The mitogen-activated protein kinase (MAPK or extracellular receptor kinase, ERK) pathway is thought to play an important role in modulating many cellular events, including cell cycle progression, cell differentiation, and the regulation of embryonic development in a variety of biological systems (Marshall, 1995; Lewis et al., 1998; Schaeffer and Weber, 1999) It is thought that mitogenic signal transduction from extracellular signals requires the MAPK signal transduction pathway through (1) initial activation of membrane bound receptors, (2) transduction of signals through serial conformational changes of cell membrane-associated intermediates, (3) activation through phosphorylation of four successive serine–threonine kinases that may then translocate to the nucleus, and (4) translocation of kinases to the nucleus. Upon entry into the nucleus, the active kinases phosphorylate transcription factors, thus activating them and inducing synthesis of transcripts that ultimately lead to G1–S phase transition and DNA synthesis (cell division; Seth et al., 1992; Whitmarsh and Davis, 1999; Roovers and Assoian, 2000; Wilkinson and Millar, 2000).
The MAPK pathway mediates mitogenic signal transduction in most cell types. It is stimulated by fibroblast growth factor (FGF) in adult somatic cell lines, leading to G1–S phase commitment, DNA synthesis, and cell division (Lovicu and McAvoy, 2001; Rappolee, 2003; Taniguchi et al., 2003). In preimplantation mouse embryos and in trophoblast stem cells (TSC) derived from them, FGF is necessary for maintenance of placental trophoblast cell division (Chai et al., 1998; Tanaka et al., 1998), but the mechanism for transducing the mitogenic signal has not been determined.
Previous studies also suggest that the MEK–MAPK pathway is important in mediating events during egg maturation (Phillips et al., 2002a, b). MEK is a required component of cytostatic factor complex (colony stimulating factor, CSF) in mammalian eggs, and the sequential inactivation of maturation promoting factor (MPF) followed by MAPK inactivation is required for normal spindle function and polar body emission (Phillips et al., 2002a, b). Thus, its function before fertilization is to maintain meiotic arrest as part of the cytostatic factor complex. However, it decreases in amount and activity after fertilization, when it may convert from its role in the egg to prevent cell cycle progression to its role in embryo (Haraguchi et al., 1998, 1999; Iwamori et al., 2000, 2002; Phillips et al., 2002a, b) and adult, to mediate G1-S phase cell cycle progression.
Temporal changes in gene transcription during mammalian preimplantation development are also likely to herald the completion of cleavage and the formation of the first differentiated cells, those of the trophectoderm (Schultz, 1986). Previous databases showed molecules in MAPK pathway, such as FGF receptor 2 (FGFR2), MAPK were present in mouse eggs and/or early embryos. But no databases existed that indicated stage-specific expression of the entire pathway of MAPK transcripts. In addition, expression of any single MAPK pathway mRNA transcript was reported for only one or two time points during the 5 days of preimplantation development. In the present study, we report that the entire MAPK pathway is present in preimplantation mouse embryos at the level of RNA, protein, and phosphoprotein.
FRS2α fibroblast growth factor receptor stimulated, lipid-anchored Grb2 binding protein GAB1 GRB2-associated binder-1 GRB2 growth factor receptor-bound protein 2 Ha-ras protooncogene GTPase HTR human trophoblast cell line (SV large T antigen transformed) Raf1/RafB Raf1/B Raf protooncogene S/T protein kinase MEK1,2,5 MAPK/ERK kinase (MAPKK) MAPK/ERK1,2,5 mitogen-activated protein kinase/extracellular signal-regulated kinase RSK1,2,3 ribosomal S6 kinase SOS1 son of sevenless guanine nucleotide exchange factor TSC trophoblast stem cells
Ten Genes Were Detected as mRNA Transcripts in Unfertilized Eggs and/or Zygotes
Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed on unfertilized eggs, two-cell, eight-cell, and blastocyst-stage mouse embryos for analysis of mRNA expression of 10 members of the MAPK pathway. Whole RNA was purified from eggs or embryos isolated in the presence of carrier rRNA and normalized sets of 10 embryos were subject to RT-PCR. Each experiment was repeated twice with similar results, and all controls were run simultaneously with egg/embryo samples. All 10 MAPK pathway genes were detected in zygotes from the morula to blastocyst stages (Fig. 1), when FGF has its necessary mitogenic effect on the embryo. Three of the genes, SOS1, RSK1, and MAPK/ERK2, had very little maternal mRNA transcripts in the unfertilized egg (for abbreviations, see list). Two of the genes, RSK3 and MAPK/ERK5 had very little mRNA transcript in the blastocyst. GAB1, Raf1, Raf β, MEK1, MEK2, MEK5, and MAPK/ERK1 were detected in the unfertilized egg and in the blastocyst. We can detect as little as 10 pg, 100 pg, and 100 pg of RNA in our positive control for SOS1, RSK1, and MAPK/ERK2, respectively, but cannot detect any band in 3.5 ng of unfertilized oocytes (Table 1; Fig. 1). For RSK3 and MAPK/ERK5 in blastocyst, the 100 pg and 10 pg can be detected in the positive control, but no band is detected in 15 ng of blastocyst.
As shown in Figure 1, Raf1 is expressed abundantly in unfertilized eggs, and its expression remains relatively constant throughout the preimplantation development. MEK1, MEK2, MEK5, MAPK/ERK1, RSK1 are detectable at the lowest level in unfertilized eggs, and gradually increase their expression throughout blastocyst stage. SOS1, GAB1 (Xie et al., in press) are also detectable at a low level in unfertilized eggs, but their expression abruptly increases beginning at the two-cell stage and throughout preimplantation development. This initial increase corresponds to zygotic genomic activation. MAPK/ERK2 was not detectable in unfertilized eggs but was detected at the two-cell stage and increased throughout preimplantation development. MAPK/ERK5 and RSK3 mRNA were detected abundantly and increasingly from unfertilized eggs to eight-cell/compaction stage, but were not detectable at the blastocyst stage. As a positive control, we also determined the expression of mRNA for glyceraldehyde-3-phosphate dehydrogenase (GAPDH), previously reported in preimplantation mouse embryos (Ercolani et al., 1988; Conquet and Brulet, 1990). GAPDH mRNA was expressed at all stages of preimplantation mouse development (Fig. 1).
Another interesting aspect arises from the temporally specific expression pattern between isoforms of a single family. At the mRNA level, RSK1 was barely detected until a low level of expression at eight-cell stage that increased at the blastocyst stage. In contrast, RSK3 was expressed abundantly and increasingly from unfertilized egg to eight-cell/compaction stage but was not detectable at blastocyst stage.
All Protein Products of the MAPK Pathway Genes Tested Are Detected in the E3.5 Blastocyst
In the present study, we found that the entire MAPK pathway is detected in preimplantation mouse embryos at both protein and mRNA level (Figs. 1–15; GAB1, GRB2, and RafB; Xie et al., in press). All antibodies yielded the expected major correct-sized band on Westerns of placental cell lines (Figs. 2–5, 7, 9, 10, 12–14). To study the 15 proteins, we used 34 antibodies and 12 proteins were detected in similar cellular and subcellular distributions by more than one antibody.
Spatial Expression of MAPK Pathway Molecules
Immunocytochemistry on E3.5 or E2.5 embryos was performed to determine the spatial expression of MAPK pathway proteins and phosphoproteins. The results showed that the position of these signal transduction gene products was in the cytoplasmic membrane, cytoplasm, and/or nucleus of embryonic and trophoblast cells in preimplantation mouse embryos, which is in agreement with the anticipated intracellular domain of function for these proteins. Each protein was detected in cells of both inner cell mass (ICM)/embryo and trophoblasts/placental lineages on E3.5.
As expected, the lipophilic sequence of the MAPK pathway was detected in the cytoplasm. FGFR2 (not shown), FRS2α (Fig. 2.), GRB2, GAB1 (Xie et al., in press), SOS1, Ha-ras (Figs. 3, 4) are all present in the cytoplasm/cell membrane of embryonic and trophoblast cells in preimplantation mouse embryo. FRS2α and Ha-ras are myristoylated and farnesylated, respectively, but do not appear to be more membrane-specific than SOS1 and GRB2, which are known to have large cytosolic subpopulations (Nimnual and Bar-Sagi, 2002), However, there may be a sizable population of cytosolic FRS2α and Ha-ras subpopulations before membrane binding.
Raf1 is known to translocate to the nucleus (Olah et al., 1991), but we did not observe this translocation in embryos at E2.5 or E3.5 (Figs. 5, 6). In fact, the phosphorylated form was detected more strongly in the cell membrane in E2.5 embryos (Fig. 6.), whereas the Raf-1 all forms is detected throughout the cytoplasm. This localization in the membrane has also been observed for activated/phosphorylated Raf1 and is thought to be mediated by 14-3-3 (Freed et al., 1994).
MEK1, MEK2, MEK5, MAPK/ERK1,2, MAPK/ERK5 all forms are also present in the cytoplasm of preimplantation mouse embryo ex vivo (Figs. 7–12). It was anticipated that the phospho form of each of these proteins would be largely nuclear, but like many other kinases and transcription factors (Rappolee et al., unpublished observations), MAPK/ERK1,2,5 were cytoplasmic at E3.5. Of these six kinases that are known to translocate to the nucleus after phosphorylation, only phospho MEK1,2 was detected in the nucleus on E3.5 and E2.5 (Figs. 7, 8) and MAPK phospho had some nuclear staining at E2.5 (Fig. 11.). However, the nuclear export inhibitor leptomycin B (LMB; Fukuda et al., 1997a, b) could trap all of these kinases in the nucleus in embryos (Liu et al., manuscript in preparation).
Phospho-MAPK/ERK1,2 was transitional in preimplantation embryos; at E2.5, there was some nuclear localization, but by E3.5, phospho MAPK/ERK1,2 was largely cytoplasmic as reported for postimplantation FGF-responsive cells (Corson et al., 2003; Figs. 10, 11). It is not known if MAPK/ERK1 or MAPK/ERK2 is predominant, but it is suggestive that MAPK/ERK1 is the predominant mRNA transcripts and p42/MAPK/ERK1 is the predominant protein in Westerns.
Beside MEK1,2 phospho, only RSK1 and RSK1 phospho show nuclear localization at all times in all cells in the preimplantation mouse embryo (Fig. 13.). RSK2 (not shown) and RSK3 (Fig. 14) are cytoplasmic in E3.5 embryos ex vivo but can be trapped in the nucleus by LMB (not shown). Of interest, phospho RSK1 is also detected at very high levels in cells in metaphase in embryos (Fig. 14). This increased phosphorylation at mitosis has been observed (Rappolee, 2003) for many signal transduction kinases, and we have observed this for a subset of six of 11 phospho kinases and four of seven phospho transcription factors in the embryo (Liu et al., manuscript in preparation).
In this study, the spatial and temporal expression of MAPK pathway molecules was examined by RT-PCR and immunocytochemistry. The results showed that the mRNA and protein gene products of the entire MAPK pathway are present in preimplantation mouse embryos. In E3.5 embryos, trophoblasts and ICM expressed all proteins at apparently similar levels. It is not surprising to find the proteins expressed similarly in ICM and trophoblasts at E3.5, a time at which lineage-specific proteins at E4.5, ICM-specific oct4 protein (Palmieri et al., 1994) and trophoblast-specific Cdx2 (Rappolee et al., unpublished observations), are expressed in both cell lineages. MAPK pathway proteins and phosphoproteins are detected at intracellular positions consistent with mitogenic function. There was also no apparent difference in the level of expression in polar or mural trophectoderm of the E3.5 embryo. This finding might also be anticipated as FGF is necessary for mitogenesis in TSC and trophoblasts in the preimplantation embryo (Chai et al., 1998; Tanaka et al., 1998), but FGF-stimulated MAPK is also important in the induction of placental lactogen (PL)-1 in differentiated trophoblasts (Peters et al., 2000).
The most surprising data from this study are the large number of signal transduction kinases that are phosphorylated but do not accumulate in the nucleus at E3.5 (Raf1, MAPK1,2,5, MEK5, RSK3). The MAPK pathway is thought to mediate the mitogenic signal from the cell surface to the nucleus in a variety of cell types and plays an important role in cell proliferation and differentiation. Nuclear translocation of MAPK is thought to be essential for mitogenesis (Lewis et al., 1998). In adult somatic cell lines, nuclear translocation of MAPK can be stimulated by FGF, which induces activation of membrane-bound receptors, serial conformational changes of lipophilic proteins, phosphorylation of serine–threonine kinases, and activation of transcription factors (Seth et al., 1992; Whitmarsh and Davis, 1999; Roovers and Assoian, 2000; Wilkinson and Millar, 2000) that ultimately lead to G1–S phase commitment, DNA synthesis, and cell division (Ercolani et al., 1988; Lovicu and McAvoy, 2001; Taniguchi et al., 2003). It has been shown recently that FGF-dependent and MEK-dependent fields of responsive cells in postimplantation mouse embryos also lack nuclear accumulation of MAPK ex vivo (Corson et al., 2003). From postimplantation days E5–E8, a mitotic field of FGF-dependent trophoblasts expressed MAPK phospho only in the cytoplasm. Our unpublished data, using the nuclear export inhibitor LMB and phosphatase inhibitor calyculin A (Ishihara et al., 1989; Fukuda et al., 1997a), suggest that there is a rapid nuclear export of kinases that is phosphatase-dependent, and this finding may account for the lack of accumulation of MAPK phospho in the nucleus in cells responding to FGF.
Previous studies also suggest that MEK-MAPK pathway is important in mediating events during mouse egg maturation (Phillips et al., 2002a, b). MEK is a required component of cytostatic factor complex (CSF) in mammalian eggs, and the sequential inactivation of MPF followed by MAPK inactivation is required for normal spindle function and polar body emission. Thus, MAPK/ERK function before fertilization is to maintain meiotic arrest at metaphase II as part of the CSF. However, MAPK/ERK decreases in amount and activity after fertilization when it may convert from its role in the egg to prevent cell cycle progression to its role in embryo (Haraguchi et al., 1998, 1999; Iwamori et al., 2000) and adult somatic tissue, to mediate G1–S phase cell cycle progression. We saw a shift in cell localization from day E2.5 to E3.5 in MAPK phospho in contrast to Raf1 phospho and MEK phospho. MAPK phospho at E2.5 is largely nuclear ex vivo, but by E3.5, MAPK phospho is cytoplasmic (Figs. 11, 12). In contrast, MEK1,2 all forms is cytoplasmic on E2.5 and E3.5 but has translocated to the nucleus upon phosphorylation on both days as reported previously in cell lines (Figs. 8, 9; Fukuda et al., 1997b; Adachi et al., 1999, 2000). Raf1 cell localization is more similar to the lipophilic segment of the pathway members, FRS2α, GRB2, GAB1, SOS1, and Ha-ras in its membrane localization before, and particularly after, phosphorylation (Figs. 5, 6). Our data suggest that a shift occurs from the maternal function of MAPK phospho in the oocyte nucleus to a cytoplasmic–nuclear import–export equilibrium that does not lead to nuclear accumulation. This process is likely controlled by MEK1,2 and is more similar to the postimplantation MAPK phospho signaling function recently observed (Corson et al., 2003).
Previous studies in mouse embryonic development showed that ICM and adjacent trophoblast cells require FGF signaling to divide beginning at the fifth cell division, which occurs before implantation (Chai et al., 1998). But how FGF transduces its mitogenic signal through intracellular pathways and how other receptor tyrosine kinases signal from the cell surface to transcription factors in the nucleus remain to be unraveled.
FGF signals are mediated by a group of four transmembrane proteins with intrinsic tyrosine kinase activity, known as FGFRs (Basilico and Moscatelli, 1992). FGFR1 through -4 have split tyrosine kinase domains and require FGF ligand-dependent dimerization for activation. Both tyrosine kinase activity and the split domain are required for mitogenic activity (Johnson and Williams, 1993). The existence of tyrosine phosphorylation targets of the FGFR, FRS2α and phospho FRS2α, and GAB1 (Xie et al., in press) suggest that necessary proteins for transduction from the receptor are present and activated. In our study, FRS2α was also shown to be present in cytoplasmic membrane of mouse blastocyst (Fig. 2.). FRS2α is myristoylated and should be in the membrane more than being cytoplasmic. This finding is in agreement with previous studies that indicated FGFRs are poorly autophosphorylated upon ligand binding. Instead, the 90-kDa (also called 89-kDa and 85-kDa) protein FRS2α (Rabin et al., 1993; Kouhara et al., 1997) is phosphorylated at multiple tyrosine sites. Consistent with this, we detected FRS2α phosphorylated on Tyr436 in embryos. It is likely that FRS2α is phosphorylated by FGFR and acts as a docking protein that recruits GRB2 and GAB1 signaling proteins to the plasma membrane during FGF signaling.
FRS2α mediates multiple FGFR-dependent signaling pathways critical for embryonic development; the disruption of the frs2α gene causes severe impairment in mouse development, resulting in embryonic lethality at E7.0–E7.5 (Hadari et al., 2001). FRS2α is also downstream of EGF and may mediate essential signals of EGFR in the implanting embryo (Threadgill et al., 1995). Experiments with FRS2α-deficient fibroblasts also demonstrate that FRS2α plays a critical role in FGF-induced MAPK stimulation and cell proliferation (Hadari et al., 2001). That the FRS2α null does create lethality until after fgf4 and fgfr2, suggests that there is redundancy in FRS2α signal transduction. This redundancy may be by means of FRS2β, SHC, or GAB1 (Rappolee, 2003).
After FGF stimulation, tyrosine phosphorylated FRS2α functions as a site for coordinated assembly of a multiprotein complex that includes GAB1 and the effector proteins that are recruited by this docking protein. FRS2α becomes tyrosine phosphorylated in response to FGF stimulation and subsequently binds to GRB2/SOS complexes, bringing them from the cytosol to the plasma membrane (Kouhara et al., 1997). The GRB2 adapter protein links receptor tyrosine kinases to the Ras/MAPK signaling pathway but does not interact directly with FGFRs (Lowenstein et al., 1992). FRS2α thus provides a link between activation of FGFRs and the Ras/MAPK pathway. Myristoylation of FRS2α is essential for membrane localization, tyrosine phosphorylation, GRB2/SOS recruitment, and MAPK activation. In agreement with these data, our studies show that FRS2α is apparently in the cell membrane and remains there, despite the nuclear trapping activity of LMB (unpublished data).
The protein GRB2 contains an SH2 domain that binds to the tyrosine phosphorylated FRS2α and two SH3 domains that link to proline-rich domains in effector proteins. One target effector is SOS, so GRB2 links receptor tyrosine kinases with the Ras signaling pathway (Holgado-Madruga et al., 1997). It is the SOS1–ras interaction that is blocked by sprouty and sef, FGF- (and EGF-) inducible negative feedback inhibitors of FGF (Minowada et al., 1999; Lin et al., 2002).
GAB1 shares amino acid homology and several structural features with insulin-receptor substrate-1. GAB1 is a substrate of the EGF and insulin receptors and can act as a docking protein for several SH2-containing proteins (Holgado-Madruga et al., 1997). GAB1 and GRB2 may be partially redundant with FRS2α and the null mutant for each is a lethal in early peri-implantation embryos. The lethality for GRB2 occurs before implantation, suggesting it transduces the signal for multiple integrin and receptor tyrosine kinase-activated pathways (Cheng et al., 1998). FRS2α, GRB2, GAB1, SOS1, and Ha-ras showed cytoplasmic/cell membrane localization in E3.5 blastocyst. None of these lipophilic proteins could be trapped in the nucleus by LMB (Liu et al., manuscript in preparation), suggesting that all were permanently localized in the plasma membrane and did not undergo nuclear translocation. In preimplantation mouse embryo, GAB1 and SOS1 mRNA were detected from the oocyte and increasing through the blastocyst stage. To our knowledge, this is the first study to show that the array of docking proteins is present in preimplantation mouse embryo.
Ras also functions primarily as a plasma membrane-localized protein and has enzymatic activity as a GTPase (Leevers et al., 1994). Raf serine/threonine kinases are considered to be the primary Ras effectors involved in the proliferation of animal cells (Avruch et al., 1994). Normally localized in the cytosol in an inactive form, Raf-1 associates with Ras at the plasma membrane after growth factor-induced Ras guanine nucleotide exchange (Leevers et al., 1994). The phosphorylation state of Raf-1 appears to modulate its kinase activity and membrane localization. We have observed this most clearly at E2.5, when the phospho–Raf-1 is mostly membrane bound and the unphosphorylated form is mostly cytoplasmic. In response to signaling events, the activation of Raf-1 is a multistep process involving a change in subcellular localization, protein interactions, and phosphorylation events, linking activated tyrosine kinases and Ras to mitogen and extracellular regulated kinase 1 (MEK1) and MAPK (Moodie and Wolfman, 1994; Marshall, 1995). Raf-1 directly phosphorylates and activates the protein kinase MEK, which in turn phosphorylates and activates MAPK/ERK1 and MAPK/ERK2.
As reported in somatic cell lines, Ras, Raf-1, Raf-Beta also showed cytoplasmic localization in E3.5 blastocysts. The expression of Raf-Beta and c-Raf (gene product: Raf-1) were detected from the oocyte through blastocyst stage at a constant level. These findings are in accordance with earlier studies (Moodie and Wolfman, 1994; Marshall, 1995; Haraguchi et al., 1998; Iwamori et al., 2000) and that used only biochemical assays, whereas the current studies showed the spatial localization of these molecules.
The MAPK pathway plays an important role in modulating many cellular events, including cell cycle progression, cell differentiation, and the regulation of embryonic development in a variety of biological systems (Ray and Sturgill, 1988; Marshall, 1995; Lewis et al., 1998; Schaeffer and Weber, 1999). This signaling cascade is initiated and progresses by the sequential phosphorylation and activation of its components, leading to the phosphorylation of a conserved Thr-X-Tyr motif of the MAPK/ERK. Activated MAPK/ERK is then translocated from the cytoplasm to the cell nucleus (Hulleman et al., 1999), where it activates nuclear transcription factors leading to changes in gene expression.
MAPK/ERK5 is a recently discovered subfamily member of the MAPK family and has been shown to be activated by oxidative stress, hyperosmolarity, and several growth factors (Kamakura et al., 1999). MEK5 is a specific activator of MAPK/ERK5 (English et al., 1995; Zhou et al., 1995). Recent reports demonstrated that the homozygous deletion of MAPK/ERK5 results in embryonic lethality with extraembryonic vascular and embryonic cardiovascular defects (Regan et al., 2002).
MEK1,2,5, and ERK1,2 mRNA expression were detected in the oocyte and increased through blastocyst stage, whereas MAPK/ERK5 was expressed at a constant level, this finding is in agreement with previous studies that show constant presence of MEK and ERK in mouse early embryos by immunoblotting (Haraguchi et al., 1998). But in contrast to the early study using only Westerns, we found by immunocytochemical (ICC) analysis that MEK1,2 (Fig. 6.), MAPK/ERK1,2 (Fig. 5D), and MAPK/ERK5 (Fig. 5D) are activated by phosphorylation. However, only MEK1,2 was clearly nuclear in the phosphorylated state in E3.5 blastocyst. Taken together, these data suggest that phospho MEK1,2 is translocated to the nucleus and may play roles in the import and fast export of MAPK, as has been reported in cell line models (Adachi et al., 2000).
MEK1, MEK2, and MEK5 showed similar patterns of expression in both mRNA and protein level. Whereas most molecules are expressed both at mRNA and protein level, MAPK/ERK5 and RSK3 are not detectable in mRNA expression at E3.5 but both show cytoplasmic localization in protein expression by immunocytochemistry. This finding suggests that the protein products persist after the RNA is no longer detectable.
P90rsk (ribosomal S6 kinase, RSK) is an additional important family of signal-transducing Ser/Thr kinases. Several lines of investigation have suggested that RSK is phosphorylated and activated by ERK1/2 isoforms (Zhao et al., 1996) and are known as MAPK-activated proteins (MAPKAPs). RSK1 has also been demonstrated to undergo translocation from the cytoplasm to nucleus after stimulation in a MAPK-dependent manner (Chen et al., 1993).
Both RSK1 and phospho RSK1 showed exquisitely nuclear localization in E3.5 mouse embryo, whereas RSK2 and RSK3 display a different pattern of protein expression, more cytoplasmic than nuclear. However, RSK2 and RSK3 are trapped in the nuclei of preimplantation trophoblasts by LMB, suggesting that they too cycle through the nucleus, but do not accumulate as much as RSK1. RSK1 mRNA was not detectable until E3.5, while RSK3 mRNA was detected in the oocyte and increased to eight-cell stage, but dropped suddenly at E3.5. The observed difference in temporal and spatial expression among RSK isoforms may indicate different responsiveness and physical association with MAPK isoforms, thus resulting in different functions for RSK1 and RSK3 in mouse embryogenesis.
In summary, we have shown for the first time that the entire MAPK pathway molecules are present in preimplantation mouse embryos in mRNA transcript, protein, and phosphoprotein. The data for cellular location and phosphorylation suggest that the proteins are functional and may have mitogenic function. There is a unique shift in MAPK phospho, but not Raf1 phospho or MEK1,2 phospho, accumulation from the nucleus to the cytoplasm from E2.5 to E3.5, suggesting a shift in MAPK regulation and function to that reported for postimplantation embryos. The present study lays a foundation to plan and analyze rational perturbation studies aimed at understanding the role of the major mitogenic pathways in embryos.
Standard techniques were used for obtaining mouse embryos (Hogan et al., 2002). Female MF-1 mice (4–5 weeks old, Harlan Sprague-Dawley, Indianapolis, IN) were injected intraperitoneally with 10 IU pregnant mare serum gonadotropin (Sigma Chemical Co., St. Louis, MO), followed by an injection of 7.5 IU of human chorionic gonadotropin (Sigma) 44–48 hr later. After the second injection, females were housed overnight with C57BL/6J × SJL/J F1 hybrid males (Jackson Laboratories, Bar Harbor, ME). Noon of the day after coitus was considered day E0.5. For ICC analysis, embryos were obtained at morula/early cavitation blastocyst (E3.5), or eight-cell/compaction stage (E2.5); for RT-PCR, embryos were collected at the following stages: unfertilized egg, two-cell stage (E1.5), eight-cell/compaction stage (E2.5), and morula/early blastocyst (E3.5). The animal use protocols were approved by the Wayne State University Animal Investigation Committee.
Indirect Immunocytochemistry and Nuclear Staining
For immunocytochemical analysis, E3.5 or E2.5 mouse embryos were fixed for 30 min in 2% fresh paraformaldehyde (pH 7.4) in phosphate-buffered saline (PBS), quenched with 0.1 M glycine (Fisher Scientific, Fair Lawn, NJ), and permeabilized for 10 min with 0.25% Triton X-100 (Sigma). The embryos were stained with primary antibodies (diluted at 1:100 in PBS-Tween 20 with 10% fetal calf serum [Gibco]). The primary antibody was followed by staining with biotinylated immunoglobulin G (Vector Labs, Burlingame, CA). Proteins were visualized with fluorescein isothiocyanate coupled to streptavidin (Vector). Nuclear counterstaining was done with Hoechst 33258 (10 μg/ml). Photomicrography was done with a Leica DM IRE2 epifluorescence microscope with a Retiga 1350 Ex cooled CCD controlled electronically by SimplePCI AI module software (Compix, Inc., Imaging Systems, Cranberry Township, PA). No neighbor deconvolution was performed by using the SimplePCI DNN module. Fluorescence photomicrographs for protein and nuclear stain were merged by using SimplePCI. Photomicrographs were formatted by using Adobe Photoshop 6.0 (San Jose, CA). All fluorescence photos were handled and analyzed in the same way. All experiments were repeated at least twice with similar results.
RNA Isolation and RT-PCR
AmpliTaq Gold DNA polymerase was purchased from Applied Biosystems. Restriction enzymes were purchased from New England BioLabs and Gibco/BRL. SUPERSCRIPT Reverse Transcriptase was obtained from Invitrogen Life Technologies. Polymerase chain reaction primers were designed by using the GeneFisher software (Bielefeld University Bioinformatics Server). PCR oligonucleotides were obtained from Integrated DNA Technologies.
Stage-specific mouse embryos were flushed, washed through six drops of M2 media (Sigma) and saved in RLT buffer (from the Qiagen RNeasy Mini Kit) at −80°C before use. Qiagen RNeasy Mini Kit (50) (Qiagen, Inc., Valencia, CA) was used to prepare total RNA from 200 embryos containing 10 μg of 16S- and 23S-ribosomal RNA (Roche) according to the manufacturer's instructions. Yields of RNA were based on carrier amount and ranged from 35% to 80%. RNA used for positive controls was isolated from mouse organs. The concentration of RNA was determined by measurement of absorbance at 260 nm. RT-PCR was performed essentially as described previously (Rappolee et al., 1988, 1989, 1994), with the use of oligonucleotide primers shown in Table 1. Briefly, RNA was reverse-transcribed with 200 units of SUPERSCRIPT Reverse Transcriptase, 0.5 mg of 12- to 18-mer oligo (dT; Invitrogen Life Technologies) in a 20-μl mixture. Reaction mixtures were heated at 37° for an hour followed by 95° for 3 min; 60 additional units of buffered SUPERSCRIPT Reverse Transcriptase were added and the reaction mixture was reincubated at 42° for another hour. The proportion of the reaction mixture that was equivalent to 10 embryos was added to sequence-specific primed PCR reaction mixture in a buffer containing 10 mM Tris-HCl (pH 8.3), 600 μM each dNTP, and MgCl2 range from 3 mM to 7.5 mM in a 50-μl reaction mixture. One microgram of organ total RNA was reverse transcribed, and 1-μl reaction mixture (or a dilution series of it) was amplified by PCR for 40 cycles on an Eppendorf Mastercycler/Mastercycler gradient PCR machine. The PCR fragments were separated on a 4% agarose gel (BMA, SeaKem LE agarose) containing ethidium bromide. Gels were photographed with a Polaroid MP-4 camera, and negative images were reversed for clarity of presentation. Fragments were verified by size and restriction enzyme mapping (Brenner et al., 1989). For restriction enzyme analysis, the fragments were precipitated with ammonium acetate directly from the PCR reaction mixture, washed twice with 70% ethanol, and digested according to the restriction enzyme manufacturer's instructions.
Cell Culture and Western blotting
TSCs were isolated from E3.5 embryos or E6.5 extraembryonic ectoderm and provided as a gift by Dr. J. Rossant (Chai et al., 1998; Tanaka et al., 1998). Briefly, TSCs were grown in mouse embryonic fibroblast conditioned medium (MEF-CM) containing 25 ng/ml FGF1 and 1 μg/ml heparin. MEF cells were cultured in RPMI 1640 (Gibco, Grand Island, NY) containing 20% fetal bovine serum (Gibco), 1% penicillin–streptomycin (Gibco), 1 mM sodium pyruvate (Gibco), 2 mM L-glutamine (Specialty Media, Lavallette, NJ), and 100 μM β-mercaptoethanol (Sigma). MEF-CM was prepared by culturing MEF cells to 100% confluency and treating MEF cells with mitomycin C (10 μg/ml, Sigma) for 4 hr. The medium was collected, centrifuged at 2,000 rpm for 10 min to remove cellular debris, filtered through a 0.22-μm filter (Corning), and stored at −80°C. Human trophoblast cells (HTR) were grown in D-MEM/F-12 (Gibco) supplemented with 10% fetal bovine serum (Gibco; Graham et al., 1993).
TSC or HTR were grown to 70–80% confluency to obtain cell lysates for Western blotting. Cells were washed twice with ice-cold PBS. Western blotting was performed by using whole-cell extracts, which were prepared by incubating the cells in cold cell lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 1 mM ethyleneglycoltetraacetic acid, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na3VO4, 1 μg/ml leupeptin, and 1 mM phenylmethyl sulfonyl fluoride; Cell Signaling) plus Phosphatase Inhibitor Cocktail (PIC1, Sigma) and PIC2 (Sigma) for 20 min. The lysates were centrifuged at 10,000 × g for 10 min, and the supernatant was stored at −80°C. The proteins in 20 μg of whole-cell extracts were separated by electrophoresis on a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel by using a Hoefer Mighty Small II SE 250 apparatus and then transferred to enhanced chemiluminescence (ECL) Hybond nitrocellulose membranes (Amersham) at 15 V for 30 min by using a Bio-Rad Semi-dry Transfer Cell. The membranes were blocked overnight with 5% non-fat milk in Tris buffered saline-Tween 20 (TTBS) and blotted with the specified primary antibodies for 1 hr followed by three washes in TTBS. And then incubated in horseradish peroxidase–conjugated secondary antibody for 1 hr followed by extensive wash with TTBS. Primary and secondary antibodies were diluted in 1% non-fat milk/TTBS. The protein bands were visualized using ECL assay system (Amersham).
We thank Dr. Randy Armant for helpful advice on fluorescent microscopy.
NOTE ADDED IN PROOF
During the submission of this paper, a novel set of data regarding the magnitude of expression of many of the genes in the MAPK pathway (Wang et al., 2004, and database cited within) was described. This gene array approach was largely in agreement with data reported here, with some differences, mainly in the MAPK/ERK family.