Molecular Human Reproduction, Vol. 6, No. 5, 454-464,
May 2000
© 2000 European Society of Human Reproduction and Embryology
Embryo development |
Surface-expressed E-cadherin, and mitochondrial and microtubule distribution in rescue of mouse embryos from 2-cell block by aggregation
1 Laboratory of Cell Morphology, Institute of Cytology, Russian Academy of Sciences, St Petersburg, Russia and 2 Universität Bielefeld, Fakultät für Biologie, Gentechnologie/Mikrobiologie, D-33501 Bielefeld, Germany
Abstract
E-cadherin (uvomorulin)-mediated cell interactions are essential for preimplantation development in mammals. We observed that E-cadherin is expressed at contact sites between blastomeres of 2-cell mouse embryos of non-blocking genotype (CBA x C57BL F1) explanted at 32 h post human chorionic gonadotrophin (HCG) and cultured in vitro, while blastomere rounding and reduced zones of contact and E-cadherin-staining were observed in embryos of a blocking strain (MF1) arrested at the 2-cell stage. Embryos of MF1 strain can be rescued by aggregation with four 2-cell embryos of the non-blocking genotype. An early event in rescue is E-cadherin expression at contact zones between adjacent embryos of different genotype in aggregation chimeras. E-cadherin-mediated signalling appears important for the rescue (including formation of adherens-like contacts, cell polarization and morphogenetic processes) since there is no rescue when E-cadherin-specific antibodies are present during phytohaemagglutinin-mediated aggregation and subsequent culture. In blocked embryos, the distribution of microtubules is disturbed and concomitantly mitochondria cluster around the nucleus. Rescue by aggregation retains normal mitochondrial distribution in the presence of a dense microtubular lattice in all blastomeres. Therefore, E-cadherin-mediated signalling and its downstream effects on cytoskeletal organization are essential in the rescue of blocking embryos by aggregation. Normal preimplantation development appears to be dependent on the polarized expression of surface E-cadherin and the microtubule-mediated dispersal of mitochondria.
2-cell block/E-cadherin/microtubules/mitochondria/preimplantation development
Introduction
Homozygous embryos of many strains of mice arrest at the 2-cell stage when cultured in certain chemically defined media from the 1-cell stage onwards in vitro (Biggers, 1971
; Cross and Brinster, 1973; Whittingham, 1974). The time of the block at the late G2/M phase of the second cell cycle coincides with the switch of maternal to zygotic gene activation (Goddard and Pratt, 1983; Telford et al., 1990). A similar arrest of preimplantation development at the time of zygotic gene activation can also be observed in many other mammalian embryos including the human (Camous et al., 1984; Braude et al., 1988). It was shown that the 2-cell block in the mouse and developmental arrest in other species can be overcome by alterations in culture conditions and composition of media, in particular those affecting energy metabolism (e.g. Lane and Gardner, 1992; Dumoulin et al., 1993; Dawson and Baltz, 1997; Devreker and Hardy, 1997; Dienhart et al., 1997). In fact, there is a direct link between energy sources in the media and developmental capacity of mammalian embryos (Brinster, 1965; Brown and Whittingham, 1992).
Developmental arrest appears also related to altered cytoskeletal organization since mitochondrial distribution and function is disturbed in blocking embryos of the mouse (Muggleton-Harris and Brown, 1988; Sekirina et al., 1997; Matsumoto et al., 1998). This may have far reaching consequences since reduced developmental capacity has been associated with mitochondrial dysfunction and low ATP in mammalian embryos (Van Blerkom et al., 1995; Barnett and Bavister, 1996; Keefe et al., 1996; Barnett et al., 1997; recently reviewed in Van Blerkom et al., 1998). Altered distribution and function of mitochondria may influence the ability of blocked embryos to scavenge free radicals, induce an oxidative stress response (Nasr-Esfahani et al., 1990), and in this way contribute to impaired development.
Sufficient pre-maturation-promoting factor (pre-MPF) appears to be present in blocking embryos of the mouse but its catalytic subunit, cyclin-dependent kinase p34cdc2, cannot be activated (Aoki et al., 1992; Haraguchi et al., 1996). Microinjection of cytoplasm of non-blocking 1-cell and 2-cell embryos into blastomeres of the blocked embryo overcomes the block (e.g. Muggleton-Harris et al., 1982), and may thereby directly provide critical maternally derived cytoplasmic factors necessary for initiating cellular activities required for full zygotic gene activation and cell cycle progression beyond the G2-phase. Recently there was evidence that injection of ooplasm containing mitochondria also supports preimplantation development of human embryos (Cohen et al., 1997; Van Blerkom et al., 1998).
We have previously shown that aggregation of blocking 2-cell embryos of the mouse with 2- or 8-cell stages of a non-blocking genotype (e.g. from the F2 generation of CBAxC57BL strains) also induces restoration of the developmental capacity (Sekirina and Neganova, 1995) and leads to the formation of chimeric embryos with blastomeres from different genotypes (Neganova et al., 1998). It is conceivable that signalling via cell–cell interactions is responsible for restoring the full developmental capacity of blastomeres of blocking embryos during aggregation, since adherens-like contacts are formed between 2-cell embryos of blocking and non-blocking genotypes (Sekirina and Neganova, 1995). In the absence of aggregation, receptor-mediated signalling is also sufficient to rescue embryos from a developmental block (e.g. Gardner and Lane, 1996; Dienhart et al., 1997; Uranga and Arechaga, 1997). When growth factors such as epidermal growth factor (EGF) or members of the transforming growth factor (TGF)-ß family are present in media after explantation, some embryos of a blocking genotype are able to develop to the blastocyst stage in vitro (e.g. Larson et al., 1992; Terrada et al., 1997).
In normal embryogenesis, cell–cell contacts involving cell surface expression of E-cadherin/uvomorulin and signalling via their intracellular partners such as E-catenin are essential for preimplantation development, in particular for cell polarization at compaction (Vestweber et al., 1987; Sefton et al., 1992; Clayton et al., 1993; Rietmacher et al., 1995; Ohsugi et al., 1997; Semb and Christofori, 1998). Maternally derived E-cadherin becomes detectable at the cell surface of activated mouse oocytes 6 h after fertilization (Clayton et al., 1993). We suspected that E-cadherin-mediated signalling and downstream changes in cytoplasmic organization and expression patterns may be important for the rescue in aggregation chimeras. Therefore, we studied the expression of this membrane protein before and after aggregation and used specific antibodies to analyse the role of E-cadherin-mediated signalling in cell polarization and the resue of blocking 2-cell embryos by aggregation.
Furthermore, we wanted to know whether there is a link between the expression of E-cadherin, rescue by aggregation and the distribution of microtubules and mitochondria. The microtubule motor protein kinesin has recently been identified as a critical cellular component for mitochondrial dispersion during early mouse embryogenesis (Tanaka et al., 1998). Developmental arrest is accompanied by disturbances in mitochondrial distribution (Muggleton-Harris and Brown, 1988; Matsumoto et al., 1998). Our observations provide evidence that there is a link between the polarized expression of E-cadherin, presence of a dense network of microtubules and dispersal of mitochondria during early preimplantation development of the mammalian embryo.
Materials and methods
Outbred MF1 mice and CBAxC57BL/6 F1 mice were initially obtained from Harlan Winkelmann (Borchem, Germany). Four to eight week old females were kept in our colony under a 12 h light/dark cycle and fed ad libitum. They were superovulated with 7.5 IU pregnant mare's serum (Intergonan, Intervet, Tönishoven) and 7.5 IU human chorionic gonadotrophin (HCG; Predalon, Organon, Oberschleißheim). After caging overnight with males of the same strain (MF1 or CBAxC57BL/6 F1 males, respectively), or homozygous CBA males, females were checked for plug formation early in the next morning, about 15 h post HCG. Isolation of 2-cell embryos from fertilized females and aggregation between non-blocking and blocking embryos was done on the next day as previously described (Sekirina and Neganova, 1995). In short, oviducts of non-blocking F1 mice were flushed with PBS 42–46 h post HCG, those of the blocking MF1 strain at 34–36 h post HCG, corresponding to the middle or early 2-cell stage, respectively. The zonae of embryos were removed by brief exposure to 0.5% pronase (Boehringer, Mannheim, Germany) and gentle pipetting (Sekirina and Neganova, 1995). Aggregation was performed in M16 medium (Whittingham, 1971) by placing one blocking embryo and four non-blocking embryos close to each other in the presence of 400 mg/ml phytohaemagglutinin (PHA; Gibco BRL, Karlsruhe) supporting physical attachment between embryos of different strains (Mintz et al., 1973). In previous experiments (Neganova et al., 1998) chimeras with transgenic mice of blocking genotype expressing ß-galactosidase were employed to follow cell fate. We did not use the transgenic LacZ animals in the current protocol since fixation of embryos for enzyme assays was incompatible with E-cadherin or tubulin immunofluorescence staining. However, the identity of individual embryos in aggregates can be determined by their different degree of pigmentation (MF1 embryos possess a more translucent cytoplasm) and by the relative positioning inside of aggregates (one 2-cell embryo of MF1 strain was surrounded by four 2-cell embryos of the non-blocking genotype). Occasionally, individual embryos of aggregation chimeras become displaced relative to each other during centrifugation/fixation procedures but were still recognizable by their distinct cytoplasmic appearance.
Aggregates between 2-cell embryos of non-blocking, blocking or both genotypes were cultured under standard conditions in M16 medium (Whittingham, 1971) with 4 mg/ml BSA (Sigma, Deisenhofen) under mineral oil for 10–12 h prior to fixation. Individual 2-cell embryos were also explanted at 34–36 h post HCG and cultured until blastocyst formation (110 h post HCG; non-blocking genotype) or until 72 h post HCG when they were in deep block (MF1 embryos). In experiments involving aggregation in the presence of antibodies, a monoclonal rat anti-E-cadherin antibody (U 3254, Sigma) or a monoclonal rat anti-tubulin antibody (YL1/2, Biozol, Eching) was present at a dilution of 1:1000 during the aggregation procedure and the following period of culture. The monoclonal rat anti-tubulin antibody was directed against the tyrosinated form of 
-tubulin (Kilmartin et al., 1982). Its specific reaction with microtubules in mouse oocytes and embryos has been shown previously (Eichenlaub-Ritter and Boll, 1989; Houliston and Maro, 1989). The monoclonal anti-uvomorulin antibody (rat IgG1 isotype) was selected against mouse cell adhesion molecule E-cadherin (L-CAM). It blocks both the aggregation of mouse embryonal carcinoma cells and the compaction of preimplantation embryos, and can be used in immunofluorescence, immunoblot or immunoprecipitation techniques with mouse tissue (Sigma, product information).
Zona-free 2-cell embryos and aggregates were rinsed once in M2 medium and placed in specially designed chambers on coverslips as previously described (Maro et al., 1984) for fixation prior to immunoreactions with anti-E-cadherin antibody. The chambers had been coated with 0.1 mg/ml concanavalin A (Sigma) in PBS before use. Embryos were centrifuged at 450 g for 10 min at 37°C onto the coverslips at the bottom of the chambers. Glass-attached, flattened embryos were then fixed for 20 min in 1% formaldehyde in PBS, followed by washing and permeabilization in 0.25% Triton X-100 in PBS for 30 min (Fleming et al., 1991). E-cadherin was detected by the monoclonal rat anti-E-cadherin antibody against mouse embryonal E-cadherin (U 3254 from Sigma) at a dilution of 1:1000, followed by washing in PBS–Tween (Fleming et al., 1991) and reaction with an anti-rat TRITC-labelled second antibody (Sigma). Second antibody labelling alone did not show any background staining when early 2-cell stages or embryos in blastocyst stage of development were viewed and when images were recorded with the same exposure times as used for documentation of anti-E-cadherin-induced images. Therefore, these unstained images are not provided as negative controls in the figures. Images obtained from blastocyst stages reacted with anti-uvomorulin antibody display the characteristic distribution and strong staining reaction between neighbouring cells, as is especially characteristic for E-cadherin/uvomorulin at this particular stage of development, as demonstrated previously (Vestweber et al., 1987; Fleming et al., 1994; Campbell et al., 1995).
For anti-tubulin immunofluorescence, embryos were extracted in a microtubule-stabilizing solution and attached to slides and fixed according to the procedure used for mouse oocytes (e.g. Eichenlaub-Ritter and Betzendahl, 1995; Yin et al., 1998b). Specificity of the monoclonal mouse anti-tubulin antibody used for immunostaining has previously been shown in mouse oocytes (e.g. Eichenlaub-Ritter and Betzendahl, 1995; Yin et al., 1998b). Again, reaction with second antibody alone provided only dark, unstained images when recorded with the same exposure times as used for the anti-tubulin-reacted samples.
DNA and chromosomes of fixed embryos were stained by DAPI (Sigma). Coverslips with samples were finally immersed in PBS plus diaminobicyclo-octane (DABCO, Sigma, Deisenhofen) as antifade agent, placed onto a slide and viewed with an Axiophot microscope and appropriate filters as previously descibed for mouse oocytes (Eichenlaub-Ritter and Boll, 1989; Eichenlaub-Ritter and Betzendahl, 1995; Yin et al., 1998b).
Distribution and activity of mitochondria was characterized in living embryos by vital staining with Mitotrack TM (Molecular Probes, Eugene, OR, USA) as previously described by Yin et al. (1998a). Nuclei and chromosomes were vital-stained by Hoechst 33342 at the same time (Yin et al., 1998a).
Results
E-cadherin in embryos of the non-blocking and blocking genotype
To follow the expression pattern of E-cadherin in embryos of different genotype developing in vivo and in vitro, we performed indirect immunofluorescence with a monoclonal rat anti-E-cadherin antibody derived by immunization of rat with mouse surface-expressed embryonal E-cadherin (Yoshida and Takeichi, 1982). Immunofluorescence showed that E-cadherin was present in the cytoplasm and some protein was expressed at the surface of in-vivo developed 2-cell embryos of both strains of mice at 34 h post HCG (Figure 1A, A') but especially bright and prominent staining was found at the cell–cell contact zones between the two blastomeres (Figure 1A'), and at the surface between blastomeres and the second polar body. When 2-cell stages of both blocking and non-blocking genotypes, MF1 and CBAxC57BL/6 F2, were observed after culture in vitro from 34 h to 48 h after HCG, E-cadherin staining was retained but was not so prominent as compared to embryos developing in vivo (Figure 1B, B'). Such embryos still possessed a second polar body (Figure 1B). Maternal E-cadherin remained enriched at the contact zones between the two blastomeres of individual embryos of both genotypes upon culture for up to 48 h post HCG (Figure 1B'). However, embryos of the non-blocking genotype (CBAxC57BL/6 F2) developed to blastocysts within 110 h post HCG and exhibited prominent expression of zygotic E-cadherin between neighbouring cells (Figure 1D, D'). In contrast, the embyos of the blocking genotype (MF1) arrested development at the 2-cell stage when explanted at 34–36 h and exhibited the characteristics of a 2-cell block thereafter. Their blastomeres attained a compact, roundish shape after 10–12 h of culture in vitro (46 h post HCG). In a deep block when embryos were cultured in vitro until 64–70 h post HCG, blastomeres touched each other only at a restricted area (Figure 1C, C'). The second polar body was no longer visible in most of the embryos in deep block (Figure 1C). E-cadherin expression at contact sites between blastomeres was still visible (Figure 1C') but was much weaker compared to the expression in non-blocking embryos (Figure 1B') or in blastocysts (Figure 1D'). This shows that alterations in expression and distribution of E-cadherin and other alterations in the cytoskeleton responsible for rounding of blastomeres and loss of cell–cell contact are associated with the blocking state.
|
E-cadherin expression in aggregation chimeras
When four of the non-blocking CBAxC57BL/6 F2 embryos after explantation at 40–42 h post HCG were aggregated by PHA to one blocking embryo of the MF1 strain explanted at 34–36 h post HCG, the embryo of the blocking MF1 strain was rescued and resumed cell cycle progression and development in vitro within the aggregation chimera, as described for other strains of mice (Sekirina and Neganova, 1995
; Neganova et al., 1998). E-cadherin staining remained bright at contact sites between sister blastomeres in all embryos of the aggregation chimera (Figure 2A, A'). In addition, expression of E-cadherin appeared also at contact zones between blastomeres of adjacent, genotypically identical non-blocking embryos about 4 h after aggregation (Figure 2A, A'). A similar increase in expression of surface E-cadherin at contact zones between adjacent embryos occurred at 8–10 h after aggregation between embryos of different genotype (Figure 2C, C'), but this was relatively delayed compared to its appearance between adjacent, genotypically identical non-blocking embryos. Upon culture for 20 h after aggregation an integrated morula formed in which all contact zones between blastomeres were strongly stained for E-cadherin while the apical outer surface of the membrane of the blastomeres exhibited no accumulation of E-cadherin (Figure 2B, B'), similar to the situation in blastocysts of the non-blocking strain (Figure 1D'
). In contrast to this, aggregation of 2-cell embryos of the blocking strain alone without others of the non-blocking strain did not lead to a rescue, there was no E-cadherin expression at contact sites between different embryos and development remained arrested at the 2-cell stage in spite of the close cell–cell apposition mediated by PHA. Therefore, polarized expression of E-cadherin at blastomere–blastomere contact sites between embryos of different genotype is one of the earliest indicators of normal development and rescue by aggregation.
|
Influence of E-cadherin antibody on rescue by aggregation
When aggregates between blocking and non-blocking embryos were produced in the presence of low concentrations of a monoclonal rat antibody directed against an intracellular protein, tyrosinated
-tubulin (YL1/2), the embryos of the blocking strain were rescued as in the controls. Therefore, the presence of rat antibodies in the culture medium did not generally appear to affect cell–cell interactions between embryos mediating rescue. There was intense staining of E-cadherin in the area of cell–cell contacts between adjacent and different embryos under these conditions, similar to what was characteristic for controls cultured without any antibody during aggregation (e.g. Figure 2A'–C').
In contrast, there was no rescue of the blocking embryos when a rat-derived E-cadherin-specific antibody was present during aggregation and embryo culture. The 2-cell embryos of the MF1 strain failed to divide and to resume development despite aggregation with CBAxC57BL/6 F2 embryos by PHA. Concomitantly, there was no enrichment of E-cadherin at contact sites between different embryos (Figure 2D, D'), although individual embryos of the non-blocking genotype still expressed E-cadherin between their own sister blastomeres.
Culture in vitro in the presence of low concentrations of the E-cadherin-specific antibody did not visibly affect development of individual embryos of the non-blocking strain early after explantation. They were able to undergo division and continued to develop during a 12–16 h period of culture after explantation comparable to embryos kept in medium without E-cadherin antibody. Embryos of the blocking strain became developmentally arrested and rounded when explanted and cultured in the presence of E-cadherin antibody, as found for embryos cultured without the antibody. Similarly, the rat-derived monoclonal antibody directed against the intracellular target, tubulin, did not affect development of individual embryos during 12–16 h of culture after explantation. Taken together, this suggests that E-cadherin antibody did not affect early preimplantation development after explantation. However, E-cadherin antibody prevented E-cadherin expression at embryo–embryo contacts and E-cadherin-mediated signalling between different embryos in aggregates. The rescue of blocking by aggregation was prevented, although embryos of different genotype remained physically attached to each other by PHA.
Distribution of mitochondria in blocking, non-blocking and rescued embryos
Similar to the human embryo (Van Blerkom et al., 1998), mitochondria were characteristically distributed in small clusters homogeneously throughout the cytoplasm of in-vivo developing embryos of mice of both blocking and non-blocking genotype at interphase. The number of mitochondria was greatest close to the nucleus but many small clusters were found in the cytoplasm between cell surface and nucleus (Figure 3A). An accumulation of mitochondria (`mitochondrial ring', Sekirina et al., 1998) was found in the perinuclear cytoplasm just before embryos progress into mitosis (at 48–50 h post HCG injection), when cytoplasmic microtubules had depolymerized and a spindle was formed upon transition from the 2- to the 4-cell stage. A dense ring of mitochondria was found at M-phase at the periphery of the spindle in both dividing blastomeres of such non-blocking embryos (Figure 3D, D'). In contrast, 2-cell embryos of the MF1 strain arrested for extended periods at the G2-phase possessed compact clusters of mitochondria also at the nuclear periphery in interphase when cultured from 32–34 h post HCG to 64 h post HCG. These blocked embryos occasionally had small clusters of mitochondria at the cell periphery of blastomeres (Figure 3B) but, unlike unblocked interphase stages, there were hardly any mitochondria found in the rest of the cytoplasm between nucleus and cell periphery in the blocked embryos (Figure 3B). The overall brightness of the mitochondrial staining with Mitotrack TM Red appeared reduced and, accordingly, mitochondrial redox state and activity was presumably lower in blocked 2-cell embryos as compared to the non-blocking genotype.
|
When the 2-cell embryos of the blocking MF1 strain were aggregated with those of the non-blocking genotype, none of the characteristic alterations in mitochondrial distribution were observed and dense accumulation of mitochondria at the nuclear periphery did not occur (Figure 3C
). All embryos of both genotypes remained brightly stained in aggregation chimeras and mitochondria retained their more or less homogeneous distribution throughout the cytoplasm of interphase blastomeres (Figure 3C, C'). Thus, aggregation appears to prevent perinuclear clustering at the 2-cell stage and to support the dispersal of small clusters of mitochondria throughout the cytoplasm of blastomeres after division has occurred in rescued embryos of the blocking MF1 strain.
Microtubules in blocking and non-blocking embryos and in aggregation chimeras
Typically, 2-cell embryos of the blocking and non-blocking strain were brightly stained by anti-tubulin antibody and possessed free tubulin and a dense, finely and homogeneously distributed microtubular network after explantation at 32–34 h post HCG (Figure 4A, A'). In these embryos the mid-body with bundles of microtubular fibres was visible at the zone of contact between sister blastomeres, where the second polar body was located (Figure 4A, A'), defining the anterior–posterior axis (Edwards and Beard, 1997; Gardner, 1997). Upon progression to the second M-phase, 14–16 h past explantation, spindles formed in both blastomeres of the non-blocking CBAxC57BL/6 F2 embryos during development in vitro (Figure 4B, B'). The mid-body from the previous cleavage and the second polar body were usually retained in non-blocking embryos.
|
In contrast, microtubules were less densely packed in the embryos of the MF1 strain blocked at the G2-stage. Tubulin staining was generally reduced in comparison to normally developing 2-cell stages, and especially in comparison to divided blastomeres of normally developing embryos. Relatively few long, individual fibres extended throughout the cytoplasm of the blastomeres when embryos were in deep block (Figure 4C'
inset). No dense ring of microtubules formed perinuclearly, as was characteristic for 4-cell embryos of the non-blocking genotype proceeding to develop in vitro during culture. The mid-body had often disappeared in blocked embryos (Figure 4C–C' inset). Usually the integrity of the polar body was also lost at this time and it was distinguished only in a few of the developmentally arrested embryos. The chromatin in the nuclei of blastomeres was usually condensed (Figure 4C'), no longer forming strings of chromomeres as was typical for non-blocking controls or blastomeres in aggregation chimeras (for comparison see Figure 2D).
However, when embryos of the blocking genotype were aggregated with those of the non-blocking genotype and induced to divide, the alterations in the cytoskeleton characteristic of the blocked state were not observed. The microtubules of all blastomeres in the aggregate retained their dense and homogeneously dispersed pattern of staining with dense arrays of microtubules close to the cell periphery as well as throughout the cytoplasm (Figure 5B,B'). In particular, when focusing on the cell centre it was obvious that a very prominent perinuclear band of microtubules existed in the divided sister blastomeres of all embryos in aggregation chimeras of the rescued, blocking and the non-blocking genotype (Figure 5A,A'), as is characteristic for normally developing embryos of the mouse at the 4–8-cell stage (Houliston et al., 1987; Maro et al., 1991). Concomitantly, the integrity of the second polar body was preserved for extended periods of development in embryos of aggregates. Mid-bodies from the previous divisions persisted (Figure 5A,A',B,B').
|
In conclusion, the study provides evidence that rescue by aggregation is associated with surface expression of E-cadherin at blastomere–blastomere contacts and increased E-cadherin staining at embryo–embryo contact sites, a homogeneous distribution of mitochondria in all interphase blastomeres before and after division and the formation of a dense, finely spaced microtubular network throughout the cytoplasm and a perinuclear microtubule ring in blastomeres of developing embryos in aggregates derived by division from the non-blocking as well as the blocking genotype.
Discussion
Cytoskeletal organization, polarization of the 2-cell embryo and surface expression of E-cadherin
Our study confirms earlier observations showing that surface-expressed E-cadherin is enriched at contact zones between blastomeres of non-blocking mammalian embryos (Vestweber et al., 1987; Fleming et al., 1994; Campbell et al., 1995). Non-random distribution of membrane components has already been recognized (Handyside et al., 1987) in mammalian embryos. Recent observations have unambigiously demonstrated that the oocyte and the 2-cell embryo of the mouse and human are already bilaterally polarized in oocytes and at the 2-cell stage (Edwards and Beard, 1997; Gardner, 1997; Antczak and Blerkom, 1998, 1999), and this appears to be important for allocation of cell fate in a mammalian embryo with high quality and normal developmental potential (Edwards and Beard, 1997; Antczak and Van Blerkom, 1999). E-cadherin and rho-like GTPases were shown to be essential for polarization of blastomeres at compaction later in development (Clayton et al., 1999). Enrichment of E-cadherin/uvomorulin at contacts between sister blastomeres also contributes to a polarized organization in the 2-cell embryo by defining a microvillus-free area of the membrane parallel to the anterior–posterior axis marked by the localization of the second polar body and the mid-body from the first mitotic division (Edwards and Beard, 1997; Gardner, 1997).
E-cadherin is one of the gene products provided initially from maternal sources for early embryonic development (Rietmacher et al., 1995). We confirmed recently in transgene-expressing mouse strains that blocked embryos arrest development at the G2-stage before full zygotic gene activation has been initiated (I.E.Neganova, unpublished results). E-cadherin is therefore presumably of maternal origin in the blocked state. E-cadherin staining was diminished in embryos arrested in development for extended periods of culture in vitro. Reductions in the concentration of membrane-associated E-cadherin may therefore have contributed to a disturbance in overall cytoskeletal organization and gene expression in these embryos (Barth et al., 1997). E-cadherin/uvomorulin-mediated cell-adhesion and signalling is mediated intracellularly by catenin, actin and other cytoskeletal components (e.g. Huber et al., 1996; Barth et al., 1997; Adams and Nelson, 1998; Adams et al., 1998). Functional ablation of E-catenin induces a similar phenotype in preimplantation embryos as seen in E-cadherin knockouts (Torres et al., 1997). Catenins have essential functions in regulation of expression of conserved genes necessary for polarization and differentiation of the vertebrate embryo later in development (for discussion see Edwards and Beard, 1997). Alterations in the concentration of surface-expressed E-cadherin and, in consequence, in free ß-catenin in blocked embryos may therefore indirectly or directly affect expression patterns at an earlier stage of embryogenesis and contribute in this way to the developmental arrest in the 2-cell stage.
Embryos of the blocking strain still possess sufficient protein to mediate tight cell adhesion with non-blocking ones when aggregation is performed directly after explantation of the blocking embryo. However, aggregation after prolonged culture of the blocking embryo does not lead to efficient rescue (data not shown). This may rely on the reduction of surface E-cadherin in blocked embryos, or may be a direct consequence of the general loss of cytoplasmic organization and cell polarity as visible by the gradual degeneration of the polar body and the disappearance of the stem body between the two sister blastomeres.
Loss of intimate cell–cell contacts between sister blastomeres in the blocked mouse embryo is accompanied by a gradual rounding up of blastomeres and other characteristic alterations in the cytoskeleton of the embryo in deep block. EGF receptor directly regulates cell–cell adhesion in some cell types by modulating the interaction of E-cadherin with the actin cytoskeleton (Hazan and Norton, 1998) and it may be this activity of EGF which rescues some but not all embryos from the block in the absence of aggregation (Terrada et al., 1997). We did not study the architecture of the actin cytoskeleton in normal and blocked embryos but can show that diminished expression of surface E-cadherin at cell-cell borders is accompanied by characteristic alterations in the microtubular cytoskeleton in embryos in deep block. An active adhesive function of E-cadherin appears dependent on an intact cytoskeleton in the mouse preimplantation embryo. Junctional communication was reduced in 8-cell embryos subjected to taxol, a microtubule-stabilizing drug (Maro and Pickering, 1984; Goodall and Maro, 1986) whereas nocodazole, a microtubule-depolymerizing drug, accelerated flattening of cells (Ducibella, 1982), suggesting a constraining effect of overstabilized microtubules on adhesion and junctional communication. The long, distinct microtubule fibres in blocked embryos are reminiscent of typical stable interphase microtubules, and by being more rigid than those of non-blocking embryos may contribute to cell rounding in blocked embryos.
The distribution of mitochondria is also dramatically altered in embryos in deep block. Mitochondria appear to be important for defining distinct subcellular compartments for calcium in mammalian embryos (Sousa et al., 1997). Clustering of mitochondria perinuclearly can therefore disturb calcium signalling as well as the local supply of ATP. Accordingly, it has previously been shown that blocked embryos suffer from oxidative stress response (Nasr-Esfahani et al., 1990; Matsumoto et al., 1998). Disturbances in mitochondrial distribution appear involved in cell cycle delays or arrest in response to chemical exposures in maturing oocytes (Yin et al., 1998a). Clustering may also contribute significantly to the reduced developmental capacity of blocking embryos and their eventual degeneration upon explantation and culture in vitro for prolonged times.
Role of E-cadherin in rescue by aggregation
Membranes at contact zones between blastomeres of embryos from different strains and those of the same strain exhibit enhanced expression of surface E-cadherin in aggregation chimeras. The initial expression of E-cadherin between blastomeres of individual embryos is retained, and sites of contact between different embryos also become enriched for E-cadherin, first between embryos of non-blocking and then between blocking and non-blocking genotype. In this way the portion of the surface which is E-cadherin-rich and microvillus-free becomes increased while the E-cadherin-poor, microvilli-carrying membrane area becomes reduced, in particular in comparison to embryos in deep block. Cadherin-mediated adhesion initiates reorganization of the cytoskeleton at compaction (reviewed by Fleming et al., 1994; Huber et al., 1996; Adams et al., 1998) and in somatic cells, and E-cadherin-mediated signalling participates crucially in the formation of junctional complexes (Adams et al., 1998). We have previously shown by electron microscopy that the regions of contact between different embryos in aggregates develop adherens-like junctional complexes and electron-dense plaques characteristic of adherens junctions (Sekirina and Neganova, 1995). The molecular components, especially the membrane proteins involved in the establishment of these specialized cell-cell connections, were unclear. The present studies suggest that E-cadherin is involved in the morphogenetic processes following aggregation. Unlike in aggregates, typical adherens junctions become expressed in unblocked control embryos only at contact zones between blastomeres at compaction, at a time when cell polarization takes place (Fleming et al., 1994). Precocious compaction/cohesion can be induced in 4-cell embryos by activation of phospholipid-dependent protein kinase (Winkel et al., 1990), and gap junction formation is initiated by the serine–threonine kinase inhibitor 6-diaminopurine (Aghion et al., 1994). One can speculate that enhanced E-cadherin expression in aggregates may influence signalling pathways affecting protein phosphorylation patterns and thereby rescue embryos from the block in development.
Autocrine factors produced by embryos themselves (e.g. platelet-activating factor; Stoddart et al., 1996; O'Neill, 1998) appear to support significantly the developmental capacity of mammalian embryos. Growth factor-supplemented media (e.g. Larson et al., 1992; Fukaya et al., 1998) and co-culture of human embryos with somatic cells (Desai et al., 1998; Wiemer et al., 1998) also improve embryo survival and have a similar beneficial influence on developmental capacity. We have previously exposed blocked 2-cell embryos to conditioned culture media from non-blocking embryos to test whether the rescue by aggregation is also initiated by secreted mitogenic/growth factors or other molecules from the non-blocking strain (Sekirina and Neganova, 1995). Since embryos remained in block, it was concluded that the rescue by aggregation is critically dependent on an active and direct cell–cell interaction. Our present observations support this notion. Rescue does not take place when embryos are aggregated by PHA in the presence of specific E-cadherin antibody. This shows that a physical contact and proximity to each other or diffusable factors transmitted by the non-blocking strain to the physically attached blocking embryo is not sufficient to obtain a rescue effect. From the present and previous findings it appears that enhanced surface expression of E-cadherin together with E-cadherin-mediated attachment and signalling and alterations in the membrane-associated as well as the cytoplasmic cytoskeleton of the embryo are essential steps in rescuing blocked embryos from the 2-cell stage.
E-cadherin expression at the cell membrane of transfected mouse mammary carcinoma cells influences the phosphorylation and/or expression of intracellular, cell cycle-regulating proteins (e.g. retinoblastoma protein and cyclin-dependent kinase inhibitors respectively; St Croix et al., 1998). Further studies have to reveal whether such proteins are also post-translationally modified or differentially expressed in rescued 2-cell embryos of the mouse. From the present observations it is clear that cell cycle progression occurs, the typical cellular organization is maintained, and gene expression and developmental capacity are restored by aggregation.
Significance of mitochondrial and microtubule distribution for the blocked state and for normal preimplantation development
Aggregation prevents the typical alterations in the microtubular cytoskeleton seen in the blocked state. Although it is known that E-cadherin-mediated signalling primarily influences the constitution of the actin cytoskeleton, signalling events via activation of cytoskeleton-associated kinases may also influence indirectly, and downstream from this, the distribution and polymerization kinetics of microtubules in blastomeres. Altered dynamics of microtubule turnover influence compaction and cell polarization in the normally differentiating and developing preimplantation embryo (Maro and Pickering, 1984; Maro et al., 1991). We show here that there is a striking correlation between mitochondrial dispersal and the state of the microtubular cytoskeleton in normally developing embryos and in those in 2-cell block. In the presence of a dense, rather homogeneously distributed microtubular network and a perinuclear ring of microtubules, mitochondria of interphase blastomeres exhibit a dispersed and active state. In contrast, mitochondria become clustered perinuclearly in the blocked state and at the periphery of the spindle in oocytes and embryos in which cytoplasmic microtubules apart from spindle tubules become depolymerized at the transition to M-phase (Van Blerkom and Runner, 1984; Yin et al., 1998a). Microtubules in blocking 2-cell stages are less dense as compared to developing embryos and do not form a perinuclear network. From these observations it is concluded that the constitution of the microtubular cytoskeleton is intimately correlated with cell shape and organelle distribution.
Spindle formation was shown to coincide with perinuclear accumulation of mitochondria in maturing mouse oocytes (Van Blerkom, 1991). Our observations suggest that a functionally intact microtubular cytoskeleton is also necessary for normal mitochondrial distribution in interphase blastomeres of the mammalian preimplantation embryo prior to compaction and trophectoderm formation. Exposure of 2-cell mouse embryos of the non-blocking strain to low concentrations of nocodazole causes depolymerization of microtubules in blastomeres, and results also in perinuclear clustering of mitochondria and a block in development at the G2-phase similar to what is observed in embryos of the blocking strain (I.E.Neganova, unpublished results). In embryos in which mitochondrial dispersal is disturbed energy substrates for critical protein phosphorylation events in cell cycle progression, for motor activities and for transcriptional regulation may not be available. Mitochondrial mutation (Keefe et al., 1995), reduced mitochondrial activity in response to anoxic conditions (Gaulden, 1995; Van Blerkom et al., 1995, 1997) or failure of mitochondria to be correctly distributed in response to environmental exposures (Yin et al., 1998a) have been associated with reduced developmental capacity of mammalian embryos of aged females or oocytes exposed to drugs, irrespective of blocking or non-blocking genotype. Disturbances in the integrity of the microtubular cytoskeleton are responsible for increases in non-disjunction during meiotic and mitotic divisions in human oogenesis and embryos of women approaching the end of their reproductive span (Eichenlaub-Ritter et al., 1988; Battaglia et al., 1996). Our present observations suggest that altered microtubule distribution and/or dynamics may also contribute to the reduced developmental capacity when microtubule and mitochondrial distribution is affected in interphase blastomeres of embryos after fertilization.
Future studies on the expression and distribution of microtubules, mitochondria and surface-expressed E-cadherin in normal and developmentally delayed or arrested human embryos may elucidate further correlations between morphogenetic processes, integrity of the cytoskeleton, gene expression and developmental potential and gain insights into the aetiology of infertility in some patients, identification of critical survival factors in embryo culture and signalling processes relevant for normal development.
Acknowledgments
Unfortunately, since contributing to this article Dr G.G. Sekirina has died. We thank Ilse Betzendahl for expert technical assistance, Matthias Peschke for help with the photographic work, and gratefully acknowledge the support by the Department of Developmental Biology of the University of Bielefeld, especially that by Dr Peter Heimann. The work has been supported by the DFG (436 RUS 113/228) and the EU (ENV4-CT97-0471). I.E.N. was a visiting guest in a joint DFG-funded project (436 RUS 113/228).
Notes
3 To whom correspondence should be addressed 
References
Adams, C.L. and Nelson,W.J. (1998) Cytomechanisms of cadherin-mediated cell–cell adhesion. Curr. Opin. Cell Biol., 10, 572–577.[ISI][Medline]
Adams, C.L., Chen, Y.T., Smith, S.J. et al. (1998) Mechanisms of epithelial cell–cell adhesion and cell compaction revealed by high-resolution tracking of E-cadherin-green fluorescent protein. J. Cell Biol., 24, 1105–1119.
Aghion, J., Gueth-Hallonet, C., Antony, C. et al. (1994) Cell adhesion and gap junction formation in early mouse embryo are induced prematurely by 6-DMAP in the absence of E-cadherin phosphorylation. J. Cell Sci., 107, 1369–1379.[Abstract]
Antczak, M. and Van Blerkom, J. (1997) Oocyte influences on early development; the regulatory proteins leptin and STAT3 are polarized in mouse and human oocytes and differently distributed within the cells of the preimplantation stage embryo. Mol. Hum. Reprod., 3, 1067–1086.
Antczak, M. and Van Blerkom, J. (1999) Temporal and spatial aspects of fragmentation in early human embryos: possible effects on developmental competence and association with the differential elimination of regulatory proteins from polarized domains. Hum. Reprod., 14, 429–427.
Aoki, F.T. Choi, T., Mori, M. et al. (1992) A deficiency in the mechanism for p34cdc2 protein kinase activation in mouse embryos arrested at 2-cell stage. Dev. Biol., 154, 66–72.[ISI][Medline]
Barnett, D. and Bavister, B. (1996) What is the relationship between the metabolism of pre-implantation embryos and their developmental competence? Mol. Reprod. Dev., 43, 105–133.[ISI][Medline]
Barnett, D., Clayton, M., Kimura, J. et al. (1997) Glucose and phosphate toxicity in hamster pre-implantation embryos involves disruption of cellular organization, including distribution of active mitochondria. Mol. Reprod. Dev., 48, 227–237.[ISI][Medline]
Barth, A.I., Nathke, I.S. and Nelson, W.J. (1997) Cadherins, catenins and APC protein: interplay between cytoskeletal complexes and signalling pathways. Curr. Opin. Cell Biol., 9, 683–690.[ISI][Medline]
Battagalia, D., Goodwin, P., Klein, N. et al. (1996) Influence of maternal age on meiotioc spindle assembly in oocytes from naturally cycling women. Hum. Reprod., 11, 2217–2222.
Biggers, J.D. (1971) New observations on the nutrition of the mammalian oocyte and the preimplantation embryo. In Blandeau, R.J. (ed.), The Biology of Blastocyst. University of Chicago Press, Chicago, pp. 319–327.
Braude, P., Balton, V. and Moore, S. (1988) Human gene expression first occurs between the four- and eight-cell stages of preimplantation development. Nature, 332, 459–461.[Medline]
Brinster, R.L. (1965) Studies on the development of mouse embryo in vitro. IV. Interaction of energy sources. J. Reprod. Fertil., 10, 227–240.[Medline]
Brown, J.J. and Whittingham, D.G. (1992) The dynamic provision of different energy substrates improved development of one-cell random-bred mouse embryos in vitro. J. Reprod. Fertil., 95, 503–511.[Abstract]
Campbell, S., Swann, H.R., Seif, M.W. et al. (1995) Cell adhesion molecules on the oocyte and preimplantation human embryo. Hum. Reprod., 10, 1571–1578.
Camous, S., Heyman, J., Menzion, W. et al. (1984) Cleavage beyond the block stage and survival after transfer of early bovine embryos cultured with trophoblastic vesicles. J. Reprod. Fertil., 72, 479–485.[Abstract]
Clayton, L., Stinchcombe, S.V. and Johnson, M.H. (1993) Cell surface localisation and stability of E-cadherin during early mouse development. Zygote, 1, 333–344.[Medline]
Clayton, L., Hall, A. and Johnson, M.H. (1999) A Role for Rho-like GTPases in the polarisation of mouse eight-cell blastomeres. Dev. Biol., 205, 322–331.[ISI][Medline]
Cohen, J., Scott, R., Schimmel, T. et al. (1997) Birth of infant after transfer of anucleate donor oocyte cytoplasm into recipient eggs. Lancet, 350, 961–962.[Medline]
Cross, P.C. and Brinster, R.L. (1973) The sensitivity of one cell mouse embryos to pyruvate and lactate. Exp. Cell Res., 77, 57–62.[ISI][Medline]
Dawson, K.M. and Baltz, J.M. (1997) Organic osmolytes and embryos: substrates of Gly and beta transport systems protect mouse zygotes against the effects of raised osmolarity. Biol. Reprod., 56, 1550–1558.[Abstract]
Desai, N.N. (1998) The road to blastocyst transfer. Hum. Reprod., 13, 3292–3294.[ISI][Medline]
Devreker, F. and Hardy, K. (1997) Effects of glutamine and taurine on preimplantation development and cleavage of mouse embryos in vitro. Biol. Reprod., 57, 921–928.[Abstract]
Dienhart, M.K., O'Brien, M.J. and Downs, S.M. (1997) Uptake and salvage of hypoxanthine mediates developmental arrest in preimplantation mouse embryos. Biol. Reprod., 56, 1–13.[Abstract]
Ducibella, T. (1982) Depolymerization of microtubules prior to compaction. Development of cell polarity and cell spreading are not inhibited. Exp. Cell Res., 138, 31–38.[ISI][Medline]
Dumoulin, J.C., Michiels, A.H., Bras, M. et al. (1993) Temporal effects of ouabain on in vitro development of mouse zygotes. Hum. Reprod., 8, 1469–1474.
Edwards, R.G. and Beard, H. (1997) Oocyte polarity and cell determination in early mammalian embryos. Mol. Hum. Reprod., 3, 863–905.
Eichenlaub-Ritter, U. and Betzendahl, I. (1995) Chloral hydrate induced spindle aberrations, metaphase I arrest, and aneuploidy in mouse oocytes. Mutagenesis, 10, 477–486.
Eichenlaub-Ritter, U. and Boll, I. (1989) Nocodazole sensitivity, age-related aneuploidy, and alterations in the cell cycle during maturation of mouse oocytes. Cytogenet. Cell Genet., 52, 170–176.[ISI][Medline]
Eichenlaub-Ritter, U., Stahl, A. and Luciani, J.M. (1988) The microtubular cytoskeleton and chromosomes of unfertilized human oocytes aged in vitro. Hum. Genet., 80, 259–264.[ISI][Medline]
Fleming, T.P., Garrod, D.R. and Elmsmore, A.J. (1991) Desmosome biogenesis in the mouse preimplantation embryo. Development, 112, 527–539.[Abstract]
Fleming, T.P., Butler, L., Lei, X. et al. (1994) Molecular maturation of cell adhesion systems during mouse early development. Histochemistry, 101, 1–7.[ISI][Medline]
Fukaya, T., Yamanaka, T., Terada, Y. et al. (1998) Growth hormone improves mouse embryo development in vitro, and the effect is neutralized by growth hormone receptor antibody. Tohoku J. Exp. Med., 184, 113–122.[ISI][Medline]
Gardner, R.L. (1997) The early blastocyst is bilaterally symmetrical and its axis of symmetry is aligned with the animal–vegetal axis of the zygote in the mouse. Development, 124, 289–301.[Abstract]
Gardner, D.K. and Lane, M. (1996) Alleviation of the `2-cell block' and development to the blastocyst of CF1 mouse embryos: role of amino acids, EDTA and physical parameters. Hum Reprod., 11, 2703–2712.
Gaulden, M. (1992) The enigma of Down syndrome and other trisomic conditions. Mutat. Res., 269, 69–88.
Goddard, M.J. and Pratt, H.P. (1983) Control of events during early cleavage of the mouse embryo: an analysis of the `2-cell block'. J. Embryol. Exp. Morphol., 73, 111–133.[ISI][Medline]
Goodall, H. and Maro, B. (1986) Major loss of junctional coupling during mitosis in early mouse embryos. J. Cell Biol., 102, 568–575.
Handyside, A.H., Edidin, M. and Wolf, D.E. (1987) Polarized distribution of membrane components on two-cell mouse embryos. Roux's Arch. Dev. Biol., 196, 273–278.
Haraguchi, S., Naito, K., Azuma, S. et al. (1996) Effects of phosphate on in vitro 2-cell block of AKR/N mouse embryos based on changes in cdc2 kinase activity and phosphorylation states. Biol. Reprod., 55, 598–603.[Abstract]
Hazan, R.B. and Norton, L. (1998) The epidermal growth factor receptor modulates the interaction of E-cadherin with the actin cytoskeleton. J. Biol. Chem., 273, 9078–9084.
Houliston, E. and Maro, B. (1989) Posttranslational modification of distinct microtubule subpopulations during cell polarization and differentiation in the mouse preimplantation embryo. J. Cell Biol., 108, 543–551.
Huber, O., Bierkamp, C. and Kemler, R. (1996) Cadherins and catenins in development. Curr. Opin. Cell Biol., 8, 685–691.[ISI][Medline]
Keefe, D., Niven-Fairchild, T., Powell, S. et al. (1995) Mitochondrial desoxyribonucleic acid deletions in oocytes and reproductive aging women. Fertil. Steril., 64, 577–583.[ISI][Medline]
Kilmartin, J.V., Wright, B. and Milstein, C. (1982) Rat monoclonal antitubulin antibodies derived by using a new nonsecreting rat cell line. J. Cell Biol., 93, 576–582.
Lane, M. and Gardner, D.K. (1992) Effect of incubation volume and embryo density on the development and viability of mouse embryos in vitro. Hum. Reprod., 7, 558–562.
Larson, R.C., Ignotz, G.G. and Currie, W.B. (1992) Transforming growth factor beta and basic fibroblast growth factor synergistically promote early bovine embryo development during the forth cell cycle. Mol. Reprod. Dev., 33, 432–435.[ISI][Medline]
Maro, B. and Pickering, S.J. (1984) Microtubules influence compaction in preimplantation mouse embryos. J. Embryol. Exp. Morphol., 84, 217–232.[ISI][Medline]
Maro, B., Gueth-Hallonet, C., Aghion, I. et al. (1991) Cell polarity and microtubule organization during mouse early embryogenesis. Development, 1 (Suppl.), 17–25.
Maro, B., Johnson, M.H., Pickering, S.F. et al. (1984) Changes in the distribution of membranous organelles during mouse early embryogenesis. J. Embryol. Exp. Morphol., 90, 287–309.
Matsumoto, H., Shoji, N., Sugawara, S. et al. (1998) Microscopic analysis of enzyme activity, mitochondrial distribution and hydrogen peroxide in 2-cell rat embryos. J. Reprod. Fertil., 113, 231–238.[Abstract]
Mintz, B., Gearhart, J.D. and Guymont, A.O. (1973) Phytohaemagglutinin-mediated blastomere aggregation and development of allophenic mice. Dev. Biol., 31, 195–199.[ISI][Medline]
Muggleton-Harris, A. and Brown, J.J. (1988) Cytoplasmic factors influence mitochondrial reorganization and resumption of cleavage during culture of early mouse embryos. Hum. Reprod., 3, 1020–1028.
Muggleton-Harris, A., Whittingham, D.G. and Wilson, L. (1982) Cytoplasmic control of preimplantation development in vitro in the mouse. Nature, 299, 460–462.[Medline]
Nasr-Esfahani, M., Johnson, M.H. and Aitkin, J.R. (1990) The effect of iron and iron chelators on the in vitro block to development of the mouse preimplantation embryo: BAT6 as new medium for improved culture of mouse embryos in vitro. Hum. Reprod., 5, 997–1003.
Neganova, I.E. and Sekirina, G.G. (1995) Block and viability of BALB/c mouse embryos after explantation during the second cell cycle of cleavage division. Ontogenez, 27, 1–12.
Neganova, I.E., Augustin, M., Sekirina, G. et al. (1998) LacZ transgene expression as a cell marker to analyse rescue from the 2-cell block in mouse aggregation chimeras. Zygote, 6, 223–226.[ISI][Medline]
Ohsugi, M., Larue, L., Schwartz, H. et al. (1997) Cell-junctional and cytoskeletal organization in mouse blastocysts lacking E-cadherin. Dev. Biol., 15, 261–271.
O'Neill, C. (1998) Autocrine mediators are required to act on the embryo by the 2-cell stage to promote normal development and survival of mouse preimplantation embryos in vitro. Biol. Reprod., 58, 1303–1309.
Rietmacher, D., Brinkmann, V. and Birchmeier, C. (1995) A targeted mutation in the mouse E-cadherin gene results in defective preimplantation development. Proc. Natl. Acad. Sci. USA, 92, 855–859.
Sefton, M., Johnson, M.H. and Clayton, L. (1992) Synthesis and phosphorylation of E-cadherin during mouse early development. Development, 115, 313–318.[Abstract]
Sekirina, G. and Neganova, I.E. (1995) The microenvironment created by non-blocking embryos in aggregates may rescue blocking embryos via cell–embryo adherent contacts. Zygote, 3, 313–324.[ISI][Medline]
Sekirina, G., Boguliobova, N.A., Antonova, N.V. et al. (1997) The behaviour of mitochondria and cell integration during somatic hybridization of sister blastomeres of a 2-cell mouse embryo. Zygote, 5, 97–103.[ISI][Medline]
Semb, H. and Christofori, G. (1998) Insights from model sytems. The tumor-suppressor function of E-cadherin. Am. J. Hum. Genet., 63, 1588–1593.[ISI][Medline]
Sousa, M., Barros, A., Silva, J. et al. (1997) Developmental changes in calcium content of ultrastructurally distinct subcellular compartments of pre-implantation human embryos. Mol. Hum. Reprod., 3, 83–90.
St Croix, B., Sheehan, C., Rak, J.W. et al. (1998) E-Cadherin-dependent growth suppression is mediated by the cyclin-dependent kinase inhibitor p27(KIP1). J. Cell Biol., 27, 557–571.
Stoddart, N.R., Wild, A.E. and Fleming, T.P. (1996) Stimulation of development in vitro by platelet-activating factor recepto