Esx1, a Novel X Chromosome-Linked Homeobox Gene Expressed In

DEVELOPMENTAL BIOLOGY 188, 85–95 (1997) ARTICLE NO. DB978640 Esx1, a Novel X Chromosome-Linked Homeobox View metadata,Gene citation and Expressed similar papers at core.ac.uk in Mouse Extraembryonic brought to you by CORE provided by Elsevier - Publisher Connector Tissues and Male Germ Cells

Yuanhao Li,* Patrick Lemaire,† and Richard R. Behringer* *Department of Molecular Genetics, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77030; and †Laboratoire de Genetique et Physiologie de De´veloppement, Institut de Biologie du De´veloppement de Marseille, Campus de Luminy, F-13288 Marseille Cedex 09, France

A novel paired-like homeobox gene, designated Esx1, was isolated in a screen for homeobox genes that regulate mouse embryogenesis. Analysis of a mouse interspeciﬁc backcross panel demonstrated that Esx1 mapped to the distal arm of the X chromosome. During embryogenesis, Esx1 expression was restricted to extraembryonic tissues, including the endoderm of the visceral yolk sac, the ectoderm of the chorion, and subsequently the labyrinthine trophoblast of the chorioallantoic placenta. In adult tissues, Esx1 expression was detected only in testes. However, Esx1 transcripts were not detected in the testes of sterile W/W v mice, suggesting that Esx1 expression is restricted to male germ cells. In situ hybridization experiments of testes indicated that Esx1 transcripts were most abundant in pre- and postmeiotic germ cells. Hybridization experiments suggested that Esx1 was conserved among vertebrates, including amphibians, birds, and mammals. During mouse development, the paternally derived X chromosome is preferentially inactivated in extraembryonic tissues of XX embryos, including the trophoblast, visceral endoderm, and parietal endoderm. In addition, the X chromosome is transiently inactivated during the meiotic stages of spermatogenesis. Thus, the identiﬁcation of Esx1 provides a molecular entry point into a genetic pathway to understand X chromosome-regulated fetal–maternal interactions and male germ cell development. ᭧ 1997 Academic Press

INTRODUCTION (ICM) of the blastocyst (Pederson, 1986). The trophectoderm cells distal to the ICM that surround the blastocoel cavity The oocytes of therian mammals do not contain sufficient (mural trophectoderm) stop proliferating and differentiate nutrition to support embryogenesis to birth. Therefore, nu- into giant cells. In contrast, the trophectoderm cells adja- trition must be provided by the mother from the uterine cent to the ICM (polar trophectoderm) continue proliferat- environment and subsequently through the placenta. The ing to form the placental precursor, the ectoplacental cone. placenta provides maternal nutrition, facilitates fetal respi- ICM cells adjacent to the blastocoel cavity form the primi- ration, eliminates fetal waste, and protects the fetus. This tive endoderm, whereas the other cells of the ICM form the transitional organ is also an important source of hormones primitive ectoderm. The primitive endoderm gives rise to and enzymes that are essential for normal embryonic devel- the visceral endoderm and the parietal endoderm (Gardner opment. Although embryos with organ defects can survive and Rossant, 1979; Gardner, 1982). The parietal endoderm to term, defects in placentation usually result in embryonic and the trophoblast giant cells form an early placental struc- death, which reflects the importance of the placenta during ture, the parietal yolk sac. Gastrulation of the mouse em- embryogenesis (Guillemot et al., 1994; Copp, 1995; Schmidt bryo results in the formation of extraembryonic mesoderm et al., 1995; Uehara et al., 1995). that, together with the visceral endoderm, forms another In eutherian mammals such as the mouse, the first em- placental structure, the visceral yolk sac. In the mouse, the bryonic cell lineages to develop are those that contribute to function of the yolk sac placenta is subsequently replaced extraembryonic tissues with placental function. In morula- by the chorioallantoic placenta at Embryonic Day 10.0 stage embryos, peripheral cells are allocated to form troph- (E10.0) (Kaufman, 1992). ectoderm and more central cells to form the inner cell mass There are diverse forms of placentation among therian

0012-1606/97 $25.00 Copyright ᭧ 1997 by Academic Press All rights of reproduction in any form reserved. 85

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba 86 Li, Lemaire, and Behringer mammals. In marsupials, implantation and placentation By using the longest cDNA insert as a probe, we screened a mouse within the uterus occur relatively late in embryogenesis, E7.5 cDNA library (generously provided by John Gearhart, Johns which is coincident with organogenesis (Tyndale-Biscoe Hopkins) and isolated two identical cDNA phage clones. The and Renfree, 1987). In most marsupials, the trophoblast does cDNA inserts were subcloned into pBluescript II KS(0) (Stratagene) and sequenced by automated DNA sequencing. not invade the uterus and an apposed placentation leads to the formation of a yolk sac placenta. In eutherian mammals, the embryo implants relatively early in development, and Northern Blot and RT-PCR Analysis then the trophoblast invades the uterus to ultimately form a chorioallantoic placenta (Cross et al., 1994). Chorioallan- For Northern blot analysis, total RNA was prepared with toic placentas are generally divided into five types based RNAzol B (Biotecx Laboratories, Inc.) according to the manufac- turer’s instructions. Northern blots of total RNA (15 mg/lane) on the extent of fetal trophoblast invasion into the uterus were performed according to standard methods and hybridized (Billington, 1971). The biological reason for this diversity is with a 972-bp cDNA fragment (nt 465–1440, Fig. 1) in Rapid- unknown, although there is speculation that the various hyb buffer (Amersham Life Science) at 65ЊC. Uniform loading of types of placentas evolved to better facilitate fetal develop- RNA was monitored by ethidium bromide staining of ribosomal ment or to drive speciation by establishing a hybridization RNA. The size of the Esx1 transcripts was determined by com- barrier (Benirschke, 1983). parison with RNA size markers (GIBCO-BRL). For RT-PCR, total Failures of the embryo to implant or the placenta to form RNA(1–3 mg) was reversed transcribed using an oligo(dT) primer account for most prenatal lethality in humans and domesti- and amplified for 30 cycles using two Esx1 primers (shown in cated animal species (Cross et al., 1994). Recent analyses of Fig. 1). The integrity of the RNA and the efficiency of the reverse mouse gene knockouts, transgenic mice, and spontaneous transcription were monitored by amplification of hypoxanthine phosphoribosyltransferase (HPRT). The sequences of the HPRT mutations also suggested that there are important develop- -primers were 5؅-GTTGAGAGATCATCTCCACC-3؅ and 5؅ .mental checkpoints around the time of implantation and AGCTATGATGAACCAGGTTA-3؅ placentation (Copp, 1995). These observations suggest that extraembryonic development may be complex and subject to precise genetic control that ensures the normal develop- In Situ Hybridization ment of mammalian embryos. However, relatively few fac- In situ hybridization was performed on embryo sections with an tors specifically expressed in the extraembryonic lineages [a-35S]UTP-labeled antisense RNA probe, corresponding to nt 465– have been identified, perhaps because much more attention 1440. Pregnant Swiss random-bred mice were killed, and their em- has been given to the embryo proper than to extraembryonic bryos were fixed in 4% paraformaldehyde in phosphate-buffered tissues. Thus, to begin to define the genetic pathways that saline for 16 hr at 4ЊC. The embryos were then dehydrated with are essential for the development and function of the extra- a series of increasing concentrations of ethanol. The dehydrated embryonic lineages in mammals, more genes expressed in embryos were cleared in xylene before embedding in paraffin. extraembryonic tissues must be identified. Seven-micrometer-thick sections were mounted onto silane-coated In a search for homeobox genes that regulate early mouse slides (Histology Control Systems, Glen Head, NY). In situ hybrid- embryogenesis, we identified a new member of the paired- ization was then performed as described by Wilkinson et al. (1987) and Sassoon et al. (1988). Hybridization was performed at 63 C for like subfamily of homeobox genes. The gene was designated Њ 16 hr. A high-stringency wash was performed with 21 standard Esx1 because it was expressed during embryonic develop- saline citrate (SSC) and 50% formamide with 1.0 mM b-mercapto- ment in extraembryonic tissues (including the endoderm of ethanol at 65ЊC for 30 min. Nonhybridized RNA probe was digested the visceral yolk sac, ectoderm of the chorion, and labyrin- with an RNase A solution (20 mg/ml RNase A in 10 mM Tris, pH thine trophoblast of the chorioallantoic placenta) and in 7.4, 500 mM NaCl, and 1 mM EDTA) at 37ЊC for 30 min. High- adult tissues in the testis. In the testis, Esx1 expression was stringency washes with 21 SSC and 50% formamide with 0.5 mM restricted to pre- and postmeiotic germ cells. Esx1 mapped b-mercaptoethanol, then 21 SSC, and finally 0.11 SSC were per- to the distal end of the X chromosome. The established formed at 65ЊC for 30, 15, and 12 min, respectively. Autoradiogra- role of, as yet unidentified, X chromosome-linked genes in phy was performed with NTB-2 Kodak emulsion, and the slides placental development suggests that Esx1 may have an im- were exposed for 6–10 days at 4ЊC. The developed slides were coun- terstained with hematoxylin. Photographs were taken with both portant role in the differentiation of specific extraembry- light- and dark-field optics. onic tissues and also in male germ cell development.

Mouse Interspecific Backcross Mapping MATERIALS AND METHODS A mouse interspecific mapping panel was obtained from the Jackson Laboratory (Bar Harbor, ME). The mapping panel was com- Isolation of Mouse Esx1 cDNA Clones posed of genomic DNA from 94 backcross progeny from an interspecific cross between (C57BL/6J 1 SPRET/Ei)F1 hybrid female and A mouse E8.5 cDNA library (generously provided by Brigid Ho- SPRET/Ei male mice. The panel has been mapped for 2370 loci gan, Vanderbilt) was screened with an [a-32P]dCTP-labeled 230-bp that are well distributed among all of the autosomes and the X reverse transcription (RT)-polymerase chain reaction (PCR) frag- chromosome (Rowe et al., 1994). C57BL/6J and SPRET/Ei genomic ment (Fig. 1). Four overlapping cDNA phage clones were isolated. DNA were digested with eight different restriction endonucleases

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba Extraembryonic and Testis Expression of ESX1 87 and analyzed by Southern blot for informative restriction-fragment- stream of the transcription initiation site were a number of length polymorphisms (RFLP) by using the initial 230-bp RT-PCR specific binding sites for known transcription factors. fragment as a probe. An RFLP was detected with this probe between The cDNA contained a long open reading frame of 945 C57Bl/6J and SPRET/Ei genomic DNA digested with SstI. The bp (Fig. 1A) with a Kozak consensus translation initiation RFLP distribution pattern of Esx1 of the backcross progeny was site (Kozak, 1987) and encoded a homeodomain-containing communicated to the Jackson Laboratory for analysis. protein of 315 amino acids. Based on its expression pattern (see below), the gene was designated Esx1 (for extraembryonic tissue–testis–homeobox gene). The homeodomain of Zoo Blot Esx1 was most closely related to that of mouse S8 (Kongsu- Genomic DNA (3 mg) from Xenopus laevis, chicken, platypus wan et al., 1988) and Drosophila aristaless (Schneitz et al., (Ornithorhynchus anatinus), tammar wallaby (Macropus eugenii), 1993) (Fig. 1B). This placed Esx1 in the paired-like subfamily grey short-tailed opossum (Monodelphis domestica), mouse, rat, of homeodomain proteins, which includes goosecoid, Oct1, and human was analyzed by Southern blotting with a random- unc-4, and ceh-10 (Bu¨ rglin, 1994). Outside the homeodo- primed [a-32P]dCTP-labeled Esx1 cDNA probe (the 230-bp fragment main, Esx1 had no sequence homology with any known initially used to screen the cDNA libraries) in Rapid-hyb buffer at protein. Esx1 had a glutamic acid-rich region N-terminal to 65ЊC. The Southern blot was washed for 30 min three times with the homeodomain (Fig. 1A), and the C-terminal region of 51 SSC and 0.1% sodium dodecyl sulfate at 65 C. Њ the predicted protein was rich in proline residues (38 of 88, 43.2%), and included a 34-amino-acid PF/NP repeat motif. These acidic and proline-rich domains are characteristic of RESULTS the activation domains of transcription factors (Mitchell and Tjian, 1989). The presence of these domains suggested Isolation of Esx1, a Novel paired-Class Homeobox that Esx1 may function as a transcriptional activator. Gene The gene structure of Esx1 was determined by comparing the sequence of Esx1 genomic DNA with the Esx1 cDNA. The initial goal of this screen was to isolate the mouse Esx1 was encoded by four exons (Fig. 2). The homeobox was homologue of the Xenopus homeobox gene Siamois (Le- encoded by the three terminal exons and was interrupted maire et al., 1995). A partial Siamois cDNA fragment con- by two introns at the same positions as those of aristaless taining the homeobox region nt 437–668 (Lemaire et al., and the mouse homeobox gene Pem (Maiti et al., 1996). The 1995) was used to screen a mouse 129/SvEv genomic library second splice site in the homeobox was at the same position at moderate hybridization stringency. Eleven positive phage as in other paired-like homeobox genes (Bu¨ rglin, 1994). clones were isolated. Sequence analysis and comparison Taken together, these data indicated that Esx1 is a novel with known sequences in the available databases indicated paired-like homeobox gene. that one clone contained a novel homeobox gene. A pair of primers corresponding to sequences that flank the homeo- Esx1 Mapped to the X Chromosome box was used to generate a cDNA subfragment by RT-PCR of total RNA from different stage mouse embryos. A pre- To determine the chromosomal location of Esx1, a mouse dicted 230-bp RT-PCR fragment was amplified from E7.5 interspecific backcross panel was analyzed. Southern blots mouse embryos. The fragment was subcloned and se- of SstI-digested DNA of the 94 backcross progeny derived quenced to confirm the presence of the novel homeobox from matings of female (C57BL/6J 1 SPRET/Ei)F1 and male sequence. The 230-bp fragment was then used as a probe to SPRET/Ei mice were obtained from the Jackson Laboratory. screen a mouse E8.5 cDNA library. Four overlapping clones The blots were hybridized with the 230-bp RT-PCR frag- were isolated from the library; together, the clones gave rise ment. SstI fragments of 2.6 and 1.65 kb were used to follow to a cDNA sequence of 1083 bp. Because Northern analysis the inheritance of the Esx1 allele from C57BL/6J and indicated that the Esx1 gene encoded a 1.65-kb transcript SPRET/Ei, respectively. The haplotypes of the backcross (see below), the longest cDNA insert from the E8.5 library panel were compared with other alleles. The mapping re- screen was used as a probe to screen an E7.5 mouse cDNA sults indicated that Esx1 was located in the distal region of library. Two identical clones were obtained. The sequences the X chromosome, approximately 15 cM distal to Xce, and of these clones extended the 1083-bp cDNA sequence 3؅ by was linked to simple sequence length polymorphisms 240 bp and included a putative polyadenylation site (AAT- DXBir15, #Mustraa, and DXMit4 (Fig. 3). The ratio of the AAA) 15 bp upstream of a long poly(A) sequence. To define total number of mice with recombinant chromosomes to -the 5؅ end of the Esx1 transcript, additional genomic se- the total number of mice analyzed for each pair of loci sug quence was obtained and analyzed. This led to the identifi- gested that the most likely gene order was centromere– cation of a TATAA box 31 bp upstream of a putative tran- DXBir15 (2/94)–Esx1 (6/94)–#Mustraa, DXMit4. The re- scription initiation site. The position of the transcription combination frequencies (expressed as genetic distances in initiation site was then confirmed by RT-PCR analysis. The centimorgans (cM) { the standard error) were DXBir15, 2.13 1681-bp length of the composite cDNA is consistent with { 1.49; Esx1, 6.38 { 2.52; #Mustraa, DXMit4. There is no the 1.65-kb transcript identified by Northern analysis. Up- known mouse mutation that maps to this region of the X

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba 88 Li, Lemaire, and Behringer

FIG. 1. Structure of the mouse Esx1 cDNA. (A) The nucleotide and deduced amino acid sequence of the mouse Esx1 cDNA. The stop codon TGA is marked with an asterisk. The intron–exon boundaries are marked by solid triangles. The homeobox is boxed. The TATA box at nt 031 and the poly(A) signal at nt 1625 are shown. The acidic charged cluster is underlined with a solid line. The PF/NP repeat is underlined with a dashed line. The primers used to generate the 230-bp RT-PCR fragment to screen the cDNA libraries and to detect differences between C57BL/6J and SPRET/Ei are indicated by two opposing arrows. The putative transcription initiation site (indicated with an arrow) at nt 363 was deduced from the genomic sequence. Transcription of this region was conﬁrmed by RT-PCR with primers corresponding to the sequence immediately after the putative initiation sites and sequences of nt 655–674 (GenBank Accession No. AF004211). (B) Comparison of the predicted amino acid sequence of the mouse Esx1 homeodomain with the sequences of the homeodomains encoded by the paired-like homeobox genes: Drosophila aristaless (Schneitz et al., 1993), mouse S8 (Kongsuwan et al., 1988), mouse Mhox (Cserjesi et al., 1992), rat Cart 1 (Zhao et al., 1994), Xenopus Mix.1 (Rosa, 1989), mouse goosecoid (gsc) (Blum et al., 1992), Xenopus Siamois (Lemaire et al., 1995), and mouse Pem (MacLeod et al., 1990); paired-class homeobox genes: Drosophila paired (Frigerio et al., 1986), mouse Pax-6 (Walther et al., 1991), mouse Pax-7 (Jostes et al., 1990), and Drosophila antenopedia (antp) (Garber et al., 1983; Scott et al., 1983). The amino acids that match those of Esx1 are indicated by a dash. The percentage of amino acid similarity of each homeodomain compared with Esx1 is stated at the right.

AID DB 8640 / 6x29$$8640 07-07-97 13:31:50 dba Extraembryonic and Testis Expression of ESX1 89

W/W v germ-cell-deﬁcient mice. In this sterile mouse, mutations in the c-kit protooncogene impair the proliferation and migration of primordial germ cells during embryogenesis (Chabot et al., 1988; Yoshinaga et al., 1991). Esx1 transcripts were not detected by RT-PCR of RNA from the adult testes of W/W v mutant mice (Fig. 4C). These results suggest that testicular expression of Esx1 is restricted to germ cells.

Spatial Expression of Esx1 in Extraembryonic Tissues To examine the spatial expression pattern of Esx1 during embryogenesis, sense and antisense Esx1 riboprobes were FIG. 2. Esx1 gene structure. (Top) Restriction map of Esx1 geno- hybridized to 7-mm-thick tissue sections of embryos within mic DNA deduced from two overlapping phage clones, G08 and G132, that span a total of Ç8 kb. Exons are indicated by open boxes. (Bottom) Details of Esx1 exon/intron structure. The nucleotide positions of the exon/intron boundaries are indicated in Fig. 1. The approximate length of each of the three introns is given in parentheses. H3, HindIII; E, EcoRI; K, KpnI; RV, EcoRV.

chromosome and produces a mutant phenotype consistent with the Esx1 expression pattern. The distal region of the mouse X chromosome where Esx1 was mapped is syntenic with the human X chromosome at Xq21.33-22. No human syndromes that map to this region of the X chromosome are consistent with the expression pattern of Esx1.

Esx1 Expression in Extraembryonic Tissues and Adult Testis Initially, Esx1 expression was analyzed by Northern analysis of total RNA from embryonic and extraembryonic tissues between E8.5 and E18.5. Weak Esx1 expression was detected in E8.5 mouse embryos (data not shown). However, from E9.5 through E18.5, Esx1 expression was not detectable in the embryo proper (data not shown). In contrast, an abundant 1.65-kb transcript was detected in total RNA from extraembryonic tissues from E9.5 to E17.5 (Fig. 4A). To more precisely delineate Esx1 expression in the extraembryonic tissues, the placenta, the endoderm layer of the parietal yolk sac, and the endoderm and mesoderm layers of the visceral yolk sac of E13.5 embryos were physically isolated FIG. 3. Mouse Esx1 mapped to the distal end of the X chromo- (Hogan et al., 1994), and total RNA extracted from these some by interspecific backcross analysis. (A) The segregation pat- tissues was subjected to Northern analysis. Esx1 transcripts tern of Esx1 and flanking loci in 94 backcross animals that were were detected in placenta and endoderm of the visceral yolk typed for all loci. Each column represents the chromosome identi- sac but not in the mesoderm of the visceral yolk sac or the fied in the backcross progeny that was inherited from the (C57BL/ endoderm of the parietal yolk sac. Thus, Esx1 was specifi- 6J 1 SPRET/Ei)F1 mother. The shaded boxes represent the C57BL/ cally expressed in a subset of extraembryonic tissues. 6J allele and the open boxes represent the SPRET/Ei allele. The number of offspring that inherited each type of chromosome is To evaluate Esx1 expression in adult tissues, we per- listed at the bottom of each column. (B) Likely gene order, with formed a Northern analysis of total RNA from liver, brain, recombination distances between loci in cM shown to the left of spleen, ovary, kidney, testis, heart, muscle, skull, bone, and the chromosome. The standard error is indicated in parentheses. skin. Esx1 expression is detected only in testis (Fig. 4B). To (C) Localization of Esx1 relative to landmark loci in the integrated determine whether Esx1 is expressed in the somatic or the X chromosome map. The human syntenic region is shown to the germ cells of the testis, Esx1 expression was examined in right.

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba 90 Li, Lemaire, and Behringer

parietal yolk sac. Thus, Esx1 was expressed in speciﬁc subsets of extraembryonic cell lineages during mouse embryogenesis.

Esx1 Is Expressed in a Developmental Stage- and Spermatogenesis Cycle-Specific Manner in Male Germ Cells To determine the temporal expression pattern of Esx1 during the first wave of spermatogenesis, Esx1 transcripts were analyzed by RT-PCR of total RNA from the testes of pre- and postpubertal mice. Esx1 expression was first detected 9 days postpartum (dpp) and at all later ages examined (Fig. 4C). In the mouse, primitive type A spermatogonia are detected in the testes of prepubertal males at 6 dpp (Bellve´ et al., 1977). Two days later at 8 dpp, type A and type B spermatogonia are present. The first wave of germ cell meiosis begins at 10 dpp as defined by the presence of preleptotene and leptotene spematocytes (Bellve´ et al., 1977; McCar- rey, 1993). The RT-PCR data suggest that testicular Esx1 expression may include type A and type B spermatogonia, preleptotene, and leptotene spermatocytes, but not primitive type A spermatogonia (Bellve´ et al., 1977). To assess the spatial expression of Esx1 in the testis, in situ hybridization analysis was performed on transverse sections of adult testis. The sense Esx1 probe did not detect signals above background levels. The corresponding Esx1 antisense probe specifically hybridized to the seminiferous FIG. 4. Northern blot and RT-PCR analysis of Esx1 expression tubules but not the interstitial cells. This suggests that Esx1 during embryogenesis and in adult tissues. (A) Extraembryonic tis- was not expressed in Leydig cells, a finding consistent with sues. A single abundant 1.65-kb transcript is detected in visceral v yolk sac (YS) and placenta (PL) from E9.5 to E17.5. At E13.5, expres- the W/W testis RT-PCR analysis. Esx1 expression was sion is specifically detected in placenta and endoderm (VE) of the most abundant in a subset of tubules (Figs. 5G and 5H). The visceral yolk sac but not mesoderm (VM) of the visceral yolk sac mouse seminiferous tubules are arbitrarily divided into 12 or endoderm (PE) of the parietal yolk sac. (B) Adult tissues. A single stages according to the specific association of germ cells abundant 1.65-kb transcript is detected only in testis. (C) RT-PCR (Russell et al., 1990). The tubules at stages VII, VIII, and of total RNA from normal mouse testis of different ages and adult early IX had the highest Esx1 expression. Varying levels of W/W v mutant mice. Esx1 expression were found in tubules at stages II–VI. No Esx1 expression was found in stage X–I tubules. In the tubules with Esx1 expression, the hybridization signals were found predominantly in the preleptotene spermatocytes at the uterus between E8.5 and E10.5. No specific hybridiza- the basal region and the round spermatids in the more cen- tion signal was observed with the sense probe at any of the tral region surrounding the tubule lumen, so that the silver stages examined (data not shown). No hybridization signal grains were in an obvious concentric pattern. Esx1 expres- was detected with the antisense probe in any of the tissues sion was not detectable in the pachytene spermatocytes and of the developing embryos. At E8.5, Esx1 transcripts were the elongated spermatids. Considering the stages of the tu- detected in the ectoderm of the chorion and the proximal bules that expressed Esx1, it appears less likely that Esx1 base of the ectoplacental cone (Figs. 5A and 5B). Thus, the is expressed in spermatogonia (Russell et al., 1990). Previous isolation of Esx1 from the E8.5 cDNA library and the weak studies have demonstrated that X-linked genes undergo hybridization signal observed in E8.5 embryos by Northern transient inactivation during meiosis (Kierszenbaum and analysis may be due to the presence of contaminating extra- Tres, 1975). The absence of Esx1 expression in the pachy- embryonic tissues. At E9.5 and E10.5, Esx1 transcripts were tene spermatocytes may reflect this transient X-linked gene abundant in the endoderm layer but not the mesoderm layer inactivation. Thus, the spatial expression pattern of Esx1 of the visceral yolk sac (Figs. 5C–5F). In addition, Esx1 was within the seminiferous tubules, together with the observa- strongly expressed in the labyrinthine trophoblast (Figs. tion that Esx1 transcripts were not detected in the testis of 5C–5F). Esx1 transcripts were detected in trophoblast giant W/W v mutant mice, indicated that Esx1 was expressed in cells at E9.5 but were not detectable in these cells at E10.5. the germ cells of the testis and its expression correlates No Esx1 expression was detected in the endoderm of the with the activation status of the X chromosome.

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba Extraembryonic and Testis Expression of ESX1 91

FIG. 5. In situ hybridization analysis of Esx1 expression in mid-gestation. (A–D) Sagittal section through conceptus within the uterus at E8.5 (A and B) and E9.5 (C and D), respectively. (E and F) Sagittal section through the placenta at E10.5. (G and H) Cross section of seminiferous epithelium of an adult mouse testis. The stages of seminiferous epithelium are indicated by Roman numbers. (A, C, E, and G) Bright-field illumination. (B, D, F, and H) Dark-field illumination. Esx1-specific hybridization signals are detected in the laminae of the ectoplacental cavity, endoderm of the visceral yolk sac (ve), and labyrinthine trophoblast of the placenta (la). Esx1 expression is shown in the stage VII and VIII but not late stage IX seminiferous epithelium. In the Esx1-expressing seminiferous epithelium, hybridization signals are found in the preleptotene spermatocytes at the basal region and the round spermatids in the more central region. No hybridization signals are found in the pachytene spermatocytes and elongated spermatids. bi, blood island; ch, chorionic plate; dc, decidua; epc, ectoplacental cone; exc, exocoelom; gc, trophoblast giant cell; pe, endoderm of the parietal yolk sac; sgc, secondary trophoblast giant cell; sp, spongiotrophoblast. Scale bars: (A and B) 100 mm, (C and F) 300 mm, (G and H) 50 mm.

07-07-97 13:31:50 dba 92 Li, Lemaire, and Behringer

and persisted until at least E17.5. Although the cell lineages that contribute to the various layers of the extraembryonic membranes of the mouse are generally well characterized (Rossant, 1986), the embryological derivation of the labyrinthine trophoblast during placentation has not been established. The polar trophoblast gives rise to the extraembryonic ectoderm-derived ectoplacental cone and chorionic ectoderm. As the chorioallantoic placenta forms, the chorion collapses onto the ectoplacental plate as the allantois fuses with the chorion. Invasion of the chorionic plate by the allantoic vessels results in the formation of the labyrinthine layer. Thus, Esx1 was expressed in the ectoplacental cone FIG. 6. Conservation of Esx1 among vertebrates. Southern blot analysis of genomic DNA of Xenopus (X), chicken (C), platypus (P), and chorion as the labyrinthine trophoblast developed and opossum (O), wallaby (W), rat (R), human (H), and mouse (M) di- may provide a unique marker for this important placental gested with EcoRI. layer. In addition, the results presented here suggest that the labyrinthine trophoblast may be derived from the ectoplacental cone and chorionic ectoderm. The labyrinthine layer of the chorioallantoic placenta is Evolutionary Conservation of Esx1 among essential for embryonic development because this is the Vertebrates layer where the embryonic and maternal circulations come into closest contact. In this layer, the trophoblast cells and To investigate the evolutionary conservation of Esx1 in the endothelia of the embryonic blood vessels create the vertebrates, a Southern blot of representative vertebrates placental barrier. As gestation proceeds, the placenta grows was analyzed. Genomic DNA from Xenopus, chicken, a until E17.5. The labyrinthine layer also grows and ulti- monotreme (the platypus), two marsupials (the gray short- mately constitutes approximately 50% of the placenta tailed opossum and the tammar wallaby), and three euthe- (Kaufman, 1992). No mutations that block the formation of rian mammals (the mouse, rat, and human) was hybridized the labyrinthine layer of the placenta have been identiﬁed. with an Esx1 cDNA probe under stringent conditions. However, in Mash2-mutant mice, the spongiotrophoblast Unique hybridizing bands were detected in most of the ver- layer of the placenta is missing and the labyrinthine layer tebrates tested. These results suggest that Esx1 is conserved is underdeveloped. The mice die at about E10.5, which sug- among vertebrates. gests that they have a defect in the development of the chorioallantois (Guillemot et al., 1994; Cross et al., 1994). Mice with a mutation in scatter factor/hepatocyte growth DISCUSSION factor (SF/HGF), a potent hepatocyte mitogen that can cause epithelial cells to disperse (Warn, 1995), die during em- We describe here the isolation of Esx1, a novel X chromo- bryogenesis because of liver and placental defects (Schmidt some-linked mouse homeobox gene whose expression was et al., 1995; Uehara et al., 1995). The number of labyrin- restricted to a subset of extraembryonic tissues and male thine trophoblasts in the placentas of the mutants is dra- germ cells of the adult testis. Sequence and gene structure matically smaller than normal. This is probably an indirect analysis of Esx1 placed this homeobox gene in the paired- effect because SF/HGF is expressed in the allantois, not the like subfamily of homeobox genes (Bu¨ rglin, 1994). Classic labyrinthine trophoblast, and the SF/HGF receptor c-met is mutations and targeted mutagenesis in the mouse have re- expressed in the labyrinthine trophoblast (Uehara et al., vealed that this subfamily of homeobox genes is essential 1995). Bone morphogenetic proteins 8A and 8B (Bmp8A and in a variety of development processes, including skeletal Bmp8B) are also expressed in the labyrinthine trophoblast morphogenesis, neural tube closure, and kidney differentia- of the placenta (Zhao and Hogan, 1996). Bmp8b mutant tion (Balling et al., 1988; Rivera-Pe´rez et al., 1995; Yamada mice have defects in spermatogenesis but no abnormalities et al., 1995; Martin et al., 1995; Torres et al., 1995; Zhao in placentation (Zhao et al., 1996b). Perhaps Bmp8a can et al., 1996b). The two putative transactivation domains compensate for the absence of Bmp8b in the labyrinthine identiﬁed in Esx1 indicate that it may function as a tran- trophoblast. The expression of Esx1 in the ectoplacental scriptional activator. Thus, the data suggest that Esx1 regu- cone and progressive restriction to and its persistence in lates a genetic pathway that functions in the development the labyrinthine trophoblast suggest that Esx1 may act cell- of a subset of extraembryonic tissues and male germ cells. autonomously in the labyrinthine trophoblast to regulate During embryogenesis, Esx1 expression was found in sub- genes involved in the establishment or maintenance of the sets of extraembryonic cell lineages, including the endo- placental barrier. derm of the yolk sac and trophoblast. Abundant Esx1 ex- In adult tissues, Esx1 expression was detected in male pression in the trophoblast was progressively restricted to pre- and postmeiotic germ cells. This expression pattern the labyrinthine trophoblast of the chorioallantoic placenta suggests that this homeobox gene has a role in regulating

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba Extraembryonic and Testis Expression of ESX1 93 genes that influence postnatal spermatogenesis. The ap- et al., 1997). Another X chromosome-linked paired-class parent downregulation of Esx1 expression in the meiotic homeobox gene called Pem, which is specifically expressed stages is consistent with Esx1 having a role in regulating in extraembryonic tissues and the adult testis, may also be spermatogenic cell progression into meiosis. Alterna- important for placental development (Wilkinson et al., tively, Esx1 may function throughout spermatogenesis if 1990; Lin et al., 1994; Lindsey and Wilkinson, 1996). The its protein product is stable or its specific activity is al- data presented here suggest that the X chromosome-linked tered, as in the case of the enzyme HPRT (Shannon and Esx1 gene may also be an important regulator of extraem- Handel, 1993). If this is true, Esx1 may be important for bryonic tissue function. maintaining spermatogenesis. Maternal and paternal X chromosomes also behave differ- Esx1 may be conserved in nonamniote vertebrates. This ently during mammalian gametogenesis (for review see may not be surprising because males of these species pos- Handel and Hunt, 1992). During oogenesis, both X chromo- sess spermatogenic cells. Perhaps an ancestral Esx1 gene somes pair and participate in recombination. Subsequently, expressed in the male germ cells of nonamniotes evolved X chromosome inactivation is reversed, and both X chromo- to have a role in extraembryonic tissues in amniotes. In somes become transcriptionally active. However, in mei- eutherian mammals, this gene may have become modified otic prophase spermatocytes, the X and Y chromosomes are to function in substructures of the chorioallantoic placenta, condensed to form heterochromatin bodies called the sex such as the labyrinthine layer. Genes that have arisen by vesicles that are transcriptionally inactive. The transcrip- duplication often have different expression patterns. The tion of X chromosome-linked genes is not detectable until Mash1 and Mash2 genes exemplify this idea: Mash1 expres- the completion of meiosis in spermatids. From these obser- sion is restricted to subsets of central nervous system neu- vations, we predict that the genes that regulate the transi- rons and Mash2 expression to the ectoplacental cone and tion from spermatogonial stem cells to meiotic spermato- chorion (Johnson et al., 1990; Guillemot et al., 1994). cytes may reside on the X chromosome. The expression X chromosome-linked genes and autosomal genes are reg- pattern of Esx1 and its potential role as a transcriptional ulated differently during mammalian embryogenesis (for re- activator suggest that it may be such a gene. view see Cattanach and Beechey, 1990). In female marsupials, the paternally derived X chromosome is preferentially inactivated in most tissues (VandeBerg et al., 1987). In con- ACKNOWLEDGMENTS trast, in female mice, X chromosome inactivation occurs randomly in the embryo (Lyon, 1961). However, the pater- We thank Brigid Hogan and John Gearhart for mouse embryo nally derived X chromosome is preferentially inactivated in cDNA libraries, Jenny Graves for the platypus genomic DNA, Mari- the primitive endoderm and trophectoderm cell lineages lyn Renfree for the tammar wallaby tissues, Ted Robinson for intro- (Takagi and Sasaki, 1975; West et al., 1977). These observa- ducing us to Monodelphis domestica, Marvin Meistrich for advice tions predict that in therian mammals, genes on the mater- on staging seminiferous tubules, Lucy Rowe of the Jackson Labora- nally derived X chromosome will have a greater effect on tory for assistance with gene mapping, and Jay Cross and Marvin the development of extraembryonic tissues than genes on Meistrich for helpful comments on the manuscript. the paternally derived X chromosome. Note added in proof: Branford et al. (1997) have isolated Spx1 The reason for genomic imprinting, including X chromo- which is equivalent to Esx1. We note that there are differences in some imprinting in mammals, is still largely unknown. the 5؅ region of the cDNAs of Esx1 and Spx1. We have identified However, there is an association between genomic im- three Esx1 transcript isoforms in extraembryonic tissues and one printing and the control of intrauterine embryonic growth. of these is also in testes. The sequences of Spx1, which extend 68 Genetic studies have demonstrated that a number of im- amino acids to the N-terminus of the deduced Esx1 sequences, printed autosomal genes are involved in the growth control represent the common transcript in extraembryonic tissues and of either the placenta or the embryo itself (Barlow et al., testis. 1991; DeChiara et al., 1991; Guillemot et al., 1995; Leigh- ton et al., 1995). The presence of an extra maternally derived X chromosome has been shown to be deleterious for extra- REFERENCES embryonic membrane development (Shao and Takagi, 1990; Takagi and Abe, 1990). Thus, it has been proposed that there Balling, R., Deutsch, U., and Gruss, P. (1988). undulated, a muta- are X chromosome-linked genes that specifically regulate tion affecting the development of the mouse skeleton has a point the development of extraembryonic tissues (Shao and Ta- mutation in the paired box of Pax1. Cell 55, 531–535. kagi, 1990). In agreement with this hypothesis, a locus con- Barlow, D. P., Sto¨ ger, R., Herrmann, B. G., Saito, K., and Schweifer, N. (1991). The mouse insulin-like growth factor type-2 receptor trolling placental growth has been localized to the proximal is imprinted and closely linked to the Tme locus. Nature 349, part of the mouse X chromosome (Zechner et al., 1996). 84–87. Furthermore, targeted mutagenesis of Xist has been shown Bellve´, A. R., Caicchia, J. C., Millette, C. F., O’Brien, D. A., Bhatna- to disrupt imprinted X chromosome inactivation in extra- gar, Y. M., and Dym, M. (1977). Spermatogenic cells of the prepu- embryonic lineages, causing the death of XX embryos that bertal mouse. J. Cell Biol. 74, 68–85. inherited the Xist mutation from their fathers (Marahrens Benirschke, K. (1983). Placentation. J. Exp. Zool. 228, 385–389.

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba 94 Li, Lemaire, and Behringer

Billington, W. D. (1971). Biology of the trophoblast. In ‘‘Advances Kierszenbaum, A. L., and Tres, L. L. (1975). Structural and tran- in Reproductive Physiology’’ (M. W. H. Bishop, Ed.), Vol. V, pp. scriptional features of the mouse spermatid genome. J. Cell Biol. 28–65, Academic Press, San Diego. 65, 258–270. Blum, M., Gaunt, S. J., Cho, K. W., Steinbeisser, H., Blumberg, B., Kongsuwan, K., Webb, E., Housiaux, P., and Adams, J. M. (1988). Bittner, D., and De Robertis, E. M. (1992). Gastrulation in the Expression of multiple homeobox genes within diverse mamma- mouse: The role of the homeobox gene goosecoid. Cell 69, 1097– lian haemopoietic lineages. EMBO J. 7, 2131–2138. Kozak, M. (1987). An analysis of 5؅-noncoding sequences from 699 .1106 Bu¨ rglin, T. R. (1994). A comprehensive classification of homeobox vertebrate messenger RNAs. Nucleic Acids Res. 15, 8125–8148. genes. In ‘‘Guidebook to the Homeobox Genes’’ (D. Duboule, Leighton, P. A., Ingram, R. S., Eggenschwiler, J., Efstratiadis, A., Ed.), pp. 27–71, Oxford Univ. Press, Oxford. and Tilghman, S. M. (1995). Disruption of imprinting caused by Cattanach, B. M., and Beechey, C. V. (1990). Autosomal and X-chro- deletion of the H19 gene region in mice. Nature 375, 34–39. mosome imprinting. Development (Suppl.), 63–72. Lemaire, P., Garrett, N., and Gurdon, J. B. (1995). Expression clon- Chabot, B., Stephenson, D. A., Chapmen, V. M., Besmer, P., and ing of Siamois, a Xenopus homeobox gene expressed in dorsal– Bernstein, A. (1988). The proto-oncogene c-kit encoding a trans- vegetal cells of blastulae and able to induce a complete secondary membrane tyrosine kinase receptor maps to the mouse W locus. axis. Cell 81, 85–94. Nature 335, 88–89. Lin, T. P., Labosky, P. A., Grabel, L. B., Kozak, C. A., Pitman, J. L., Copp, A. J. (1995). Death before birth: Clues from gene knockouts Kleeman, J., and MacLeod, C. L. (1994). The Pem homeobox gene and mutations. Trends Genet. 11, 87–93. is X-linked and exclusively expressed in extraembryonic tissues Cross, J. C., Werb, Z., and Fisher, S. (1994). Implantation and the during early murine development. Dev. Biol. 166, 170–179. Science placenta: Key pieces of the development puzzle. 266, Lindsey, J. S., and Wilkinson, M. F. (1996). Pem: A testosterone- 1508–1518. and LH-regulated homeobox gene expressed in mouse Sertoli Cserjesi, P., Lilly, B., Bryson, L., Wang, Y., Sassoon, D. A., and cells and epididymis. Dev. Biol. 179, 471–484. Olson, E. N. (1992). MHox: A mesodermally restricted homeodo- Lyon, M. F. (1961). Gene action in the X chromosome of the mouse main protein that binds an essential site in the muscle creatine (Mus musculus L.). Nature 190, 373. kinase enhancer. Development 115, 1087–1101. MacLeod, C. L., Wilkinson, M. F., Fong, A. M., Seal, B. S., and DeChiara, T. M., Robertson, E. J., and Efstratiadis, A. (1991). Paren- Walls, L. (1990). Isolation of novel murine thymocyte cDNA tal imprinting of the mouse insulin-like growth factor II gene. clones: One encodes a putative multiple membrane spanning Cell 64, 849–859. protein. Cell Growth Differ. 1, 271–279. Frigerio, G., Burri, M., Bopp, D., Baumgartner, S., and Noll, M. (1986). Structure of the segmentation gene paired and the Dro- Maiti, S., Doskow, J., Li, S., Nhim, R. P., Lindsey, J. S., and Wilkin- sophila PRD gene set as part of a gene network. Cell 47, 735– son, M. F. (1996). The Pem homeobox gene. J. Biol. Chem. 271, 746. 17536–17546. Garber, R. L., Kuroiwa, A., and Gehring, W. J. (1983). Genomic and Marahrens, Y., Panning, B., Dausman, J., Strauss, W., and Jaenisch, cDNA clones of the homeotic locus Antennapedia in Drosophila. R. (1997). Xist-deficient mice are defective in dosage compensa- EMBO J. 2, 2027–2036. tion but not spermatogenesis. Genes Dev. 11, 156–166. Gardner, R. L., and Rossant, J. (1979). Investigation of the fate of Martin, J. F., Bradley, A., and Olson, E. N. (1995). The paired-like 4-5 day post-coitum mouse inner cell mass cells by blastocyst homeobox gene MHox is required for early events of skeletogen- injection. J. Embryol. Exp. Morphol. 52, 141–152. esis in multiple lineages. Genes Dev. 9, 1237–1249. Gardner, R. L. (1982). Investigation of cell lineage and differentia- McCarrey, J. R. (1993). Development in the germ cell. In ‘‘Cell and tion in the extraembryonic endoderm of the mouse embryo. J. Molecular Biology of the Testis’’ (C. Desjardings and L. L. Ewing, Embryol. Exp. Morphol. 68, 175–198. Eds.), pp. 58–89. Oxford Univ. Press, Oxford. Guillemot, F., Nagy, A., Auerbach, A., Rossant, J., and Joyner, A. L. Mitchell, P. J., and Tjian, R. (1989). Transcriptional regulation in (1995). Essential role of Mash-2 in extraembryonic development. mammalian cells by sequence-specific DNA binding proteins. Nature 371, 333–336. Science 245, 371–378. Guillemot, F., Caspary, T., Tilghman, S. M., Copeland, N. G., Gil- Pedersen, R. A. (1986). Potency, lineage, and allocation in preim- bert, D. J., Jenkins, N. A., Anderson, D. J., Joyner, A. L., Rossant, plantation mouse embryos. In ‘‘Experimental Approaches to J., and Nagy, A. (1995). Genomic imprinting of Mash2, a mouse Mammalian Embryonic Development’’ (J. Rossant and R. Ped- gene required for trophoblast development. Nature Genet. 9, ersen, Eds.), pp. 3–33. Cambridge Univ. Press, Cambridge. 235–242. Rivera-Pe´rez, J. A., Mallo, M., Gendron-Maguire, M., Gridley, T., Handel, M. A., and Hunt, P. A. (1992). Sex-chromosome pairing and and Behringer, R. R. (1995). Goosecoid is not an essential compo- activity during mammalian meiosis. BioEssays 14, 817–822. nent of the mouse gastrula organizer but is required for craniofa- Hogan, B., Beddington, R., Costantini, F., and Lacy, E. (1994). ‘‘Ma- cial and rib development. Development 121, 3005–3012. nipulation of the Mouse Embryo,’’ pp. 160–162. Cold Spring Har- Rosa, F. (1989). Mix.1, a homeobox mRNA inducible by mesoderm bor Laboratory Press, Cold Spring Harbor, NY. inducers, is expressed in the presumptive endodermal cells of Johnson, J. E., Birren, S. J., and Anderson, D. J. (1990). Two rat ho- Xenopus embryos. Cell 57, 965–974. mologues of Drosophila achaete-scute specifically expressed in Rossant, J. (1986). Development of extraembryonic cell lineages in neuronal precursors. Nature 346, 858–861. the mouse embryo. In ‘‘Experimental Approaches to Mammalian Jostes, B., Walther, C., and Gruss, P. (1990). The murine paired box Embryonic Development’’ (J. Rossant and R. Pedersen, Eds.), pp. gene, Pax7, is expressed specifically during the development of 97–120. Cambridge Univ. Press, Cambridge. the nervous and muscular system. Mech. Dev. 33, 27–37. Rowe, L. B., Nadeau, J. H., Turner, R., Frankel, W. N., Letts, V. A., Kaufman, M. H. (1992). ‘‘The Atlas of Mouse Development.’’ Aca- Eppig, J. T., Ko, M. S. H., Thurston, S. J., and Birkenmeier, E. H. demic Press, San Diego. (1994). Maps from two interspecific backcross DNA panels avail-

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba Extraembryonic and Testis Expression of ESX1 95

able as a community genetic mapping resource. Mamm. Genome Walther, C., and Gruss, P. (1991). Pax-6, a murine paired box gene, 5, 253–274. is expressed in the developing CNS. Development 113, 1435– Russell, L. D., Ethlin, R. A., Sinha Hikim, A. P., and Clegg, E. D. 1449. (1990). ‘‘Histological and Histopathological Evaluation of the Warn, R. (1995). Growth factors: A scattering of factors. Curr. Biol. Testis.’’ Cache River Press, Clearwater, FL. 4, 1043–1045. Sassoon, D. A., Garner, I., and Buckingham, M. (1988). Transcripts West, J. D., Frels, W. I., Chapman, V. M., and Papaioannou, V. E. of a-cardiac and a-skeletal actins are early markers for myogen- (1977). Preferential expression of the maternally derived X chro- esis in the mouse embryo. Development 104, 155–164. mosome in the mouse yolk sac. Cell 12, 873–882. Schmidt, C., Bladt, F., Goedecke, S., Brinkmann, V., Zschiesche, Wilkinson, D. G., Bailes, J. A., Champion, J. E., and McMahon, W., Sharpe, M., Gherardi, E., and Birchmeier, C. (1995). Scatter A. P. (1987). A molecular analysis of mouse development from 8 factor/hepatocyte growth factor is essential for liver develop- to 10 days post coitum detects changes only in embryonic globin ment. Nature 373, 699–702. expression. Development 99, 493–500. Schneitz, K., Spielmann, P., and Noll, M. (1993). Molecular genetics Wilkinson, M., Kleeman, J., Richards, J., and MacLeod, C. L. (1990). of aristaless, a prd-type homeobox gene involved in the morpho- A novel onco-fetal gene expressed in a stage-specific manner in genesis of proximal and distal pattern elements in a subset of murine embryonic development. Dev. Biol. 141, 451–455. appendages in Drosophila. Genes Dev. 7, 114–129. Yamada, G., Mansouri, A., Torres, M., Stuart, E. T., Blum, M., Scott, M. P., Weiner, A. J., Hazelrigg, T. I., Polisky, B. A., Pirrotta, Schultz, M., De Robertis, E. M., and Gruss, P. (1995). Targeted V., Scalenghe, F., and Kaufman, T. C. (1983). The molecular orga- mutation of the murine goosecoid gene results in craniofacial nization of the Antennapedia locus of Drosophila. Cell 35, 763– defects and neonatal death. Development 121, 2917–2922. 776. Yoshinaga, K., Nishikawa, S., Ogawa, M., Hayashi, S.-I., Kunisada, Shannon, M., and Handel, M. A. (1993). Expression of the Hprt T., Fujimato, T., and Nishikawa, S.-I. (1991). Role of c-kit in gene during spermatogenesis: Implications for sex-chromosome mouse spermatogenesis: Identification of spermatogonia as a spe- inactivation. Biol. Reprod. 49, 770–778. cific site of c-kit expression and function. Development 113, Shao, C., and Takagi, N. (1990). An extra maternally derived X 689–699. chromosome is deleterious to early mouse development. Devel- Zhao, G. Q., Eberspaecher, Seldin, H. M. F., and de Crombrugghe, opment 110, 969–975. B. (1994). The gene for the homeodomain-containing protein Takagi, N., and Sasaki, M. (1975). Preferential inactivation of the Cart-1 is expressed in cells that have a chondrogenic potential paternally derived X chromosome in the extraembryonic mem- during embryonic development. Mech. Dev. 48, 245–254. brane of the mouse. Nature 256, 640–642. Zhao, G. Q., and Hogan, B. L. M. (1996). Evidence that mouse Takagi, N., and Abe, K. (1990). Detrimental effects of two active Bmp8a (Op2) and Bmp8b are duplicated genes that play a role X chromosomes on early mouse development. Development 109, in spermatogenesis and placental development. Mech. Dev. 57, 189–201. 159–168. Torres, M., Go´ mez-Pardo, E., Dressler, G. R., and Gruss, P. (1995). Zhao, G. Q., Deng, K., Labosky, P. A., Liaw, L., and Hogan, B. L. Pax-2 controls multiple steps of urogenital development. Devel- (1996a). The gene encoding bone morphogenetic protein 8B is opment 121, 4057–4065. required for the initiation and maintenance of spermatogenesis Tyndale-Biscoe, H., and Renfree, M. (1987). ‘‘Reproductive Physiol- in the mouse. Genes Dev. 10, 1657–1669. ogy of Marsupials.’’ Cambridge Univ. Press, Cambridge. Zhao, Q., Behringer, R. R., and de Crombrugghe, B. (1996b). Prena- Uehara, Y., Minowa, O., Mori, C., Shiota, K., Kuno, J., Noda, T., and tal folic acid treatment suppresses acrania and meroanencephaly Kitamura, N. (1995). Placental defect and embryonic lethality in in mice mutant for the Cart1 homeobox gene. Nat. Genet. 13, mice lacking hepatocyte growth factor/scatter factor. Nature 373, 275–283. 702–705. Zechner, U., Reule, M., Orth, A., Bonhomme, F., Strack, B., Guenet, VandeBerg, J. L., Robinson, E. S., Samollow, P. B., and Johnston, P. J.-L., Hameister, H., and Fundele, R. (1996b). An X-chromosome (1987). X-linked gene expression and X-chromosome inactiva- linked locus contributes to abnormal placental development in tion: Marsupials, mouse and man compared. In ‘‘Isozymes: Cur- mouse interspecific hybrids. Nat. Genet. 12, 398–403. rent Topics in Biological and Medical Research,’’ Vol. 15, ‘‘Genet- ics, Development, and Evolution,’’ pp. 225–253. Academic Press, Received for publication April 9, 1997 San Diego. Accepted May 22, 1997

AID DB 8640 / 6x29$$$101 07-07-97 13:31:50 dba