia,” which means that no mature sperm cells were detectable in the patients’ semen fluid. Since then, the presence of AZF in Yqll was confirmed by numerous studies at the cytogenetic level (Sandberg, 1985)and the molecular level (Anderson et a]., 1988; Bardoni et al., 1991; Vogt, 1992; Reijo et al., 1995). However, the genetic complexity of AZF could not be revealed by these analyses. How many spermatogenesis genes in Yql1 are present and which of them cause azoospermia if disrupted? A rough estimation of the DNA length in Yq 1 1 ranges between 7-1 0 million nucleotides. This is large enough to accommodate more than one spermatogenesis gene. After the detection of different interstitial microdeletions in Yq 1 1 occurring de novo in azoospermic patients with a cytogenetically normal Y chromosome (Vogt et al., 1992) and mapping the extension of those deletions in a detailed Yql 1 subinterval map (Ma et al., 1992) the first evidence was gained that indeed more than one spermatogenesis gene may be present in Yql 1 causing azoospermia if deleted. This view got further support by an extensive analysis of the histology in testis tissue sections in sterile males with different cytogenetically visible Yql 1 anomalies (Vogt et al., 1993), and then from the results of an extensive molecular screening program on Yql 1-microdeletions in more than 300 sterile males with idiopathic azoospermia and severe oligospermia (Vogt eta]., 1996). From these analyses it is now evident that at least three AZF loci (AZFu, AZFb, and AZFc) are present in Yql 1. Every Y gene expressed in human testis tissue and located in Yql 1 are potential AZF candidate genes. Three were recently published: the RBMgene family (formerly YRRM, Ma et al., 1993), the SCMY gene (Agulnik et al., 1994a), and the homologous DAZ/SPGYZ genes (DAZ: Reijo et al., 1995; SPGYZ: Maiwald et al., 1996) here defined as the SPGY locus. Three Y genes, SRY (Sinclair et al., 1990), ZFY(Page et al., 1987), and TSPY (Arnemann et al., 1991; Zhang et al., 1992), also expressed in human testis, were not located in Yql 1 , but to different
PETER H. VOGT
194
subintervals on the short arm of the human Y chromosome. However, no spermatogenesis locus has yet been mapped to Yp and since it is known that numerous genes transcribed in testis do not have any biological function (Willison and Ashworth, 1987) a candidate status of SRY, ZFI: and TSPY as spermatogenesis genes in Yp cannot be given, but also cannot be excluded. This review focuses on the germ cell function of the human Y chromosome located in Yq I 1 now designated as AZFa, AZFb, and AZFc. It aims to summarize its present state of research and to discuss ideas of possible AZF functional sites in the germ line. Since strong evidence exists that, besides the expression of protein coding AZF genes, the intactness of the Y chromosome structure in Yq 1 1 may also be important during human male germ cell development, its dynamic behavior in the male germ line will be described in a separate section and discussed as an additional phenotype of AZF. A discussion of potential switch signals along the human germ cell differentiation pathway, in which the expression of AZF genes may be involved, is not possible without a profound knowledge of human gametogenesis itself. Therefore, I felt it necessary to include in this review a section describing an actual overview on the current knowledge of human gametogenesis. Humans are not experimental species. Therefore, most results analyzing the germ cell function of AZF in Yql 1 were collected from sterile patients consulting infertility clinics. From the scientific point of view my experience with this research area was therefore often discouraging since it was rather difficult and time consuming to get testis tissue samples from every sterile male with an identified AZF mutation. And quite often, important experiments like the analysis of meiotic chromosomes or preparing of testis tissue sections in special fixatives were omitted or not available. From the clinical point of view my experience was just the opposite. Sterile males screened successfully for an AZF mutation in Yql 1 were told the reason for their sterility. This helped them a lot in coping with their sterility and saved them from further long and often useless medical treatments.
11. HUMAN GAMETOGENESIS The human species is not an experimental system. Consequently, knowledge of human gametogenesis can usually be collected only from testis tissue samples, which are accessible for other medical reasons, like
Human Y Chromosome and Spermatogenesis
195
spontaneous or induced abortion events, surgical gonadal corrections during childhood, or surgical treatments of testis tumors or obstructive spermatogenesis in the adult. Studying these tissues for analysis of gametogenesis would be, however, only meaningful if one were able to recognize and subtract any destruciive effects on the patient’s gametogenesis caused by his disease. But this is obviously only possible after the biological pathway of human gametogenesis is known in detail, The summary looks like a vicious circle, and one possibility to escape--used most o f t e e i s to analyze testis tissue samples and gametogenesis in an adequate animal as a model system. This has been done extensively in Drosophila (Hennig, 1988) and different mammals like mouse, rat, pig, and bull (Guraya, 1987; Hilscher and Hilscher, 1989; Ciba Foundation Symposium, 1994). But, doing these experiments, it is important to keep in mind that any extrapolation of the results obtained in one species to another species, even if closely related, may be a hazardous procedure and may not be justified (Baker, 1978). Human spermatogenesis was long thought to be highly disorganized in comparison to other mammals (Roosen-Runge and Barlow, 1953) and even healthy and fertile men ejaculate a high percentage of spermatozoa with abnormal morphology and motility. First Clermont (1 963) then demonstrated that human spermatogenesis is not a disorganized process, but that six stages (I-VI) could be clearly distinguished by different nuclear morphologies of germ cells and their topographical arrangements in testis tubuli sections. During proliferation and differentiation of the germ cells it always takes about 16 days from one germ cell layer to the next. This is called one cycle (Heller and Clermont, 1964). Since four cycles are needed until the development of a germ cell to a spermatozoon the time of the total process of spermatogenesis could be estimated to be about 64 days. The different stages of Clermont were later recognized to follow a helical pathway (Hilscher, 1979) with a human specific tubular geometry (Schulze, 1982) called human wave of spermatogenesis. It could best be visualized by marking the arrangement of the different types of spermatocytes approaching the lumen of the tubules with advancing development (Figure I). This picture now provides the base for a rational understanding of the arrangement of the human germinal epithelium. Its description during the embryonic, prepuberal, and postpuberal phase of gametogenesis is given in the following three sections (IIAIIC).
Figure 7. Cross section of a human seminiferous tubule, kindly provided by Professor W. Schulze (Departmentof Andrology, University of Hamburg). The arrangement of the primary spermatocytes are explained by a section through a model at the level of the dotted arrow. LEIlate leptotene/early zygotene; PI mid-pachytene; P2 late-pachytene.
Human Y Chromosome and Spermatogenesis
197
A. Male Germ Cell Development During Embryogenesis
Human primordial germ cells (PGC) are derived from the epiblast and first observed in the yolk sac close to the hindgut and allantois region by day 2 1 up to day 26 post conception‘(p.c.; Politzer, 1933; Witschi, 1948). During the fifth week of pregnancy they migrate to the genital ridges, which then differentiate into an ovary or testis between day 36 and day 42 p.c. Migration of PGC is actively stimulated in an ameboid manner, of which the mechanism is unknown (Falin, 1969). PGC are found as isolated germ cells in contrast to their descendants-ogonia and M prospermatogonia-which are arranged in pairs and clusters and connected by intercellular bridges. M prospermatogonia were multiplying from the PGC, which can be observed in cross sections of the sex cords by an increase of germ cells from about 22 in the eighth week of pregnancy up to 70 in the twenty-second week (Hilscher and Engemann, 1992). Thereafter, the number of germ cells decreased clearly in contrast to the pre-Sertoli cells. M prospermatogonia are replaced by T prospermatogonia, which are sometimes also called fetal spermatogonia (Wartenberg et al., 1971). They are observed in the cords up to six months after birth. T prospermatogonia complete the period of prespermatogenesis and initiate spermatogenesis. Initially, in rodent experiments, Hilscher ( I98 1) supposed that T prospermatogonia were the first exponents of the male pathway of the germ line. Recently, however, it was shown that the first differences between male and female germ cells can be identified earlier, at the end of the first proliferation wave of oogonia and M prospermatogonia (Hilscher and Hilscher, 1990). The G2 phase of M prospermatogonia is about one hour shorter than that of oogonia. This indicates that at least in the last generation of M prospermatogonia late replicating chromatin domains are present in their cell nucleus, which do not exist in oogonia. During this proliferation wave one X chromosome is inactivated in oogonia. In M prospermatogonia the state of the X chromosome is not known. The first cell lineage showing a sex-specific differentiation is the “supporting cell lineage” (McLaren, 1991). It surrounds and supports the primordial germ cells, giving rise to pre-Sertoli cells in the male and follicle cells in the female. Pre-Sertoli cells are different from mature Sertoli cells in adults because of their capacity to proliferate and their lack of tight junctions.
198
PETER H. VOGT
In response to a cascade of genetic activity initiated by the gene product(s) of SRY (the sex-determining region gene on the Y chromosome; Koopman et al., 1991) pre-Sertoli cells aggregate to form seminiferous cords, containing the prospermatogonia. From studies in the mouse, experimental evidence is presented that the supporting cell lineage will develop toward Sertoli cells only in the presence of the Y chromosome (Burgoyne et al., 1988). Expression of SRYin this cell type is therefore assumed to be an important switch signal toward the development of testis. Pre-Sertoli cells also influence germ cells. They block the proliferation of T prospermatogonia and inhibit their meiosis (McLaren, 1994; Pelleniemi et al., 1993). Dark and light pre-Sertoli cells were observed in fetal testis by Wartenberg (1978). It was argued that one Sertoli cell type exercises a stimulating effect on germ cell proliferation and the other type an inhibitory effect, both by means of immediate cellular contacts and direct substantial influence (Wartenberg, 1978). Meiotic inhibition may be induced by a different chromosome behavior in the male and female germ line at this stage. A chromosomal condensation period followed by decondensation was observed only in fetal human oocytes. In prospermatogonia a long condensation period is observed (Luciani et a]., 1977). Aggregation of fetal Sertoli cells into seminiferous cords and secretion of anti-Mullerian hormone (AMH) gives first evidence of their role in testicular organogenesis (Tran et al., 1987). Normal formation oftestis tubules can take place in the absence of germ cells, but not in the absence of Sertoli cells. Basement membrane formation during tubulogenesis is a complex process, in which Sertoli cells and peritubular myoid cells contribute cooperatively (Fritz and Tung, 1986). Other cell types in the testis, such as the cells of the tunica albuginea and the Leydig cells that produce testosterone differentiate later and may indeed be dependent upon prior Sertoli cell differentiation. The development of fetal Leydig cells reaches its maximum between the twelfth and fourteenth week of pregnancy. Their secretion of testosterone at that time interval is important for development of the male genitale. In summary, one may conclude that the presence of pre-Sertoli cells induces prespermatogenesis and the differentiation process of the Leydig cells. It has been shown that fetal Leydig cells in human are highly differentiated (Holstein et al., 1971).
Human Y Chromosomeand Spermatogenesis
199
B. Male Germ Cell Development Before Puberty Immediately after birth, the seminiferous tubules are composed by T prospermatogonia, spermatogonia, and Sertoli cells. Prospermatogonia are located mainly in the center of the tubules with a tendency to move toward the basement membrane. Spermatogonia are of the so-called A p and Ad type. A p spermatogonia display an excentrically located nucleus and homogenously distributed chromatin. A d spermatogonia can be distinguished by a characteristically lighter zone generally located in the middle of the nucleus (Hadziselimovic, 1977). The most common cell in the first year is the Sertoli cell type Sa (Figure 2), an oval or polarized cell, which is always in contact with the basement membrane (HadziseIimovic, 1983). The volume of Ap spermatogonia is described to be two to four times bigger than the Sa-type Sertoli cell. The wall of the seminiferous tubules is composed by one layer of a basement membrane, a collagen fiber zone and fibroblasts. The interstitium contains mainly neonatal Leydig cells. Secretion of testosterone at that time coincides with manifestation of the psychological part of the male sex and the development of A spermatogonia (Bidlingmaier and Hilscher, 1989). Leydig cell activity then declines until puberty. At the age of four years the ultrastructural appearance of the seminiferous tubules has changed. Prospermatogonia are no longer visible. In addition.to the A-type spermatogonia, B-type spermatogonia and primary spermatocytes are encountered for the first time (Hadziselimovic, 1983). B spermatogonia are smaller and rounder than the A-type from which they develop. Primary spermatocytes appear in only in 5-25% of the seminiferous tubules (Nistal and Paniagua, 1984). They do not progress to spermatozoa although occasionally some spermatids were seen. An explanation for the premature presence of spermatocytes was given by the observation of a marked germ cell proliferation at the same time intervals: at the end of infancy (about 3 4 years of age), and at 8-9 years of age (Mancini et al., 1960). This has been interpreted as two premature attempts of spermatogenesis, which became somewhat more successfid in some tubuli giving rise to spermatocytes or even spermatids but never complete spermatogenesis. Only at age 13 do primary spermatocytes occur with increasing numbers. Prepuberal spermatocytes have no contacts with the basement membrane but form intercellular bridges and are surrounded by Sertoli cells. The centrally located nucleus has a round shape and 1-3 nucleoli, indicating genetic activity. Its chromatin is dispersed irregularly. Nuclear granules and synaptonemal complexes
200
PETER H. VOGT
Figure 2. Different types of Sertoli cells. The Sf fetal Sertoli cell i s transformed after birth into the Sa type cell, which give rise to the Sb type cell. In puberty, Sa and Sb cells transform into the Sc type due to increased gonadotropin and testosterone stimulation (Hadziselimovic, 19.83). look similar to those in adult primary spermatocytes (Hadziselimovic, 1977). The nuclear membrane is in contact with the endoplasmic reticulum. The Sertoli cells have completed their transformation from fetal cells into the Sa- and Sb-type cells (Figure 2). Simultaneously with the appearance of B spermatogonia and primary spermatocytes, Sb-type Sertoli cells are found in increasing numbers (Figure 3). As in the fetal period different Settoli cell types may exercise a regulating effect on germ cell proliferation and their meiotic activity by means of immediate cellular contacts and direct substantial influence (Wartenberg, 1989).
Human Y Chromosome and Spermatogenesis
201
figure 3. Section of a seminiferous tubule of a five-year-old boy showing primary spermatocytes (Sp) and Sertoli cells of the Sb type. Occurrence of increasing numbers of Sb Sertoli cells is observed to be connected to premature attempts of spermatogenesis before puberty (Hadziselimovic, 1983).
The isolation of mitogenic factor(s) from prepuberal rat Sertoli cells stimulating germ cell proliferation in v i m seems to support this view (Kancheva et al., 1990). The number of Sertoli cells per tubulus crosssection decreased continuously in normal testicles from the first year until puberty (Hadziselimovic, 1977). Leydig cells before puberty are found in groups of 2-6 cells in the interstitium. The chromatin is homogenously distributed while the heterochromatin is found on the nuclear periphery. The well developed nucleolus is attached to the nuclear membrane. They are smaller than the neonatal Leydig cells (Hadziselimovic, 1977).
202
PETER H. VOGT
C. Male Germ Cell Development During and After Puberty The differentiation process of A type spermatogonia to B spermatogonia, spermatocytes, spermatids, and mature spermatozoa is accompanied by a large increase in testicular size and lasts 64 days (Heller and Clermont, 1963). It is still an open question by which the spermatogonia stem cell population (As) is balanced (spermatogoniogenes~s).Obviously, the diminishing population of Ad spermatogonia-due to their differentiation process-must be refilled. Parallel mitotic activities of Ad spermatogonia or Ap spermatogonia dividing into Ad and B spermatogonia can be discussed. The primary stem cell population would then be derived from Ad or Ap spermatogonia or both. It has been estimated that in adult testis one spermatogonia stem cell per 100-1,000 differentiating spermatogonia is preserved (Clermont, 1966; Clermont and Hermo, 1976). The mechanism of keeping them quiescent is not known. The main morphological features used in distinguishing the different spermatogonia types at the light microscope level are the shape and staining characteristics of the nucleus, the placement of the nucleolus, and the presence or absence of glycogen in the cytoplasm (stained by periodic acid Schiff [PAS] reaction). So Ad spermatogonia have a dark karyoplasm with a typical pale hallow structure in the middle and Ap spermatogonia are only stained weakly (pale). B spermatogonia are clearly distinguishable from both A type spermatogonia. In young adults a third A type spermatogonia was described first by Rowley et al. (1971) as A long (AL).A L spermatogonia are very flat with the largest contact to the basal lamina. In the electron microscope, Schulze (1 978) additionally observed a so-called A cloudy type of spermatogonia, recognized later as DNA synthesizing spermatogonia. The most comprehensive analysis of spermatogonias in adult testis was done by Hilscher (1981) studying a complete series of 250 semithin sections for the following parameters: shape of the nucleus, arrangements of nucleoli, retention of dyes in the karyoplasm and cytoplasm, cytoplasmatic structure, form of attachment to the basement membrane, and cluster sizes of the same and different spermatogonia types. In that way she was able to distinguish 13 phenotypes of spermatogonia (Figure 4). Pale spermatogonia with light, absolutely homogenous, karyoplasm with or without a large contact to the basal membrane (Ap or A L ) were found to be the only cell types grouped in only small clusters. The size of the clusters increased with the degree of the spermatogonia differentiation process. From analysis of the spermatogonial population in different males with oli-
Human Y Chromosome and Spermatogenesis
203
Huiiian Spermatogoniogenesis
figure 4. Distribution frequency of the sizes of clusters of different germ cells in the basal compartment in man; content of a tubule with length of about 1,000 um (Hilscher, 1981).At the right an example of the histological appearance of each cell type is given (courtesy of B. Hilscher).
gospermia or after treatment with antiandrogens, it is suggested that spermatogoniogenesis in adults uses mainly A p spermatogonia as stem cells. This view got support by the observation that Ap spermatogonia can be present alone in testis tubules, but that Ad spermatogonia are only
204
PETER H . VOGT
observed in combination with Ap spermatogonia (Schulze, 1978).Therefore, spermatogoniogenesis is most likely to start with stem cells derived from spermatogonia of the pale Ap or AL type. Germ cells involved in meiosis are the primary and secondary spermatocytes. B spermatogonia change their morphology by starting DNA replication which doubles their DNA content and increases the size of the nucleus. Long chromosome threads became visible (leptotenephase) followed by pairing of homologous chromosomes, frequently from the chromosome ends (zygotene phase). During this premeiotic phase, the sex chromosomes also start to pair at the tip oftheir short arms. The exact molecular mechanism of chromosome pairing is not yet known. The paired homologues are connected by the so-called synaptonemal complex. After completion of pairing, the chromosomes begin to condense (pachytene) and the complex of the X and Y chromosome becomes visible as a separate structure called sex vesicle (Figure 5). It is interesting to note that between X-Y complex and chromosome arms, which carry rDNA genes and express the structure of the nucleolus, a distinct association is repeatedly observed (Figure 6). Since most human Robertsonian translocation chromosomes and sex/autosome translocations involve the short arms of nucleolar chromosomes, it is assumed that breakage and fusion events between them preferably take place during this germ cell stage (Berrios and Fernandez-Donoso, 1990). If this holds true, the decondensed chromatin domains observed at this germ cell stage for the X and the Y chromosome may provoke these rearrangements (Speed et al., 1993; Armstrong et al., 1994; see also Section VIII). Decondensed chromatin domains at this stage must also be present in the ribosomal RNA gene loci of the acrocentric chromosomes. At leptotene, new nucleoli were formed increasing in size at zygotene due to their high transcription rate (Stahl et al., 1991). When a longitudinal cleft becomes visible in each chromosome pair, four chromatids of each kind are seen side by side (diplotene). At this phase, chiasmata between non-sister chromatids also become visible. After first meiotic division, a second division cycle of the secondary spermatocyte nucleus leads to a reduction of the diploid chromosome number preserving each germ cell with a haploid chromosome number and an X or a Y chromosome. The cytoplasm of the primary spermatocytes is more electron-dense than that of spermatogonia and contains evenly scattered polysomes and ribosomes reflecting their high genetic activity.
Human Y Chromosome and Spermatogenesis
205
Figure 5. Human chromosomes at late pachytene. The X-Y pairing complex is distinctly visible as an extra chromatin structure (arrow).
Postmeiotic germ cells with a haploid chromosome number are called spermatids. They first look morphologically similar to the secondary spermatocytes but have a smaller nuclear diameter. Their developmental differentiation process to motile spermatozoa lasts 21 days and includes three major events: formation of the acrosome, development of the flagellum, and condensation of the nucleus. Due to the presence of different morphological features observed in the light microscope, six different cell types could be distinguished (Clermont, 1963) They are shown with the resolution of the electron microscope in Figure 7. With emphasis on the development of the so-called mid-piece in the flagellum, Holstein observed eight different cell stages in the electron microscope (Holstein, 1976). Their morphological details and staging aspects are
206
PETER H. V O C T
Figure 6. A partial electron microscope series of an early-pachytene nucleus showing an association between the XY bivalent and two nucleolar bivalents. Bar indicates 1 pm. Tp nucleolar telomere, XY sex bivalent, f-Nu fibrillar zone, FC fibrillar center, CTR centrioles, arrows in h point to attachment of the sex chromosome axes to the nuclear envelope (for a detailed description see Berrios and Fernandez-Donoso, 1990).
207 -7-
-7-
s
W
L
0
W
0
Figure 7. Electron microscopic features of the six stages of human spermiogenesis according to the definition of Clermont (1963).
208
PETER H.VOGT
described extensively elsewhere (de Kretser, 1969; Holstein, 1976; Bustos-Obregon et al., 1975; Holstein and Roosen-Runge, 1981; de Kretser and Kerr, 1988). Condensation of the nuclear shape, volume, and position is associated with condensation of its chromatin configuration. The acrosome is produced by the Golgi complex of the young spermatid in the form of a series of granules which aggregate within the acrosomal vesicle. The vesicle is closely applied to the nucleus. At the same stage, the formation of the axial filament is already visible arising from one pair of centrioles. Just before delivery of the mature spermatid from the germinal epithelium into the tubular lumen, the nucleus reaches its paddle-shaped slightly conical form and is covered by a cap-like acrosome. The karyoplasm is highly condensed. The flagellum consists of a distinct "neck" region, a mitochondria-filled middle piece, and the main principal piece. Sperm motility was achieved by passage through the epididymis (Yanagimachi, 1994). With the onset of puberty the Sertoli cells change to the Sc-type cells (Figure 2) and increase innumber again (Cortes et al., 1987). This change takes place very quickly until all cells are of the Sc-type. The Sc cell is five times bigger than the Sa cell. The nuclear membrane is irregular with one deep invagination and the chromatin is fine and evenly distributed. During puberty changes in the ultrastructure of the Sertoli cells indicates their involvement in the production of steroids. No mature sperm is found until the characteristic ultrastructure of the Sertoli cell has been developed. This morphology could be distinguished in the electron microscope by five different patterns of infranuclear inclusions (Schulze et al., 1976), most likely reflecting different metabolic situations. A correlation to different stages of spermatogenesis was not detectable. Once Sertoli cells have completed their differentiation, they cease dividing and establish a stable population at puberty. It is interesting to note that the Y chromosome in the mouse and human becomes decondensed in Sertoli cells after puberty (Guttenbach et al., 1993; Speed et al., 1993). In the human these Sertoli cells should be of the Sc type. A particularly important feature of postpuberal Sertoli cells is the development of tight junctions between them. This results in the establishment of basal and adluminal compartments within the seminiferous tubule. Tubular lumen opening also occurs at that time. The junctions are considered to be the morphological basis for the blood-testis barrier, which restricts passage of macromolecular materials into the interior of the seminiferous tubules. A and B spermatogonia and DNA synthesizing
Human Y Chromosome and Spermatogenesis
209
spermatocytes @releptotenephase) are located in the basal compartment of the seminiferous tubule. More mature germ cells are located in the adluminal compartment (Fawcett, 1975). The ionic and protein composition of the extracellular milieu of the adluminal compartment is different from that in the basal compartinent. Thus, Sertoli cells are largely responsible for producing an environment suitable for germ cell survival, meiotic divisions, and maturation in the adluminal compartment. Sertoli cells are considered to have a nutrient and holding function for germ cells. Additionally, they can phagocytize degenerating cells and the so-called “residual bodies” occurring during the maturing process of spermatids (Figure 7). They contain all the ribosomes plus remnants of mitochondria and other discarded spermatid organelles accounting for approximately one-third of the cytoplasmic contents in spermatids (Morales and Clermont, 1993; Russell, 1993). Analysis of the different functions of the adult Sertoli cells were mainly done in rats. Some of them seem to be dependent on the presence of certain germ cell stages; some of them are independent of the presence of germ cells. Sertoli cell functions which are dependent on germ cells may need direct germ cell contacts or gap junctions or the secretion of paracrine factors facilitating intercellular communication (McGuiness and Griswold, 1994). This should be particularly important during the maturing of spermatids after their nuclei become condensed repressing their own genetic activity. Since Sertoli cells are associated with a particular combination of germ cells at any particular stage of the germ cell cycle, it can be assumed that Sertoli cell products can be channeled to different germ cell types. Specialized junctional contacts observed between Sertoli cells and the various classes of germ cells support this view (Russell, 1993). Numerous growth factors, proteases and protease inhibitors, and transporter proteins, can be secreted by Sertoli cells. Potential pathways of growth factors to the other cells of the testis are schematically drawn in Figure 8. The transporter protein which is likely to be most important for germ cell divisions and subsequent development is transferrin moving ferric ions from the lymph through the Sertoli cells to spermatocytes and spermatids (Trowbridge and Omary, 1981). Sertoli cell germ cell contacts, which are needed premeiotically at the primary spermatocyte stage, are important for increasing the secretion of the androgen-binding protein (ABP) induced by the follicle stimulating hormone (FSH; Galdieri et al., 1984). Paracrine interactions between the germ cell and Sertoli cell can, however, also be triggered by the germ cell (Fritz, 1994). This has been
PETER H. VOCT
21 0
Sertoli
figure 8. The paracrine and endocrine relationships among Sertoli cells and other testicular cells and the pituitary. Abbreviations: IL-1 , interleukin1 ; ICF, insulin-like growth factor; TGF, transforming growth factor; NGF, nerve growth factor; bFGF, basic fibroblast growth factor; FSH, follicle stimulating hormone, T, testosterone. Peritubular cells are the myoid-like cells surrounding the seminiferous tubules. The arrows depict the proposed direction of the paracrine or endocrine action (McCuiness and Griswold, 1994).
especially shown for spermatocytes (Djakiew and Dym, 1988; d’Agostino and Stefanini, 1990). Differentiation of the adult generation of Leydig cells takes place during the prepubertal period under gonadotropin control. This is, in fact, a redifferentiation process resembling the earlier phase in the fetus but developing now more gradually over a longer time period. It is most likely that they derived from mesenchymal cells in the interstitial regions responding to rising gonadotropin levels by increasing in size and cytoplasmic complexity and developing the capacity for steroid hormone production. The fully differentiated Leydig cells are characterized by round, eccentric nuclei, abundant smooth endoplasmic reticulum, tubulovesicular mitochondria, and the presence of prominent Golgi areas. Small groups of Leydig cells are seen at 1 0 - 1 1 years, with larger aggregates beginning to appear at 12-1 3 years. A marked increase in the responsiveness of Leydig cells to gonadotropins occurs at the time of
Human Y Chromosome and Sperniatogenesis
211
puberty. This correlates with a sharp rise in testosterone secretion and gonadotropin production. Testosterone production is essential for meiotic gametogenesis (Steinberger, 1971). The function of adult Leydig cells is regulated by three hormones from the adenohypophysis: LH (luteinizing hormone), FSH (follidle stimulating hormone), and prolactin. LH induces its biosynthesis of testosterone (de Kretser et al., 1983). The peritubular tissue at puberty becomes a complex structure consisting of several layers of basal lamina, fibroblasts, and collagen fibers. The fibroblastic cells develop elaborate cytoplasmic extensions containing bundles of filaments and numerous pinocytotic vesicles. Its development is under hormonal control. It has been argued that the,first wave of spermatogenesis is rather different from later ones (Fritz, 19941, because, in that first wave, spermatogonia don’t have any spermatocytes next to them and spermatocytes don’t have any spermatids next to them, and so on. Lack of temperature sensitivity and the time required to develop from the last S phase of spermatogonia to the late stage of primary meiosis appear similar to those of oogenesis (Fritz, 1982). In summary, one may conclude that human gametogenesis includes certainly specific differentiation aspects such as the specific morphogenesis of the sperm head, but that the genetic program expressed especially at critical germ cell stages (coding switch signals) may be evolutionarily highly conserved.
Ill. Y-MUTATIONS INTERFERING W I T H MALE FERTILITY Before the era of chromosome banding, it was difficult to detect structural anomalies of the Y chromosome at all. Due to its well-recognized length-polymorphism in normal males, it even seemed impossible to relate phenotypical abnormalities of the male causally to a detected “abnormality” of its Y chromosome. Very often abnormal Y chromosomes were found in combination with a 45,XO cell line. But there exists a great variability in the phenotypes of these XO/XY cases including a different sex. Different types of gonadal dysgenesis including the development of intersexual gonads were observed in human males and females with XO/XY mosaic karyotypes (Sohval, 1964). This also raised the question earlier of whether the observed phenotype variability is related to an anomaly of the Y chromosome at all, and whether only the XO complement causes the phenotypes observed (for review see Pfeiffer et
PETER H. VOGT
21 2
al., 1968). Distortion of spermatogenesis in males with Y chromosome anomalies seemed therefore first to be only a secondary effect of distortion in gonad development. This picture then changed when banding analysis of the human Y chromosome made it possible to distinguish different Y chromosomal rearrangements observed de novo only in males with sterility. They are described in this review for the short (Yp) and long (Yq) chromosome arm, separately. A. Y-Mutations in Yp Interfering with Male Fertility
Cytogenetically visible mutations of the Y chromosome in Yp interfering with male germ cell development always include deletions of at least part of the Yql 1 region. It is therefore impossible at the moment to define a spermatogenesis locus in Yp not including Yql 1 . But a spermatogenesis locus in Yp is likely since different Y genes, SRY, ZFY, and TSPY were isolated, present in Yp, and expressed in human testis. Most attractive as a candidate gene for a spermatogenesis locus in Yp is TSPY since it is expressed only in testis (Arnemann et al., 1991; Zhang et al., 1992). It is conserved on the Y chromosome during evolution between homo and bos (Jakubiczka et al., 1993). The overview of Y mutations in Yp interfering with male fertility will present the actual state of knowledge in this research field. Yp;Xp-Translocation Chromosomes
All males with X;Y translocations with a Y breakpoint in Yp are azoospermic. Cytogenetically this translocation chromosome is often not visible, and the karyotype of these males is frequently recognized as 46,XX (Petit et al., 1987). This type of Y rearrangement is an abnormal interchange between parts of the short X and Y arm, loss of distal Xp, proximal Yp, and the whole long Y arm. Azoospermic XX males resemble in their general masculine appearance sterile males with a 47,XXY karyotype and Klinefelter syndrome (de la Chapelle, 1981); the lesser body height, smaller tooth size, and the general absence of mild mental retardation can, however, differentiate them. But like the Klinefelter patients, XX males have small testes, weak or abnormal secondary sexual characteristics, and normal to low androgen level. Only Sertoli cells and no germ cells are currently analyzed in their testis tissue.
Human Y Chromosome and Spermatogenesis
213
If the breakpoint in the Y chromosome is in distal Ypl1, the translocation becomes cytogenetically visible and the karyotype is recognized as: 46,X,t(X;Y). Phenotypically, sterile males with the 46,X,t(X;Y) karyotype resemble those with the 46,XX karyotype as expected. In three cases the breakpoint of the X chromosome was described as being in Xp22.3. All had short stature, small testis, and, in one case, severe mental retardation. In one patient with azoospermia, the testicular histology showed Leydig cell hypertrophy and nodular hyperplasia. Seminiferous tubules were reduced in number and size showing hyalinization, involuting Sertoli cells and the absence of germ cells. In comparison to the testicular histology described for sterile males with a Yq-, Ynf, or ring-Y chromosome, distortion of early testis development seems to be more severe in males with X,Y translocation chromosomes. One may assume, therefore, that the dosis of X-chromosomal genes located in distal Xp may contribute to early testis and germ cell development. Inactivation of the X chromosome is nonrandom in males with X,Y translocation chromosomes. It is easy to imagine that the process of X inactivation during prespermatogenesis (see Section 11) may be hampered by a breakpoint in Xp and its fusion with the short Y arm. Since mosaic karyotypes between 46,X,t(X;Y) and 47,XXY have been described, it has been discussed that the 46,X,t(X;Y) karyotype is only present in fibroblasts and lymphocytes derived from a 47,XXY karyotype in the germ line (Bernstein, 1985). Yp;Yq-TranslocationChromosomes
Translocations between two Y chromosomes are difficult to examine and can only be differentiated from iso(Yq) chromosomes by the different lengths of both Y arms. The seven cases described were azoospermic or oligospermic (Wahlstrom, 1985). Their testicles are small and soft. In only one case was an analysis of the testis histology done. A maturation arrest at the primary spermatocyte stage suggests that the Y/Y translocation chromosome in this male interferes with development of the pairing complex ofthe Y to the X chromosome during the pachytene stage. Since no studies are done to determine whether one Y breakpoint is located in Yq, the question is open whether AZF in Yql 1 may be distorted in one of the rearranged Y chromosomes or perhaps a double dose of AZF is present in those males. Often their karyotype is a mosaic with a 45,XO cell line.
PETER H. VOGT
214
B. Y-Mutations in Yq Interfering with Male Fertility
When it became possible to mark the heterochromatic block in the long arm of the human Y chromosome by its brilliant quinacrin fluorescence, it was recognized that the “normal looking” Y chromosome in sterile males with a mosaic XO/XY karyotype had lost this heterochromatic block completely (Caspersson et al., 1971). They were defined as “Ynf’ (non-fluorescent) Y chromosomes. Later, the “Ynf’ chromosomes were recognized as dicentric iso(Yp) chromosomes with a breakpoint in Yq 1 1 and inactivation of the second centromere (Magenis and Donlon, 1982). At the same time, large deletions in the long Y-arm including the distal part of the euchromatic Yql 1 region were found in various azoospermic males (Neu et al., 1973; Tiepolo and Zuffardi, 1976; Yunis et al., 1977) leading Tiepolo and Zuffardi to postulate the presence of Y-chromosomal spermatogenesis gene(s) in this Y region. Since all sterile males with Yqll anomaly were azoospermic, they defined this Y function as azoospermiafactor (AZF). Five different karyotypes with a deletion of parts of Yql I are known today in sterile males (Sandberg, 1985; Vogt, 1992). Not all of those males were azoospermic. Molecular analysis in Yql1 recently revealed interstitial microdeletions as an additional class of mutations in azoospermic males (Vogt et al., 1992; Nagafuchi et al., 1993; Najmabadi et al., 1994). These Y-mutations will be discussed in connection to their clinical phenotype and testis histology separately in the following sections. Monocentric Y9- Chromosomes
A terminal deletion of distal Yq 1 1 (Yq-) in six azoospermic males was first described by Tiepolo and Zuffardi (1976). Their external genitalia were normal and their testicles hypoplastic. When the authors analyzed testis tissue sections of three patients it was revealed that one (case 1) had a low number of spermatogonia and Sertoli cells, the second (case 3) had only Sertoli cells, and the third (case 4) neither germ nor Sertoli cells in his seminiferous tubules (Tiepolo and Zuffardi, 1976). In two patients (cases 1 and 4)a testicular fibrosclerosis was observed. These variable phenotypes were confirmed by analyzing testis tissue sections of a series of other azoospermic males with a 46,XYq- karyotype (Yunis et al., 1977; Hartung et al., 1988). Some authors additionally describe an abnormal thickness of the basal lamina of the patients’ testis tubules. Hartung et al. ( 1 988) observed spermatogonia and spermato-
Human Y Chromosome and Spermatogenesis
21 5
cytes in the testis tubules only in cases where the thickness of the lamina was normal. Absence of germ cells and a severe reduction of pachytene nuclei in some tubuli were described by Bardoni et al. (1991) for one azoospermic male in a meiotic chromosome study. Dicentric iso(Yp) Chromosomes
A karyotype with a dicentric Yp isochromosome in azoospermic males is often associated with a 45,XO cell line. Its proportion is highly variable and may be different in the patient’s fibroblasts and lymphocytes (Daniel, 1985). The rearranged Y chromosome is expected to have a size similar to the normal Y chromosome since it is hard to distinguish its phenotype in unstained metaphase plates from a normal Y chromosome. But after staining the heterochromatic block in the long arm of the normal human Y chromosome by quinacrine, it can be recognized that the rearranged “normal looking” iso-Y chromosome has lost the fluorescent heterochromatic Yql2-block. It is therefore also called “nonfluorescent” Y chromosome (Ynf). The dicentric chromosome structure of Ynf chromosomes is unstable during cell divisions as long as it contains two active kinetochore regions (Daniel, 1985). Karyotypes with Ynf chromosomes sometimes include, therefore, more than two cell lines with “different looking” Y chromosomes (Chandley et al., 1986; Fryns et al., 1978; Diekmann et al., 1992). Recently, two karyotypes with an Ynf chromosome were described as including no mosaic cells (Kohler and Vogt, 1994). FISH analysis with different Y-specific DNA probes revealed a large interstitial deletion in one of the two centromer regions in both cases. Due to these deletions, the phenotypes of both Ynf chromosomes in metaphase plates became “abnormal” leading earlier investigators to describe them as monocentric Yq- chromosomes (Chandley et al., 1989; Ma et al., 1992). In all cases where a testis biopsy was performed, spermatogenesis is blocked at different phases before meiosis. Only Sertoli cells were observed in the studies of Chandley et al. (1 986) and Affara et al. (1986). Degenerating premeiotic germ cells and Sertoli cells are observed in the study by Kaluzewski et al. (1988) and an arrest at the pachytene stage is described for one case by Affara et al. (1986). Since testis histology observed in sterile males with a Ynf chromosome and no XO cells is similar as described for those with a mosaic karyotype (Chandley et al., 1989), it can be concluded that Ynf chromosomes induce azoospermia. This may be possible by disruption of one spermatogenesis gene in Yql 1
21 6
PETER H. VOGT
or by disruption of the pairing structure of the Y chromosome in Yp to the X chromosome in Xp before formation of the sex vesicle in the primary spermatocyte nucleus. Iso(Yp) chromosomes have two pseudoautosomal regions in distal Up, which should be able to pair to each other. Indeed a high amount of selfpairing of an iso(Yp) chromosome was observed in a meiotic chromosome study (Figure 9). If this happens, it would interfere with pairing of Yp to the pseudoautosomal region in Xp. The XpNp pairing process seems to be an essential prerequisite for normal spermatogenesis (Chandley, 1988), since its failure induces azoospermia (Gabriel-Robez et al., 1990; Mohandas et al., 1992). Ring-Y Chromosomes
Sterile males with a ring-Y chromosome are often phenotypically quite normal (Daniel, 1985). This is surprising since in most cases the karyotype of those males is a mosaic with 45,X cells in their lymphocyte cultures. As sterile males with a nonmosaic ring-Y karyotype have been also described (Wilson et al., 1976), the question is raised whether a mosaic karyotype with a ring-Y chromosome exists in vivo, at all (Steinbach et al., 1979; Daniel, 1985). In v i m cells with ring chromosomes can generate 45,X cells (without a ring chromosome) because of media-induced sister chromatid exchanges (Daniel, 1985). The ring constitution of an abnormal Y chromosome is difficult to display in metaphase plates, but is visible in meiotic chromosome pictures (Chandley and Edmond, 197I). Disruption of spermatogenesis can take place at meiotic stages (Chandley and Edmond, 1971) or before the development of spermatogonia (Steinbach et al., 1979). The patient with a spermatogenic disruption at meiosis had a high amount of 45,X cells in his lymphocytes (30%). It can be expected that ring formation of the Y chromosome will interfere with the X-Y pairing process during pachytene as described in iso(Yp) chromosomes. Yq;Xp-Translocation Chromosomes
Sterile males with Y;X translocations and breakpoint in Yql1 are rare. Total Yp and proximal Yql 1 is lost and the distal part of the long Y arm is translocated to the X chromosome. The karyotype includes a normal Y chromosome: 46,Y,t(X;Y). All sterile males with a 46,Y,t(X;Y) karyotype had inherited their translocation chromosome from the mother (Bernstein, 1985). Recent molecular analysis of different X DNAprobes
N
1:
c
I\
U
Figure 9. Pachvene spread of the X and Y chromosome of an azoospermic male with a dicentric Up chromosome and a non-mosaic karyotype, 46,X,dic(Yp) as described by Kohler and Vogt (1 994). The chromosomes are stained with silver nitrate. The self-paired Y-Y chromosome displays a balloon-like structure with a hairpin (arrow). This Y chromosome was earlier thought to be monocentric (Chandley et al., 1989).
218
PETER H. VOGT
located in Xp22 and proximal Yql1 revealed that X-Y homologous DNA blocks in those chromosome regions may be involved in the etiology of these Y;X translocation events (Ballabio et al., 1989). Anomalies in the sexual development of 46,Y,t(Y;X)-males are also described. Phenotypically, they had a small penis, small testes, and a hypoplastic scrotum. In two adult brothers with azoospermia, the histology of their testes revealed a decrease of spermatocytes reaching diakinesis and the absence of secondary spermatocytes and sperms (Yamada et al., 1982). In a similar case, it has recently been shown that the spermatogenic failure in those males is not most likely due to a dysfunction of AZF, but to a distortion in the pairing structure between the tips of the short arms of both sex chromosomes (Gabriel-Robez et al., 1990). Yq;A- Translocation Chromosomes
Autosomal translocations of the Y chromosome are observed in a balanced or non-balanced karyotype. In most familial cases with a balanced Y-A translocation, the distal heterochromatic part of the Y chromosome is translocated to the short arm of an acrocentic chromosome (Smith et al., 1979). This chromosomal preference may be explained by the observed non-random association of those chromosomes to the X-Y pairing complex in the sex vesicle (Stahl et al., 1984; see also Section I1 and Figure 6). The Y-heterochromatin, attached to an autosome, can easily be detected by fluorescent staining. Balanced or reciprocal Y-A translocations were observed in fertile and sterile males (Fryns et al., 1985). It is assumed that in fertile males the breakpoint of the Y chromosome is in Yq12, the genetically inert heterochromatic region. In sterile males the breakpoint is assumed to be adjacent of Yq12 in the distal euchromatic Yql1 region. A family with the same balanced Y;A-translocation in fertile and sterile males of its pedigree (Reitalu, 1973) suggests that the primary breakage event on the Y chromosome occurs in Yq12 but is unstable during further germ cell development in the father or during early embryogenesis of the son. Azoospermic males with a balanced Y;A-translocation chromosome have normal external genitalia or hypogonadism. The histology of their testis shows an arrest of spermatogenesis at the spermatocyte stage before or during meiosis (Viguii: et al., 1982; Fryns et al., 1985)or during the formation of spermatids (Laurent et al., 1982; Faed et al., 1982). In two cases where an analysis of meiotic chromosomes was performed, the condensation of the X and Y chromosome during their premeiotic
Human Y Chromosome and Spermatogenesis
21 9
pairing process was shown to be partly inhibited (Gonzales et al., I98 1 ; Viguie et al., 1982). Unbalanced Y,A translocation chromosomes have lost the acentric heterochromatic chromosome part (Yq 12). Earlier they were described as mosaic cases with a rare 46,XY’cell line, necessary to explain their sexual phenotype (de la Chapelle et al., 1986). Azoospermic males with this karyotype are rare. However, molecular analysis of the Y chromosome in those males had revealed that the male determining Yp chromosome part is present and attached to an autosome. The Y-breakpoint is in the Yqll region (Disteche et al., 1986a; Gal et al., 1987; Andersson et al., 1988) and the Y centromere is assumed to be inactive. Phenotypically, their testis can be normal but also hypoplastic and small. Testis histology reveals an early disruption of germ cell development, showing only Sertoli cells and immature seminiferous tubules.
Y91 1 lnterstifial Microdeletions Interstitial microdeletions in Yq 1 1 are too small to be observed in the microscope. Sterile patients with this Yql1 mutation are therefore described with a normal karyotype, 46,XY. They were detected with the aid of hybridization experiments on genomic DNA blots of the patient using DNA probes located in Yql 1. The first interstitial microdeletion in Yq 1 1 was described by Johnson et al. ( 1 989) in a sterile male with germ cell aplasia. But this study only includes one DNA probe, 50fz (DYS7), known to be polymorphic (Nagafuchi et al., 1993; Disteche et al., 1986b) and the authors did not test the presence of this probe in the Y chromosome of the patient’s father. This was done for the first time in an extensive molecular screening program with 30 DNA probes located in Yq 1 1 on 19 males with non-obstructive azoospermia (Vogt et a]., 1991a, 1992) Two different microdeletions were detected. Both deletions were identified by a DNA probe of the pY6H sequence family (Figure 10). Probe Y6HP35 located in proximal Yql 1 was deleted in JOLAR. Probe Y6HP52 located indistal Yq 1 1 was deleted in KLARD (the mapping procedure of DNA probes in Yql1 is described in Section IV). This result was intriguing since pY6H sequences contain some homology to dhMiFl , a fertility gene sequence on the Y chromosome of Drosophilu hydei (Vogt et al., 1991b). Both microdeletions were proofed to be de novo (i.e., only present in the Y chromosome of the sterile patient and not in the Y chromosome of his father or fertile brothers), and confirmed by additional deletions of
PETER H.VOGT
220
flanking DNA probes: 12f3 in proximal Yqll and 50f2/C,Y367/B,
GMGY18,RBF8,FRI5-II,pl16/21,Y6BS65/E,p49f, GMGYS indistal Yql 1 (Figure 10). Later screening programs in my laboratory and others have confirmed the occurrence of interstitial deletions in Yql 1 in larger series of patients with idiopathic sterility (Nagafuchi et al., 1993; Najmabadi et al., 1994; Reijo et al., 1995; Vogt et al., 1996). The second major screening program in my laboratory was done in collaboration with six infertility clinics from Germany and abroad including 370 sterile men with azoospermia or severe oligospermia (<2 million sperms per ml ejaculate). Its result is described in detail elsewhere (Vogt et al., 1996). Beside deletions in proximal and distal Yql 1, the extensions of which were similar to JOLAR and KLARD, respectively, a new class of microdeletions was detected in the middle of Yql 1 again by probes of the pY6H sequence family: pY6D14, pY6BaH34, pY6PHc54, and pY6BS65K (Vogt et al., 1996). Sterile males with Yql 1 microdeletions have different testis volumes, normal or high FSH levels, beside a normal LH and testosterone level. The histological testis picture of individuals with a deletion in proximal Yql 1 revealed a Sertoli-cell-only-syndrome. Males with deletion in middle Yqll have an arrest of their spermatogenesis at the primary spermatocyte stage. Males with a deletion in distal Yql 1 displayed a heterogenous low amount of spermatogonia, spermatocytes, and spermatids in their testis tubuli beside tubuli with only Sertoli cells (Vogt et al., 1991). Some males were oligospermic producing a low amount of mature spermatozoa. JOLAR
I zn
KURD
YKUC.Y367llJ. G M C Y l l RBFS, FR15-11,pll(Vz1 YfiBSfiYE, p4Vf. CMCYS
Figure 70. identification of different interstitial deletions in two azoospermic males, coded IOLAR and KLARD by probes of the pY6H sequence family described in Vogt et al. (19911. Both deletions could be confirmed by the additional deletion of other DNA probes as indicated. For a description of probes see Ma et al. (1 992).
221
Hiiman Y Chromosome and Spermatogenesis
46.X.dlclYpl
amblglous genltalla
atrophlc tubules
45.XO
hypospadlas
only Scrloll nlls
cryptwrchldlsm
sprmatogonla
small penls
rare sprmatocytrr arrest a1 pachytenr
46. X . r M
hypoplastlc tesuclcs
only Scrtoll cells
45.XO
hypospadlas
arrest at paehytcnr
~vrptwrchldlsm
q
46.Y.tWp Yq)
hypoplasuc testleks
decrease of spermatocytcs
small p n l s
46. XYq-,(-.(IA.Yqhl
normal or
arrest at pachytcnr
hypoplastlc tesucles
a m s t durlng melosls spermaud dlR arrest
ollgospermra
46.X.tlkYl
Immature semlnlf tub
46.XY
P
D
hypoplastic t r s t c l r s
an& S e r l O l l d l s
normal or
only Scrtoll cells
hypoplasuc tcsurlrs
arrcat at pachylme rare spermalogonla and rare spmnatwytrs ollgosprm,la
Figure 11. Gonadal phenotype and testicular histology observed in sterile males with different Yql1 anomalies as indicated and described in detail in the text.
222
PETER H. VOGT
Thus, the histological pictures observed were clearly dependent from the subregion in Yq I I , which was deleted. The different mutations in Yql 1, detected in sterile males, strongly support the presence of at least one spermatogenesis locus in Yqll, Its definition as azoospermia factor (AZF) by Tiepolo and Zuffardi (1976) was linked to the sterile phenotype most often observed. The description and comparison of the testis histology in AZF-males with different Yq 1 1 mutations summarized in this review (Figure 11) suggest that the AZF locus is active during different phases of human spermatogenesis. After the detection of different interstitial Yql 1 microdeletions, it became attractive to think of different and perhaps independent AZF loci in Yq I 1, which, if disrupted, all express the azoospermia phenotype, but, looking in the testis tubules, disrupts spermatogenesis at different phases. To distinguish between pleiotropic mutation effects of one AZF locus and mutation of different AZF loci, asks for isolation and analysis of the AZF gene(s) themselves. As a first step toward this aim a detailed mapping study ofAZF in Yql 1 was performed.
IV. MOLECULAR DELETION MAPPING OF THE AZOOSPERMIA FACTOR (AZF) IN Yql 1 Mapping of the azoospermia factor AZF in Yql1 by linkage analysis could not be applied, due to the lack of regularly crossing over events between the Y and X chromosome in Yql I . Mapping of AZF in Yql 1 also seemed not to be possible by differentiating the breakpoint in Yql 1 in cytogenetically detectable Yq 11 deletions using the microscope. Several research groups tried this approach with the aid of high resolution banding (e.g., Oosthuizen et al., 1990). But due to the different chromosomal rearrangements in Yq 11 described above, mapping by karyotype analysis is inherently difficult. Before the detection of microdeletions in Yq 1 1, genomic DNA samples of males and females with cytogenetically visible Yql1 anomalies were therefore collected in different laboratories for mapping of AZF by molecular deletion mapping. With the aid of a series of Y-chromosomal DNA probes and DNA blotting experiments analyzing the presence or absence of these probes on the Y chromosome of individuals with Yql I aberrations, several so-called deletion maps were established (Vergnaud et al., 1986; Ferguson-Smith et al., 1987; Bardoni et al., 1991). They displayed a linear order of the probes used in the analysis by putting their position to so-called intervals along the human Y chromosome from pter
Human Y Chromosome and Spermatogenesis
223
to qter. The boundary of each interval was marked by the Y breakpoint positions of the individuals tested. The most detailed study of this kind for Yql 1 was done first by Ma et al. (1992). The study included 40 Y-specific DNA probes and 21 individuals with Yql 1 anomalies, dividing Yqll in 14 subintervals (1-XIV). Obviously, the state-of-the-art study would be the creation of an interval map in which each Y-DNA probe tested would mark one distinct Y interval. lfthe breakpoints on the Y chromosome are statistically distributed, this should be achieved by analysis of more individuals with more Y probes. Using PCR analysis, this goal was strived for by Vollrath et al. (1992). His study succeeded in mapping 132 Y chromosomal loci (sY loci) along the whole Y chromosome using the Y breakpoints of 96 individuals. But only 43 Y intervals could be established suggesting that the human Y chromosome has preferential breakage domains on both chromosome arms. The first detailed molecular AZF mapping experiment was done by Bardoni et al. (1991). They located AZF to subinterval K of their interval map, that is, to distal Yql 1 . The testis histology of their sterile males with Yql 1 deletions was described to be different and not all of them were azoospermic, but also oligospermic males were observed. We therefore compared the breakpoint position of a series of sterile males with cytogenetically visible Yql1 anomalies using the map of Ma et al. (1992) and included an analysis of their testis histology (Vogt et al., 1993). In this study, all males with a proximal breakpoint in Yql I revealed in testis histology an early disruption in spermatogenesis before or after the proliferation phase of spermatogonia (phase a). In three males with a breakpoint in middle Yq 1 1, however, arrest of spermatogenesis was observed at the spermatocyte stage @hase b). The karyotype of one of the three males with an arrest duringplme b (JOWAL) contained two different cell lines. The major cell line (ca. 90%) had a ring-Y chromosome with a break and fusion point in proximal Yql 1 (Vogt et al., 1993). The minor cell line (ca. lo%), only detectable by PCR mapping analysis, had a ring-Y chromosome with a break and fusion point in middle Yq 11 (Henegariu et al., 1994). The primary breakage event in the Y chromosome of JOWAL must, therefore, have happened in middle Yql 1. Most likely, further Yql 1 deletions then happened in vitro during 20 years of cell culturing. The most important result of this study was a distinct relation of a specific picture in testis histology to a specific breakpoint region in Yql 1 indicating two different spermatogenesis loci in Yql 1, which, if dis-
224
PETER H.VOCT
rupted, cause azoospermia. We tentatively defined these loci AZFa and AZFb (Vogt et al., 1993). Both loci should include at least one functional spermatogenesis gene. But using only patients with cytogenetically visible Yql 1 anomalies for AZF mapping, we had, of course, to keep in mind that arrest at the spermatocyte stage in the three males studied could also simply reflect a failure of the Xp/Yp pairing process in the sex vesicle of their spermatocyte nuclei. One male had a ring-Y chromosome. The other males had iso(Yp) chromosomes. Since X-Y pairing distortions in the spermatocyte nucleus would be expected, especially in males with these Y constitutions (see Section IIIB), assumption of a second AZF locus expressed before or during meiosis (AZFb) would be not needed for explanation of their sterility. Therefore, to prove the existence of different AZF loci by molecular mapping in comparison of breakpoints in Yql 1 , it would be better to use karyotypes of non-mosaic monocentric Yq- chromosomes or A;Yq translocation chromosomes (see Figure 1 1). Unfortunately, however, these karyotypes are very rare in sterile males and, of the patients available, Yql 1 breakpoints all mapped to proximal Yql 1 (Gal et al., 1987; Anderson et a]., 1988; Vogt et al., 1993). Whether one or more AZF loci are present in Yql 1 became, therefore, only feasible to study after the detection of sterile males with different microdeletions in Yql 1, including an analysis of their testicular histology (Reijo et al., 1995; Vogt et al., 1996). A comparison of the testis histology was not possible with the first two males detected with an interstitial microdeletion in proximal (JOLAR) and distal (KLARD) Yql l, respectively (Figure 10). Therefore, we started a new collaborative study with different infertility clinics screening more than 300 males with nonobstructive azoospermia or severe oligospermia and a normal karyotype for more microdeletions in Yql 1.In order to increase the resolution of our interval map for the mapping experiments as much as possible, the interval map of Ma et al. (1992) was extended with more DNA probes and a selected set of SY DNA loci from the map of Vollrath et al. (1992). In this way, Yql 1 could be subdivided into 25 subintervals (Dl-D25; Figure 12). A PCR-multiplex procedure was developed for a rapid screening protocol (Henegariu et al., 1994). Three different subregions in Yq 11 were found to be deleted in different individuals. They were mapped in our interval map to Yqll (D3-D6), Yqll (D13-D16), and Yqll (D20-D22; Figure 13).
Human Y Chromosome and Spermatogenesis
225
molecular analysis of breakpoint positions in Yq 1 1 Figure 12. Schematic view of molecular interval map in Yqll based on the map of Ma et al. (1992), extended by Henegariu et al. (1994), and Vogt et al. (1996) mapping additional a seriesof sYDNA loci in Yql1 . Adetailed description of this map including 76 DNA loci is presented by Vogt et al. (1996).
lntentitial deletions in YqI 1 occurring de nuvu in sterile males with azoospermia/oligozoospermia
2 n m p c r m i c patienk with interstitial deletions in proximal Y q l l (D3-W)
4 0 7 1 DB I DQlDl~lDllID12IDl~DI4IDl5IDtMD~OlBID19lDXD2llO221D2~D211D2 3D31D4I 051 D d
N N
6
7 rrimrpcrrnic or ~~ligimnirpcrrnir prlimta wilh intrntitivl deletionr in distal Yyl I (VZO-I)22) Dl1 021 D 3 l 041 D S l 0 6 l D 7 1 OBI DQlDlOlDlllOl2IDl~Dl4lDl5lDtslD~Dl~lD~~~ @2~21l022j
figure 13. Survey of the position of microdeletions found de n o w in sterile males with azoospermia or severe oligospermia. The extension of the deleted subregions (I, //,and //I) are drawn schematically according to the interval map displayed in Figure 12.
Human Y Chromosome and Spermatogenesis
227
Males with only Sertoli cells in their testis tubules had an interstitial deletion in proximal Yql 1 (D3-D6). Males with arrest at the spermatocyte stage had an interstitial deletion in middle Yqll (D13-D16) and males with a variable arrest pattern in their tubuli had an interstitial deletion in distal Yqll (D20-D22). All oligospermic males had an interstitial deletion in distal Yql 1. Thus, males with a different position of their interstitial Yql 1-microdeletion also revealed a different testis histology and males with the same interstitial Yq 11-microdeletion revealed the same testis histology. A detailed description of the patients studied is presented in Vogt et al. (1996). With this study, the assumption of different Y spermatogenesis loci in Yqll causing azoospermia or severe oligospermia (AZF loci) now seemed to get strong experimental support. . Therefore, I like to confirm and actualize the hypothesis of different spermatogenic loci in Yql 1 ,which if disrupted cause azoospermia (Vogt et al., 1993), by adding AZFc besides AZFa and AZFb. The functional phase ofAZFc during spermatogenesis should be related to the histological testis tissue pictures seen in azoospermic or oligospermic males with an interstitial deletion in distal Yql 1 . This picture is distinctly different from the histological pictures in patients with a deletion in proximal or middle Yq 11. In our first analysis, comparing only the histology of patients with cytogenetically visible Yqll deletions, the phase ofAZFc could not be distinguished from AZFb, since no sterile males with arrest at postmeiotic stages or oligospermia were available for our mapping experiments. We, therefore, described at that time the phase of AZFb as “highly variable” (Vogt et al., 1993). After detection of the interstitial microdeletion class in middle Yql i, the phase of AZFb can now be distinctly restricted premeiotically to the spermatocyte stage. The phase of dZFc remains variable. Since the histological picture of patients with deletions in distal Yql 1 can also be oligospermic, one may assume that the fknctional phase of expression of the putative AZFc locus is postmeiotic during the maturation process of the spermatids. Another possibility may be that besides AZFc additional spermatogenesis loci are located in distal Yq 1 1 (i.e., AZFd, AZFe, . . .?). The molecular extensions of the interstitial deletions in distal Yql 1 are rather large and can easily accommodate more than one spermatogenesis gene. They were estimated to be between 2,00&3,000 kb (Kirsch et al., 1996). Different types and amounts of germ cells in different testis tubuli including tubuli with only Sertoli cells in AZFc-males may, however, also only reflect gradual degeneration processes of postmeiotic germ
228
PETER H. VOCT
cells starting shortly after puberty, when the first waves of postmeiotic germ cell development were initiated. Patients with sterility are usually observed in the clinic 15-20 years after puberty. This would be time enough for an age-dependent gradual degeneration of germ cells to different stages, as observed in sterile Klinefelter patients with nonmosaic 47,XXY karyotypes (Foss and Lewis, I97 1). Therefore, it may be that males with distal deletions in Yqll are all oligospermic after puberty. In summary, I would like to conclude that mapping results in Yql 1 for the Y chromosomal azoospermia factor AZF have succeeded in dividing this locus into at least three different spermatogenesis loci, tentatively defined as AZFa, AZFb, and AZFc, since all three can cause azoospermia. The location of these loci coincides with the position of DNA loci of the pY6H sequence family (Figure 14) supporting the idea of our initial experimental approach to select AZF sequences by Y fertility gene sequences ofDrosophiZa (Vogt et al., 1991b). This coincidence may just be accidental. But it may also indicate a similar functional role of Y spermatogenesis genes in DrosophiZa and human. Correspondingly,AZF loci may be functional because they contain gene(s) coding for testis specific protein(s) and/or RNA molecules and/or a specific DNA domain coding for a testis-specific chromatin folding structure (Vogt, 1990; Hennig, 1993). We and others used YACs and cosmid clones, mapped to the different interstitial Y deletions, for screening cDNA libraries of poly A+ testis RNA in order to isolate the proposed spermatogenesis genes (AZFa, AZFb, and AZFc candidate genes). Three candidate genes have been published until now: RBM, aka YRM(Ma et al., 1993), SMCY(Agu1nik et al., 1994b) and DAZ/SPGY (Reijo et al., 1995; Maiwald et al., 1996). Their isolation and expression pattern with view to AZFa, AZFb, and
AZFa
AZFb
AZFc
Y6HP3.5
Y6D14 Y6BaH34 Y6PHCH
Y6HP.52 Y6BS6yE
Y(IBS6WC
Figure 14. Schematic view of position of the three defined azoospermia loci AZFa, AZFb, and AZFc in Yqll mapped by the extensions of different interstitial microdeletions in Yqll as displayed in Figure 13.
Human Y Chromosome and Spermatogenesis
229
AZFc will be described in the following section. Since strong evidence exists that, besides the expression of protein coding genes, the intactness of the Y chromosome structure in Yql 1 may also be important during male germ cell development, its dynamic behavior in the male germ line is described in a separate section.
V. AZFCANDIDATE GENES EXPRESSED IN TESTIS A. The R B M Gene Family
Molecular screening experiments with the genomic repetitive Y probe Y-367 on a cDNA library from poly A+ adult testis RNA succeeded in isolating a gene family expressed in human testis (Ma et al., 1993). It was called RBM aka YRRM due to homology of their putative gene product(s) to poly(A) RNA-binding proteins containing a specific RNA recognition motif. The authors claimed that RBM genes are candidates for the AZF locus, since one gene copy was deleted in azoospermic males with an interstitial deletion in distal Yql 1. Using the isolated RBM gene copy, MK5 (RBMI), as a hybridization probe on genomic male DNA blots, a repetitive hybridization pattern was revealed, quite similar to the Y-367 pattern as expected (Muller et al., 1989). A large genomic Y fragment (RBMZIB,) was mapped to subinterval IX (where Y-367IA also resides), and a fragment of about 4 kb (RBMIIE) was mapped to subinterval XI11 of the Ma map, where Y-367IB resides (Ma et al., 1993). Like Y-367, RBM gene copies were also present around the Y centromer and in proximal Yp (Inglis et al., 1994). But besides this coincidence between the position of Y-367 and RBM copies, RBM gene copies were also shown to be present in subinterval 111 (RBMZIG), subinterval VIII (RBMIIA,F), and subinterval X (RBMIIC) of the Ma map in Yql1 (Ma et al., 1993). Obviously, it is not possible to map the RBMl gene copy to a distinct Y-interval by molecular deletion mapping on DNA blots. Therefore, 1 disagree with Ma et al. ( 1993), who claimed to have mapped RBMI to distal Yq 1 1. Using a RBMI specific primer pair in a highly stringent PCR experiment with different samples of genomic DNA from patients with a breakpoint and deletion in Yq 1 1, we were able to map the RBMI gene copy to subinterval D 16of our interval map (see Figure 12). This interval corresponds to subinterval IX of the Ma map. It is, therefore, most likely that RBMI resides on the genomic Y fiagment RBMIIB2 in this subin-
230
PETER H . VOGT
terval, where Y-367/A also resides, and not in subinterval XII-XV (D20-D22 in Figure 12) as claimed by Ma et al. (1993). Subinterval D16 is included in the Y deletion of azoospermicpatients with an arrest of spermatogenesis at the pachytene stage defining AZFb (Yqll-DI2-Dl6; Figure 13). Mapping of RBMZ to this subinterval, therefore, suggests the involvement of RBMZ expression in the function of AZFb. This thesis is supported by in situ hybridization experiments of RBMZ to testis tissue sections revealing its expression in primary spermatocytes (Chandley and Cooke, 1994). The RBMgene copy in distal Yql 1 may be functional as well, and this could be tested after its isolation from testis poly A’cDNA libraries. But it may be also possible that not all RBM gene copies displayed by crosshybridization of RBMZ on DNA blots are indeed functional in spermatogenesis. The genomic distribution of RBM gene copies on the human Y chromosome reflects recent evolutionary amplification steps of certain sequence blocks, since a similar RBM distribution pattern was only observed on the Y chromosome of the gorilla (Ma et al., 1993). Do other mammals need only one RBM gene copy for spermatogenesis? Besides RBMI only one additional human RBM gene copy was isolated until now (MK29, called RBM2; Ma et al., 1993). It reveals a highly conserved exon structure in comparison to RBMZ, but could be distinguishedby seven nucleotide substitutionsand a 5 bp deletion in the translated 3’ region. This changed the 3’end of the sequence of the putative RBMZ protein in 34 amino acids and reduced its length by 76 amino acids due to an earlier stop codon (compared to RBM1). Using highly stringent PCR-conditions with RBMZ specific primers, RBM2 was mapped to Yqll-subinterval DI 8 (Figure 12). This Y region corresponds to subinterval XI of the Ma-map, corresponding to the genomic RBMI/C fragment (Ma et al., 1993).The RBM2 gene copy does not seem to be essential for spermatogenesis since PCR analysis with RBM2 gene specific primers could show that this gene copy is a null allele of the RBM gene family in fertile Japanese males and polymorphic in Caucasian and Negro males (Nakahori et al., 1994). So it is quite possible that the expressions of other RBM gene copies are also not biologically important or are not expressed at all. On the other hand, the RBM gene family is conserved on the Y chromosome in different mammals and its expression appears to be highly specific to adult testis. This suggests an important fbnction of at least one member of this gene family in human spermatogenesis. And this may be RBMZ, defined here as a strong candidate gene for the expression of AZFb, but
Human Y Chromosome and Spermatogenesis
231
not AZFc, as stated by Ma et al. (1993; Vogt et al., 1996). At least two other RBM gene copies are deleted in azoospermic patients with an interstitial deletion in middle Yql 1 . They correspond to the genomic Y fragments RBMI/A and RBMZ/F in interval D15,respectively, subinterval VIII of the Ma map. It is not ye't known, whether these gene copies are functioning. But how many RBM gene copies are functioning? If deletion of only one or a few gene copies is sufficient to cause oligo- or azoospermic phenotypes, it would imply a biological requirement for multiple RBM gene copies in human spermatogenesis. But it may also be that the gene products of only a subset of RBM gene copies are functional and that these gene copies are clustered in middle Yq 11. If they are spread in Yq 1 1 and Yp, one may assume, that the sterile phenotype of the male may then be dependent on the amount of deleted functional RBM gene copies ranging from oligospermia to azoospermia phenotypes (Figure 15). An
I
100%
/
I
I
I 1 1
/
/
I I f I
I
20
/
/
/
/
10
0 mill
reduction of sperm number figure 15. Schematic view of how a gradual deletion of RBMgene copies spread in Yqll may cause a gradual reduction of sperm number. Deletion of RBM gene copies in Yqll may reduce the number of sperm scarcely (----I, proportionally (----I, overproportionally (----I. The model suggests that only the deletion of all functional R5M gene copies may lead to the azoospermia phenotype.
PETER H.VOCT
232
azoospermic phenotype is then only caused, when all functional RBM gene copies are deleted, for example, after a total deletion of Yql 1 . In summary, it is concluded that any analysis of the RBM gene family content in sterile males with azoospermia or oligospermia must include an analysis of its potential polymorphic structure in the Y chromosome of the father and brother(s). This view is supported by recent SSCP analysis of exonic PCR products prepared with primer pairs from the RBMZ gene structure (Inglis et al., 1994). It revealed seven classes of RBM gene copies on the Y chromsome of fertile males, some of which had multiple members found in Yp and Yq. When they applied this SSCP analysis to a sterile male population, individuals were found lacking a class of genes. Other individuals had reduced numbers of members of a class and others revealed a single base pair difference in one member of one class. Unfortunately, the authors did not include a study of the patients’ fertile fathers. The polymorphic background of RBM gene copies in these families could therefore not be evaluated. 6 . The SMCY Gene The SMCY gene was mapped to middle Yql 1 with the aid of the DNA loci sY123 and sY124 flanking the gene (Agulnik et al., 1994b). Since this DNA region corresponds to Yql 1 subinterval D 13-D 14 (Figure 12) its position overlaps with the proposed position of AZFb, like RBMI located distal to it. Therefore SMCY is another AZFb candidate gene. The candidate status is stressed by a similar position of its mouse homolog Smcy to deletion interval 2 of the Y chromosomal Sxv region (Agulnik et al., 1994a). Genetically this deletion interval includes the spermatogenesis locus Spy, and a Y gene family involved in the expression of the male specific minor histocompatibility antigen H-Y (Hyu) and the serologically detected male antigen SDMA (Sdmu). These mapping locations on the Y chromosome are similar in the human, since recently, the human H-Y antigen locus was mapped also to middle Yqll (Cantrell et al., 1992; O’Reilly et al., 1992), that is, in the neighborhood of the SMCY gene. This of course raised the question of whether the expression of H-Y antigen is related to a Y chromosomal spermatogenesis function, which is conserved in the mouse and human. For a long time it was thought that the H- Y locus was the primary sex determining locus (Wachtel et al., 1975). But mapping of H-Y to Yql 1 and isolation of the SRY gene now have falsified this hypothesis. However, expression of H-Y in spermatogenesis is not yet excluded. Its gene
Human Y Chromosome and Spermatogenesis
233
product is present in every male cell except in prespermatogonia, germ cells before puberty, and erythrocytes (Zenzes et al., 1978). However, it is only secreted in Sertoli cells. And receptor protein(s) binding to H-Y antigen are only present in gonads (Ohno et al., 1979). Therefore, morphogenetic function of the protein during development of the somatic cells in the testis tubuli was suggested by Ohno. It has been argued that the male specific H-Y antigen is already functional before spermatogenesis and needed as male specific switch signal in early embryogenesis for increasing the growth rate of XY cells toward XX cells (Mittwoch, 1969, 1989). A Y chromosomal effect on the growth rate of early embryonic mouse cells has recently been observed (Burgoyne et al., 1993). The H-Y locus was shown to be expressed in preimplantation mouse embryos (Krco and Goldberg, 1976). It is, therefore, a good candidate locus for expression of the male specific growth rate. Besides Smcy two other genes on the short mouse Y arm, Sty and Z b , are also expressed in preimplantation embryos (Zwingman et al., 1993), suggesting that perhaps the whole Sxr region is decondensed during early embryonic development. If this holds true, the thesis put forward by Mittwoch (1969, 1989) postulating male-specific growth promoting factor(s) of the Y preceding gonadal differentiation would come into focus. Since SMCY is expressed in all human male tissues tested and in preimplanatation embryos, it is suggested that it is a strong candidate gene for expression of the H-Y locus. As this locus is also expressed in germ cells at puberty (Zenzes et al., 1978) a functional relation to expression of the spermatogenesis locus AZFb remains possible. As in mice, the SMCY gene has a homolog on the X chromosome SMCX. It has been mapped to proximal Xp (Xpll.l-Xp11.2; Agulnik et al., 1994b). Both gene copies are also expressed in preimplantation embryos and their position on the sex chromosomes is conserved in marsupials, that is, for about 120 million years (Agulnik et al., 1994a). A balance between the dosis of Smcx and Smcy expression as an important genetic switch signal during the first embryonic cell divisions toward or against the male developmental pathway can, therefore, be discussed. A double dose of Smcx would confirm the female pathway as the default decision (McLaren, 1994). A single dose of Smcy and Smcx would initiate the male pathway. If one accepts this view and assumes a similar function of SMCY/SMCX in human embryogenesis, one has to postulate that the SMCY locus must have been present during the embryonic life of
PETER H. VOCT
234
azoospermicmales with a deletion in middle Yql 1. Since the SMCY gene is located in this deletion, it would imply that this deletion event must have happened afier the expression period of SMCX/SMCY in early embryonesis. If the SMCY gene is also part of the spermatogenesis locus AZFb, SMCY and SMCXgene products may both be essential before or during puberty. This can now be tested by analysis of gene-specific mutations in idiopathic sterile males indicating disruption of AZFb by analysis of their testis histology. C. The SPGY Locus
Recently, exon trapping experiments with cosmids selected from the AZFc region in distal Yqll succeeded in the isolation of exon probes expressed specifically in adult human testis (Reijo et al., 1995). Consequently, they were used to isolate homologous “full-length” cDNA clones from a human poly(A) testis cDNA library. Using additionally 5’ RACE experiments to find the messenger 5’ end, Reijo et al. presented a cDNA sequence of 1,64 1 nueleotides. It contains the 5’ end of a Y gene encoding a novel RNA binding protein, but not the 3‘ end. The 3‘ untranslated cDNA sequence part was assumed to be truncated as it was not polyadenylated although the clone was isolated from a poly(A) cDNA library. As an exon probe of DAZ was shown to hybridize to a testis poly(A) 3.5kb-RNA population, the untranslated 3’ end of the corresponding Y gene may include roughly 2,000 nucleotides not yet isolated (Reijo et al., 1995). They designated it as “DAZ” (Deleted in AZoospermia) as the whole sequence was deleted in azoospermic men with deletion of AZFc (formerly defined as AZF in Reijo et al., 1995). The DAZ protein contains one RNA Recognition Motif (RRM) in its N-terminal part and a tandem repetitive peptide domain in its central part with seven copies of a 24-amino acid basic repeat. Each repeat can be marked by its pattern of single nucleotide substitutions. DAZ was first described as being encoded by a single copy gene in the AZFc region (Reijo et al., 1995). However, the analyses of more cDNA sequences homologous to DAZ, but not identical, indicates that this cannot be true. One cDNA sequence (CT52Y; Maiwald et al., 1996) with distinct homology to DAZ, but not identical, contained 1 1 complete copies of the 24-amino acid basic repeat in its central region. Although the consensus sequence of the peptide repeats was identical in DAZ and CT52Y, their structural composition in both repetitive peptide domains
Human Y Chromosome and Spermatogenesis
C
235
..-
Figure 16, SPGYl transcripts are expressed at different stages of spermatogenesis. In situ hybridization with a 35S labeled riboprobe of part of the SPGYl sequence to frozen testis tissue sections indicates three reproducible sites of expression in germ cells. Two examples of each stage are displayed: (a) expression in the mid (left picture) and late (right picture) pachytene in the spermatocyte nucleus; (b) expression in young (left picture) and late (right picture) round spermatids; and (c) expression in spermatids at the onset of elongation and condensation (this experiment was performed by Robert Maiwald).
was different. Each repeat unit seems to be encoded by a repetitive 72 bp exon unit. However, as the exon-intron structure of the Y gene encoding DAZ and CT52Y is not yet known, the molecular reasons for the observed heterogeneities in the DAZ homologous cDNA sequences are unclear. Besides alternative splicing and exon skipping mechanisms, the possibility exists that DAZ and CT52Y are encoded by different members of a second Y gene family located in Yql 1 and expressed specifically in adult testis tissue. This assumption got support from recent
236
PETER H. VOGT
molecular deletion experiments with DAZ and CT52Y specific primer pairs. Both were mapped to the AZFc region, but CT52Y was mapped proximal to DAZ (Vogt et al., 1996). It can be shown that CT52Y is expressed at different germ cell stages (Figure 16). Therefore, it now has got the GDB symbol SPGYl and is defined as a transcript of a gene copy of the Spermatogenesis Gen Y locus “SPGY.” DAZ and SPGYZ are the only cDNA sequences of the SPGY locus published until now. It is intriguing to note that there obviously exists a functional and structural similarity between DAUSPGYl and RBM genes. All encode testis-specific RNA binding proteins and contain a structure with a single RRM (RNA Recognition Motif) domain and a series of near-perfect tandem repeats (RBMZ contains four tandem repeats of a 1 1 1-nucleotide unit; Ma et a]., 1993). However, their coding sequences suggest only a family relationship between DAZ and SPGYI, but exclude it to RBMI.
VI. AZFAND Y CHROMOSOME STRUCTURE IN MALE GERM LINE In earlier replication banding experiments a specific chromatin structure in Yqll was described (Limon et al., 1979), and it was speculated that this chromosomeregion mediates or stabilizes the gradient of the pairing structure of the Y chromosome to the X chromosome by a specific chromatin code like the “collochores” of Drosophila (Coyne, 1985).One may, therefore, speculate as to whether an AZF locus is not only expressed by a specific protein or RNA molecule, but also by a specific chromatin folding structure coded by specific genomic DNA domain(s) in Yql 1 as discussed earlier (Vogt, 1990). During the leptotene phase, the Y chromosome starts to pair to the X chromosome at the tip of the short arms (Xp/Yp), where both chromosomes have a homologous DNA region, called “pseudoautosomal”(Burgoyne, 1982). Disrupture ofthis pairing process results in the breakdown of spermatogenesis during meiosis. This was shown by genetic analysis in mice (Matsuda et al., 1991; Burgoyne et al., 1992) and in humans (Gabriel-Robez et al., 1990; Mohandas et al., 1992). X-Y pairing can, therefore, be defined as essential for spermatogenesis and its disruption as an X/Y chromosomal azoospermia factor (AZF-XY), suggested already by Buhler (1 985). It has been observed repeatedly that X-Y pairing during pachytene is a dynamic process. It follows a time gradient including more than 20% of the short Y arm (i.e., more than the pseudoautosomalregion; Chandley
Human Y Chromosome and Spermatogenesis
237
et al., 1984) and develops a complex pairing picture (Solari, 1980) defined as a sex vesicle in late pachytene (Figure 5). By labeling the X and Y kinetochores with autoimmune CREST serum followed by a second antibody labeled with collodial gold particles, X-Y pairing beyond their centromeric region to Yql1 was distinctly observed in the electron microscope (Sumner and Speed, 1987). Interesting to note was the observation that the chromatin structure of the long Y arm displayed a dynamic behavior (Sumner and Speed, 1987). A dynamic behavior of the Y chromatin structure in the germ line was also visualized later by FISH analysis with different Y-specific DNA probes. Marking the heterochromatic part of the long Y arm (Yq12) by in situ hybridization with a specific fluorescent DNA probe (DYZ2), it could be demonstrated that this Y part is highly decondensed in germ cells at the spermatocyte stage especially at zygotene and spermatogonia before puberty (Speed et al., 1993). At mid-pachytene, the FISH Y signal becomes condensed and forms a dot at late pachytene in the sex vesicle. After puberty the Y chromosome decondensed again but now in Sertoli cells. This has been shown in mice (Guttenbach et al., 1989, 1993) and human (Speed et al., 1993) and can therefore be discussed as an important functional feature of the Y chromosome structure in the male germ line. Since in the human the same condensation gradient was visible with the Y-specific painting probe pBSY, it was assumed that the whole Y chromosome structure expressed a dynamic condensation cycle during its X-Y pairing process (Speed et al., 1993). A similar decondensation cyle of the X chromosome at zygotene to late pachytene was recently shown by FISH analysis with an X chromosomal painting probe (Armstrong et a]., 1994). Thereby the tips of the short and long arm of both sex chromosomes were in a distinct neighborhood as shown in Figure 17. Although similar X-Y spreading structures were repeatedly observed, pairing in XqYq was long thought to display an experimental artifact until a homologous DNA region of about 400 kb in XqYq was recently revealed by FISH experiments (Pedicini et al., 1991). Its recombination frequency now defined it as a second pseudoautosomal region of the sex chromosomes (Freije et al., 1992). The obvious involvement of the whole Y chromosome structure in the X-Y pairing process until formation of the sex vesicle now raised the question of whether AZF-His not only essential for its time-dependent formation (Speed and Chandley, 1990) but also at least one spermatogenesis locus in Yq 1 I ?
238
PETER H. VOGT
Figure 77. Electron-microscopic picture of the X-Y pairing complex at
stage / I according to Solari (1980),that is, early during meiosis. The length of the pairing structure in Xp-Xp (marked X-Y) corresponds to about one-third of the length of the Y chromosome. Pairing i s also observed in Xq-Yq. Granular structures are associated in Yqll and distinct part of the X chromosome (the picture was kindly provided to me by Peter Goetz).
Human Y Chromosome and Spermatogenesis
239
As seen in Figure 16 large protein granular structures are often observed at early pachytene attached to the long arm of the Y chromosome (e.g., Chandley, 1988), suggesting genetic activity in Yqll at this germ cell stage. According to our definition of the three phases of AZF activity, this stage overlaps distinctl’y with phase b. Expression of similar granular structures, but visible already in the light microscope are pictures of the genetic activity of Y chromosomal fertility genes is well known in different Drosophila species (Hennig et al., 1987). They also occurred before meiosis in the primary sperrnatocyte nucleus and, according to their distinct morphology, were called lampbrush structures (Meyer et al., 1961). In this context it is interesting to note that the position of the spermatogenesis locus AZFb in middle Yql 1 included five members of the pY6H sequence family: pY6H14, pY6H34, pY6H54, pY6H64, and pY6BS65K (Figure 14) isolated by using dhMiF1, a genomic DNA clone from a fertility gene structure on the Y chromosome of Drosophila hydei (Vogt et al., 1991b). In the primary spermatocyte nucleus of Drosophila hya’ei dhMiFl is transcribed and part of the lampbrush structure “cones.” The same DNA domain contains clusters of (CA), blocks (Figure 17). Since all pY6H sequences include at least one (CA), block and pY6BS65 is repetitive (Vogt et al., 1991b) Yql I-subinterval D13 should contain at least 13 (CA), blocks (one copy in pY6Hl4, pY6H34, and pY6H54, respectively, and 10 copies in pY6BS6YC; Kohler, 1994). Although the molecular extension of subinterval D 13 is not yet known, nor is the distance between the different pY6H loci, the assumption of a cluster of (CA), blocks in Yq 1 1-D 13 would draw a speculative parallel to the (CA), block enriched DNA domain of the lampbrush structure “cones” on the Y chromosome of Drosophila hydei, transcribed in the primary spermatocyte nucleus (Figure 18). Analysis of transcription of pY6H sequences in human testis tissue in the primary spermatocyte nucleus is more difficult to analyze than in the Drosophila species. Homologous poly-A’ RNA species have not been detected in Northern blots (Lewe and Keil, unpublished results) indicating that the amount of nuclear RNA in the human (if present) is certainly lower than in the primary spermatocyte nucleus of Drosophila. But dynamics in a DNA folding structure may not necessarily be linked to nuclear transcription process, but can also be triggered by specific nuclear proteins binding to distinct DNA regions as observed for the W chromosome in chicken (Mizuno et al., 1988). Specific nuclear
240
PETER H. VOGT
Figure 18. Transcript in situ hybridization of a ’H-CA or b 3H-GT to spermatocyte nuclei of D. hydei. a,c Phase contrast photographs, b,d light field photographs to visualize the label. In a, b, the lampbrush structure “cones“ (Co) laterally associated with the lampbrush structure “pseudonucleolus” (Ps) are strongly labeled with 3H-CA, indicating a clustering of (CA), nucleotides in the cones DNAdomain of the D. hydeiY chromosome (Huijser et al., 1987). Bar represents 10 urn.
proteins attached to pY6H DNA loci in middle Yql 1 have been detected (Kohler, 1994) and may be involved in such processes. Sequence analyses of different pY 6H sequences indicated numerous sequence blocks with simple sequence complexity. Such sequence compositions usually have a high potential to form non-B-helical secondary structures and suprahelical curvatures (Vogt, 1990). In summary, one gets the idea that the genomic DNA structure in middle Yq1 1 includes at least one distinct DNA domain (in subinterval
Human Y Chromosome and Spermatogenesis
241
D13), which may be exceptional in its chromatin conformation and susceptibility for changes. Therefore, I would like to postulate that one component of the expression ofAZFb may be the ability of the chromatin folding structure in middle Yq I 1 to change quickly its conformation for triggering or stabilizing the X-Y pairing structure in the premeiotic sex vesicle (AZFbchromatin code). This may be achieved by a stage-specific transcription process as visible in the primary spermatocyte nucleus of Drosophi/u and/or an inherent coding potential of the genomic DNA structure to develop a locus-specific folding structure. This hypothesis does not exclude the proposed spermatogenesis function of RBM gene copies in this Y region, but on the contrary, it may support it. If one assumes that RBM gene copies are already functional at the transcript level, they may not only help to decondense the chromatin structure in middle Yql 1, but also along the whole Yql 1 region and in proximal Yp, since they are spread in these Y regions. This may be a prerequisite functional element for the timely regulated initiation of the X-Y pairing process as observed by Speed and Chandley (1990). A similar function for ribosomal RNA genes as X-Y pairing sites during male meiosis was observed in Drosophila (McKee and Karpen, 1990). Analysis of the X-Y pairing structure in sterile males with interstitial deletions in middle Yql 1 now offers the first experimental possibility to test my hypothesis.
VII. SUMMARY AND CONCLUSIONS Since in human males with deletions in the short Y arm but presence of the SRY gene impairment of spermatogenesis is often observed in connection with other gonadal dysmorphic phenotypes such as hypospadia, micropenis, or cryptoorchidism, disruption of spermatogenesis may not always be the primary effect of a Y gene expressed in testis tissue. Disruption of the differentiation pathway of a somatic cell line in testis development by a Y gene defect may cause a disruption of spermatogenesis as a secondary effect as well. Y genes expressed in testis and located in the short chromosome arm (SRI: ZFI: and TSPY) may be candidates for this (for discussion see Bogan and Page, 1994). Y genes expressed in testis and located in Yql 1 may be different in this aspect and be functional only in spermatogenesis. In this aspect AZFu, AZFb, and AZFc as described in this review may be comparable to the Y chromosomal spermatogenesis loci as defined by Burgoyne (199 1).
242
PETER H. VOGT
A function of the human Y chromosome in male germ cell development was first postulated in 1976 by Tiepolo and Zuffardi. They recognized a terminal deletion of the long Y arm by analyzing the lack of the heterochromatic chromosome block (Yq 12) in six azoospermic males. At that time, no one believed in a genetic fimction of heterochromatin. And, to postulate a Y-chromosomal azoospermia factor (AZF), they had to assume that these deletions included at least the distal part of the adjacent euchromatin (Yqll). Since similar terminal deletions in Yq could then be found repeatedly in azoospermic males in other laboratories, the literature of AZF deletions increased rapidly and AZF became fixed to distal Yqll (Daniel, 1985; Bardoni et al., 1991; Ma et al., 1993; Reijo et al., 1995). At the time when my research interest shifted from the fertility function of the Drosophila Y chromosome to the fertility function of the human Y chromosome, I recognized that everyone thought ofAZFas one spermatogenesis locus. And many thought of AZF as only one gene expressed early in spermatogenesis. This was remarkable as it was obvious that the histology of testis in sterile males with deletions in Yql 1 was variable, ranging from the “Sertoli-cell-only” picture to the “oligospermic” picture (see Section IIIB and Figure 11). One key patient who (for me) pointed distinctly to the presence of more than one AZF locus in Yqll was a sterile male with a ring-(Y) chromosome and arrest of spermatogenesis at the pachytene stage (Chandley and Edmond, 1971). This male, later coded JOWAL, had certainly lost the terminal part of the long Y-arm including distal Yql 1. Thus, according to the definition of Tiepolo and Zuffardi, AZF should be deleted and depletion of germ cells in most tubuli was expected. But despite deletion ofAZF, spermatogenesis in this male seemed to remain undisturbed premeiotically. The testis tubuli of JOWAL were filled with spermatogonia and spermatocytes. If AZF located in distal Yql 1 was expressed early, during, or before spermatogonia proliferation, only complex rearrangements of the Y-chromosome in JOWAL would be able to explain the presence of the requested AZF locus in this male (Vogt et al., 1993). We then noticed that the Y breakpoints in azoospermic males with arrest of spermatogenesisduring spermatogoniaproliferation cluster in proximal Yql 1 and of those with arrest at the pachytene stage in middle Yql 1 (Vogt et al., 1993). This was intriguing but not sufficient to prove the presence of different AZF loci in Yq 1 1 due to the presence of AZF-Has defined in this review (section VI). But the assumption of the presence of different AZF loci in Yql 1 then got definite support after the
Human Y Chromosome and Spermatogenesis
243
detection of different interstitial microdeletions in Yql 1, occurring de now and in the same Yql I subregions in sterile patients with non-obstructive azoospermia or severe oligospermia (Figure 13). Since the phase of spermatogenic disruption visible in the testis tissue of patients with a deletion in proxinial Yql i is always different from the disruption phase in patients with a deletion in middle Yql 1 and distal Yql 1, respectively, and since the phase of disruption is the same for all males with the same Yqll microdeletion, I think it is now justified to extend and to differentiate the original definition of the Y-chromosomal azoospermia factor AZF given by Tiepolo and Zuffardi (1976) to the definition of at least three Y-chromosomal azoospermia factors (AZFa, AZFb, and AZFc). This thesis can now be tested experimentally by isolating the corresponding spermatogenesis genes which are supposed to be located in the various microdeletions observed. Analyzing the molecular pathology of these genes in the testis tissue will then have to prove or disprove whether their disruption is able to cause azoospermia. Three AZF candidate genes were already isolated (RBM. Ma et al., 1993; SMCY: Agulnik et al., 1994; DAUSPGYI: Reijo et al., 1995; Maiwald et al., 1996).The RBMlocus is a multi-copy locus on the human Y chromosome with polymorphic sequence copies (Nakahori et al., 1994; Inglis et al., 1994b; DAZ/SPGY Reijo et al., 1995; Maiwald et al., 1996). Therefore, if RBM can cause azoospermia at all, the amount of R B k gene copies essential for human spermatogenesismay be variable. In this review I tried to present some evidence that functional RBMgene copies are clustered in middle Yq 11and not in distal Yq 1 1. This Y position would relate RBMgene(s) expression to the expression ofAZFb, and not AZFc as stated by Ma et al. (1993). A candidate status for AZFb is also suggested by the expression of RBMZ in spermatocytes (Chandley and Cooke, 1994). This does, of course, not exclude that further functional RBMgene copies are spread in proximal and distal Yql 1 and Yp. Due to their mapping positions in proximal Yq 1 1, they seem not to be involved in the expressionofAZFa. Apossible contribution ofone RBMgene copy mapped in distal Yq 1 1 to the expression of AZFc remains to be analyzed. I could present some evidence that the second AZFb candidate gene, SMCY, may be more a candidate gene for the HY locus than for AZFb, although a fhnctional relationship between both cannot be excluded at the moment. Since the HY locus is supposed to be involved early during embryogenesisbefore gonadogenesisand spermatogenesis,the question is raised as to whether there is not a functional relation between the HY locus and the Y-chromosomal gonadoblastoma locus GBY, defined re-
244
PETER H. VOGT
cently by Page (1987). The function of GBY is assumed to be expressed before spermatogenesis. In 30% of individuals with XY gonadal dysgenesis GBY expression causes gonadoblastomas. This percentage is even raised to 55% in HY antigen positive patients (Verp and Simpson, 1987).Arelation between the tumorization potential ofgonadal remnants and the level of expression ofH- Yantigen is therefore suggestive(Amice et al., 1986). At least some cases of gonadal dysgenesis in XY females with deletion of SRY would then be explicable by function of the HY locus in early embryogenesis as discussed above, leading to a “wrong growth rate” of female gonad cells. It is interesting to note that the GBY locus was mapped to different sites in Yq and Yp (Yq12: Arce et al., 1991; distal Yqll: Sekine et al., 1992; middle Yqll: Nagafuchi et al., 1992; close to the centromere in Yq or Yp: Petrovic et al., 1992). Expression of the GBY locus may therefore also be related to expression of one AZF locus in Yql 1. His proposed position in Yp was discussed as indicative for a functional relation to expression of the TSPY gene family in proximal Yp (Tsuchiya et al., 1994). As SRY is discussed as a genetic switch-signal for male development, the expression of AZF genes may be discussed as important genetic switch-signals for spermatogenesis. Expression of AZFu is expected to be involved in the proliferation of spermatogonia and/or differentiation before and during puberty. The testis of AZFa patients are usually very small, suggesting disruption of spermatogenesis before puberty. Since only Sertoli cells were observed in their testis tubuli, it is dificult to assess whether deletion of AZFu disrupts spermatogenesis before the proliferation phase of A spermatogonia after birth or already in embryogenesis during prespermatogenesis. However, AZFa patients usually come to the infertility clinic 15-20 years after puberty. Therefore, the observed depletion of spermatogonia may be also only a secondary age-dependentdegenerationeffect. Afunction ofAZFu as spermatogonia stem-cel1 regulator (see Section IIC) controlling the level ofAp spermatogonia after their further differentiationat puberty is an attactive speculation. Also possible is its involvement as initiation factor in the spermatogonia differentiation process. A study of these possibilities, should be possible by analyzingthe stage specific morphology of sertoli cells as displayed in Figure 2. In any case it is most likely that the expression of AZFu indeed serves as an important premeiotic switchsignal in human spermatogenesis. The function of a switch-signal could also be attributed to the expression ofAZFb.Whatever its gene product or function may be, it seems to
Human Y Chromosome and Spermatogenesis
245
serve as a switch signal for germ cells to enter meiosis. Transcripts of RBM genes and a specific chromatin structure in middle Yqll were discussed as most likely essential components of the AZFb function. Another component is what I called the AZF-XY function, controlling the time and initiation of the XY pairing structure in Xp and Yp at leptotene. Which genetic signal will act first and whether the components are interdependent or independent from each other can now be tested experimentally. The function of the spermatogenesis locus AZFc in distal Yqll is difficult to discuss at the moment since testis histology does not reveal a distinct disruption phase, but the presence of nearly all germ cell stages until motile spermatozoa, although in very low amounts. SPGYl is shown to be expressed in postmeiotic germ cells (Figure 16). In our patient collective a microdeletion in distal Yqll overlapping to the location of AZFc was also found in the father of a sterile male with deletion of AZFc (Vogt et al., 1996). It may therefore be possible that premutations in the AZFc locus are already happening in the patient’s father, leading to a gradual reduction of his number of spermatozoa, but remaining still high enough for producing his offspring. Due to a second mutation in the same gene or perhaps another gene of the AZFc locus, most likely occurring during early embryogenesis of the son, sperm numbers are reduced more or are deleted totally at spermatid stages. It is well known in human spermatogenesis that reduction in the amount of mature sperms can cause hrther degenerative effects on spermatids and premeiotic germ cells (Roosen-Runge, 1977). This is known especially for germ cells with abnormal chromosome numbers, like those expected in males with a 47,XXY karyotype (Klinefelter syndrome). These effects are age-dependent (Foss and Lewis, 1971). Males with deletions of AZFc must therefore be analyzed after birth or at least at puberty in order to elucidate their primary disruption phase. But as mentioned at the beginning of this review, humans are not an experimental species. Thus, it may take some time to collect these key patients and it may even take more time to analyze their testis histology and to develop a molecular AZF gene(s) therapy.
ACKNOWLEDGMENTS I would like to thank all my friends and colleagues,especiallyDr. Ann Chandley, Octavian Henegariu, Stefan Kirsch, Franklin Kiesewetter, Mike Kohler, Frank Kohn, Bob Speed, Professor E. Nieschlag, Professor R. A. Pfeiffer, and Profes-
PETER H. VOCT
246
sor W. B. Schill whose ideas and experiments have contributed so much to the picture of the human Y chromosomal spermatogenesis function as presented in this review. I am indebted to Barbara and Werner Hilscher for sharing and discussing their experiences and insights in human gametogenesis with me in numerous phone calls and privatissime lectures and for the critical reading o f this manuscript. Mrs. K. Vogt is acknowledged for her support in preparing the final version and Mrs. M. Lebkuchner and Mrs. A. Wiegenstein are acknowledged for excellent photographic assistance.
REFERENCES Affara, N.A., Florentin, L., Morrison, N., Kwok, K., Mitchell, M., Cooke, A., Jamieson, D., Glasgow, L., Meredith, L., Boyd, E., & Ferguson-Smith, M.A. (1986). Regional assignment of Y-linked DNA probes by deletion mapping and their homology with X-chromosome and autosomal sequences. Nucl. Acids Research l4( 13), 53535373. Agulnik,A.I., Mitchell,M.J., Lerner,J.L., Woods,D.R.,&Bishop,C.E.( 1994a).Amouse Y chromosome gene encoded by a region essential for spermatogenesis and expression of male-specific minor histocompatibility antigens. Hum. Mol. Genet. 3(6), 873-878. Agulnik, A.I., Mitchell, M.J., Mattei, M.-G., Borsani, G., Avner, P.A., Lemer, J.L., & Bishop, C.E. (1994b). Anovel X gene with a widely transcribed Y-linked homologue escapes X-inactivation in mouse and human. Hum. Mol. Genet. 3(6), 87W384. Amice, V., Amice, J., Bercovici, J.-P., Riviere, D., & Corolleur, M.-J. (1986). Gonadal tumor and H-Y antigen in 46,XY pure gonadal dysgenesis. Cancer 57, 1313-13 17. Anderson, M., Page, D.C., Pettay, D., Subrt, I., Turleau, C., Grouchy, J., & de la Chapelle, A. (1988). Yautosome translocations and mosaicism in the aetiology of 45,X maleness: Assignment of fertility factor to distal Yq 11. Hum Genet. 79, 2-7. Armstrong, S.J., Kirkham, A.J., & Hulten, M.A. (1994). XY chromosome behaviour in the germ-line of the human male: A FISH analysis of spatial orientation, chromatin condensation and pairing. Chromosome Research 2,445-452. Arnemann, J., Jakubiczka, S., Thiiring S., & Schmidtke, J. (1991). Cloning and sequence analysis of a human Y-chromosome-derived, testicular cDNA, TSPY. Genomics 11, 108-114. Baker, T.G. (1978). Effects of ionizing radiations on mammalian oogenesis: A model for chemical effects. Environ. Health Perspect 24,3 1-37. Ballabio, A., Carrozzo, R., Gil, A., Gillard, B., Affara,N., Ferguson-Smith, M.A., Fraser, N., Craig, I., Rocchi, M., Romeo, G., & Andria, G. (1989). Molecular characterization of human X N translocations suggests their aetiology through aberrant exchange between homologous sequences on Xp and Xq. Ann. Hum. Genet. 53, S14. Bardoni, B., Zuffardi, O., Guioli, S., Ballabio, A,, Simi, P., Cavalli, P., Grimoldi, G., Fraccaro, M., & Camerino, G. (1991). A deletion map of the human Yql I region: Implications for the evolution of the Y chromosome and tentative mapping ofa locus involved in spermatogenesis. Genomics 11,443-45 1.
Human Y Chromosome and Spermatogenesis
247
Bernstein, R. (1985). X;Y chromosome translocations and their manifestations. In: The Y Chromosome, Part B: Clinical Aspects of Y Chromosome Abnormalities (Sandberg, A.A., Ed.). Alan R. Liss, Inc., New York, pp. 171-206. Berrios, S., & Fernandez-Donoso, R. (1990). Nuclear architecture of human pachytene spermatocytes: Quantitative analysis pf associations between nucleolar and XY bivalents. Hum. Genet. 86, 103-1 16. Bidlingmaier, F., & Hilscher, W. (1989). Endocrinological and morphological development of the human testis during the first two years of life. In: Reproductive Biologv and Medicine (Holstein, A.F., Voigt, K.D., & Grasslin, D., Eds.). Diesbach Verlag, Berlin, pp. 3443. Bogan, J.S., & Page, D.C. (1994). Ovary? Testis?-A mammalian dilemma. Cell. 76, 603-607. Bonaccorsi, S., & Lohe, A. (1991). Fine mapping of satellite DNA sequences along the Y chromosome of Drosophila melanogaster: Relationships between satellite sequences and fertility factors. Genetics 129, 177-189. Buhler, E.M. (1985). Clinical and cytologic impact of Y-chromosome abnormalities. In: The Y Chromosome. Part B: Clinical Aspects of Y Chromosome Abnormalities (Sandberg, A.A., Ed.). Alan R. Liss, Inc., New York, pp. 61-93. Burgoyne, P. (1982). Genetic homology and crossing over in the X and Y chromosomes of mammals. Hum. Genet. 61,85-90. Burgoyne, P.S. (1993). A Y-chromosomal effect on blastocyst cell number in mice. ,Development II,34 1-345. Burgoyne, P.S., Buehr, M., Koopman, P., Rossant, J., & McLaren, A. ( 1988).Cell-autonomous action of the testis determining gene: Sertoli cells are exclusively XY in XX-XY chimaeric mouse testes. Development 102,443-450. Burgoyne, P.S., Mahadevaiah, S.K., Sutcliffe, M.J., & Palmer, S.J. (1992). Fertility in mice‘requires X-Y pairing and a Y-chromosomal “spermiogenesis” gene mapping to the long arm. Cell 71,39 1-398. Bustos-Obregon, E., Courot, M., Flechon, J.E., Hochereau-de-Reviers, M.T., & Holstein, A.F. (1975). Morphological appraisal of gametogenesis. Spermatogenetic process in mammals with particular reference to man. Andrologia 7(2), 141-163. Cantrell, M.A., Bogan, J.S., Simpson, E., Bicknell, J.N., Goulmy, E., Chandler, P., Pagon, R.A., Walker, D.C.,Thuline, H.C., Graham, J.M. Jr., DeLaChapelle,A., Page, D.C., & Disteche, C.M. (1992). Deletion mapping of H-Y antigen to the long arm of the human Y chromosome. Genomics 13, 12551260. Caspersson, T., Zech, L., Johansson, J., Lindsten, J., Therkelsen, A., & Zech, L. (1971). Translocations causing non-fluorescent Y chromosomes in human XO/XY mosaics. Hereditas 68,281-304. Chandley, A.C. (1988). Meiosis in man. TIG 4(3), 79-84. Chandley, A.C., & Cooke, H.J. (1994). Human male fertility-Y-linked genes and spermatogenesis. Human Molecular Genetics 3, 1449-1452. Chandley, A.C., & Edmond, P. (1 97 I). Meiotic studies on a subfertile patient with a ring Y chromosome. Cytogenetics 10,295-304. Chandley, A.C., Ambros, P., McBeath, S., Hargreave, T.B., Kilanowski, F., & Spowart, G. (1986). Short arm dicentric Y chromosome with associated statural defects in a sterile man. Hum. Genet. 73,350-353.
248
PETER H. VOGT
Chandley, A.C., Gosden, J.R., Hargreave, T.B., Spowart, G., Speed, R.M., & McBeath, S. (1989). Deleted Yq in the sterile son of a man with a satellited Y chromosome (Yqs). J. Med. Gen. 26, 145-153. Chandley, A., Goetz, P., Hargreave, T.B., Joseph, A.M., & Speed, R.M. (1984). On the nature and extent of XY pairing at meiotic prohase in man. Cytogenet. Cell Genet. 38,241-247. Ciba Foundation Symposium. (1994). Germline Development, Vol. 182. John Wiley & Sons, Chichester, New York, Brisbane, Toronto, Singapore. Clepet, C., Schafer, A.J., Sinclair, A.H., Palmer, M.S., & Lovell-Badge, R. (1993). The human SRY transcript. Hum. Mol. Genet. 2,2007-2012. Clennont, Y. (1963). The cycle of the seminiferous epithelium in man. Am. J. Anat. 112, 35-51. Clermont, Y. (1966). Renewal of spermatogonia in man. Am. J. Anat. I 18,509-524. Clermont, Y., & Hermo, L. (1976). Spermatogonial stem cells and their behaviour in the seminiferous epithelium of rats and monkeys. In: Stem Cells of’Renewing Cell Population (Cairnie, A.B., Lala, P., & Osmond, D.G., Eds.). Academic Press, New York, pp. 273-286. Cortes, D., Muller, J., & Skakkebaek, N.E. (1987). Proliferation of Sertoli cells during development of the human testis assessed by stereological methods. International Journal ofAndrology 10, 589-596. Coyne, J.A. (1985). The genetic basis of Haldane’s rule. Nature 314,736738. d’Agostino, A., & Stefanini, M. (1990). An intercellular adhesion protein of rat testis: Isolation and preliminary characterization. Mol. Reprod. Dev. 25, 172-176. Daniel, A. (1985). Y isochromosomes and rings. In: The Y Chromosome, Part B: C h i c a l Aspects of Y Chromosome Abnormalities (Sandberg, A.A., Ed.). Alan R. Liss, Inc., New York, pp. 105-135. de Arce, M.A., Costigan, C., Gosden, J.R., Lawler, M., & Humphries, P. (1992). Further evidence consistent with Yqh as an indicator of risk of gonadal blastoma in Y-bearing mosaic Turner syndrome. Clin. Genet. 41,28-32. de Kretser, D.M., & Ken; J.B. (1988). The cytology of the testis. In: The Phzysiology of Reproduction (Knobil, E., & Neill, J., Eds.). Raven Press Ltd., New York, pp. 837-93 2. de Kretser, D.M. ( 1969). Ultrastructural features ofhuman spermiogenesis. Z. Zellforsch. 98,477-505. de Kretser, D.M., Burger, H.G., & Hudson, B. (1983). The Pmitaiy and Testis. Clinical and Experimental Studies. Springer Verlag, Berlin. de la Chapelle, A. (198 1). The etiology of maleness in XX men. Hum. Genet. 58, 105-1 16. de la Chapelle, A., Page, D.C., Brown, L., Kaski, U., Parvinen, T., & Tippett, A. (1986). The origin of 45,X males. Am. J. Hum. Genet. 38, 330-340. Diekmann, L., Palm, K., Pfeiffer, R.A., Trautmann, U., Scholz, W., Schroers, E., Vogt, I?, & Kohler, M. (1992). Multiple minute marker chromosomes derived from Y identified by FISH in an intersexual infant. Hum. Genet. 90, 181-183. Disteche, C.M., Brown, L., Saal, H., Friedman, C., Thuline, H.C., Hoar, D.I., Pagon, R.A., & Page, D.C. (1986a). Molecular detection of a translocation (Y; 15) in a 45,X male. Hum. Genet. 74, 372-377.
Human Y Chromosome and Spermatogenesis
249
Disteche, C.M., Casanova, M., Saal, H., Friedman, C., Sybert, V., Graham, J., Thuline, H., Page, D.C., & Fellous, M. (1986b). Small deletions of the short arm of the Y chromosome in 46,XY females. Proc. Natl. Acad. Sci. USA 83,7841-7844. Djakiew, D., & Dym, M. (1988). Pachytene spermatocyte proteins influence Sertoli cell function. Biology of Reproduction 39,, 1193-1 205. Faed, M.J.W., Lamont, M.A., & Baxby, K. (1982). Cytogenetic and histological studies of testicular biopsies from subfertile men with chromosome anomaly. J. Med. Genet. 19,4%56. Falin, L.I. (1969). The development of genital glands and the origin of germ cells in human embryogenesis. Acta Anat. 72, 195-232. Fawcett, D.W. (1975). Ultrastructure and function of the Sertoli cell. In: Male Fertility and Sterility (Mancini, R.E., & Martin, L., Eds.). Academic Press, New York, pp. 13-36. Ferguson-Smith, M.A., Affara, N.A., & Magenis, R.E. (1987). Ordering of Y specific sequences by deletion mapping and analysis of X-Y interchange males and females. Development I01 (Supplement),4 1-50. Foss, G.L., & Lewis, F.J.W. (197 I). A study of four cases with Klinefelter’s syndrome, showing mobile spermatozoa in their ejaculates. J. Reprod. Fert. 25,401-408. Freije, D., Helms, C., Watson M. S., & Donis-Keller, H. (1 992). Identification ofa second pseudoautosomal region near the Xq and Yq telomeres. Science 258, 178k1787. Fritz, I.B. (1982). Comparison of granulosa and Sertoli cells at various stages of maturation: Similarities and differences. Adv. Exp. Med. Biol. 147, 357-384. Fritz, I.B. (1994). Simatic cell-germ cell relationships in mammalian testes during development and spermatogenesis. In: Cellular lntersctions in Mammalian Testis. Fritz, I.B., & Tung, P.S. ( I 986). Role of interactions between peritubular cells and Sertoli cells in mammalian testicular functions. In: Gametogenesis and the Early Embryo (Gall, J., Ed.). Alan R. Liss, New York, pp. 151-173. Fryns, J.P., Cassiman, J.J., & Van den Berghe, H. (1978). Unusual in vivo rearrangements of the Y chromosome with mitotic instability in vitro. Hum. Genet. 44, 349-355. Fryns, P., Kleczkowska, A,, & Van den Berghe, H. (1985). Clinical manifestations of Ylautosome translocations in man. In: The Y Chromosome, Part B: Clinical Aspects of YChromosome Abnormalities (Sandberg, A.A., Ed.). Alan R. Liss, Inc., New York, pp. 2 1S 2 4 3 . Gabriel-Robez, O., Rumpler, Y., Ratomponirina, C., Petit, C., Levilliers, J., Croquette, M.F., & Couturier, J. (1990). Deletion of the pseudoautosomal region and lack of sex-chromosome pairing at pachytene in two infertile men carrying an X;Y translocation. Cytogenet. Cell. Genet. 54, 38-42. Gal, A., Weber, B., Neri, G., Serra, A., Miiller, U., Schempp, W., & Page, D.C. (1987). A 45,X male with Y-specific DNA translocated onto chromosome 15. Am. J. Hum. Genet. 40,477-488. Galdieri, M., Monaco, L., & Stefanini, M. (1984). Secretion of androgen binding protein by Sertoli cells is influenced by contact with germ cells. J. Androl. 5 , 4 0 9 4 15. Gonzales, J., Lesourd, S., & Dutrillaux, B. (198 I). Mitotic and meiotic analysis of a reciprocal translocation t(x3) in an azoospermic male. Hum. Genet. 57, 111-1 14. Guraya, S.S. ( 1987). Biologv of Spermatogenesis and Spermatozoa in Mammals. Springer Verlag, Berlin.
250
PETER H. VOGT
Guttenbach, M., Schmid, M., Jauch, A., & Vogt, P. (1989). The Y chromosome of the mouse is decondensed in Sertoli cells. Chromosoma 97,42%433. Guttenbach, M., Winking, H., & Schmid, M. (1993). Organization of the Y chromosome in testis cells offetal, subadult and adult mice as determined by in situ hybridization. Chromosoma 102,618-622. Hadziselimovic, F. ( 1977). Cryptorchidism. Ultrastructure of normal and cryptorchid testicular development. In: Advances in Anatomy, Embryology and Cell Biology, Vol. 53. Springer Verlag, Berlin, pp. 1-72. Hadziselimovic, F. ( 1983). Cryptorchidism. Management and Implications. SpringerVerlag, Berlin, Heidelberg, New York. Hartung, M., Devictor, M., Codaccioni, J.L., & Stahl, A. (1988). Yq deletion and failure of spermatogenesis. Ann. Genet. 31,2 1-26. Heller, C.G., & Clermont, Y. (1963). Spermatogenesis in man: An estimate of its duration. Science 140, 184-1 85. Heller, C.G., & Clermont, Y. (1964). Kinetics of the germinal epithelium in man. Recent Progress. Horn. Res. 20, 54S575. Henegariu, O., Hirschmann, P., Kilian, K., Kirsch, S., Lengauer, C., Maiwald, R., Mielke, K., & Vogt, P. (1994). Rapid screening of the Y chromosome in idiopathic sterile men, diagnostic for deletions in AZF, a genetic Y factor expressed during spermatogenesis. Andrologia 26,97-106. Hennig, W. (1989). The Y chromosome of Drosophila. In: Chromosomes: Eucaiyotic, Procaryotic and Ural, Vol. 1 (Adolph, K., Ed.). CRC Press, pp. 2 13-238. Hennig, W. (1 993). Conventional protein coding genes in the Drosophila Y chromosome: Is the puzzle of the fertility gene function solved? Proc. Natl. Acad. Sci. USA 90, 10904-10906. Hennig, W., Brand, R.C., Hackstein, J., Huijser, P., Kirchhoff, C., Kremmer, H., Lankenau, D.H., & Vogt, P. (1987). Structure and function of Y chromosomal genes in Drosophila. In: Chromosomes Today, Vol. 9 (Stahl, A,, Luciani, J.M., & VagnerCapodano, A.M., Eds.). Allen & Unwin, Boston, Sidney, Wellington, pp. 48-58. Hilscher, B. ( 1979). Spermatogoniogenesis, an interacting proliferation process between stem cell spermatogonia and differentiating Spermatogonia. Fohringer Symp., M i d l u d Alkersum. Hilscher, B., & Engemann, A. (1992). Histological and morphometric studies on the kinetics of germ cells and immature Sertoli cells during human prespermatogenesis. Andrologia 7, 7-10. Hilscher, W. ( 1981). T-1 -prospermatogonia (primordial spermatogonia of Rauh): The ameiotic counterpart of early oocytes. Adv. Androl. 7.2 1-32. Hilscher, W., & Hilscher, B. (1989). Some aspects of orthology and pathology of the Keimbahn. Reproductive Biology and Medicine, 2 1-33. Hilscher, W., & Hilscher, B. (1990). Details of the female and male pathway of the Keimbahn determined by enzyme histochemical and autoradiographic studies. Basic Appl. Histochem. 34,21-34. Holstein, A.F. (1976). Ultrastructural observations on the differentiation of spermatids in man. Andrologia 8, 157-165. Holstein, A.F., & Roosen-Runge, E.C. ( 1981). Atlas oj Human Spermatogenesis. Grosse Verlag, Berlin.
Human Y Chromosome and Spermatogenesis
251
Holstein, A.F., Wartenberg, H., & Vossmeyer, J. (1971). Zur Zytologie der pranatalen Gonadenentwicklung beim Menschen. 111. Die Entwicklung der Leydigzellen im Hoden von Embryonen und Feten. Z. Anat. Entwicklungsgesch. 135,4346. Huijser, P., Hennig, W., & Dijkhof, R. (1987). Poly(dC-dA/dG-dT) repeats in the Drosophila genome: A key function for dosage compensation and position effects? Chromosoma 95,209-2 15. Inglis, J., Prosser, J., Kun, M., Condie, A., Speed, B., Taylor, K., Hargreave, T.,Chandley, A., & Cooke, H. (1994). Structure, chromosomal distribution and variants of the YRRM genes. Cytogenet. Cell. Genet. 67,41. Jakubiczka, S., Schnieders, F., & Schmidtke, J. (1 993). A bovine homologue ofthe human TSPY gene. Genomics 17,732-735. Johnson, M.D., Tho, S.P.T., Behzadian, A., & McDonough, P.G. (1989). Molecular scanning ofYq I 1 (interval 6) in men with Sertoli-cell-only syndrome. Am. J. Obstet. Gynecol. 161,1732-1737. Kaluzewski, B., Jakubowski, L., Debiec-Rychter, M., Grzeschik, K.H., Limon, J., & Gibas, Z. (1988). Two mosaic cases with nonfluorescent Y chromosome analyzed with Y-specific DNA probes. Am. J. of Med. Gen. 31,489-503. Kancheva, L.S., Martinova, Y.S., & Georgiev, V.D. (1990). Prepubertal reat Sertoli cells secrete a mitogenic factor(s) that stimulates germ and somatic cell proliferation. Mol. Cell. Endocrinology 69, 12 1-127. Kirsch, S., Edelmann, A,, Keil, R., Henegariu, O., Hirschmann, P., Le Paslier, D., & Vogt, P.H. (1996). Quantitative blot analysis of DNA loci in YAC clones reveals acomplex repetitive DNA stiucture in middle Yq 1 1 and amplification events of different DNA domains in distal Yql I.In: Report of the Second International Workshop on Y chromosome mapping, Asilomar 1995; Cytogenet. Cell Genet. Forthcoming. Kohler, M.R. (1994). Molekulare und cytogenetische Analyse des pY6H65-DNA-Locus in Yq 11 mit Blick auf die Genstruktur von AZF, einem Fertilitatsfaktor im Y-Chromosom des Menschen. Thesis. Ruprecht-Karls-Universitat Heidelberg. Kohler, M.R., & Vogt, P.H. (1994). Interstitial deletions of repetitive DNA blocks in dicentric human Y chromosomes. Chromosoma. 103,324-330. Krco, C., & Goldberg, E.H. (1976). H-Y (male) antigen: Detection on eight-cell mouse embryos. Science 193, 1134-1 135. Laurent, C., Chandley, A.C., Dutrillaux, B., & Speed, R.M. (1982). The use of surface spreading in the pachytene analysis of a human t(y17) reciprocal translocation. Cytogenet. Cell Genet. 33,3 12-3 18. Limon, J., Gibas, Z., Kaluzewski, B., & Moruzgala, T. (1979). Demonstration of two different regions of lateral asymmetry in human Y chromosomes. Hum. Genet. 51, 247-252. Luciani, J.M., Devictor, M., & Stahl, A. ( 1977). Preleptotene chromosome condensation stage in human foetal and neonatal testis. J. Embryol. Exp. Morph. 38, 175-1 86. Ma, K., Sharkey, A., Kirsch, S., Vogt, P., Keil, R., Hargreave, T.B., McBeath, S., & Chandley, A.C. (1992). Towards the molecular localization of the AZF locus: Mapping of microdeletions in azoospermic men within 14 subintervals of interval 6 of the human Y chromosome. Hum. MoI. Genet. I, 2%33. Ma, K., Inglis, J.D., Sharkey, A., Bickmore, W.A., Hill, R.E., Prosser, E.J., Speed, R.M., Thomson, E.J., Jobling, M., Taylor, K., Wolfe, J., Cooke, H.J., Hargreave, T.B., & Chandley, A.C. (1993). A Y chromosome gene family with RNA-binding protein
252
PETER H. VOGT
homology: Candidates for the azoospermia factor AZF controlling human spermatogenesis. Cell 75, 1287-1295. Magenis, E., & Donlon, T. (1982). Nonfluorescent Y chromosomes. Cytological evidence of origin. Hum. Genet. 60, 133-1 38. Maiwald, R., Seebacher, T., Edelmann, A,, Hirschmann, P., Kohler, M.R., Kirsch, S., & Vogt, P.H. ( 1996). A human Y chromosomal gene mapping to distal Yql I is expressed in spermatogenesis. In: Report of the Second International Workshop on Y chromosome mapping, Asilomar, 1995; Cytogenet. Cell Genet. Forthcoming. Mancini, R. E., Narbaitz, R., & Levieri, J.C. (1960). Origin and development of the germinative epithelium and Sertoli cells in the human testis: Cytological, cytochemical and quantitative study. Anat. Rec. f36,477490. Matsuda, Y., Hirobe, T., & Chapman, V.M. (1991). Genetic basis of X-Y chromosome dissociation and male sterility in interspecific hybrids. Proc. Natl. Acad. Sci. USA 88,4850-4854. McGuinness, M.P., & Griswold, M.D. (1994). Interactions between Sertoli cells and germ cells in the testis. Developmental Biology 5 , 6 1 4 6 . McKee, B.D., & Karpen, G.H. (1990). Drosophilu ribosomal RNA genes function as an X-Y pairing site during male meiosis. Cell 61, 6 1-72. McLaren, A. (1991). Development of the mammalian gonad: The fate of the supporting cell lineage. Bioessays 13, 151-156. McLaren, A. ( 1994). Germline and soma: Interactions during early mouse development. Developmental Biology 5 , 4 3 4 9 . Meyer, G.F., Hess, O., & Beermann, W. ( 196I). Phasenspezifische Funktionsstrukturen in Spermatocytenkernen von Drosophilu melanogaster und ihre Abhangigkeit vom Y-Chromosom. Chromosoma 12,676-7 16. Mittwoch, U.(1969). Do genes determine sex? Nature 221,44&448. Mittwoch, U.(1989). Sex differentiation in inamrials and tempo of growth: Probabilities vs. switches. J. Theor. Biol. 137,445455. Mittwoch, U., & Mahadevaiah, S.K. (1992). Unpaired chromosomes at meiosis: Cause or effect ofgametogenic insufficiency? CytogeneticsandCell Genetics 59,274-279. Mizuno, S., Saitoh, H., & Harata, M. (1988). High affinity binding of W-protein to the bent, repetitive DNAsequences from avian W chromosomes. J. Cell. Biol. 107,313a. Mohandas,T.K., Speed, R.M., Passage,M.B.,Yen,P.H.,Chandley, A.C.,&Shapiro, L.J. (1992). Role of the pseudoautosomal region in sex-chromosome pairing during male meiosis: Meiotic studies in a man with a deletion of distal Xp. Am, J. Hum. Genet. 51,526-533. Morales, C., & Clermont, Y. (1 993). Structural changes of the Sertoli cell during the cycle ofthe seminiferous epithelium. In: TheSerfoliCell(Russel1, L.D., & Griswold M.D., Eds.). Cache River Press, Clearwater, FL, pp. 305-329. Muller, U., Fontaine, D., Adinolfi, M., Fraccaro, M., & Lalande, M. ( 1989). Analysis of complex Y chromosome aberrations using a single DNA probe (Y-367). Cytogenet. Cell Genet. 50, 161-164. Nagafuchi, S., Namiki, M., Nakahori, Y., Kondoh, N., Okuyama, A., & Nakagome, Y. (1993). A minute deletion of the Y chromosome in men with azoospermia. J. Urol. 1.50, 1155-1157.
Human Y Chromosomeand Spermatogenesis
253
Nagafuchi, S., Tamura, T., Nakahori, Y., et al. (1992). The majority of the marker chromosomes in Japanese patients with stigmata of Turner syndrome are derived from Y chromosomes. Hum. Genet. 89,59&592. Najmabadi, H., Stratheam, M., Naseeruddin, S., Sethi, S., Kouri, D., de Kretser, D., Baker, G., McLachlan, R.I., Loveland, K., Silber, S., & Bhasin, S. (1994). Substantial prevalence of Yq microdeletions in azoospermic men detected by a sequence tagged site (STS)-based mapping strategy. First Intemat. Y chromosome workshop, Cambridge; conference report in: Cytogenet. Cell Genet. Nakahori, Y., Kobayashi, K., Komaki, R., Matsushita, I., & Nakagome, Y. ( 1994). Alocus of the candidate gene family for azoospermia factor (YRRM2) is polymorphic with a null allele in Japanese males. Human Molecular Genetics 3, 1709. Neu, R.L., Matthew, M.S., Barlow Jr., M.D., & Gardner, L.I. (1973). A 46,XYq- male with aspermia. Fertility and Sterility 24, 81 1 4 13. Ngo, K.Y., Vergnaud. G., Johnson, C., Lucotte, G., & Weissenbach, J. (1986). A DNA probe detecting multiple haplotypes of the human Y chromosome. Am. J. Hum. Genet. 38,407-41 8. Nistal, M., & Paniagua, R. (1984). Occurrence of primary spermatocytes in the infant and child testis. Andrologia 16, 532-536. O’Reilly, A.J., Affara, N.A., Simpson, E.. Chandler, P., Goulmy, E., & Ferguson-Smith, M.A. ( 1 992). A molecular deletion map of the Y chromosome long arm defining X and autosomal homologous regions and the localisation of the HYA locus to the proximal region of the Yq euchromatin. Human Molecular Genetics I, 37%385. Ohno, S., Nagai, Y., Ciccarese, S., & Iwata, H. (1979). Testis organizing H-Y antigen and the primary sex-determination mechanism of mammals. Recent Prog. Horm. Res. 35,449. Oosthuizen, C.J.J., Herbert, J.S., Vermaak, L.K., Brusnicky, J., Fricke, J., du Plessis, L., & Retief, A.E. ( 1990). Deletion mapping of 39 random isolated Y-chromosome DNA fragments. Hum. Genet. 85, 205-210. Page, D.C. (1987). Hypothesis: A Y chromosomal gene causes gonadoblastoma in dysgenetic gonads. Development 101 (Supplement), 151-1 55. Page, D.C., Mosher, R., Simpson, E.M., Fisher, E.M.C., Mardon, G., Pollack, J., McGillivray, B., de la Chapelle, A., & Brown, L.G. (1987). The sex-determining region of the human Y chromosome encodes a finger protein. Cell 51, 1091-1 104. Pedicini, A., Camerino, G., Avarello, R., Guioli, S., & Zuffardi, 0. (1991). Probe St35-239 (DXYS64) reveals homology between the distal ends of Xq and Yq. Genomics 11,482483. Pelliniemi, L.J., Frojdman, K., & Paranko, J. (1993). Embryological and prenatal development and function of Sertoli cells. In: The Sertoli Cell (Russell, L.D., & Griswold, M.D., Eds.). Cache River Press, Clearwater, FL, pp. 365309. .,& J. (1987). Petit,C.,delaChapelle,A., Levilliers, J., Castillo, S.,Noel, €IWeissenbach, An abnormal terminal X-Y interchange accounts for most but not all cases ofhuman XX maleness. Cell 49, 59S602. Petrovic, V., Nasioulas, S., Chow, C.W., Voullaire, L., Schmidt, M., & Dahl, H. (1992). Minute Y chromosome derived marker in a child with gonadoblastoma: Cytogenetic and DNA studies. J. Med. Genet. 29, 542-546.
254
PETER H. VOGT
Pfeiffer, R.A., Lambertz, B., Friederiszick, F.K., Distel, H., Pawlowitzki, I.H., Nicole, R., Ober, K.G., & Ruckes, J. (1968). Die nosologische Stellung des XO/XY-Mosaizismus. Archiv f Gynakologie 206,36%410. Politzer, G. (1933). Die Keimbahn des Menschen. Z. Anat. Entw. Gesch. iOO,33 1-361. Reijo, R., Lee, T-Y., Salo, P., Alagappan, R., Brown, L.G., Rosenberg, M., Rozen, S., Jaffe, T., Straus, D., Hovatta, O., de la Chapelle, A., Silber, S., & Page, D. (1995). Diverse spermatogenic defects in humans caused by Y chromosome deletions encompassing a novel RNA-binding protein gene. Nature Genet. 10,383-393. Reitalu, J. (1973). A familial Y-22 translocation in man. Hereditas 74, 155460. Roosen-Runge, E.C., & Barlow, F.D. (1953). Quantitative studies on human spermatogenesis. I. Spermatogonia. Am. J. Anat. 93, 143-170. Roosen-Runge, E.C. (1977). Degeneration ofgerm cells. Polymorphism of spermatozoa. Genetic control. In: The Process ofSpermatogerresis. Cambridge University Press, pp. 145-153. Rossi, P., Dolci, S., Albanesi, C., Grimaldi, P., & Geremia, R. (1993). Direct evidence that the mouse sex-determing gene Sry is expressed in the somatic cells of male fetal gonads and in the germ cell line in the adult testis. Mol. Reprod. & Dev. 43,369-373. Rowley, M.J., Berlin, J.D., & Heller, C.G. (1971). The ultrastructure of four types of human spermatogonia. Z. Zellforsch. 112, 139-1 57. Russel, L.D. (1993). Morphological and functional evidence for Sertoli cell-germ cell relationships. In: The Sertoli Cell (Russell, L.D., & Griswold, M.D., Eds.). Cache River Press, Clearwater, FL. pp. 365-390. Sandberg, A.A. (1985). Clinical aspects of Y chromosome abnormalities. In: Progress and Topics in Cvtogenetics, Vol. 6, Part B. Alan R. Liss Inc., New York. Schulze, C., Holstein, A.F., Schirren, C., & Korner, F. (1976). On the morphology of the human Sertoli cells under normal conditions and in patients with impaired fertility. Andrologia 8(2), 167-178. Schulze, W. (1 978). Zuin Problem der morphologischen Charakterisierung von Spennatogonientypen beim Erwachsenen. Verh. Anat. Ges. 72,539-540. Schulze, W. ( 1 982). Evidence of a wave of spermatogenesis in human testis. Andrologia 14(2), 200-207. Sekine, T., Fukutani, K., Motegi, T., Hayakawa, H., Tamura, T., Nagafuchi, S., Nakahori, Y., & Nakagome, Y. (1992). DNA analysis of two patients with a non-fluorescent Y chromosome. Jpn. J. Human Genet. 37, 157-162. Sinclair, A.H., Berta, P., Palmer, M.S., Hawkins, J.R., Griffiths, B. L., Smith, M. J., Foster, J. W., Frischauf, A-M., Lovell-Badge, R., & Goodfellow, P. N. (1990). Agene from the human sex-determing region encodes a protein with homology to a conserved DNA binding motif. Nature 346,240-244. Smith, A., Fraser, I.S., & Elliot, G. (1979). An infertile male with balanced Y;19 translocation. Review of Yautosome translocations. Ann. Genet. 22, 189-1 94. Sohval, A.R. ( 1 964). Hermaphroditism with “atypical” or “mixed” gonadal dysgenesis. Amer. J. Med. 36,281-292. Solari, A.J. (1980). Synaptonemal complexes and associated structures in microspread human spermatocytes. Chromosoma 8 i , 3 1 S 3 3 7 . Speed, R.M., & Chandley, A.C. (1990). Prophase of Meiosis in human spermatocytes analyzed by EM microspreading in infertile men and their controls and comparisons with human oocytes. Hum. Genet. 84,547-554.
Human Y Chromosome and Spermatogenesis
255
Speed, R.M., Vogt, P., Kohler, M.R., Hargreave, T.B., & Chandley, A.C. (1 993). Chromatin condensation behaviour of the Y chromosome in the human testis. Chromosoma 102,421427. Stahl, A., Hartung, M., Devictor, M., & Berge-Lefranc, J.-L. (1984). The association of the nucleolus and the short arm of actocentric chromosomes with the XY pair in human spermatocytes: Its possible role in facilitating sex-chromosome acrocentric translocations. Hum. Genet. 68, 173-1 80. Stahl, A., Wachtler, F., Hartung M., Devictor, M., Schofer, C., Mosgoller, W., de Lanversin, A., Fouet, C., & Schwarzacher, H.G. (1991). Nucleoli, nucleolar chromosomes and ribosomal genes in the human spermatocyte. Chromosoma 101, 23 1-244. Steinbach, P., Fabry, H., & Scholz, W. (1979). Unstable ring Y chromosome in an aspermic male. Hum. Genet. 47,227-231. Steinberger, E. (1 97 1). Hormonal control of mammalian spermatogenesis. Physiol. Rev. 51, 1-22. Sumner, A.T., & Speed, R.M. (1 987). Immunocytochemical labelling of the kinetochore of human synaptonemal complexes, and the extent of pairing of the X and Y chromosomes. Chromosoma 95,359-365. Tiepolo, L., & Zuffardi, 0. (1976). Localization of factors controlling spermatogenesis in the nonfluorescent portion of the human Y chromsome long arm. Hum. Genet. 34, 11%124. Tran, D., Picard, J.Y., Campargue, J., & Josso, N. ( 1987). Immunocytochemical detection of anti-Mullerian hormone in Sertoli cells of various mammalian species including human. J. Histochem. Cytochem. 35,733-743. Trowbridge, I.S., & Omary, M.B. (1 98 1). Human cell surface glycoprotein related to cell proliferation is the receptor for transferrin. Proc. Natl. Acad. Sci. USA 78, 30393043. Tsuchiya, K., Sultana, R., Donlan, M., & Disteche, C.M. (1994). Gonadoblastoma: Molecular definition of the susceptibility region on the Y chromosome and role of TSPY. Am. J. Hum. Gen. 55 (supplement), 397 Vergnaud, G., Page, D.C., Simmler, M.-C., Brown, L., Rouyer, F., Noel, B., Botstein, D., de la Chapelle, A., & Weissenbach, J. (1986). A deletion map of the human Y chromosome based on DNA hybridization. Am. J. Hum. Genet. 38, lOW24. Verp, M.S., & Simpson, J.L. (1987). Abnormal sexual differentiation and neoplasia. Cancer Genet. Cytogenet. 25, 191-218. Viguie, F., Romani, F., & Dadoune, J.P. ( 1982). Male infertility in a case of ( y 6 ) balanced reciprocal translocation. Mitotic and meiotic study. Hum. Genet. 62, 22S227. Vogt, P. (1990). Potential genetic functions of tandem repeated DNA sequence blocks in the human genome are based on a highly conserved “chromatin folding code.” Hum. Genet. 84,301-336. Vogt, P.( 1992). Y chromosome function in spermatogenesis (review). In: Spermatogenesis-Fertilization-Contraception. Moleciila~Cellular and Endocrine Events in Male Reproduction, Vol. 4 (Nieschlag, E., & Habenicht, U.F., Eds.). Springer Verlag, Heidelberg, Berlin, New York, pp. 226-257. Vogt, P., Chandley, A.C., Hargreave, T.B., Keil, R., Ma, K., & Sharkey, A. (1991a). Towards the molecular localization of AZF, a male fertility gene on the human Y
256
PETER H. VOCT
chromosome by comparative mapping of microdeletions in the Y chromosome of men with idiopathic infertility. Am. J. Hum. Genet. 49 (supplement), 390. Vogt, P., Keit. R., Kohler, M.R., Lengauer, C., Lewe, D., & Lewe, G. (1991b). Selection of DNA sequences from interval 6 of the human Y chromosome with homology to a Y chromosomal fertility gene sequence of Drosophila hydei. Hum. Genet. 86, 34 1-349. Vogt, P., Chandley, A.C., Hargreave, T.B., Keil, R., Ma, K., & Sharkey, A. (1992). Microdeletions in interval 6 of the Y chromosome of males with idiopathic sterility point to disruption of AZF, a human spermatogenesis gene. Hum. Genet. 89, 491-496. Vogt, P., Keil, R., & Kirsch, S. (1993). The AZF-function of the human Y chromosome during spermatogenesis. In: Chromo.somes Today, Vol. 11 (Sumner, A.T., & Chandley, A.C., Eds.). Chapman & Hall, London, pp. 227-239. Vogt, P.H., Edelmann, A,,. Kirsch, S.,. Henegariu, O., Hirschmann, P., Kiesewetter, F., Kohn, EM., Schill, W.B., Farah, S., Ramos, C., Hartmann, M., Hartschuh, W., Meschede, D., Behre, H.M., Castel, A. Nieschlag, E., Weidner, W., Grone, H.-J., Jung, A., Engel, W., & Haidl, G. (1996). Human Y chromosome azoospermia factors (AZF) mapped to different subregions in Yql I . Hum Mol. Genet. 5,933-944. Vollrath, D., Foote, S., Hilton, A., Brown, L.G., Beer-Romero, P., Bogan, J.S., & Page, D.C. (1992). The human Y chromosome: A 43-interval map based on naturally occurring deletions. Science 258,52-59. Wachtel, S.S., Ohno, S., Koo, G.C., & Boyse, E.A. (1975). Possible role for H-antigen in the primary determination of sex. Nature 257,235-236. Wahlstrom, J. (1 985). Y N translocations and their cytologic and clinical manifestations. In: The Y Chromosome, Part B: Clinical Aspects of’Y Chromosome Abnormalities (Sandberg, A.A., Ed.). Alan R. Liss, Inc., New York, pp. 207-212. Wartenberg, H. (1978). Human testicular development and the role of the mesonephros in the origin of a dual Sertoli cell system. Andrologia 10, 1-2 1. Wartenberg, H. (1989). Differentiation and development of the testes. In: The Testis, 2nd ed. (Burger, H., & de Kretser, D., Eds.). Raven Press. Ltd., New York. Wartenberg, H., Holstein, A.F., & Vossmeyer, J. (1971). Zur Cytologie der pranatalen Gonadenentwicklung beim Menschen. 11. Elektronenmikroskopische Untersuchungen bber die Cytogenese von Gonocyten und fetalen Spermatogonien im Hoden. Z. Anat. Entwicklungesch. 134, 165-185. Willison, K., & Ashworth, A. (1 987). Mammalian spermatogenic gene expression. T I G 3,351-355. Wilson, M.G., Stein, R.B., & Towner, J.W. (1976). Ring Y chromosome without mosaicism. Birth Defects 12, 105-1 12. Witschi, E. (1948). Migration of the germ cells of human embryos from the yolk sac to the primitive gonadal folds. Cont. Embryol. 209,67-80. Yamada, K., Nanko, S.,Hattori, S., & Isurugi, K. (1982). Cytogenetic studies in a Y-to-X translocation observed in three members of one family, with evidence of infertility in male carriers. Hum. Genet. 6 0 , 8 5 9 0 . Yanagimachi, R. ( 1994). Mammalian fertilization. In: The Physiology ofReproduction, 2nd ed. (Knobil, E., & Neill, J.D., Eds.). Raven Press, Ltd., New York. Yunis, E., Garcia-Conti, L., Torres de Caballero, O.M., & Giraldo, A. (1977). Yq deletion, aspermia, and short stature. Hum. Genet. 39, I 17-1 22.
Human Y Chromosome and Spermatogenesis
257
Zenzes, M.T., Miiller, U., Aschmoneit. I., & Wolf, U. (1978). Studies on H-Y antigen in different cell fractions of the testis during pubescence. Hum. Genet. 45,297. Zhang, J.S., Yang-Feng, T.L., Miiller, U., Mohandas, T.K., deJong, P.J., & Lau, Y-F. C. (1992). Molecular isolation and characterization of an expressed gene from the human Y-chromosome. Hum. Mot. Genet. I, 7 17-726. Zwingman, T., Erickson, R.P., Boyer, T., & Ao, A. (1993). Transcription of the sexdeterming region genes Sty and Z& in the mouse preimplantation embryo. Proc. Natl. Acad. Sci. USA90, 814817.
This Page Intentionally Left Blank
INDEX
60A gene, 52-53 Abdominal-B gene, 85, 116 Abdominal-BI gene, 116 Abdominal-BII gene, I16 Acrosome, 206,208 Activin, 70, 151-152, 169-173 Activin A protein, 170-171 Activin BA, 169 Activin type I receptor, 62-63 Activin type I1 receptor, 62, 169171 Adrenal four-binding protein (Ad4BP), 14 Alleles, 50-51, 105-106 of dpp group genes, 45-47 Amnioserosa, 31, 37,45,47-48, 54 formation of, 60-61 Androgens, 4-5, 1 I Anterior terminal region, 129-131 Anterior-posterior axis, 27, 84 formation of, 152-156 polarity, 163-164 Anti-Mullerian hormone (AMH), 5 Area opaca, 160 Area pellucida, 160, 162 Armadillo (arm) gene, 85, I 19-120 Aromatase activity, 6 AZF, 193-194, 214, 229-236,242245 deletion mapping of, 222-229
AZFa, 194,224,225,241,243-244 AZFb, 194,224,225,239,241,243245 AZFc, 194,225,241,243,245 AZF-SY, 245 Azoospermia, 193 Azoospermia factor (see also “AZF...”) bHLH proteins, 39 Bicoid gene, 129-131, 164 Blastocyte, 153-155 BMP-1, 50, 126 BMP-2,44, 50, 63 BMP-4,44, 50, 70-72 BMP-6, 51 BMPs, 44,50,70, 126-127 Bone morphogenetic proteins (see also “BMPs.. .”) Brachyury, 173, 174-176, 180 Cactus (cact) gene, 35, 85 Cap ‘n’ collar (cnc) gene, 85, I 17 Caudal (cad) gene, 86, 117 Cell fates, 84 specification of, 110-1 13, 124-131 terminal, 100-101 Cellularization, 58-59 Central nervous system, 31 Central region of termini, 124-125
259
260
Chick embryo, 153-156 and fate maps, 156-160 and gene expression, 172-173 and goosecoid gene, 164-165 hypoblast of, 160-165 Chordin, 58, 72 Connective tissue growth factor (CTGF), 59 Co-repressor, 4 1 Corkscrew (csw) gene, 86, 106-107 Cripto factor, 171, 180 CT25Y, 234-236 Cubitis interruptus gene, 120 Cubitus interruptus/ Cell (ci/ Ce) gene, 86 Cut (ct) gene, 87, 117 Cuticle, 47, 120 Cuticular structures, 3 1 DAZ protein, 234-236 Decapentaplegic gene (see also “Dpp group genes.. .”) Decidua, 177 Decondensation of chromosomes, 237,241 Diplotene, 204 Dishevelled (dsh) gene, 87, 119 D1 gene (see also “Dorsal (dl) gene.. .”) DNA synthesizing spermatogonia, 202 Dorsal (dl) gene, 87, 127 binding sites, 38-40 gradient, 37-42,65-70 morphogen, 34-37 protein, 37-38 Dorsal epidermis, 31, 37,48 Dorsal group genes, 34-37 Dorsal midline cells, 58-59 Dorsal morphogen gradient, 34-37 Dorsal patterning of embryo, 4043,45-48, 51-52 Dorsal stripe, 60
INDEX
Dorsal ventral pattern, 67-70 Dorsalization, 48, 125 Dorsalmost region, 30, 54-55 Dorsoventral axis, 27,67, 84, 110, 125 formation of, 149 Dorsoventral genes, 32-33, 125-128 Dorsoventral patterning of embryo, 29-34,44 Downstream of receptor kinases (drk) gene, 88, 104-105 Dpp group genes, 40,42411, 126127, 128 activity gradient, 48, 65-70 receptors, 62-65 in vertebrates, 70-73 D-RAF, 106-107 Drosophila and dpp group genes, 27-73 and terminal genes, 84-132 Dsorl gene, 88, 107 Dsrc gene, 88, 107-108 Ectoderm, 159, 166, 181 development of, 111-114 embryonic, 180 Eed mutants, 174 Effectors, zygotic, 109-123 Embryo (see also “Specific types of embryos.. .”) axis, 84 and maternal factors, 177- I78 triple mutant, 68-70 Embryogenesis, 28, 34-37, 177-178 and primordial germ cell development, 197-198 Embryonic visceral endoderm, 172, 178-179 Empty spiracles (ems) gene, 88, 115 Endoderm, 158 development of, 111-121 formation of, 149 and maternal factors, 177
Index
Engrailed (en) gene, 89, 119 Epiblast, 153-163, 170, 180-181 Even-skipped (eve) gene, 89,115, I 18 Expression domains, 4 1-42 Extracellular signaling molecule, 44-45 Extraembryonic visceral endoderm, 178-179 Fate maps, 156-160 Feminization, 9-10, 12-13 Fertility, male, 21 1-222 Fetal hormones and sexual development, 3-6 Fetus and exposure to Mullerian inhibiting substance (MIS), 8 FGFs, 170-171 FISH analysis, 237 FKBP-12, 19,65 Folded gastrulation (fog) gene, 89 Forkhead (frk) gene, 89, 116, 117, 120-I2 1 Freemartinism, 8 Frog embryo and gastrulation, 148-152 and goosecoid gene, 164 fs(1)Nasrat (fs(1)N) gene, 89, 102104 fs(1)polehold (fs(1)ph) gene, 90, 102-104 Ftz-FI, 18, 19 Fused (fu) gene, 90, 120 Fushi tarazu factor 1 (Ftz-Fl), 1415 Fushi tarazu (ftz) gene, 90, 1 15, 118 Gametogenesis, human, 194-211 Gap1 gene, 90, 105 Gastrulation, 30, 34,43, 68, 128, 148-181 folded, 122 and mutations, 173-176
261
short, 54-58 twisted, 58-61 GBY, 243-244 Genes in the embryo, 31-73 expression of, 37-42 mutations of, 54,84-132 targeting in mice, 10-12 Genetic hierarchies, 84-132 Genital duct systems development of, 18 differentiation of, 2 Germ cells, 225,228 development in males, 192-245 Germ wall, 158, 160-162, 172-173 Germband, 121 retraction, 121-123 Giant (ft) gene, 90, 115 Gonadal differentiation, 14-15 Gonadotropin, 210-21 I Gonads, 15 Gooseberry (gsb) gene, 91, 120, 164 Goosecoid gene, 164-165, 173, 175176, 178 expression of, 167-169 GRB2, 107 GTPase activating protein (GAP), 105
GTP-binding proteins, 108 Hairy (h) gene, 91, 118 Haplo-lethal sensitivity, 47 Heat shock induction, 6 1 Hedgehog (hh) gene, 91, 120 Heterochromatin, 193 Heterodimers, 52 Hindsight (hnt) gene, 91, 123 HMG (high mobility group) proteins, 13-14, 41 HNF-4 (hepatocyte nuclear factor4) gene, 91, 117, 121, 166, 174
262
HNF3B (hepatocyte nuclear factor-3B) gene, 167-168, 175-176, 180-181 HNKl, 161 Hopscotch (hop) gene, 92, 119 Huckebein (hkb) gene, 92, 111-1 14, 116, 122 Hunchback (hb) gene, 92, 114 H-Y antigen, 233 H-Y locus, 232-233,243-244 Hypoblast, 160-165, 167, 178 Infertility of males, 11-12, 192-245 Insects and dorsalventral axis, 72 Interstitial microdeletions, 2 19-222 Knock-out mice, 10-12 Koller’s sickle, 160-165 Kruppel (Kr) gene, 92, 117 Lampbrush structures, 239-240 Leptotene phase, 204 Leydig cells, 201 differentiation of, 198, 210-21 1 function of, 10, 12 Lines (In) gene, 93, 114 l(1)polehole (l(1)ph) gene, 92, 102, 106 LOCUS, 50-51 Malpighian tubules, formation of, 116-117 MAP kinase cascade, 108-109 Marginal zone, 158, 160-161 Maternal dorsoventral genes, 127 Maternal factors, 177-178 Medea (Med) gene, 53-54 Meiosis, 192 Mesectoderm, 31, 34, 39,68-70 Mesoderm, 29,37, 68-72, 179 differentiation of, 38-40 formation of, 149-151 inducers of, 149-152, 156-160
INDEX
induction of, 169-172 mutations, 174-176 Microdeletions, 222-229, 243 Midgut, 120-121 MIS (see also “Mullerian inhibiting substance (MIS). ..”) MIS regulatory element 1 (MISRE-I), 14 Mitotic domains, 29-3 1 Morphogenesis, 121-1 23 Morphogens, 28,48-49 gradients, 27-28 Mothers against dpp (Mad) genes, 53-54 Mouse embryo, 153-156 and expression of factors, 169-172 and fate maps, 156-160 and mutations, 173-176 and primitive streak, 179-181 and visceral endoderm, 165-169, 178-179 Msd locus, 176 Mullerian ducts, 17-18 regression of, 2-5, 7 Mullerian inhibiting substance (MIS) assay for, 6-7 functional analysis of, 9-13 history of, 3-6 mutations in, 8-9 receptor for, 15-18 regulatory factors for, 13-15 sexual expression of, 7-8 Mutations, 47, 54-61,63-65, 68-70, 173-178 Naked (nkd) gene, 93, 120 Neuroectoderrn, 31,37,67-68 Nieuwkoop center, 149-151, 158159, 162-164 Nodal factor, 169, 171-172, 180 and mutations, 174-176 Nuclear proteins, 239-240
Index
Odd-paired (opa) gene, 93, 118 Odd-skipped (odd) gene, 93, 118119 Oogonia, 197 Organizer, 149, 151, 163-165, 181 inducers of, 156-160, 169 Ovaries in transgenic mice, 9 Oviducts, 10, 11, 13 PI transcript, 128 P2 transcript, 128 Pachytene phase, 204,206 Paired (prd) gene, 93, 119 Pair-rule genes, 1 17-1 18 Parasegments, 109-110 Patched (ptc) gene, 94, 119 Pellucida, 153, 158 Persistent Mullerian duct syndrome (PMDS), 8-9 Phenotypes m’utant, 85-99 ventralized, 45-48 Pointed gene, 128 Polyingression, 160 Porcupine (porc) gene, 94, I20 Posterior germ wall, 163 Posterior marginal zone epiblast, 163 Posterior terminal region, 129-131 Pre-Sertoli cells, 197-198 Primitive streak, 153-155, 158-159, 162, 163-164, 167-169, 173, 175, 179-181 Primordial germ cells (PGC), 197 Procollagen, 57 Prospermatogonia, 199 M prospermatogonia, 197 T prospermatogonia, 197 Protease encoding, .49-51 Pseudohermaphroditism, male, 8-9, 11 pY6H sequence family, 219-220, 239
263
RAS protein, 105 RASI protein, 105 RASl gene, 94 RBMl gene, 229-230 RBM2 gene, 230 RBM gene family, 193,229-232, 236,241 Receptor type I, 62 Receptor type II,62 Receptor type III,62 Receptors, 62-65 Reproductive tract, abnormalities of, 10-12 Rho gene, 39, 54-55, 68-70 Ring-Y chromosomes, 216 RNA injections, 48 Rolled (rl) gene, 94, 107 Runt (run) gene, 94, 119 Sax gene, 63-64 Sax gene receptors, 63 SCMY gene, 193 Sc-type cells, 208 SCWgenes, 49, 51-53 Segment polarity genes, 117-120 Seminiferous tubules, 199-201 Serine/ threonine kinases, 17 Serpent (srp) gene, 95, 120 Sertoli cells, 7,8,18,198-201,208-210 Sevenless protein, 104 Sexual differentiation, 2-20 Sexual dimorphic development, 1820 SF-1, 13-15, 19 SH2 domain, 104-105 Shaggy/zeste white 3 (sgg/zw3) gene, 95, 119 Short gastrulation (sog) gene, 3940,49, 54-58,66,95, 128 Shrew gene (see also “Srw gene.. .”) Signal transduction, 17-18,62, 6465, 100, 103, 104-108 conservation of, 108-109
264
Signaling molecule, 52 Slater (sla) gene, 63 Sloppy paired (slp) gene, 95, 119 SMCXgene,233-234 SMCYgene,232-234 Sna gene, 38-39,68 Son of sevenless (Sos) gene, 96, 105-106 Spalt (sal) gene, 96, 115-1 16 Spatzle protein, 35 Sperm, 11-12 Spermatids, 205-208,235,245 Spermatocytes, 196, 199-200, 204, 223 Spermatogenesis, 192-245, 207 human wave of, 195,211 Spermatogonia, 199, 223, 244 Ad spermatogonia, 199,202-204 AL spermatogonia, 202 Ap spermatogonia, 199,202-204 B spermatogonia, 199-200, 202, 204 Spermatogoniogenesis, 202-2 1 1 Spermatozoa, 245 SPGY IOCUS, 193,234-236 Spi gene, 69-70 Srw gene, 49,53 SRY gene, 13-14, 193-194, 198,244 Steroidogenic factor-1 (see also "SF-1.. .") Stripe of expression, I18 Supporting cell lineage, 197-198 Switch signals, 192, 198, 233, 244245 Symmetrization, 149, 153 Tailless (tll) gene, 96, 106-107, 11 I116, 122, 129 Tailup (tup) gene, 96, 123 Tenascin"/ odd Oz (ten"/ odz) gene, 97, 119 Terminal pathway, 84-132 maternally encoded, 100-109
INDEX
Testicular descent, 10-11 Testicular feminization, 12-13 Testis analysis of, 194-195 fetal, 3-6 formation of, 198 histology, 221-224 Testosterone, 5-6, 10, 19, 21 1 Tfm mutant mice, 12-13 TGF-B superfamily, 44, 51-53, 6465 cross talking, 15-17, 18 and receptor binding, 62-65 Tkv gene, 63-64 Tkv gene receptors, 63 Toll (tl) receptor, 35 Tolloid (tld) gene, 40,42,49-52,97, 126-127 Tolloid-related-1 (tlr-I) gene, 5 1 Tolloid-related-2 (tlr-2) gene, 5 1 Torso receptor, 100, 125 Torso receptor ligand, 102-104 Torso (tor) gene, 97, 100-104, 127 Torso'"' alleles, 127-128 Torsogofphenotypes, 111-1 15, 122123 Torso-like (tsl) gene, 98, 102-104 Torso'ofphenotypes, 125 Tramtrack (ttk) gene, 98, 115 Transcription, 57, 58-60,6467 Transcription factors, 108-120 Transcriptional activator, 38-40 Transcriptional repressor, 40-42 Transcripts, 63-64 Transferrin, 209 Transgenic mice and Mullerian inhibiting substance (MIS), 9-10 Transmembrane receptor, 100-102 T-related (trg) gene, 96, 116 Trophectoderm, 154-155 Trunk (trk) gene, 98, 102-103 TSPY gene, 193-194
Index
Tumors, suppression of, 12 Twi gene, 38-39,68 Twisted gastrulation (tsg) gene, 5860,98, 128 Tyrosine kinase, 84, 100-102, 125 Unpaired (upd) gene, 99 U-shaped (ush) gene, 99, 123 Uterine decidua, 170-171 Uterus in transgenic mice, 9 Ventral epidermis, 3 1, 34,45, 54 Ventral furrow, 29-30 Ventral pattern elements of embryo, 38-40 Ventral repression, 40-41 Ventralization, 64 of terminal region cells, 125 Vertebrates and dorsal ventral patterning, 70-13 Vg-l,70-71, 151-152, 169 Virilization of ovaries, 9-10 Visceral endoderm, 159, 165-169, . 174, 178-179, 180-181 Wingless (wg) gene, 99, 119 Wnt-11, 151-152 Wnt gene, 151, 169 Wolffian ducts, 2-6 Xenopus, 70-73 X-Y chromosome pairing, 236-238, 241, 245 translocation, 2 12-213
Y chromosome, 192-245 in male germ line, 236-24 1 microdeletions of, 222-229 mutations in, 21 1-222
265
Y-A translocation chromosomes, 2 18-219 Ynf chromosomes, 214,215 Yolk mass, 156-158 and storage of factors, 172 Yp chromosome arm, 212-213 Yp;Xp translocation chromosomes, 2 12-213 Yp;Yq translocation chromosomes, 213 Y q l l , 193, 242-243 microdeletions of, 222-229 mutations in, 219-222 Yq12,218 Yq chromosome arm, 214-222 monocentric, 214-215 Yq isochromosomes, dicentric, 215216 Yq;A translocation chromosomes, 218-219 Yq;Xp translocation chromosomes, 216-218 Y;X translocation chromosomes, 2 16-218 Y;Y translocation chromosomes, 213 Zerknullt (Zen) gene, 40, 58-61,99, 127 mutations of, 47 ZFY gene, 193-194 Zygotene phase, 204 Zygotic genes, 100, 110-1 11, 114123, 129
J A I P R E S S @
Advances in Developmental Biology Edited by Paul Wassarman, Department of Cell and Developmental Biology, Roche Institute of Molecular Biology Volume 1,1992, 192 pp. ISBN 1-55938-348-8
$109.50
CONTENTS: Introduction. Y Chromosome Function in Mammalian Development, Paul S. Burgoyne. A Super Family of Putative Developmental Signaling Molecules Related to the Proto-OncogeneWnt-l/int-1, Andrew P. McMahon and Kenneth R. Howard. Pattern Formation in Caenorhabditis Elegans, Min Han and Paul W. Sternberg. Gap Junctional Communication During Mouse Development, Norton 19.Gilula, Miyuki Nishi, and Nalin Kumar. Lens Differentiation and Oncogenesis in Transgenic Mice, Heiner Westphal. Subject Index.
Volume 2,1993, 195 pp. ISBN 1-55938-582-0
$109.50
CONTENTS: Preface, Paul M. Wassarman. The Sty Gene and Sex Determinationin Mammals, Blanche Capeland Robin Love//-Badge. Molecular and Genetic Studies of Human X Chromosome Inactivation, Carolyn J. Brown and Huntington f . Willard.Genomic Imprinting and Regulationof Mammalian Development, Coiin L. Stewart. Cell Interactions in Neural Crest Cell Migration, Marianne Bronner-fraser. Enzymes and Morphogenesis: Alkaline Phosphataseand Control of Cell Migration, Saul L. Zackson. Subject Index.
Volume 3, 1994, 194 pp. ISBN 1-55938-865-X
$109.50
CONTENTS: Preface, Paul M. Wassarman. Mechanisms of Neurogenesisin Drosophiia melanogaster, Jose A. CamposOrtega.The Role of Growth Factors in Mammalian Pregastrulation Development, Daniel A. Rappolee andZena Werb. Retinoid Signaling in Mouse Embryos, Elwood Linney and Anthony-Samuel LaMantia. RNA Localization During Oogenesis in Drosophila, Elizabeth R. Gavis and Ruth Lehmann. Actin as a Tissue-Specific Marker in Studies of Ascidian Development and Evolution, William R. Jeffery. Index.
Advances in DeveloDmental Biochemistry Edited by Paul Wassarman, Department of Cell and Developmental Biology, Roche Institute
of Molecular Biology, Volume 1, 1992,256 pp. ISBN 1-55938-347-X
$109.50
CONTENTS: Introduction. Organelle Assembly and Function in the Amphibian Germinal Vesicle, Joseph G. Gall. DNA Replication and the Role of Transcriptional Elements During Animal Development, Melvin L. DePamphilis. Transcriptional Regulation During Early Drosophila Development, K. Prakash, Joanne Topol, C.R. Dearolf, and Carl S. Parker. Translational Regulation of Maternal Messenger RNA, L. Dennis Smith. Gut Esterase Expression in the Nematode Caenorhabditis elegans, James D. McGhee. Transcriptional Regulation of Crystallin Genes: Cis Elements, Trans-factors and Signal Transduction Systems in the Lens, Joram Piatigorksy and Peggy S.Zelenka. Subject Index.
Volume 2, 1993, 237 pp. ISBN 1-55938-609-6
$109.50
CONTENTS: Preface, Paul M. Wassarman. Drosophila Homeobox Genes, William McGinnis. Structural and Functional Aspects of MammalianHOX Genes, Denis Duboute. Develop.mental Control Genes in Myogenesis of Vertebrates, Hans Henning-Arnold. Mammalian Fertilization: Sperm Receptor Genes and Glycoproteins, Paul M. Wassarman.The Fertilization Calcium Signal and How It Is Triggered, Michael Whitaker. Subject Index.
Volume 3, 1994, 188 pp. ISBN 1-55938-865-X
$109.50
CONTENTS: Preface, Paul M. Wassarman. Expression and Functionof Protein Kinases During Mammalian Gametogenesis, Deborah L. Chapman and Debra J. Wolgemuth. Regulation of the Dopa Decarboxylase Gene During Drosophila Development, Martha J. Lundell and Jay Hirsh. Transcription Factors in Mammalian Development: Murine Homeobox Genes, S.Steven Potter. Expression and Function of C-Mos in Mammalian Germ Cells, Geoffrey M. Cooper. Regulationof Pigmentation During Mammalian Development, Friedrich Beermann, Ruth Ganss, and Gunther Schutz. Index.
J A I
P R E S S @
J A I P R E S S
@
Advances in Molecular and Cell Biology Edited by E. Edward Bittar, Department of Physiology, University of Wisconsin, Madison Cell biology is a rapidly developing discipline which brings together many of the branches of the biological sciences that were separate in the past. The interrelations between cell structure and function at the molecular and subcellularlevel isthe central connecting theme of this series. “As the twentieth century nears its close”, writes Kenneth Miller, “the develop ment of an enormous range of tools and techniques, some physical, some chemical, some biological, has changed the situation forever. Cell biology today crosses the boundary, links the molecule with the organelle, associates the cellular response with the larger organism.” Most of the contributors are cell and molecular biologists who are currently engaged in research and have made important contributions to our present understanding of cell function. Volume 14, Physiological Functions of Cytochrome P450 in Relation to Structure and Regulation 1996,408 pp. $109.50 ISBN 0-7623-0113-9
Edited by Colin R. Jefcoate, University of W/sconsin, Madison CONTENTS: Preface and Overview, Colin R. Jefcoafe. Electron Transfer Proteins of Cytochrome P450 Systems, lsrael Hanukoglu.The Molecular Structureof P450’s. The Conserved and Vanable Elements, Sandra Graham-Lorence and Julian A. Peterson. Diverse Molecular Mechanisms Regulate the Expression of Steroid Hydroxylase Genes Required for Production of Ligands for Nuclear Receptors, Michael Waterman and Diane S. Keeney. Cytochrome P450S C. The Conversion of Cholesterol to Pregnenolone. Hormonal Regulation through Multiple Routes, Colin Jefcoate. Regulation of the Formationof Glucocorticoids and MinerealocorticoidsIn Vivo, Jean-Guy LeHoux, et a/. Physiology and Molecular Biology of P450c21 and P450cl7, Bon-chu Chung. The Aromatase Reaction, Evan Simpson, eta/. Regulation of Calcium Metabolism by the Vitamin D Hydroxylases, James Armbrecht, et a/. Regulation of Cholesterol Conversion to Bile Acids, John Chiang and Reno Vlahcevic. Eicosanoid Metabolism and Bioactivation by Microsomal Cytochrome P450, Jorge Capdevilla. Steroid Hormones and Other Physiologic Regulators of Liver Cytochromes P450: Metabolic Reactions and Regulatory Pathways, David J. Waxman.
Also Available: Volumes 1-13 (1987-1995)
$109.50 each
Advances in Structural Biology Edited by Sudarshan Malhotra, Department of Z ~ l mUniversity , of Alberta Advances in Structural Biology as a new serial publication may require some explanation. First, structural biology now has a very broad range and is highly diversified; and second, research papers on structure, biochemistry, and function often appear in widely disparate journals. Consequently, investigators in specialized areas often find it difficult to maintain an integrated view of their own field of work, let alone an overview of advances in the field. The editor and publisher recognize their responsibility to specialists and generalists in providing a dedicated forum for a wide range of representative current subjects in structural biology.
J A I
-From the Preface
Volume 3,1994, 311 pp.
$1 09.50
ISBN 1-55938-627-4
CONTENTS: Preface, Sudarshan K. Malhotra. Early Appearance of Neuropeptidesduring Development: Neuromodulatory and Neurotrophic Roles, Susan Kentroti and Antonia Vernadakis. Role of Membrane Calpain (CA2I"ActivatedNeutral Proteinase) in Central Nervous System, Naren L. Banik and Edward L. Hogan. Synaptic Ribbons of the Mammalian Pineal Gland: Enigmatic Organelles of Poorly Understood .Function, Kunwar P. Bhatnagar. Structural Characteristics of the Transcription Factor Tfiiia and Its Interaction With Nucleic Acids, MarkL. Brown andDavid P. Bazett-Jones.The Nuclear Estrogen Receptor of the Mammalian Uterus: Role in Transcriptional and Post-TranscriptionalControl of Gene Expression, Raghava V. Thampan. Peroxisomes in Biology and Medicine, lnderjit Singh. Biological Basis of Hypertrophic Scarring, Paul G.Scott, Aziz Ghahry, Margaret H. Chambers, and Edward E. Tredget. Multiple Origins of lntravacuolar Protein Accumulation of Plant Cells, Eliot M. Herman. Controls of Germination in Noncereal Monocotyledons, Darleen A. DeMason. Also Available: Volumes 1-2 (1991-1993)
P R E S S
$109.50 each
@
Advances in Gene Technology
J A I
P R E S S @
Edited by Peter J. Greenaway, Division of Pathology, PHLS Centre for Applied Microbiology and Research, Porton Down, England The techniquesassociatedwith genetic manipulationare now being applied to a large number of research and industrial problems. This new series aims to keep abreast of developments in the general fields of molecular biology and biotechnology by prov iding state-of-the-art reviews from scientists with specialist experience in particular areas. Volume 2, 1991, 250 pp. ISBN 1-55938-268-6
$109.50
CONTENTS: Preface, Pefer J. Greenaway. The Polymerase Chain Reaction, P.A. Kitchin and A. Almond. Constructionof Ordered Gene Libraries, N.G. Sfoker. Diagnosisand Preventionof Exotic Virus Diseases: Application of Recombinant DNA Technology, J.C. S. Clegg. Structure, Organization and Function of Transfer RNA Genes from the Cellular Slime and Mold Dictyostelium discoideum, R. Marschalek and T.Dingermann. Optimizing Post-TranslationalSteps of Gene Expression in Escherichia coli, J.EG. Mecarthy. Resistance of Herpesvirusand HIV to Antiviral Drugs, G. Darby and 5.A. Larder. The Molecular Genetics of Host Specificity in Plant Pathogenic Bacteria, A. Vivian. The Cloning and Structure of Genes from the Autotrophic Biomining Bacterium, Thiba cillus ferrooxidans, D.E. Rawlings, D.R. Woods,andN.P. Mjoli. Direct Sequencing of RNA Using Dideoxynucleotides, M.Kaatfinen, M.L. Solin, and 0.Makela. Extrachromosomal Elements of the Genus Chlamydia, M. Lusher, C.C. Storey and S.J. Richmond. The Cytomegalovirus Major Immediate Early Promoter and Its Use in Eukaryotic Expression Systems, G.W.G. Wilkinson and A. Akrigg.
Also Available: Volume 1 (1990)
$109.50
FACULTY/PROFESSIONAL discounts are available in the US. and Canada at a rate of 40% off the list price when prepaid by personal check or credit card and ordered directly from the publisher.
JAI PRESS INC.
55 Old Post Road No. 2 - P.O. Box 1678 Greenwich, Connecticut 06836-1678
Tel: (203) 661- 7602
Fax: (203) 661-0792
