Portal de Periódicos da CAPES

Serum response factor is essential for mesoderm formation during mouse embryogenesis

1998; Springer Nature; Volume: 17; Issue: 21 Linguagem: Inglês

10.1093/emboj/17.21.6289

1460-2075

Sergei Arsenian, Birgit Weinhold, Michael Oelgeschläger, Ulrich Rüther, Alfred Nordheim,

Genomics and Chromatin Dynamics

Article2 November 1998free access Serum response factor is essential for mesoderm formation during mouse embryogenesis S. Arsenian S. Arsenian Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Search for more papers by this author B. Weinhold B. Weinhold Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Search for more papers by this author M. Oelgeschläger M. Oelgeschläger Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Search for more papers by this author U. Rüther U. Rüther Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Search for more papers by this author A. Nordheim Corresponding Author A. Nordheim Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Institut für Zellbiologie, Abteilung Molekularbiologie, Eberhard-Karls-Universität Tübingen, Auf der Morgenstelle 15, 72076 Tübingen, Germany Search for more papers by this author S. Arsenian S. Arsenian Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Search for more papers by this author B. Weinhold B. Weinhold Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Search for more papers by this author M. Oelgeschläger M. Oelgeschläger Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Search for more papers by this author U. Rüther U. Rüther Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Search for more papers by this author A. Nordheim Corresponding Author A. Nordheim Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany Institut für Zellbiologie, Abteilung Molekularbiologie, Eberhard-Karls-Universität Tübingen, Auf der Morgenstelle 15, 72076 Tübingen, Germany Search for more papers by this author Author Information S. Arsenian1, B. Weinhold1, M. Oelgeschläger1, U. Rüther1 and A. Nordheim 1,2 1Institut für Molekularbiologie, Medizinische Hochschule Hannover, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany 2Institut für Zellbiologie, Abteilung Molekularbiologie, Eberhard-Karls-Universität Tübingen, Auf der Morgenstelle 15, 72076 Tübingen, Germany ‡S.Arsenian, B.Weinhold and M.Oelgeschläger contributed equally to this work *Corresponding author. E-mail: [email protected] The EMBO Journal (1998)17:6289-6299https://doi.org/10.1093/emboj/17.21.6289 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The transcription factor serum response factor (SRF), a phylogenetically conserved nuclear protein, mediates the rapid transcriptional response to extracellular stimuli, e.g. growth and differentiation signals. DNA–protein complexes containing SRF or its homologues function as nuclear targets of the Ras/MAPK signalling network, thereby directing gene activities associated with processes as diverse as pheromone signalling, cell-cycle progression (transitions G0–G1 and G2–M), neuronal synaptic transmission and muscle cell differentiation. So far, the activity of mammalian SRF has been studied exclusively in cultured cells. To study SRF function in a multicellular organism we generated an Srf null allele in mice. SRF-deficient embryos (Srf−/−) have a severe gastrulation defect and do not develop to term. They consist of misfolded ectodermal and endodermal cell layers, do not form a primitive streak or any detectable mesodermal cells and fail to express the developmental marker genes Bra (T), Bmp-2/4 and Shh. Activation of the SRF-regulated immediate early genes Egr-1 and c-fos, as well as the α-Actin gene, is severely impaired. Our study identifies SRF as a new and essential regulator of mammalian mesoderm formation. We therefore suggest that in mammals Ras/MAPK signalling contributes to mesoderm induction, as is the case in amphibia. Introduction The transcription factor serum response factor (SRF) (Prywes and Roeder, 1987; Schröter et al., 1987; Treisman, 1987; Norman et al., 1988) and SRF-directed gene activity have become one of the best characterized model systems for understanding the molecular mechanisms underlying signal-dependent gene regulation (Johansen and Prywes, 1995; Treisman, 1995). SRF directs the signal-induced activity of ‘immediate early’ genes (IEGs) by binding to the serum response elements (SREs) of IEG promoters (Treisman, 1986, 1987, 1992, 1995; Herrera et al., 1989; Herschman, 1991). Functional SRF binding sites have been identified in the promoters of some 30 different genes so far (Cahill et al., 1995), including c-fos, Egr-1 and various α-Actin genes. SRF, a MADS-box-containing transcription factor (Schwarz-Sommer et al., 1990; Shore and Sharrocks, 1995), has been characterized extensively, both in structural (Pellegrini et al., 1995) and functional (Treisman, 1995) terms. DNA recognition by SRF is directed by CC(A/T)6GG sequences, called CArG-boxes, found within SREs. SRF is able to recruit additional proteins to SREs. Such accessory factors comprise the ternary complex factors (TCFs) (Shaw et al., 1989a; Treisman, 1994), which belong to the Ets family of transcription factors (Hipskind et al., 1991; Dalton and Treisman, 1992). TCFs include the proteins Elk-1 (Hipskind et al., 1991; Dalton and Treisman, 1992), Sap-1 (Dalton and Treisman, 1992) and Net/ERP/Sap-2 (Giovane et al., 1994; Lopez et al., 1994; Price et al., 1995). Other SRF-interacting proteins are the Ets protein Fli-1 (Magnaghi-Jaulin et al., 1996), the homeodomain-protein Phox-1 (Grueneberg et al., 1992, 1997), the HTLV-1 protein Tax1 (Fujii et al., 1992), the p65 subunit of NF-κB (Franzoso et al., 1996), the myogenic bHLH heterodimers myogenin–E12 and MyoD–E12 (Groisman et al., 1996), and the cardiogenic homeodomain protein Nkx-2.5 (Chen and Schwartz, 1996). SRF-containing transcription factor complexes are nuclear targets of intracellular signalling cascades, primarily the cascades of the MAP kinase network (Treisman, 1996). The Ets/TCF proteins represent direct targets of the three best characterized types of MAP kinase, i.e ERKs, Jnk/SAPK2 and p38/SAPK1 (Gille et al., 1992, 1995; Janknecht et al., 1993; Marais et al., 1993; Hipskind et al., 1994; Hill and Treisman, 1995; Zinck et al., 1995; Cahill et al., 1996; Price et al., 1996; Raingeaud et al., 1996; Treisman, 1996). SRF, a phospho-protein itself, is targeted by direct (Janknecht et al., 1992; Marais et al., 1992; Rivera et al., 1993; Miranti et al., 1995) and, possibly, indirect signalling mechanisms (Hill and Treisman, 1995; Hill et al., 1995). Given the transcriptional induction of SRF-regulated genes (i.e. IEGs) during the mitogen-induced G0–G1 transition (Herschman, 1991), an essential involvement of SRF has been assumed in the control of proliferation and cell-cycle progression (Johansen and Prywes, 1995). Additionally, SRE-regulated gene activity during the G2–M transition of K562 human erythroleukaemia cells (Liu et al., 1994), in light of the essential function performed by the Saccharomyces cerevisiae SRFhomologue Mcm1 at the G2–M transition (Althöfer et al., 1995; Maher et al., 1995), further suggested that SRF participates in cell-cycle control. In addition to regulating genes during the cell cycle, SRF and related MADS-box factors have been demonstrated as being essential for post-replicative cell-type-specific gene regulation, namely neuronal- (Ghosh and Greenberg, 1995) and muscle-specific gene expression (Vandromme et al., 1992; Buckingham, 1994; Soulez et al., 1996; Firulli and Olson, 1997). The α-Actin genes have served as paradigms for SRF-directed myocyte-specific gene expression (Mohun et al., 1989; Sartorelli et al., 1990; Moss et al., 1994; Chen and Schwartz, 1996; Sepulveda et al., 1998). This includes cardiac, skeletal and vascular muscle Actin genes. For example, the combinatorial action of SRF, Nkx-2 and GATA-4, as part of a multi-component transcriptional regulatory complex, was shown to regulate the cardiac α-Actin gene in early cardiac progenitor cells (Sepulveda et al., 1998). SRF expression studies support the proposed role for SRF in post-replicative neuronal and muscle gene expression. In the adult rat nervous system, SRF immunoreactivity was present in the vast majority of neurons in the forebrain, cortex, striatum, amygdala and hippocampus, and in some scattered neurons in the medulla and spinal cord (Herdegen et al., 1997). In the chicken, SRF expression was found to be restricted to tissues of mesodermal and neuroectodermal origin (Croissant et al., 1996). During chicken embryogenesis and the progression of gastrulation, strongly localized Srf mRNA expression was observed in the primitive streak, the neural groove, lateral plate mesoderm, Hensen's node, the precardiac splanchnic mesoderm, the myocardium and the somites. Strong SRF protein expression was seen in the myocardium of the developing chicken heart and the myotomal portion of the somites (Croissant et al., 1996). In the mouse, highest Srf mRNA levels were seen in adult skeletal and cardiac muscle. During mouse embryonic development, Srf transcripts were found to be enriched in smooth muscle media of the vessels, the myocardium of the heart and myotomal portions of somites (Belaguli et al., 1997). Further light was shed on the biological role of SRF in a living organism by the identification of the Drosophila melanogaster genes pruned (Guillemin et al., 1996) and blistered (Montagne et al., 1996) as two alleles of the Drosophila SRF homologue (DSRF) (Affolter et al., 1994). The corresponding mutant phenotypes revealed SRF functions in the development of the wing disc and the tracheal system. We sought to expand the genetic analysis of SRF function into the vertebrate system and therefore generated, using homologous recombination, Srf null alleles in embryonal stem (ES) cells and in the mouse. Our analysis revealed an essential function of SRF for inductive gene regulatory events leading to mesoderm formation during gastrulation. Results Expression pattern of SRF during early mouse development To guide our functional analysis of SRF we first investigated the expression pattern of the Srf gene, at both RNA and protein levels, during the early stages of mouse development. Staining of sectioned embryos with an SRF-specific antiserum revealed expression at E6.5 in ectoderm as well as endoderm, both embryonic and extra-embryonic (Figure 1A). At E7.5, SRF protein could be seen in all three germ layers of wild-type (wt) embryos (Figure 1B). Interestingly, at E8.5, this ubiquitous distribution became a regionally localized one and SRF protein expression was found to be high in the developing heart (Figure 1C–E), but barely detectable in other tissues. This expression was specific for the myocardium (Figure 1E). At E10.5 we also detected distinct SRF protein expression in the developing myotome (Figure 1F). Northern blotting of E8.5–E12.5 embryonal RNA preparations detected two mRNA species (Figure 2) which possibly represent two differently polyadenylated variants (Norman et al., 1988; Belaguli et al., 1997). Embryonic (E8.5) protein extracts also showed SRF-associated specific DNA-binding activity toward SRE sequences. These studies confirm and extend the analysis of Belaguli et al. (1997) and provide, for the first time, insight on SRF protein expression during mouse early embryogenesis. Figure 1.SRF expression from E6.5 to E10.5, as detected by immunohistochemistry using an SRF-specific antiserum. (A) At E6.5 SRF was detectable in the embryonic as well as extra-embryonic ectoderm and endoderm. (B) At E7.5 SRF is expressed in all three germ layers (ectoderm: □; mesoderm: ▴; endoderm: ▵) in both the embryonic and extra-embryonic parts of the embryo. Instead, at E8.5 (C and D), SRF was found selectively and highly expressed in the developing heart (▴) and was hardly detectable in other regions of the embryo; e.g. no expression could be detected in the developing somites (▵). Expression in the heart was specific for the myocardium (E). Expression of SRF in the myotome (▵) was first detected at E10.5 (F). Panel (C) represents the histological stain of the identical embryo stained with anti-SRF in panel (D). Download figure Download PowerPoint Figure 2.Northern blot analysis of Srf expression during wt embryonal development. Download figure Download PowerPoint This expression analysis shows that strong embryonic SRF expression occurs ubiquitously in all germ layers at times before and after the onset of mesoderm formation. Interestingly, subsequent to the onset of organogenesis, domains of localized, strong SRF protein expression are found in specific mesodermal tissues, namely the heart myocardium and the myotome. These SRF expression patterns in mouse embryos are congruent with those found in chicken embryos (Croissant et al., 1996). Targeted disruption of Srf in ES cells by homologous recombination In order to analyse the function of SRF in vertebrates by genetic means, we generated a null mutation of Srf by homologous recombination in embryonic stem (ES) cells. To construct the recombination vector we cloned and structurally characterized the genomic Srf locus (Figure 3 and data not shown), as performed in parallel by Belaguli et al. (1997). Our Srf targeting vector (for details of the construction see Materials and methods) was designed to delete sequences encoding essential functions of SRF, namely dimerization and DNA binding (Figure 3) (Pellegrini et al., 1995). Twenty-three independent ES cell clones were identified as having undergone correct recombination at one Srf allele, as determined by genomic PCR and Southern blotting (not shown). ES cells heterozygous for the mutated Srf allele showed no phenotypic abnormalities. Two independent ES cell clones were used for blastocyst injections to generate chimeric mice, followed by subsequent breeding to obtain germline transmission and establishment of heterozygous Srf+/− mouse strains. Figure 3.Targeted disruption of the mouse Srf gene. Maps of part of the wild-type Srf locus, the targeting vector and the recombined allele. The genomic sequences cloned into the recombination vector, as well as their corresponding positions in the wt and recombined alleles, are marked as shaded bars (sizes indicated). An 880 bp segment is deleted in the targeted allele (top). This segment covered parts of exon 1 and intron 1, including SRF sequences encoding amino acid residues 1–167. Arrows (A, B and C) indicate primers used for embryo genotyping by PCR. Download figure Download PowerPoint Embryonal lethality of embryos lacking SRF Like the genotypically identical ES cells, mice heterozygous for the mutated Srf allele showed no detectable phenotypic abnormalities. In contrast, upon breeding heterozygous Srf+/− animals no Srf−/− offspring were born, indicating that the Srf mutation was lethal during embryogenesis (Figure 4A and B). To determine the time of embryonic lethality, embryos at different stages were analysed by morphological criteria and by genotyping. For embryos up to E9.5 Mendelian distribution of the mutated alleles was still observed (Figure 4B); however, after E12.5 Srf−/− embryos could no longer be detected. Thus, the embryonally lethal phenotype observed here for Srf−/− mouse embryos clearly reveals an essential requirement for SRF activity during the early stages of murine embryogenesis. Figure 4.(A) Genotyping of E9.5 embryos by PCR. Primers A to C (see Figure 3) were used to amplify DNA segments from embryo genomic DNAs. Allele-specific PCR fragments (wt allele, Srfwt; mutated allele, Srfneo) were separated by agarose gel electrophoresis. (B) Ratios of genotypes found with new-born animals and pre- and post-implantation embryos. (C) Electrophoretic mobility shift assays using an SRE probe and extracts from E8.5 embryos derived from Srf+/− matings. Embryos 1–4 differed somewhat in size but all showed primitive streak and head fold formation. Embryos 5–7 all lacked primitive streak and head fold formation and, therefore, displayed Srf−/− phenotypic appearance. SRF indicates the use of partially purified SRF protein, whereas SRF* indicates the additional presence of antibody blocking peptide. α-SRF indicates the use of a specific anti-SRF antiserum. No corresponding inter-embryonal differences in band-shift activity were displayed by these extracts with a DNA probe containing an Ets protein binding site (not shown), which served as a control for equal protein recovery. Download figure Download PowerPoint The mutated Srf allele represents a bona fide null allele Homodimerization and specific DNA-binding to SRE sequences are essential to SRF's function as a transcription regulator (Johansen and Prywes, 1993; Sharrocks et al., 1993). The targeted Srf allele was intended to have deleted Srf coding sequences contributing to both these functions (Pellegrini et al., 1995). Indeed no SRF-derived SRE-binding activity (Figure 4C) or SRF protein (Figure 5E) was found in phenotypically or genotypically identified Srf−/− embryos. RT–PCR studies also failed to detect any normal or aberrant Srf transcripts in these embryos. Figure 5.Histology of paraffin sections from wt and Srf−/− embryos, including anti-SRF staining patterns. Wt embryos of E7.5 (A) and E8.5 (C) are compared with Srf−/− embryos of E7.5 (B) and E8.5 (D). Additionally, an E6.5 Srf−/− embryo is shown, displaying normal histological appearance (F) while revealing the complete lack of anti-SRF antibody reactivity (E). Sections (A) and (C) were photographed as 100× views of the microscopic fields, (D) as 200× view, and the other sections were photographed as 400× enlargements. Note the pyknotic cells in (B), marked by the open triangle. Download figure Download PowerPoint Being unable to detect either wt SRF protein activity or any aberrant gene product derived from the targeted Srf locus in Srf−/− embryos, we conclude that our targeting strategy achieved the generation of an Srf null allele, and that the observed phenotype in Srf−/− embryos is a direct consequence of the lack of SRF. Impaired gastrulation, lack of primitive streak formation, and absence of mesodermal cells in Srf−/− embryos No phenotypic abnormalities were apparent upon comparing wt with mutated embryos at E6.5 (compare Figure 1A with Figure 5E or F). However, as early as E7.5 Srf−/− embryos could be distinguished from heterozygous or wt embryos by their reduced size. At E7.5, Srf−/− embryos showed delayed development and displayed late egg cylinder stage morphology (compare Figure 5A and B). They did not form a primitive streak and, histologically, no mesodermal cells were apparent. Instead, pyknotic cells could be detected in the embryonic cavity. At E7.5 some mutant embryos already displayed aberrant folding of both embryonic ectoderm and endoderm, which was seen more strongly at E8.5 in all Srf−/− embryos (compare Figure 5C and D). Disintegrating Srf−/− embryos started to appear from E8.5 and after E12.5 no such embryos could be detected. These observations demonstrate that the lack of SRF causes an early embryonic phenotype at E7.0–E7.5, shortly after the normal onset of gastrulation. The apparently normal development up to E6.5 reveals that SRF is not essential for the proliferation of embryonal cells, at least up to E6.5. However, the observed phenotype of Srf−/− embryos implies an essential requirement for SRF during the process of mesoderm formation, possibly for mesoderm induction itself. Absence of mesodermal marker gene expression in Srf−/− embryos Next, we characterized in more detail the types of cell present and absent in Srf−/− embryos. Immunohistological stainings on paraffin-embedded sections of E7.5 embryos were used to identify cellular marker proteins of ectoderm (Oct-6 and cytoplasmic LEF-1) (Behrens et al., 1996; Huber et al., 1996; Zwart et al., 1996), mesoderm (Brachyury and nuclear LEF-1) (Wilkinson et al., 1990) and endoderm, which was specifically stained by an anti-HNF3β antiserum (Sasaki and Hogan, 1994). In wt embryos, primitive ectoderm specifically expressed the transcription factor Oct-6 (Figure 6A), as did anterior definitive ectoderm and the chorion after closure of the proamniotic channel (Zwart et al., 1996). Staining for the transcription factor HNF-3β was specific for parietal and visceral endoderm of wt embryos (Figure 6B; see Materials and methods). In these wt embryos, Brachyury was highly expressed in the ectoderm surrounding the node and in the migrating mesoderm (Figure 6C). LEF-1 staining was found in the cytoplasm of ectodermal and, in contrast, in the nuclei of mesodermal wt cells (Figure 6D). This differential intracellular localization of LEF-1 possibly reflects nuclear translocation of LEF-1/β-catenin complexes after mesoderm formation (Behrens et al., 1996; Huber et al., 1996). In Srf−/− embryos strong staining for Oct-6 and HNF-3β was seen (Figure 6E and F), whereas, significantly, no Brachyury or nuclear LEF-1 could be detected at all (Figure 6G and H, respectively). Since the stainings were performed on successive sections of the same homozygous embryos we conclude that Srf−/− embryos consist of primitive ectoderm and endoderm, while clearly lacking mesoderm. Figure 6.Antibody staining of developmental marker proteins in E7.5 wt and Srf−/− embryos. Wt (A–D) and Srf−/− embryo stainings (E–H) were performed with antisera directed against Oct-6 (A and E), HNF-3β (B and F), Brachyury (C and G) or LEF-1 (D and H). Sections from identical embryos are each represented in (A and D), (B and C) and (E–H), respectively. All sections were photographed as 400× views of the microscopic fields. In wt embryos, the three germ layers are marked (ectoderm, □; mesoderm, ▴; endoderm, ▵). Download figure Download PowerPoint Expression of developmental marker genes in Srf−/− embryos The lack of mesoderm in Srf−/− embryos was substantiated by RT–PCR studies that revealed absent or strongly impaired expression of developmental marker genes. In Srf−/− embryos ranging from E7.5 to E9.5, no transcripts could be detected for Bra, Shh and the TGFβ-related genes Bmp2 and Bmp4 (Figure 7A and B, and data not shown) (Wilkinson et al., 1990; Chiang et al., 1996; Zhang and Bradley, 1996). Significantly reduced mRNA levels were measured for Gsc, Fgf-5 and Nodal (Figure 7B and data not shown) (Lemaire and Kodjabachian, 1996). This expression pattern of developmental marker genes correlates with an arrest in development of Srf−/− embryos during gastrulation, at the stage when mesoderm is being formed. Figure 7.Expression analysis by RT–PCR of embryonal marker genes in wt and mutated Srf embryos. (A) RT–PCR patterns representing marker gene expression in E9.5 embryos of the Srf+/+, Srf+/− and Srf−/− genotypes. (B) Summary representation of developmental marker gene expression levels determined by RT–PCR. Not shown are expression patterns of Bmp4 and Fgf-5 which displayed the same patterns as the ones of Bmp2 and Nodal, respectively. Semi-quantitative expression levels are indicated by the symbols +, − and (+). nd, not determined. Download figure Download PowerPoint Reduced expression levels in Srf−/− embryos of the SRE-regulated genes c-fos, Egr-1 and α-Actin In tissue culture cells SRF contributes significantly to the transient induction of immediate early genes, such as Egr-1 and c-fos. This gene activation during the cellular G0–G1 transition is directed by SRE sequences and occurs efficiently upon activation of the MAP kinase signalling network (Herschman, 1991; Cahill et al., 1996; Treisman, 1996). Figure 7A and B (and data not shown) show that in Srf−/− embryos both the c-fos and Egr-1 genes are drastically reduced in their expression levels. It is important to note, however, that the activity of both genes was not abolished completely in the Srf−/− embryos, suggesting that basal expression levels of these genes are modulated additionally by factors other than SRF. The α-Actin genes, i.e. those encoding the skeletal, cardiac and vascular smooth muscle α-actins, also represent well-characterized direct SRF target genes (Muscat et al., 1988; Taylor et al., 1989; Treisman, 1992; Moss et al., 1994). In wt embryos, α-Actin expression was already observed at times before (E7.5) and concomitant with (E8.5–E9.5) the onset of myogenesis (Figure 7A and B). Heterozygous Srf+/− embryos displayed Actin expression levels comparable with wt. However, no α-Actin gene expression could be detected by our RT–PCR analysis in Srf−/− embryos. This applied equally to skeletal, cardiac and smooth muscle α-Actin genes (Figure 7A and B, and data not shown). This analysis demonstrates that, in mouse embryos, SRE-regulated genes are severely impaired in their expression in the absence of SRF. Furthermore, this expression analysis confirms that SRF is essential for the transcriptional activation of these genes in vivo. This indicates that SRE control sequences are indeed the DNA-binding sites through which SRF exerts its transcriptional regulatory function in living organisms (Treisman, 1987; Herrera et al., 1989). Discussion By generating a null allele of the Srf gene we identified a new function for the transcription factor SRF in mouse embryogenesis. SRF is essential for mesoderm formation during gastrulation. Mouse embryos lacking SRF develop normally until day 6.5 of embryogenesis. However, formation of the mesodermal germ layer does not occur and, consequently, SRF-negative embryos die in utero. This represents the first genetic analysis of SRF function within vertebrates. SRF is not essential for cellular proliferation SRE/SRF-directed gene activation has been observed at different stages of the cell cycle, i.e. at the G0–G1 transition (IEG activation; Gauthier-Rouvière et al., 1991a; Herschman, 1991), during G1 (Gauthier-Rouvière et al., 1991b), and at the G2–M transition (Liu et al., 1994). Accordingly, an essential role for SRF in the regulation of cell-cycle progression has been assumed. In support of this notion, the S.cerevisiae homologue of SRF, Mcm1, was shown to be essential for the G2–M transition in yeast cells (Althöfer et al., 1995; Maher et al., 1995). In contrast, the phenotype of the Srf−/− embryos revealed that the lack of mammalian SRF did not prevent cell proliferation per se, since the Srf−/− embryos developed normally up to E6.5 (Figure 5E) and continued to grow even in the absence of mesoderm (Figure 5B and D). Therefore it is unlikely that general proliferative defects formed the basis of the defective mesoderm formation in Srf−/− embryos. Similarly, preliminary studies with ES cells homozygous for the mutated Srf allele failed to detect any severe consequences on cell proliferation (B.Weinhold, S.Arsenian, A.Nordheim and U.Rüther, unpublished observations). Interestingly, Roch et al. (1998) have shown that cells homozygous for the blistered mutation, representing an allele encoding a defective DSRF, are also not affected in their proliferation but rather in their capacity to differentiate into vein or intervein tissue in the developing Drosophila wing. Taken together, it can be concluded that SRF is not essential for normal progression of the cell cycle, nor are the products of the SRF-regulated genes c-fos (Wang et al., 1992) and egr-1 (Lee et al., 1996), the absence of which did not reveal any general proliferative defects (Field et al., 1992). Impaired expression of SRE-regulated genes in Srf−/− embryos SREs have been strongly implicated in directing the transient, growth-factor-induced transcriptional activation of the immediate early genes (IEGs) c-fos and Egr-1, and the muscle-specific regulation of α- and γ-Actin genes. Whereas these genes are expressed efficiently at E7.5–E8.5 in wt embryos, we see a drastic impairment of their activation in Srf−/− embryos. Thus, our data provide strong evidence that these SRE-containing genes are indeed regulated by SRF in the living organism, at least during mouse embryogenesis at E7.5–E8.5. The observed phenotype of Srf−/− embryos is not likely to be due to impaired expression of c-fos and Egr-1, since individually, none of these display severe gastrulation defects when mutated (Wang et al., 1992; Lee et al., 1996); a corresponding double null mutation has yet to be generated. Impaired Actin gene expression may well contribute to the phenotype (see below). Although the expression of c-fos and Egr-1 was found to be drastically impaired, it was not prevented fully. This may reflect the contribution of other promoter elements to the expression of these genes. Alternatively, it may hint at a dual role of SRF in vivo, in that SRF may contribute to the basal repression of target genes (Shaw et al., 1989b), in addition to mediating their transient transcriptional induction. SRF is essential for mesoderm formation in the mouse embryo Although molecular events of gastrulation, specifically mesoderm induction and patterning, are much better characterized in amphibians than in mice (Smith, 1995), gene ‘knockout’ strategies have provided important insights into vertebrate early embryogenesis (St-Jacques and McMahon, 1996; Tam and Behri