|Home | About | Journals | Submit | Contact Us | Français|
Specification of primordial germ cells in mice depends on instructive signalling events, which act first to confer germ cell competence on epiblast cells, and second, to impose a germ cell fate upon competent precursors. fragilis, an interferon-inducible gene coding for a transmembrane protein, is the first gene to be implicated in the acquisition of germ cell competence.
Here, we describe four additional fragilis-related genes, fragilis2–5, which are clustered within a 68 kb region in the vicinity of the fragilis locus on Chr 7. These genes exist in a number of mammalian species, which in the human are also clustered on the syntenic region on Chr 11. In the mouse, fragilis2 and fragilis3, which are proximate to fragilis, exhibit expression that overlaps with the latter in the region of specification of primordial germ cells. Using single cell analysis, we confirm that all these three fragilis-related genes are predominant in nascent primordial germ cells, as well as in gonadal germ cells.
The Fragilis family of interferon-inducible genes is tightly associated with germ cell specification in mice. Furthermore, its evolutionary conservation suggests that it probably plays a critical role in all mammals. Detailed analysis of these genes may also elucidate the role of interferons as signalling molecules during development.
Germ line determination in the mouse is thought to occur through instructive signalling in the gastrulating post-implantation embryo [1,2]. First, proximal epiblast cells acquire germ cell competence at E6.5, partly in response to extraembryonic ectoderm-derived signalling molecules. A subset of these competent cells then acquire a primordial germ cell (PGC) fate and a population of approximately 45 founder germ cells are detected in the posterior proximal region of the embryo at the base of the incipient allantoic bud on E7.5 [1,2]. The secreted signalling molecules, BMP4, BMP8b and BMP2 as well as components of the BMP signal transduction pathway, including Smad1 and Smad5, appear to be involved in the specification of PGCs [3-7]. However, in vitro culture studies and analysis of BMP4-deficient mice suggest that an additional signal may also be required for the acquisition of PGC fate, but its identity is yet unknown [2,3].
We recently identified fragilis, a putative interferon-inducible gene, which codes for a transmembrane protein that is apparently associated with the acquisition of germ cell competence by epiblast cells . Extraembryonic ectoderm is able to induce fragilis expression in epiblast tissue, and BMP4 is required for this induction . fragilis is expressed in proximal epiblast at E6.5, the region in which PGC-competent cells reside according to clonal analysis . As these proximal cells move to the posterior proximal region during gastrulation, fragilis expression increases within a community of cells at the base of the incipient allantoic bud. Cells with the highest expression of fragilis initiate the germ cell-characteristic expression of TNAP and stella/PGC-7 [8-10]. These nascent founder PGCs with high expression of fragilis also show repression of Hox genes, including Hoxb1 .
In view of the strong association of fragilis with PGC specification, we have started to investigate further how this gene may be regulated and what precise function it serves during germ cell development. Towards this objective, we now report that fragilis belongs to a novel murine gene family, comprising five members, which code for five highly similar transmembrane proteins. Importantly, the genes are clustered within a 68 kb genomic region. As we found several homologues of the Fragilis family in human, cow and rat, they seem to be evolutionarily conserved amongst mammalian species. Most if not all homologous genes have been reported to be responsive to interferon signalling, which is in agreement with the presence of conserved interferon stimulable response elements (ISREs) within at least the murine and human loci. Furthermore, our in situ hybridisation and single cell expression analysis reveal that the two members located close to fragilis, fragilis2 and fragilis3, are also expressed in founder PGCs, although their overall expression pattern in post-implantation embryos in other respects is distinct. Studies on the Fragilis family of genes could therefore be crucial for our understanding of PGC specification, especially since their homologues have been implicated in mediating homotypic cell adhesion and lengthening of the cell cycle time [11,12]. These studies may also show how interferons act as signalling molecules, which has hitherto not been considered in the context of embryonic development.
Using the cDNA sequence of fragilis as a template to search the Ensembl genome browser http://www.ensembl.org, we identified eight mouse genes with moderate to high DNA sequence similarity to fragilis (45–74%). ESTs from a variety of embryonic and adult tissues have been reported for five of these genes, of which four possess a two-exon structure similar to fragilis. Analysis of the genomic location of the latter revealed that the four genes cluster around the fragilis locus within a 68 kb region on the distal tip of mouse Chr 7 (F5). We therefore named the four novel genes fragilis2-5, reflecting their genomic location, their similarity to fragilis and their germ cell associated expression pattern (see below; Fig. Fig.1).1). The four remaining putative genes that we detected have few or mostly no reported ESTs and are coded by a single exon unlike fragilis. We therefore consider them to be pseudogenes.
To determine whether the Fragilis genes are evolutionarily conserved, we searched the human genome using Ensembl and found four homologues on Chr 11 (p15.5), a region which is indeed syntenic to the Fragilis family locus on mouse Chr 7 (Fig. (Fig.1).1). Three of these genes, Ifitm1 (9–27), Ifitm2 (1-8D) and Ifitm3 (1-8U), share 58–65% similarity to the fragilis gene cluster and are located within an 18 kb genomic stretch . They are responsive to type1/2 interferons and code for interferon induced transmembrane (Ifitm) proteins, involved in antiproliferative signalling and homotypic cell adhesion [11,12,14,15]. The fourth gene, ENSG142056, a novel gene with two exons, is highly similar to mouse fragilis4 (83% DNA sequence similarity) and neighbours Ifitm2. The human Fragilis family homologues hence form a similar genomic cluster as the five Fragilis genes in the mouse (Fig. (Fig.1).1). Phylogenetic tree analysis suggests however, that only two Fragilis genes, fragilis4 and either fragilis, fragilis2 or fragilis3, have been conserved from mouse to human (data not shown). Subsequent gene duplications may therefore have occurred independently in both species. We also identified two Fragilis family-like genes in cow (bovine 1-8U, bovine 9–27) and four genes in rat (P26376, JC1241, NP110460, AAD48010). While the rat genes have been annotated as putative interferon-inducible, the two bovine genes that are similar to the human Ifitm genes, have been reported to respond to interferon signalling [16,17]. Due to limited mapping information of the cow and rat genomes, we cannot, at this stage, deduce whether these homologous genes are also organised in a cluster. Interferon stimulable response elements (ISREs, GGAAAN(N)GAAAC) within the human Ifitm locus confer the responsiveness of the three human Ifitm genes to interferons [11,18]. Similar ISRE consensus sequences are also found within the Fragilis family cluster in the mouse, associated in particular with fragilis, fragilis2 and fragilis5 (Fig. (Fig.11).
The murine family of fragilis and related genes code for five highly similar transcripts of 104 to 144 amino acids, each containing two predicted transmembrane domains (Fig. (Fig.2).2). The sequence similarity to human, cow and rat fragilis-like genes is equally high (overall 68% amino acid similarity). It should be noted, that the first transmembrane domain as well as the following stretch to the beginning of the second transmembrane domain constitute the regions of highest intra- and inter-species conservation.
We analysed the expression pattern of the five Fragilis family genes by whole mount in situ hybridisation using probes that span the 3' region (150–200 bp) of the corresponding mRNAs. These probes show no significant cross-hybridization between members of the Fragilis family as judged by dotblot analysis (data not shown). As reported, we saw expression of fragilis restricted to the epiblast at E5.5 and E6.5. More importantly, around E7.5, expression of fragilis is intense within a population of cells at the base of the allantois in the region where PGC specification occurs (Fig. 3a,3b,3c) . fragilis2 and fragilis3 are also expressed within the epiblast of E5.5 embryos (Fig. (Fig.3g,3g, data not shown). While expression of fragilis2 is thereafter significantly downregulated, fragilis3 remains expressed at a similar level in the embryonic tissues. At E7.5, fragilis2 is detected in the posterior mesoderm, while fragilis3 expression is seen throughout the epiblast. More significantly, like fragilis, both fragilis2 and fragilis3 show high expression in the region where the cluster of founder PGCs originates (Fig. 3i/i',3n/n'). Thus, these three members of the Fragilis family show significant expression at the time and site of PGC specification.
At E8.5, fragilis expression is seen in cells at the base and within the proximal third of the allantois (Fig. (Fig.3d).3d). Additionally, a signal is detected in the latero-anterior aspects of the developing brain (Fig. (Fig.3e).3e). At this stage, fragilis2 is expressed in the mesoderm in the caudal half of the embryo (Fig. 3j,3k), whereas fragilis3 appears present throughout the entire embryo (Fig. 3p,3q,3r). It is noteworthy, that expression seems significantly stronger in cells at the base and within the proximal third of the allantois at this stage (Fig. (Fig.3q).3q). At E9.5, fragilis signal is seen in the mesoderm and the pharyngeal arches (Fig. (Fig.3f).3f). At this stage, fragilis2 expression appears restricted to the tailbud, the mesoderm caudal to the 15th somite and the developing lung tissue (Fig. (Fig.3l3l).
In order to further analyse fragilis2 expression, we carried out in situ hybridisation on transverse and sagittal sections of E9.5 and E10.5 embryos, respectively. This confirmed the pattern seen in whole mount preparations, namely presence of staining in the presomitic, somatic and bodywall mesoderm in the caudal regions of the embryo at E9.5 (Fig. 4a,4b,4c,4d). It appears that fragilis2 expression is lost in differentiating mesodermal or somatic tissue, hence explaining the caudal-to-rostral gradient of staining seen in the caudal region of the whole mount embryos (see Fig. Fig.3l).3l). In addition, the sections show expression in the ventral neural tube, as well as in single cells within the wall of the hindgut. As the sagittal sections of E10.5 embryos demonstrate, these latter cells seem to belong to a population of fragilis2 mRNA positive cells, which are migrating along the hindgut and ventral to the dorsal aorta (Fig. 4e,4g). Their location suggests strongly that these cells are at least in part PGCs, migrating from their site of origin into the developing genital ridges. Furthermore, fragilis2 expression is seen in the developing lung tissue posterior to the embryonic heart and ventral to the dorsal aorta (Fig. 4e,4f).
In contrast to the first three members of the family, neither fragilis4 nor fragilis5 showed expression at early post-implantation stages (E7.0-E8.5, data not shown). Consequently, only the three genes at the centre of the family cluster, that is fragilis, fragilis2 and fragilis3 are expressed in the embryo between E5.5 and E9.5. While their expression pattern is distinct, there is a striking overlap within the region where founder germ cells are located. This could suggest that the three neighbouring genes, fragilis, fragilis2 and fragilis3, may share regulatory elements that might be present within the cluster. Such regulatory elements may be responsible for the genes' overlapping expression pattern specifically around the region of PGCs specification.
To obtain more precise information on the expression of the new Fragilis family members in the context of germ cell specification, we tested single cell cDNAs from PGCs and surrounding somatic cells sited at the base of the incipient allantoic bud in E7.5 embryos. Both fragilis2 and fragilis3 were expressed in founder PGCs, which show transcription of the germ cell marker stella/PGC7 (Fig. (Fig.4a)4a) [8,10]. The two Fragilis family members were also detected in surrounding somatic cells that lack expression of stella/PGC7 . Notably, semi-quantitative analysis using Southernblotting showed that fragilis2 and fragilis3 are expressed predominantly and at higher levels in PGCs compared to the neighbouring somatic cells (Fig. 4b,4c). This mimics the pattern seen for fragilis, although expression of the latter is more specific to germ cells. Combined with the in situ hybridisation data, these observations further support the notion that certain common control elements may be involved in the upregulated expression of the three Fragilis genes in the founder PGCs.
During the developmental stages directly subsequent to PGC specification, all three Fragilis family genes are expressed in a population of cells associated with the allantois and in a location where premigrating PGCs are thought to reside (Fig. 3,3d,3k,3q). The precise gene expression during migration of PGCs is not clear at this stage from our analysis. However, using in situ hybridisation and PCR analysis of cDNAs from single cells within the genital ridge, we found clear expression of fragilis, fragilis2 and fragilis3 in the gonadal germ cells at E11.5/12.5 (Fig. (Fig.5).5). While fragilis3 expression extends to the mesonephros, fragilis and fragilis2 signal was restricted to the genital ridge. A punctate staining pattern was seen for fragilis, mimicking the germ cell restricted expression of stella/PGC7 (Fig. (Fig.5b).5b). This pattern in addition to the PCR analysis suggests that fragilis is expressed predominantly if not solely in germ cells at E11.5. As was the case in earlier embryos, neither fragilis4 nor fragilis5 were detected in gonadal germ cells (data not shown).
In this study we describe the identification of the murine Fragilis gene family, which appears to be conserved amongst mammalian species, and whose members code for five highly similar transmembrane proteins. Three members of the Fragilis family, fragilis, fragilis2 and fragilis3, exhibit expression, which is associated with germ cell specification and development. Located at the cell membrane, the Fragilis proteins may be important for mediating interactions amongst germ cells and their surrounding neighbours. While the three genes are expressed earlier at E5.5 and thereafter to a varying extent, they all show upregulation of expression within nascent PGCs. It is possible that a cis control element exists within the locus that is required for this expression, which continues within gonadal PGCs. Future studies will elucidate whether such control elements exist, where they are located and how they could regulate expression of the fragilis-related genes.
Although the five Fragilis family members are clustered within a small genomic region, it appears that neither fragilis4 nor fragilis5 show expression in early embryos or embryonic germ cells. It is striking that these two members are located at the periphery of the cluster in contrast to the centrally located fragilis, fragilis2 and fragilis3 genes. This lack of expression may be due to the presence of boundary elements, which might restrict the action of control elements to genes present within the centre of the cluster. The four human homologues of the Fragilis family are also arranged in a genomic cluster. However, sequence comparison suggests that gene duplications at the mouse Fragilis and the human Ifitm locus may have occurred independently in the two species. Hence, it appears that a certain evolutionary constrain may exist on duplication and maintenance of the duplicated genes within immediate neighbourhood. Since the human homologues are clustered in the syntenic region and are highly similar to the family genes, it is tempting to speculate that they may also serve similar functions as in mouse and man.
The presence of several interferon stimulable response element (ISRE) consensus sequences within the Fragilis locus, together with the similarity of the genes to their interferon-inducible human and bovine counterparts, suggest very strongly that fragilis and the fragilis-related genes are responsive to interferons or possibly other cytokines. Indeed, the ISRE tandem repeat present in the 5' flanking region of human Ifitm1, Ifitm2 and Ifitm3 genes is also present in the 5' flanking region of fragilis exon 1 . Interferons, as secreted signalling molecules, have so far been implicated mainly in the process of immune response, the inhibition of cellular growth and the control of apoptosis . Although interferons are expressed in the post-implantation embryo, their role during development has not been addressed in detail [20,21]. Our studies have pointed to a possible involvement of interferons in germ cell development. Future work will determine whether the Fragilis genes respond to interferon signals in all or some instances where the genes are expressed.
We have identified the Fragilis family of interferon-inducible genes, which code for transmembrane proteins. The five members are arranged in a cluster within a genomic region of 68 kb in the mouse that also contains ISRE elements. The centrally located fragilis, fragilis2 and fragilis3 genes are of particular interest, because they are expressed in the region and at the time where germ cell specification occurs. The family is evolutionarily conserved amongst mammalian species where it may serve similar functions. Detailed studies of the Fragilis family may also show what role interferons have in embryonic development.
Nucleotide sequence data reported are available in the DDBJ/EMBL/GenBank databases as primary sequences under the accession numbers: AY082484 (fragilis), AY082486 (fragilis3), AY247203 (fragilis5); and in the Third Party Annotation Section under the accession numbers: BK001123 (fragilis2), BK001124 (fragilis4).
Ensembl and partly NCBI genome browsers have been used for data retrieval. Embryos and genital ridges used for in situ hybridisation experiments came from 129/SvEv × 129/SvEv or B6CBAF1 × Oct4-GFP matings. Embryos and genital ridges used for single cell analysis came from 129/SvEv × 129/SvEv or MF1 × Oct4-GFP matings, respectively. The noon of the day of the vaginal plug was designated as E0.5. Embryos were staged according to Downs and Davies . Work on animals was performed under Home Office project licence PPL80/1706.
3'-fragments of fragilis and fragilis2–5 cDNAs were PCR amplified using the primers described below, and cloned into pGEM-T vector (Promega). DIG-labelled antisense RNA probes were synthesized using DIG RNA labelling kit (Sp6/T7; Roche). In situ hybridisation on whole embryos and urogenital ridges was performed as described [23,24]. Hybridisation was carried out using 1 μg/ml DIG-labelled RNA probe in hybridisation buffer (50% formamide, 1.3 × SSC (pH 5), 5 mM EDTA (pH 8), 50 μg/ml yeast RNA, 0.2% Tween-20, 0.5% CHAPS, 100 μg/ml heparin in DEPC treated H2O) at 70°C over night. In situ hybridisation on sections was performed as described . Prior to cryosectioning at 12–15 μm thickness, embryos were fixed in 4% paraformaldehyde in PBS and cryoprotected in 20% sucrose in PBS. Hybridisation was carried out using 200–400 μg/ml DIG-labelled RNA probe in hybridisation buffer (50% formamide, 0.1 mg/ml yeast total RNA, 10% dextran sulphate, 1 × Denhardt's, 0.2 M NaCl, 5 mM EDTA, 10 mM Tris (pH 7.5), 5 mM NaH2PO4 × 2H2O, 5 mM Na2HPO4 in dH2O). Hybridised probe was detected using alkaline phosphatase conjugated anti-DIG Fab fragments (Roche) and BM Purple alkaline phosphatase substrate (Roche) for whole mounts and 0.35 mg/ml NBT, 0.175 mg/ml BCIP, 0.24 mg/ml levamisole in 0.1 M Tris (pH 9.5), 0.1 M NaCl, 50 mM MgCl2 for sections, respectively.
Early bud stage embryos (E7.5) and genital ridges (E11.5) were isolated in DMEM/10% fetal calf serum/25 mM HEPES (pH 7.4). Fragments bearing primordial and gonadal germ cells, respectively, were dissected out and dissociated into single cells. The latter were picked using mouth pipettes and their cDNAs were amplified as described previously . The following primers were used in order to PCR amplify stella cDNA, parts of fragilis2 cDNA and 3'-fragments of fragilis and fragilis3–5 cDNAs (25 cycles of amplification): stella: 5'CTCACAGCTTGAGGCTTCTAA3', 5'GCGATTCAGATGTCTCTGCAC3', fragilis: 5'GTTATCACCATTGTTAGTGTCATC3', 5'AATGAGTGTTACACCTGCGTG3'; fragilis2: 5'CCTTCCTTATTCTCACTCTG3', 5'GTTGCAAGACATCTCACATC3'; fragilis3: 5'GATCTTCAGCATCCTTATGGTC3', 5'GAAGGTAACATTTGCATACGCG3'; fragilis4: 5'AACTTGGAGGCTGCAAGGCAG3', 5'CTCGGAACTCTTAGTTATAGTC3'; fragilis5: 5'TGCTCTGGTCATCTCCCTCA3', 5'CAGGATAAGGGGCAACTCTG3'. PCR products were run on 1.5% agarose/TBE electrophoresis gels. For Southernblot analysis, single cell cDNAs were blotted onto Hybond-N+ membranes (Amersham) and probed with 32αP dCTP-labelled DNA probes comprising the 3' regions of fragilis, fragilis2 (primers for amplification: 5'AGCTCCCTGTTCTTCACCAT3', 5'GCAAATGGTCAGGACTAAGTC3') and fragilis3 cDNAs and full length stella cDNA. The 3' fragment of GAPDH cDNA was used as loading control (primers for amplification: 5'TTGGGCTACACTGAGGACCAG3', 5'GATAGGGCCTCTCTTGCTCAG3'). Blotting signal was detected using a Fuji film FLA 5000 scanner. Signal strength was semi-quantified in relation to GAPDH signal, whereby relative gene expression was calculated as ratio of gene signal to GAPDH signal and this ratio was subsequently normalized by division through the highest hybridisation signal per blot (as described previously ). For dotblot analysis, full length fragilis cDNAs were blotted and probed with 32αP dCTP-labelled 3' probes.
U.C.L. carried out the in situ hybridization studies, PCR and Southernblot analysis, sequence alignment and genomic cluster identification, participated in the database searches and drafted the manuscript. M.S. constructed the single cell cDNAs of E7.5 embryos, participated in the in situ hybridisation studies and the database searches. P.S.W. constructed the single cell cDNAs of the genital ridges of E11.5 embryos. S.C.B. participated in the construction of single cell cDNAs. U.C.L., M.S. and M.A.S conceived of the study, participated in its design and coordination and were primarily involved in writing the manuscript.
We thank B. Payer for helpful comments on the manuscript. U.C.L. acknowledges the Wellcome Trust for a PhD student grant (grant no: 065601). Work in the laboratory of M.A.S. is funded by the BBSRC and the Wellcome Trust (grant nos, RG35410 and RG23643, respectively).