We previously showed that Tgfbr1 null mutant embryos fail to activate markers labeling the hindlimb and GT primordia (the VLM) (Dias et al., 2020). We now assessed the defects of the Tgfbr1 mutants at the axial level at which these genes become active in wild-type embryos. Transverse sections through this area indicated abnormal morphology of the LPM (Figure 1). In wild-type embryos, the somatic and splanchnic LPM layers converge at the posterior region of the trunk at both sides of the developing endoderm at E9.5, ending the celomic cavity. In Tgfbr1 mutant embryos, however, the two LPM layers, and, consequently, the coelomic cavity, continued extending from the trunk region of the embryo until its posterior end (Figure 1b’, d’). Despite this apparent extension of the trunk LPM into the prospective hindlimb and VLM regions, molecular analyses suggested that the LPM properties in this area differed from those observed in the trunk. This was best illustrated by the expression of Irx3 and Foxf1, somatic and splanchnic LPM markers, respectively (Funayama et al., 1999; Mahlapuu et al., 2001). Both markers were expressed following normal patterns in the trunk of Tgfbr1^−/− embryos (Figure 1Aa, Bb, Cc, Dd). However, in contrast to wild-type controls, in Tgfbr1 null mutant embryos these markers were not downregulated at the level of the trunk to tail transition, being Foxf1 even clearly upregulated in this embryonic region (Figure 1B, D). In addition, their expression was no longer restricted to their respective LPM layers, and instead expanded to encompass the entire mesodermal tissue surrounding the celomic cavity. This feature was more clearly observed for Foxf1 (Figure 1Dd’). The expression pattern of the IM marker Pax2 at E10.5 revealed that the pronephric ducts also extended into the posterior end of the embryo instead of merging with the cloaca (Figure 1—figure supplement 1c, c’ , d”). Furthermore, pronephric ducts were bifurcated in the posterior embryonic end in some mutant embryos (2/4) (Figure 1—figure supplement 1D-d”).

Another major mesodermal derivative affected by the absence of Tgfbr1 was the main vascular tree, particularly the region connecting embryonic and extraembryonic circulation. The four orifices surrounding the hindgut in E9.5 embryos observed in transverse sections, and diagnostic of the rDA (Zakin et al., 2005), were not detected in Tgfbr1 mutants, in which only a single expanded vessel was visible on each side of the gut tube (Figure 1Aa’, Bb’, Cc’, Dd’). Pecam1-aided 3D reconstruction of the main blood vessels revealed that the DA of Tgfbr1 mutant embryos was elongated posteriorly, reaching the tip of the gut tube. The posterior portion of the DA formed a vessel of enlarged diameter, as observed in the histological sections, that contrasts with the curvature characteristic of the rDA of wild-type embryos (Figure 2). The DA of the Tgfbr1 mutants still merged with the umbilical artery, although following patterns different to those observed in control embryos. For instance, while in wild-type embryos the paired rDAs remained as two independent vessels, merging just before their connection with the umbilical artery (Figure 2b), the two paired arteries of Tgfbr1^−/− embryos were fused along most of their path ventral to the endodermal tube, from the posterior embryonic tip to the umbilical artery. In addition, the allantois appeared to protrude perpendicularly to the embryo instead of following its curvature, as observed in wild-type embryos (Figure 2d).

Together, the above observations indicate that in the absence of Tgfbr1 the mesodermal tissues derived from the lateral mesoderm fail to execute the normal differentiation routes associated with the trunk to tail transition.

Splanchnic Foxf1 expression depends on endodermal Shh activity (Astorga and Carlsson, 2007; Tsiairis and McMahon, 2009). The Foxf1 expression observed in the lateral layer of the expanded LPM of Tgfbr1 mutant embryos is separated from the endoderm by the celomic cavity, suggesting that it is likely to have a Shh-independent origin. Such Foxf1 expression has been detected in the posterior PS/allantois of E8.5 embryos (Astorga and Carlsson, 2007; Tsiairis and McMahon, 2009), where it plays a role in vasculogenesis (Astorga and Carlsson, 2007). Reorganization of the vascular system to connect the embryonic and extraembryonic circulation occurs within the emergent VLM as the allantois becomes displaced anteriorly during the trunk to tail transition (Downs and Rodriguez, 2020; Rodriguez and Downs, 2017). From a functional perspective, Foxf1 expression in the VLM is likely to derive from its expression domain in the posterior PS/allantois (Figure 2—figure supplement 1). Foxf1 expression in the posterior PS/allantois is not affected by the absence of Tgfbr1 (Figure 3A, B), consistent with the apparent dispensability of Tgfbr1 activity in axial tissues before the transition to tail development. Given the absence of VLM in Tgfbr1^−/− embryos, the Foxf1-positive cells in the extended LPM might represent the derivatives of posterior PS/allantois cells that failed to enter their normal fates, which instead become intermingled with LPM cells extending from the trunk. Indeed, the abnormal development of the posterior aortae in Tgfbr1 mutant embryos might result from compromised development of the Foxf1-positive posterior PS/allantois.

To assess whether the PS contributes to the prospective pericloacal region, we generated a transgenic line (T-str-creERT) expressing the tamoxifen-inducible cre recombinase in the PS under the control of an enhancer of Brachyury (currently known as Tbxt) (Clements et al., 1996) and inducing cre-mediated ROSA26-derived reporter expression (Soriano, 1999; Srinivas et al., 2001; Figure 3—figure supplement 1). Tamoxifen administration at E8.0, which activates cre activity within the 10–12 hr corresponding to E8.5 (Figure 3—figure supplement 2), thus coincident with the start of the trunk to tail transition, resulted in labeling of the VLM of E9.5 embryos, and of the pericloacal mesoderm at E10.0 (Figure 3c). Although this experiment does not allow regional sub localization within the PS, it is consistent with the Foxf1-positive posterior PS/allantois being a functional target of Tgfbr1 activity in the LPM during the trunk to tail transition.

Reporter, gain and loss of function experiments suggested that Isl1 might be functioning downstream of Tgfbr1 signaling in the LPM during the trunk to tail transition (Dias et al., 2020; Jurberg et al., 2013). It has been shown that Isl1 first becomes activated in axial tissues at the PS/allantois junction, just prior to the transition to tail development (Cai et al., 2003; Wymeersch et al., 2019), and Isl1-positive cells later contribute to the VLM, the hindlimbs and GT (Yang et al., 2006). Analysis of the role of Isl1 during the trunk to tail transition might thus provide an independent test of the functional relevance of the posterior PS/allantois for Tgfbr1 activity in the LPM during the transition. We generated Isl1 null mutant embryos using the null allele resulting from the insertion of the cre recombinase replacing the Isl1 gene in the Isl1-cre strain (Srinivas et al., 2001). Isl1 null embryos were embryonic lethal between E9.5 and E10.5, showing malformations in different embryonic structures. Regarding axial development, Isl1 mutants halted their growth around the stage of the trunk to tail transition, as estimated by the number of somites generated (Figure 4D–F, I, J), thus reminiscent of the Tgfbr1 mutant phenotype. However, in contrast to Tgfbr1 mutants, Isl1 mutants generated a structure resembling the tail bud which expressed Sox2 and Tbxt in domains comparable to wild-type embryos (Figure 4A, B, D, E). The presence of a tail bud in Isl1 mutants indicates that this gene might not be involved in the activity of the NMC cells, as suggested by its expression pattern (Cai et al., 2003). However, we observed major morphological and molecular alterations in the region corresponding to the LPM. Interestingly, some of those alterations are comparable to the defects observed in Tgfbr1 mutants. The two layers of the lateral mesoderm, as well as the celomic cavity, extended to the posterior extremity of the embryo (Figure 4H, h’, h’’). Foxf1 expression was also upregulated in the posterior of the embryo, showing a spatial distribution that starts at the dorsal border between the splanchnic and somatic LPM layers, extending to fully cover the somatic lateral mesoderm at more posterior embryonic regions (Figure 4h’, h’’). In addition, the DA were expanded into two globular vessels on either side of the gut tube reaching the caudal embryonic end where they merged ventrally (Figure 4h,h’, Figure 5C,c,D,d). The connection between umbilical artery and the expanded DAs was highly disorganized (Figure 5D). Together, these observations indicate that Isl1 acts downstream of Tgfbr1 to regulate the processes associated with the trunk to tail transition in the LPM, including the main vascular system, and are consistent with the Foxf1-positive area of the posterior PS/allantois being a functional target of Tgfbr1 signaling to reorganize the LPM during the trunk to tail transition.

Molecular analyses showed that Tbx4 was not expressed in the posterior part of Isl1 mutants (Figure 4I, I’, J, J’), indicating that Tbx4 is downstream of Isl1 in the regulatory network controlling the hindlimb/external genitalia, consistent with previous observations using a conditional mutant for Isl1 (Itou et al., 2012; Kawakami et al., 2011).

Analysis of transverse sections of the Tgfbr1 mutant embryos suggested abnormal morphogenesis of the gut tube. Consistent with this, expression of two endodermal markers, Foxa2 and Shh (Ang et al., 1993; Echelard et al., 1993) at E9.5 showed abnormal morphology at the posterior end of the gut tube (Figure 6A, a, B, b, Figure 6—figure supplement 1). This abnormal morphology was clearer in mutant embryos immunostained with the epithelial marker Keratin 8 (Runck et al., 2014), where it was observed that the endodermal tube finished contacting the ventral ectoderm forming a structure reminiscent of the cloacal membrane (Figure 6C, D). Also, in contrast with what was observed in wild-type controls, Tgfbr1^−/− embryos lacked the endodermal widening characteristic of the developing cloaca and failed to extend the endodermal tube caudal to the cloacal membrane to form the tail gut (Figure 6C, D). Remarkably, expression of the endodermal marker Apela Hassan et al., 2010 followed abnormal patterns in Tgfbr1 mutants. Contrary to wild-type embryos, in which the tail endodermal tube was strongly positive for Apela (Figure 6E, e, F, f), in the Tgfbr1 mutants most of the endodermal tube was negative for this marker, its expression being observed only in a few cells in the dorsal part of the gut tube (Figure 6E’, e’, F’, f’). Surprisingly, Apela-positive cells were found mixed with the cells of the expanded LPM (Figure 6E’, F’), suggesting that endodermal progenitors were produced but misrouted, failing to enter the gut tube. Consistently with the endodermal origin of the Apela-positive cells, we observed Keratin 8 staining scattered within the extended LPM of the mutant embryos (Figure 6c, d).

It has been shown that the visceral endoderm contributes to the formation of the embryonic gut, with the hindgut being particularly populated by this extraembryonic tissue (Kwon et al., 2008). To understand whether the absence of cloacal and tail gut structures in the Tgfbr1 mutants resulted from the inability of the posterior visceral endoderm cells to become incorporated into the embryonic definitive endoderm, becoming instead mixed with the mis-patterned LPM cells, we introduced the visceral endoderm reporter Afp-GFP transgene (Kwon et al., 2008) into the Tgfbr1 mutant background. Analysis of E7.5 Afp-GFP^+/0:Tgfbr1^−/− embryos indicated that the dispersal of visceral endodermal cells was not affected by the absence of Tgfbr1 (Figure 7—figure supplement 1). In addition, Afp-GFP-positive visceral endoderm cells were observed in the embryonic endoderm of Tgfbr1 mutant embryos at E8.5, in a distribution comparable to wild-type embryos (Figure 7A, A’, B, B’). At later stages of development, however, when embryos engage in tail development, distinct distributions were observed in the wild-type and mutant embryos. Both at E9.5 and E10.5, the entire endodermal tube of Tgfbr1^−/− embryos contained GFP-positive cells, also showing the premature end at the ventral surface of the embryo and the absence of further posterior extension to form the tail gut (Figure 7D–D’’, F, F’). Importantly, we did not observe Afp-GFP signal mixed with the extended LPM (Figure 7F’’, F’’’). These data thus indicate that the absence of Tgfbr1 does not affect recruitment of visceral endodermal cells to the embryonic gut, and that the abnormal Apela patterns observed in Tgfbr1 mutant embryos are unlikely to derive from misrouting of visceral endodermal cells.

Interestingly, while we observed a significant contribution of Afp-GFP-positive visceral endoderm cells to the embryonic endoderm of wild-type embryos at E8.5 (Figure 7A, A’), we could detect just a few cells in the tail gut of E9.5 embryos (Figure 7C–C’’) and virtually none at E10.5 (Figure 7E, E’). This indicates that, while the visceral endoderm contributes significantly to the embryonic gut up to the region of the cloaca, as previously reported (Kwon et al., 2008), it is likely to play a minor role in the extension of the endodermal tube growing into the tail. The Apela expression patterns indicate that this gene is active in the newly generated endodermal tissues, becoming progressively downregulated after they are part of the gut tube (Hassan et al., 2010). The strong Apela expression restricted to the posterior portion of the endodermal tube within the developing tail (Figure 8A, B, Bb), suggests that the tail gut grows from the addition of cells at the tip of this structure. Interestingly, analysis of the T-str-creERT:ROSA26-R-gal reporter activity upon tamoxifen administration at E8.0 showed the presence of β-galactosidase-positive cells in the tail gut (Figure 3d). A similar finding has also been independently reported using a different T-creERT strain (Anderson et al., 2013). Of note, more anterior regions of the gut were negative for β-galactosidase under these conditions, despite the presence of labeled cells in adjacent mesodermal tissues (Figure 3c). This suggests that the most caudal region of the gut tube grows through the addition of cells generated from a structure derived from the PS located at the end of the growing tail. Consistent with this hypothesis, we observed an Apela-positive structure adjacent to the tip of the gut at the posterior end of the growing tail (Figure 8—figure supplement 1C). A similar structure was observed in embryos immunostained for Keratin 8 (Figure 8; Figure supplement 1A–b).

We further tested whether this Apela- and Keratin 8-positive region at the tip of the tail bud contributes to the gut tube using a DiI-mediated cell tracing approach ex vivo. E9.5 embryos injected with DiI that showed no label in the gut tube just after injection (Figure 8—figure supplement 2A-c3) were analyzed after 20 hr in culture (Figure 8E–h’, Figure 8—figure supplement 2). In the three embryos that filled this criterium, DiI-positive cells were observed in the gut tube epithelium. Additionally, staining was observed in the surrounding mesenchyme and in the neural tube, possibly due to DiI labeling also becoming incorporated into the NMC population. Even considering the possible leakage of the DiI into the NMC compartment, the presence of DiI signal in the gut epithelium further supports the existence of a region at the tip of the tail bud able to generate cells that become incorporated into the tail gut.

The Tgfbr1^+/− (Dias et al., 2020), Alf-GFP (Kwon et al., 2008), ROSA26-R-gal (Jackson Labs stock #003474, RRID:IMSR_JAX:003474; Soriano, 1999), ROSA26-R-EYFP (Jackson Labs stock #006148, RRID:IMSR_JAX:006148; Srinivas et al., 2001), and Isl1-cre Jackson Labs stock #024242, RRID:IMSR_JAX:024242; Yang et al., 2006 used in this work have been previously described. T-str-creERT was generated by cloning the creERT cDNA, containing the SV40 polyadenylation signal obtained from the Cdx2-creERT construct (Jurberg et al., 2013), under the control of the PS enhancer of the Brachyury (Tbxt) gene (Clements et al., 1996). The construct was used to generate transgenic animals by pronuclear microinjection according to standard protocols (Hogan et al., 1994). Mutant and transgenic lines were genotyped from ear or digit biopsies incubated in 50 μl of PBND buffer (50 mM KCl, 10 mM Tris–HCl, pH 8.3, 2.5 mM MgCl₂, 0.1 mg/ml gelatin, 0.45% NP40, 0.45% Tween-20) supplemented with 100 μg/ml of proteinase K at 55°C overnight. Samples were incubated at 95°C for 15 min to heat-deactivate proteinase K. 1 μl of genomic DNA was used in PCR reaction with the relevant primers specified in Table 1.

Tgfbr1 null and Isl1 null embryos were generated from intercrosses between Tgfbr1^+/− and Isl1^+/cre mice, respectively. Embryos obtained from heterozygous crossings were genotyped from their yolk sacs. Yolk sacs were collected to 50 μl of lysis buffer (50 mM KCl, 10 mM Tris–HCl, pH8.3, 2 mM MgCl₂, 0.45% Tween-20, 0.45% NP40) supplemented with 100 μg/ml of proteinase K and incubated at 55°C overnight. Samples were heat-deactivated as described above. PCR was performed using 1 μl of genomic DNA using the primers specified in Table 1.

All animal procedures were performed in accordance with Portuguese (Portaria 1005/92) and European (directive 2010/63/EU) legislations and guidance on animal use in bioscience research. The project was reviewed and approved by the Ethics Committee of ‘Instituto Gulbenkian de Ciência’ and by the Portuguese National Entity ‘Direcção Geral de Alimentação Veterinária’ (license reference: 014308).

Embryos were fixed in 4% paraformaldehyde in PBS (PFA) overnight, then dehydrated through a 25%, 50%, and 75% series of methanol in PBT (PBS, 0.1% Tween-20), then incubated in 100% methanol. Embryos were then rehydrated through a reverse methanol/PBT series and incubated three times in PBT for at least 5 min each at room temperature. Embryos were then bleached for 1 hr in 6% hydrogen peroxide in PBT and permeabilized in 10 μg/ml of proteinase K (Roche #3115801001) in PBT for a time period that depended on the embryo size. The reaction was then quenched with a 2-mg/ml solution of glycine in PBT, washed twice in PBT and postfixed in a 4% PFA and 0,2% glutaraldehyde mix for 20 min, followed by two washes in PBT. Hybridization was performed at 65°C overnight in hybridization solution (50% formamide, 1.3× saline sodium citrate (SSC) pH 5.5 [20× SSC is 3 M NaCl, 300 mM sodium citrate], 5 mM EDTA, 0.2% Tween-20, 50 μg/ml yeast tRNA, 100 μg/ml heparin) containing the relevant digoxigenin-labeled antisense RNA probes. RNA probes were in vitro transcribed from the linearized vector for 3 hr at 37°C with the corresponding RNA polymerase and DIG RNA Labeling Mix (Roche #11277073910). The reaction product was verified in 0.8% agarose gel and diluted in hybridization solution for further use. After hybridization, embryos were washed twice at 65°C with hybridization solution without tRNA, heparin, and the RNA probe and then in a 1:1 mix of hybridization solution and TBST (25 mM Tris–HCl, pH 8.0, 140 mM NaCl, 2.7 mM KCl, 0.1% Tween-20) for 30 min at 65°C. Embryos were then washed three times with TBST at room temperature, equilibrated in MABT (100 mM maleic acid, 150 mM NaCl, 0.1% Tween-20, pH 7.5) and blocked in MABT blocking buffer [MABT containing 1% blocking reagent (Roche #11096176001)] with 10% sheep serum for 2.5 hr at room temperature. Embryos were then incubated overnight at 4°C with a 1:2000 dilution of alkaline phosphatase-conjugated anti-digoxigenin antibody (Roche #11093274910, RRID:AB_514497) in MABT blocking buffer with 1% sheep serum. After extensive washes with MABT at room temperature, embryos were equilibrated in NTMT buffer (100 mM Tris–HCl, pH 9.5, 50 mM MgCl₂, 100 mM NaCl, 0.1% Tween-20) and developed with a 1:50 dilution of NBT/BCIP solution (Roche #11681451001) in NTMT at room temperature in the dark. Stained embryos were mounted in a 0.45% gelatin, 27% bovine serum albumin (BSA), 18% sucrose mix, jellified with 1.75% glutaraldehyde and sectioned at 35 μm on a Leica Vibratome VT 1000 S. At least three embryos of each relevant genotype were analyzed with each probe.

Embryos were fixed in 4% PFA on ice for 2 hr and then dehydrated through a 25%, 50%, and 75% methanol/PBST (PBS, 0.1% Triton X-100) series followed by 100% methanol. Embryos were then rehydrated through a reverse methanol PBST series, washed with PBST and permeabilized in 0.5% Triton X-100 in PBS for 1 hr and incubated in 1 M glycine in PBST for 30 min to reduce unspecific binding. After several washes in PBST embryos were blocked in 1% BSA, 3% donkey serum in PBST at 4°C overnight. Embryos were then incubated for 72 hr at 4°C with the following dilutions of the primary antibodies in blocking buffer: Pecam1 1:50 (Abcam #ab28364, RRID:AB_726362), Keratin 8 1:100 (Troma1, developed by Dr Brulet and Dr Kemler, obtained from the NICHD Developmental Studies Hybridoma Bank maintained by the University of Iowa, RRID:AB_2891089). Secondary antibodies were diluted 1:1000 in blocking buffer and embryos incubated for 48 hr at 4°C. After extensive washes in PBST embryos were stained with a 1:10,000 DAPI dilution in PSBT at 4°C overnight. Embryos were then mounted on a depression slide with RapiClear 1.49 clearing reagent (SunJin lab). Embryos were imaged on a Prairie Multiphoton microscope using an Olympus 20× 1.0 NA W objective. Stacks were then digitally stitched in Fiji using the Grid/Collection stitching plugin. After removing the outliers, tissues were segmented using Amira Software. Three biological replicates were performed per genotype.

E7.5 embryos were fixed for 20 min in 4% PFA at room temperature, washed three times in PBST, and then counterstained in 5 μg/ml Hoechst and 5 U/ml phalloidin before imaging. E8.5 embryos were permeabilized in 0.5% Triton X-100 in PBS for 20 min, washed three times in PBST and incubated in blocking buffer containing 5% donkey serum (Jackson Labs) and 1% BSA in PBST for 1 hr at 4°C. Embryos were then incubated overnight at 4°C with Epcam (Biolegend, #118202, RRID:AB_1089027) (1:100) and GFP (Aveslabs #GFP-1020) (1:500) in blocking buffer. After three washes in PBST, embryos were incubated with secondary antibody (1:500) overnight at 4°C, and then washed again three times in PBST and counterstained in 5 μg/ml Hoechst. For E9.5 and E10.5 embryos, samples were fixed for 1 hr in 4% PFA at room temperature, washed three times in PBS, then dehydrated through a 25%, 50%, and 75% methanol/H₂O series followed by 100% methanol. Embryos were stained with antibodies against Epcam (1:100) and GFP (1:500) using the iDISCO+ tissue clearing protocol as previously described (Renier et al., 2014) (updated at https://idisco.info/). E7.5 embryos were mounted in PBS and imaged on a Zeiss LSM880 using a Plan-Apo 20×/NA0.8 M27 objective. E8.5 embryos were mounted in FocusClear clearing reagent and imaged on a Zeiss LSM 880 using a Plan-Apo 20×/NA0.8 M27 objective. E9.5 embryos were mounted in dibenzyl ether and imaged on a Zeiss LSM 880 using a Plan-Apo 20×/NA0.8 M27 objective and an EC Plan-NEOFLUAR 10×/NA0.3 objective. E10.5 embryos were imaged on a Luxendo MuVi SPIM light-sheet microscope using a Nikon Plan-Apo 10×/NA0.8 Glyc objective. Raw image data were processed in ZEN (Zeiss), Luxendo Image Processor, or Imaris (Bitplane, http://www.bitplane.com/) software.

E9.5 embryos were dissected out in ice cold media [DMEM (Gibco, Life Technologies #11965092)/15% fetal bovine serum (FBS; Gibco, Life Technologies, #A5670701)/11 mM HEPES (Sigma #7365-45-9)]. Stock DiI solution was prepared by dissolving aliquot of powdered CellTracker CM-DiI (Life technologies, #C7000) in 10 μl of 96% ethanol. Working labeling solution was a 1:10 dilution of the stock solution in 0.3 M sucrose (Sigma). Embryos were injected with labeling solution in the ventral tail bud by mouth pipetting using a glass capillary and collected individually in a 24-well plate on ice. Injected embryos were first imaged on SteREO Lumar.V12, Zeiss and then cultured for 20 hr at 37°C in a rotator bottle culture apparatus (B.T.C. Engineering, Milton, Cambridge, UK) at 37°C, in a 65% O₂ atmosphere. Each embryo was cultured individually in a tube with 1.5 ml of DMEM/F-12, GlutaMAX (Gibco, Life Technologies, #31331-028) containing 15% FBS, 11 mM HEPES (Sigma #7365-45-9) and supplemented with a penicillin/ streptomycin mixture. Cultured embryos were washed with PBS, imaged on SteREO Lumar.V12, Zeiss and fixed in 4% PFA for 1 hr at room temperature. Next, embryos were permeabilized with 0,3% Tween in PBS for 1 hr at room temperature. Nuclei were stained with a 1:5000 solution of DAPI in PBS for 16 hr at 4°C with rotation. Embryos were then washed in PBS, mounted on depression slides, and cleared with RapiClear 1.49 clearing reagent (SunJin lab). Imaging was done on a Prairie Multiphoton microscope using an Olympus 20× 1.0 NA W objective.

Pregnant females from T-str-creERT and either ROSA26-R-βgal or ROSA26-R-YFP intercrosses were treated with tamoxifen (200 μl of a 1 mg/ml solution in corn oil) by oral gavage at different times from E7.5 to E8.5 and harvested at E9.5 or E10.5. When the ROSA26-R-YFP reporter was used, embryos were dissected on PBS and observed with SteREO Lumar.V12, Zeiss. When the ROSA26-R-βgal reporter was used embryos were fixed with 4% PFA at 4°C for 30 min, then washed three times in PBS containing 0.02% Tween-20 for 10 min each at room temperature and developed with 0.4 mg/ml X-gal in 5 mM K₃Fe(CN)₆, 5 mM K₄Fe(CN)₆·3H₂O, 2 mM MgCl₂, 0.02% NP40, 0.02% Tween-20, in PBS for several hours at 37°C in the dark. The reaction was stopped with wash buffer, fixed with 4% PFA overnight and sectioned as described for the whole-mount in situ-stained embryos. To test the time required to observe reporter activity upon tamoxifen treatment, pregnant females from T-str-creERT and ROSA26-R-YFP intercrosses were treated with a single tamoxifen dose and embryos recovered after 6, 8, or 10 hr after treatment and evaluated for YFP signal.