In eukaryotes, most secretory and membrane proteins are synthesized in the endoplasmic reticulum (ER) and are then transported through a protein-conducting channel in the ER membrane for secretion or membrane integration. The elements involved in the translocation machinery have been characterized collectively as the translocon (Beckmann et al.,1997; Johnson and van Waes,1999; Ménétret et al.,2000; Hegde and Kang,2008). The Sec61 complex is a major component of the translocon that contributes to protein translocation. Other components, such as the translocon-associated protein (TRAP, also called the signal sequence receptor) complex are generally required for efficient protein translocation (Gorlich et al.,1992; Fons et al.,2003; Hartmann et al,1993; Ménéétret et al,2005). The TRAP complex comprises four subunits, α, β, γ, and δ, localized in the ER membrane and associating with the Sec61 translocon as a heterotetramer. The α, β, and δ subunits contain a single ER transmembrane region, whereas the γ subunit spans the membrane four times. A large part of each subunit localizes to the ER lumen. The TRAP complex is thought to be a functional component of the translocon that acts in a substrate-specific manner to facilitate the initiation of protein translocation (Fons et al.,2003). During translocation through the ER, synthesized proteins undergo various modifications, such as folding, processing, and maturation. The TRAP complex is also involved in protein translocation through the ER, as well as in posttranslational modification of those proteins during translocation. Nagasawa et al. (2007) found that all four TRAP subunits are induced simultaneously by ER stress and that the TRAP complex binds preferentially to misfolded proteins rather than correctly folded wild-type substrates. Subsequently, the TRAP complex induced by the unfolded protein response (UPR) pathway might discriminate ER-associated degradation substrates from correctly folded substrates, which results in accelerating degradation (Nagasawa et al.,2007). One of the four TRAP subunits, Trap δ also binds misfolded dismutase-1 superoxide, but not the wild-type form, with the ubiquitin-protein isopeptide ligase and this forms an ubiquitinated protein complex (Kunst et al.,1997; Miyazaki et al.,2004). Therefore, the TRAP complex might be involved in the UPR pathway to maintain cellular homeostasis when cells are exposed to unfavorable conditions, such as oxidative stress, viral infections, and glucose deprivation.
The mammalian placenta provides a large surface area for exchanging nutrients and gases between the mother and embryo. During gestation, the placenta supports embryonic growth, and placental insufficiency can result in intrauterine embryo growth retardation (IUGR). Placental growth accompanies fetal–maternal interactions. Intercellular autocrine/paracrine stimuli are required for promoting trophoblastic proliferation, differentiation, and invasion during placental development and for establishing the intricate vascular networks to supply nutrients and gases to the embryo. Several secretory proteins, such as growth factors, cytokines, and membrane proteins including their receptor proteins, that are expressed in the embryonic part of the placenta are essential for placental development (Rossant and Cross,2001; Watson and Cross,2005). Furthermore, recent molecular evidence suggests that efficient protein production is important for placental development. For example, protein synthesis inhibition secondary to ER stress has provided an explanation for the “small placenta” phenotype and subsequent IUGR (Burton et al,2009).
We previously identified XenopusTrap-γ encoding the γ-subunit of the TRAP complex by screening a cDNA library constructed from the lateral mesoderm of Xenopus neurulae (Li et al.,2005). Using morpholinos, depletion of Trap-γ expression in the lateral marginal zone of 32-cell-stage Xenopus embryos resulted in a failure of tubulogenesis during pronephric development (Li et al.,2005). To investigate the role of Trap-γ in mouse development, we generated Trap-γ knockout mice (Ssr3, Mouse Genome Informatics). Here we report that genetic inactivation of Trap-γ in mice results in impaired placental formation and IUGR. On embryonic day (E) 13.5, mutant placentae showed vascular network malformation with poor proliferation of embryonic endothelial cells in the chorionic plate region and impaired labyrinth zone development along with increased apoptotic cell death, whereas the embryos were unlikely to show an apparent morphological abnormality at this stage. On E18.5, mutant placentae showed extensive growth retardation with significant reduction of the blood space and of the vascular networks in the labyrinth zone, while mutant pups showed severe IUGR, especially with collapsing alveoli in the lung, and they all died soon after birth. Therefore, our results suggest that Trap-γ is required for vascular network formation during placental development and acts during intrauterine embryonic growth by regulating efficient protein translocation and/or modification in the ER.
Generation of Mice With Targeted Mutagenesis in the Trap-γ Locus
Mice carrying a null allele of Trap-γ were generated through homologous recombination in the TT2 ES cell line (C57BL/6 × CBA strain cross). The gene-targeting scheme used to generate ES cells with a modified Trap-γ locus on chromosome 3 is shown in Figure 1. We generated recombinant ES cells harboring the Trap-γ-null allele by replacing a part of the first exon of Trap-γ before the initiation codon with a lox71-LacZ-pA-frt-Pro-Neo-frt-loxP-pA cassette to monitor Trap-γ expression with reporter LacZ (Fig. 1A). The recombination was confirmed by Southern blotting with either 3′ or 5′ probes (Fig. 1B). The recombinant ES cells were injected into 8-cell-stage embryos to generate chimeric mice. These were crossed with C57BL/6 mice to obtain Trap-γ heterozygous offspring and the Trap-γ heterozygous mutants (designated Trap-γ+/LacZ) mice appeared to be normal and fertile. To generate homozygous mutant (Trap-γLacZ/LacZ) mice, Trap-γ+/LacZ mice were interbred and the offspring genotyped by genomic polymerase chain reaction (PCR) amplification (Fig. 1C). Reverse transcription (RT)-PCR analysis showed that Trap-γ mRNA was absent from the E18.5 Trap-γLacZ/LacZ embryos (Fig. 1D) and depletion of Trap-γ protein from the E18.5 Trap-γLacZ/LacZ embryos was also confirmed by Western blotting using an anti-Trap-γ antibody (anti-SSR3, Sigma-Aldrich; Fig. 1E).
Expression of Trap-γ in the Embryo and Placenta
By crossing a Trap-γ+/LacZ male with a wild-type female, we obtained Trap-γ+/LacZ embryos and placentae. Using staining for LacZ reporter expression, Trap-γ appeared to be expressed widely in the E10.5 Trap-γ+/LacZ embryos (Fig. 2A, 2D′). Histologically, LacZ staining was observed ubiquitously in the E10.5 embryo (Fig. 2A′). This was consistent with results annotated in the Gene Expression Nervous System Atlas database (http://www.ncbi.nlm.nih.gov/projects/gensat/) that Trap-γ expression is found ubiquitously in the E10.5 entire embryo by in situ hybridization on sections (GENSAT Image 73015 and 73040). In the E18.5 Trap-γ+/LacZ embryo, LacZ staining was also observed widely in several organs, such as the heart, lung, and kidney (Fig. 2B, C). Histochemically, LacZ staining was observed ubiquitously in the entire lung (Fig. 2B′).
In contrast, strong LacZ reporter expression was found in the vascular networks in the chorionic plate region of the placenta (Fig. 2D–F). Histochemically, specific LacZ staining was observed in embryonic vascular endothelial cells in the chorionic plate region but not in other tissues of the E10.5 placenta (Fig. 2 F′, F′′; arrowheads), compared with the wild-type counterpart (Fig. 2E, E′). The Trap-γ-null allele that we used in this study contains the Neo-cassette behind the reporter LacZ gene and the presence of the Neo-cassette is known to inhibit the endogenous promoter activity for reporter LacZ expression. We then confirmed Trap-γ expression in the placenta by in situ hybridization. Predominant Trap-γ expression was found not only in the vascular endothelial cells in the chorionic plate region (Fig. 2G, G′, arrowheads) but also in the spongiotrophoblast layer of the E13.5 placenta (Fig. 2G), where the gene for trophoblast-specific protein a (Tpbpa), a spongiotrophoblast-specific marker (Lescisin et al.,1998), is expressed (Fig. 2H). In situ hybridization also revealed depletion of Trap-γ expression from the Trap-γLacZ/LacZ placenta (Fig. 2I). Therefore, these results suggest that Trap-γ is widely expressed in the entire embryo and is expressed predominantly in the embryonic vascular endothelial cells in the chorionic plate region and in the spongiotrophoblast layer of the placenta.
Depletion of Trap-γ Results in Intrauterine Embryonic Growth Retardation and Trap-γLacZ/LacZ Embryos Die Shortly After Birth
By intercrossing Trap-γ+/LacZ mice, Trap-γLacZ/LacZ embryos were obtained at the expected Mendelian ratio (34/128 embryos; 27%). Mutant pups were alive at birth, but all of them died after a couple of hours, although they were breathing and their hearts were beating. On E18.5, just before birth, we found that Trap-γLacZ/LacZ embryos showed significant IUGR (Fig. 3A, B). The Trap-γLacZ/LacZ embryos (1.02 ± 0.10 g body weight; n = 8) were significantly lighter than the Trap-γ+/LacZ embryos (1.35 ± 0.07 g; n = 8; Table 1; P < 0.001 by Student's t-test). Consistently, organs such as the heart, lung, and kidney of the Trap-γLacZ/LacZ embryos were smaller than those of the Trap-γ+/LacZ embryos seen by whole-mount views (Fig. 3C–E). Further, histological sections also showed retarded growth phenotypes in the mutant organs, such as in the heart, lung, and kidneys. Thus, the principal organs in the Trap-γLacZ/LacZ embryo were formed but were smaller than in the Trap-γ+/LacZ embryos (Fig. 2F–I). Notably in the lung, alveoli were formed but were small and likely to be collapsed (Fig. 2F–G′), which might account for the neonatal deaths of Trap-γLacZ/LacZ embryos.
Table 1. The Mean Weights of Homozygous Trap-γLacZ/LacZ Embryos and Placentae Compared With Those for Heterozygotes, Trap-γ+/LacZ, on Embryonic Day 18.5a
Mean weight of embryos (g) ± SD
Mean weight of placentae (g) ± SD
There were statistically significant differences between the different genotypes in the embryo and placental weights (P < 0.001 by Student's t-test). SD, standard deviation; n, number of embryos and placentae examined.
1.35 ± 0.077 (n = 8)
0.096 ± 0.01 (n = 8)
1.02 ± 0.10 (n = 8)
0.072 ± 0.01 (n = 8)
Depletion of Trap-γ Results in Insufficient Placental Growth
We next examined placental development in the mutants and found that the E18.5 Trap-γLacZ/LacZ placentae were significantly lighter and smaller than in the Trap-γ+/LacZ conceptuses (Table 1; P < 0.001 by Student's t-test; Fig. 4A, B). Histological analysis also showed insufficient placental growth (Fig. 4C–D′). For example, there was a significant reduction of the blood space at the border of the trophoblast cell layer and maternal decidua in the E18.5 Trap-γLacZ/LacZ placenta (Fig. 4E, F, arrows). The mean surface area of the blood space in the E18.5 Trap-γLacZ/LacZ placentae (2.1 ± 0.01 mm; n = 3) was significantly smaller than in the Trap-γ+/LacZ placentae (5.9 ± 0.05 mm; n = 4; P < 0.001 by Student's t-test; Fig. 4K). Using immunostaining with an endothelial cell-specific and hematopoietic progenitor cell-specific anti-CD34 antibody, there was a clear reduction in the size of CD34-positive entire labyrinth zone and extensive reduction of CD34-positive vascular networks in the labyrinth zone in the E18.5 Trap-γLacZ/LacZ placenta (Fig. 4G–H′, arrowheads). Further, we examined the formation of embryonic blood vessels in the chorionic plate region and the spongiotrophoblast layer, where Trap-γ is strongly expressed (Fig. 2D–J′′). There was a significant reduction in the numbers of embryonic blood vessels in the chorionic plate region of the E18.5 Trap-γLacZ/LacZ placentae (Fig. 4G′′, H′′). The mean number of blood vessels in the chorionic plate region in a section of the E18.5 Trap-γLacZ/LacZ placenta (4.2 ± 0.1; n = 5) was significantly less than in the Trap-γ+/LacZ placenta (9.3 ± 0.11; n = 3; P < 0.001 by Student's t-test; Fig. 4L). On the other hand, the formation of the spongiotrophoblast layer was unlikely to be affected in the E18.5 Trap-γLacZ/LacZ placenta. Using in situ hybridization, the Tpbpa-positive spongiotrophoblast layer was observed in the Trap-γLacZ/LacZ placenta, similar to the wild-type embryo (Fig. 4I–J′).
Significant Malformation Already Exists in the E13.5 Trap-γLacZ/LacZ Placenta by Mid-Gestation
At mid-gestation (E13.5), the Trap-γLacZ/LacZ embryos were unlikely to show IUGR (Fig. 5A, B; in the lung; see Supp. Fig. S1, which is available online), whereas the Trap-γLacZ/LacZ placenta already showed the “small placenta” phenotype (Fig. 5A, B, arrow). In the Trap-γLacZ/LacZ placenta, the embryonic vascular network formation in the chorionic plate region was remarkably impaired, with a reduced number of embryonic blood vessels and little expansion compared with the Trap-γ+/LacZ conceptuses (Fig. 5C–D′). Malformation of the vascular network formation in the Trap-γLacZ/LacZ placenta was confirmed by histology. Endothelial cells were visualized using immunostaining for CD31 (platelet endothelial cell adhesion molecule 1; PECAM-1) and the CD31-positive labyrinth zone in the Trap-γLacZ/LacZ placenta was already smaller than the Trap-γ+/LacZ placenta at this stage of gestation (Fig. 5E–F′). Immunostaining also revealed the malformation of embryonic vascular networks, so that the CD31-positive vascular networks in the labyrinth zone of the Trap-γLacZ/LacZ placenta were sparser than in the Trap-γ+/LacZ placenta (Fig 5E′–F′′). Moreover, the embryonic blood vessel network formation in the chorionic plate region was reduced in the Trap-γLacZ/LacZ placenta compared with the Trap-γ+/LacZ placenta (Fig. 5E′′, F′′, arrowheads).
To further examine the impact of Trap-γ depletion on placental hematopoiesis, we performed immunostaining for CD41 (Integrin αIIb), a hematopoietic cell marker, and endothelial cells were visualized using immunostaining for CD144 (VE-cadherin). In the E14.5 Trap-γLacZ/LacZ placenta, there were CD41-positive hematopoietic cells in the vascular networks similar to the Trap-γ+/LacZ embryo, although the surface area of the CD144-positive vascular networks was reduced in the Trap-γLacZ/LacZ placenta (Supp. Fig. S2A–F′).
Increased Apoptosis in the Labyrinth Zone and in the Junctional Zone of the E13.5 Trap-γLacZ/LacZ Placenta
Using terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assays for apoptosis, we found that programmed cell death was greater in the labyrinth zone of the E13.5 Trap-γLacZ/LacZ placenta than in the Trap-γ+/LacZ placenta (Fig. 6A, B, E). Increased apoptotic cell death was also revealed by staining for cleaved Caspase-3, an apoptosis marker, in the labyrinth zone of the E13.5 Trap-γLacZ/LacZ placenta, compared with the Trap-γ+/LacZ placenta (Fig. 6C–D′). This finding was consistent with the labyrinth zone malformation in the Trap-γLacZ/LacZ placenta shown by our previous histological analyses (Fig. 4C–H′ and Fig. 5E–F′′).
Using TUNEL assays, we also found that apoptotic cell death was greater in the junctional zone at the border of the trophoblast cell layer and maternal decidua, but not in the spongiotrophoblast layer in the E13.5 Trap-γLacZ/LacZ placenta (Fig. 6A′–B′′, E), compared with the Trap-γ+/LacZ placenta. This probably accounted for the significant reduction of the blood space at the border of the trophoblast cell layer and maternal decidua in the E18.5 Trap-γLacZ/LacZ placenta (Fig. 4E, F, arrows, and K).
Proliferation Defects in Embryonic Vascular Endothelial Cells in the Chorionic Plate Region of the E13.5 Trap-γLacZ/LacZ Placenta
Poor proliferation was also found by staining for phospho-histone H3 (pHH3) in the embryonic vascular endothelial cells in the chorionic plate region of the E13.5 Trap-γLacZ/LacZ placenta (Fig. 6F, G, arrowheads). The number of pHH3-positive cells was significantly reduced in the CD34-positive embryonic endothelial cells in the Trap-γLacZ/LacZ chorionic plate (P < 0.005 by Student's t-test; Fig. 6H). Thus, in the Trap-γ+/LacZ chorionic plate there was a mean of 0.0087 ± 0.00002 pHH3-positive cells per CD34-positive endothelial cells (n = 2695), whereas in the Trap-γLacZ/LacZ chorionic plate this was only 0.0013 ± 0.0001 (n = 1,628). Similar reduction in the number of pHH3-positive cells was found in the CD144-positive embryonic endothelial cells in the Trap-γLacZ/LacZ chorionic plate. By contrast, the number of pHH3-positive cells in the CD41-positive hematopoietic cells that had integrated into the vascular tubules in the Trap-γLacZ/LacZ chorionic plate was unchanged compared with the Trap-γ+/LacZ embryos (Supp. Fig. S2G–J).
Our results suggest that depletion of Trap-γ inhibits the proliferation of embryonic vascular endothelial cells in the chorionic plate region and increases apoptotic cell death in the labyrinth zone and the junctional zone. This vascular network malformation could lead to insufficient exchange for nutrients and gases in the Trap-γLacZ/LacZ placenta and adversely affect embryo development.
Several genetic studies have demonstrated that vascular network formation and labyrinth development in the placenta are regulated by a number of intercellular signals such as Fgf, Egf, Notch, LIF, Pdgfb, and Wnt (Threadgill et al.,1995; Ware et al.,1995; Xu et al.,1998; Ohlsson et al.,1999; Krebs et al.,2000; for review see Watson and Cross,2005). Furthermore, inhibition of protein synthesis secondary to ER stress can cause the “small placenta” phenotype and subsequent IUGR (Burton et al.,2009). Inhibition of TRAP complex function by depletion of the Trap-γ subunit might reduce the efficient translocation of secretory and membrane proteins in the embryonic vascular endothelial cells in the chorionic plate region and spongiotrophoblast cells. This could lead to insufficient stimulation of vascular network formation and/or cell survival in the labyrinth and junctional zones of the developing placenta (Figs. 4, 5, 6). This defect in proliferation of the embryonic vascular endothelial cells in the chorionic plate region could also account for increased apoptotic cell death in the placenta and reduced blood space in the Trap-γLacZ/LacZ placenta, because it would lead to insufficient blood supply from the mother.
The TRAP complex is also reported to be involved in the UPR pathway to maintain cellular homeostasis. This is activated when cells are exposed to hostile conditions such as oxidative stress, viral infection and glucose deprivation (Kunst et al.,1997; Miyazaki et al.,2004; Nagasawa et al.,2007). Inhibition of TRAP complex function by depletion of the Trap-γ subunit might influence not only efficient protein translocation but also the UPR pathway. Inactivation of the UPR pathway–related gene IRE1 also resulted in poor placental growth with impaired vasculogenesis in the labyrinth zone (Iwawaki et al.,2009). Therefore, this study provides a potential cue for understanding the molecular pathway between ER stress and placental insufficiency caused by an impaired vascular network formation. In general, upregulation of the UPR pathway is thought to provide a growth advantage to tumor cells. In the hemochorial placenta such as seen in humans and rodents, trophoblast cells actively invade the maternal tissues accompanied by dynamic angiogenesis, using similar molecular pathways to those activated in metastatic tumor cells (Graham and Lala,1992; Cross et al.,1994; Tanaka et al.,1998). However, in contrast to cancer metastasis, trophoblastic invasion is strictly controlled in the placenta during organogenesis. Therefore, this suggests that regulation of UPR activity is likely to be important for controlling angiogenesis and/or trophoblastic invasion in the developing mouse placenta.
Fetuses that are associated with the small labyrinth type mutant placenta usually show IUGR because of insufficient metabolic exchange between fetal and maternal tissues (Watson and Cross,2005). Trap-γLacZ/LacZ embryos died soon after birth with the IUGR phenotype, especially with retarded lung development (Fig. 3F–G′). We speculate that the Trap-γLacZ/LacZ embryos die from IUGR caused by placental insufficiency. However, at present we cannot rule out the possibility that the Trap-γLacZ/LacZ embryo growth retardation depends on placental insufficiency alone. Although mutant placentae, but not embryos, showed poor growth at the mid-gestation stage (E13.5; Fig. 5, Supp Fig. S1), it is possible that Trap-γ is also required for embryonic development. Thus, the conditional inactivation of Trap-γ in specific embryonic tissues might provide us with a method to test whether Trap-γ plays an important and direct role in embryonic development.
The TRAP complex could have distinct functions depending on the combination of each subunit, because depletion of either the α or the γ subunit from the TRAP complex produces different phenotypes in mouse embryogenesis. Insertional mutation of Trap-α, which encodes the α subunit, resulted in serious growth retardation with severe cardiac defects in mouse embryos from E14.5 and the mutant pups died at birth (Mesbah et al.,2006). By contrast, Trap-γLacZ/LacZ embryos did not show such severe growth retardation and cardiac defects and they were still alive at birth (Figs. 3, 5, Supp. Fig. S1). Absence of Trap-α also influences substrate-specific protein secretion so that secretion of gamma interferon and atrial natriuretic peptide are impaired, but preprolactin is not, in Trap-α−/− embryonic fibroblasts (Mesbah et al.,2006). This also suggests that individual TRAP subunits might have a distinct function in protein translocation and/or modification in a substrate-specific manner. Depletion of both Trap-α and Trap-γ (or in combination with other subunits) will allow us to understand the function of each TRAP complex subunit in mouse embryonic and placental development.
In summary, by generating Trap-γLacZ/LacZ knockout mice, genetic inactivation of Trap-γ led to insufficient placental growth together with vascular network malformation and neonatal death of the mutant embryos, along with IUGR. Our finding highlights the impact of the inhibition of efficient protein translocation and modification in the ER on placental vascular network formation. It also provides a new insight into the molecular pathways involved in the “small placenta” phenotype and IUGR, which might be induced when pregnant mothers are exposed to an unfavorable environment.
Generation of Trap-γ Mutant Mice
The gene-targeting scheme used to generate embryonic stem (ES) cells with a modified Trap-γ locus is shown in Figure 1. Our strategy for the generation of knockout mice was essentially as described (Murata et al.,2004). In brief, recombination in TT2 ES cells (C57BL/6 × CBA strain cross) harboring the Trap-γ-null allele and the replacement of a part of the first exon of Trap-γ before the initiation codon with a lox71-LacZ-pA-frt-Pro-Neo-frt-loxP-pA cassette (http://www.cdb.riken.jp/arg/cassette.html) was confirmed by Southern blotting with either 3′ or 5′ probes, as indicated in Figure 1B. Two independent recombinant ES cells were injected into 8-cell-stage embryos to generate chimeric mice. Chimeras were crossed with C57BL/6 mice to obtain Trap-γ heterozygous offspring and these heterozygotes, which we designated as Trap-γ+/LacZ mice (Acc. No. CDB0456K; http://www.cdb.riken.jp/arg/mutant%20mice%20list.html), were interbred to generate homozygous mutants (Trap-γLacZ/LacZ). The offspring were genotyped by genomic PCR amplification (Fig. 1C) using primer sets either for the wild-type allele (F1 and R1) or for the mutant allele (F1 and LR2). Primer sequences were as follows: F1, 3′–tatcca agcctgtctctggagactgagg–5′; R1, 3′–aaatc ctgaaggagcaggtcctcctcg –5′ and LR2, 3′– ttccatggttgtggcaagcttgatggg–5′. Primer positions are indicated in Figure lA as arrows. PCR amplification was performed using the Go-Tag system (Promega, Madison, WI) with 32 cycles of 30 sec at 95°C, 30 sec at 55°C, and 30 sec at 72°C. Homozygous mutant mice from the two independent ES cell clones (nos. 76 and 155, indicated in Fig. 1B) exhibited the same phenotypes and these animals were crossed with C57BL/6 mice to maintain a mixed genetic background (C57BL/6 × CBA) for this study.
Western blotting was performed essentially as described (Tanaka et al.,2005). In brief, detection of the Trap-γ protein in the mutant embryos was performed using an anti-SSR3 primary antibody (1:2,000; Sigma-Aldrich, St. Louis, MO) and an alkaline phosphatase-conjugated secondary antibody (1:10,000; Kirkegaard & Perry Laboratories, Gaithersburg, MD) according to the manufacturer's instructions. The immunogen for anti-SSR3 antibody spans the second to fourth exons of Trap-γ (83–133 amino acids). The detection of γ-tubulin was performed using an alkaline phosphatase-labeled anti-tubulin antibody (1:10,000; Sigma-Aldrich). Western blotting was performed using an XCell II Blot Module (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Alkaline phosphatase activity was detected using NBT/BCIP substrates (Roche, Indianapolis, IN).
RNA Extraction and RT-PCR
Total RNAs were extracted using Trizol (Invitrogen) from E18.5 whole embryos. Reverse transcription was carried out with 0.2 μg RNA using Superscript III (Invitrogen). PCR amplification was performed using an Ex-Taq system (Takara) for Trap-γ with 30 cycles of 30 sec at 95°C, 30 sec at 55°C, and 30 sec at 72°C. The primer set for Trap-γ mRNA amplification was 3′–tatccaagcctgtctgtgggctg agg–5′ and 3′–aaatcctgaaggagcaggtcct cctc–5′ (spanning base pairs 57–547). PCR amplification for Hprt was as described (Yamaguchi et al.,2006).
Histology, Immunochemical Staining, and In Situ Hybridization
For histology, embryos and placentae were fixed with 4% paraformaldehyde (PFA) in phosphate buffer saline (PBS; pH 7.4). After dehydration in a graded ethanol series, tissue specimens were cleared in xylene, embedded in paraffin wax (Paraplast; Oxford Labware, St. Louis, MO), and 10-μm sections were cut for immunochemical staining or hematoxylin and eosin staining. For immunostaining, specimens were treated with a blocking solution (2.5% fetal bovine serum in PBS) and then incubated with either anti-CD31 (PECAM-1, 1:100; PharMingen, San Jose, CA), anti-CD34 (1:50; Abcam, Cambridge, MA), anti-CD41 (Integrin αIIb, 1:200; Santa Cruz Biotechnology, Santa Cruz, CA), anti-CD144 (VE-cadherin, 1:50; Millipore, Billerica, MA), anti-cleaved-caspase-3 (1:400; Cell Signaling, Danvers, MA), or anti-phospho-histone H3 (1:500; Upstate, East Syracuse, NY) antibody overnight at 4°C. The secondary antibody was either anti-rabbit Alexa488, anti-rat Alexa488, anti-rat Alexa350, or anti-rabbit Alexa594 (1:1,000; Molecular Probes, Eugene, OR). Nuclear staining with 4′, 6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich) was performed essentially as described (Tanaka et al.,1998). For the detection of β-galactosidase activity by X-gal staining in Trap-γ+/LacZ embryos and placentae, tissues were fixed with 1% PFA in PBS with 0.02% NP40 for 30 min at 4°C. After washing with PBS with 0.02% NP40, specimens were incubated with a staining solution containing 0.1% X-gal, 2 mM MgCl2, 5 mM K3Fe(CN)6 (potassium ferricyanide), and 5 mM K4Fe(CN)6·6H2O (potassium ferrocyanide), 0.01% deoxycholate, and 0.02% NP40 (all chemicals were purchased from Sigma-Aldrich) in PBS overnight at 37°C. For the sections, X-gal-stained embryos were fixed with 4% PFA overnight, embedded in paraffin wax, and sectioned at 10 μm. Fluorescent imaging was performed using an Olympus IX51 microscope and a DP71 image capturing system. In situ hybridization was carried out using an automated Discovery System (Ventana, Roche) according to the manufacturer's protocols. A cDNA fragment of Trap-γ (57–547 bp) spanning the open reading frame was used as a riboprobe. A specific probe for Tpbpa (69–527 bp) was generated by RT-PCR according to Lescisin et al. (1998). The procedure for the TUNEL assay was essentially as described by Tanaka et al. (2000). Quantification of the blood space area was done using an Olympus DP71 image capturing system.
We thank Dr. Patrick Tam for helpful comments, Dr. Takeshi Terabayashi for useful discussion, and Sayoko Fujimura, Junko Nakai, and Maho Kumagai for technical assistance. This work was supported in part by Grants-in-Aid for scientific research from the Ministry of Education, Science, Sports and Culture of Japan, by the Japan Society for Promotion of Science and by the Global COE Program (Cell Fate Regulation Research and Education Unit) of the Ministry of Education, Science, Sports and Culture of Japan (20200071, 20057021, 21028016 to S.S.T. and 22770220 to Y.L.Y.).