Changes in the distribution of membranous organelles during mouse early development

Many differentiated cells manifest a highly asymmetric organization that is dependent partly upon continuing cell contact, within for example an epithelial layer, and is partly intrinsic to the structure of the cell (e.g. Ziomek, Schulman & Edidin, 1980). Various approaches to the study of how this cell asymmetry is developed and maintained are available including the use of cell lines, such as the MDCK cell line, which can modulate some epithelial properties reversibly in vitro (Van Meer & Simons, 1982), and the examination of polarized, epithelial cells after their isolation and manipulation (Ziomek et al. 1980). However in these model systems polarity is not generated de novo from a truly symmetric precursor cell. De novo polar organization of cells develops first early in embryogenesis with the formation of the primary epithelial germ layers and the delamination of extraembryonic epithelia. The earliest evidence of this process in the mouse embryo has been detected at the 8-cell stage, during which elements of the cell surface (Handyside, 1980), cytoskeleton (Johnson & Maro, 1984), and endocytotic processing pathway (Reeve, 1981; Fleming & Pickering, 1985) undergo a radical reorganization to convert a non-polar cell to a highly polarized cell over a period of 8–10 h (Ziomek & Johnson, 1980). Elements of this polarity are conserved at division (Johnson & Ziomek, 1981), and the polarity is elaborated and stabilized at the 16- and 32-cell stages to generate the definitive trophectodermal epithelium (Fleming, Warren, Chisholm & Johnson, 1984; Fleming & Pickering, 1985). In this paper, we report on the immunocytochemical localization of various antigens specific to membranous organelles (endoplasmic reticulum, the acid vesicle/ lysosome compartment, coated vesicles, the Golgi apparatus) concerned with endocytotic and biosynthetic activity and on the changes that occur during the early stages of cell polarization.

MFI female mice (3–5 weeks; Olac) were superovulated by injections of 5i.u. of pregnant mare’s serum gonadotrophin (PMSG; Intervet) and human chorionic gonadotrophin (hCG; Intervet) 48 h apart. The females were paired overnight with HC-CFLP males (Hacking & Churchill) and inspected for vaginal plugs the next day. Unfertilized and fertilized eggs were recovered from females at 14–16 h post hCG; 2-cell and 4-cell embryos were recovered at 46–50h post hCG; 8-cell embryos were derived by overnight culture of 2- to 4-cell embryos; early 16-cell embryos were recovered at 65–70 h post hCG.

2-cell embryos were recovered at 48 h post hCG and cultured in Medium 16 containing 4 mg ml⁻¹ BSA (M16+BSA) (Whittingham & Wales, 1969) under oil for 13 h at 37°C in 5 % CO₂ in air. All 4-cell embryos were then exposed briefly to acid Tyrode’s solution (Nicolson, Yanagimachi & Yanagimachi, 1975) to remove the zona pellucida, rinsed in Medium 2+BSA (Fulton & Whittingham, 1978), and placed in Ca²⁺-free M2+6mgml⁻¹ BSA for 5–45 min, during which time they were disaggregated to single 4-cell blastomeres (1/4 cells) using a flame-polished micropipette. Cells were cultured on Sterilin tissue culture dishes in drops of M16+BSA under oil at 37°C in 5 % CO₂ in air. The cultures were inspected hourly for evidence of division to 2/8 pairs, and couplets were removed, designated Oh old, and cultured in M16+BSA as natural 2/8 pairs.

Late 8-cell embryos were recovered at 64 h post-hCG and were disaggregated to single 8-cell blastomeres (1/8 cell) as described above. Couplets of 16-cell blastomeres (2/16 natural pairs) were selected as above. In some experiments, 2/16 pairs were cultured in the presence of a monoclonal antibody to cadherin (Yoshida-Noro, Suzuki & Takeichi, 1984; also called uvomorulin, L-CAM) in order to avoid the envelopment of the apolar cell by the polar cell that would otherwise occur (Ziomek & Johnson, 1981).

In one series of experiments, whole 8-cell embryos were freed from their zonae, disaggregated to single or paired blastomeres in Ca²⁺-free medium and the blastomeres analysed immediately.

Surface polarity was assessed by incubation of cells or embryos in 50μgml⁻¹ tetramethylrhodamine-labelled succinyl Concanavalin A (50μgml⁻¹ M2+BSA: TMRTC-S-ConA, Polysciences) for 5 min at room temperature, followed by two to three washes in M2+BSA. Labelled cells were then placed in specially designed chambers exactly as described previously (Maro, Johnson, Pickering & Flach, 1984) for fixation with 3·7 % formaldehyde followed by extraction with 0·25 % Triton X-100. After washing, cells were incubated with affinity-purified polyclonal, rabbit antibody to clathrin, the major coat protein of coated vesicles (Louvard etal. 1983), to a 135kD antigen associated with the Golgi apparatus (Louvard, Reggio & Warren, 1982), to an endoplasmic reticulum antigen (Louvard et al. 1982) or to a 100kD protein associated with the acid vesicle/lysosomal fraction (Reggio et al. 1984). A second layer of fluorescein-labelled anti-rabbit immunoglobulin was used to visualize the bound antibody. The detailed characteristics of the procedures have been reported previously (Maro etal. 1984).

Samples were mounted in Citifluor (City University, London) in order to reduce fading of fluorescent labels and viewed on a Leitz Ortholux II microscope with filter set L2 for FITC-labelled reagents and N2 for TMRTC-labelled reagents. Photographs were taken on Kodak Tri X film using a Leitz Vario-orthomat photographic system.

We first studied the distribution of organelle antigens in permeabilized whole embryos from the unfertilized egg to the 16-cell stage. Each antigen examined showed a pattern of distribution that varied characteristically with time.

Clathrin was distributed in a diffuse granular pattern throughout early development (Fig. 1b,c,d,e) but with two exceptions. First, when meiotic (Fig. la) or mitotic (Fig. lh,i) cells were examined, clathrin distribution corresponded closely with that of the spindle, and remaining areas of the cytoplasm were relatively free of clathrin. Second, during the late 8-cell stage clathrin appeared to be more concentrated in the apical region of the blastomeres (Fig. If). However, this localization was difficult to resolve clearly in whole mounts.

The Golgi antigen was also distributed diffusely throughout the cytoplasm of early embryonic cells (Fig. 2a–e). However, at all stages the granular foci of Golgi antigen tended to be larger than those observed with the anti-clathrin antibody, and this was particularly marked at the 8-cell and 16-cell stages. Moreover at 16cell stages there appeared to be a greater concentration of Golgi-antigen at the centre of the embryo (Fig. 2e), a location corresponding to the totally enclosed inside cells and/or to the basal regions of outer cells.

The antiserum to endoplasmic reticulum antigen stained the nuclear membrane region intensely, together with a diffuse granular staining throughout the cytoplasm that was somewhat weaker at the 1-cell stage than subsequently (Fig. 2f–h). Otherwise, no convincing evidence of temporal change or spatial asymmetry in staining pattern was observed with use of this antiserum, although in many cases the most peripheral areas of the blastomeres can appear to stain more weakly. However, it was difficult to be certain that this appearance was not simply due to the greater density of positive cytoplasm at the centre of the embryo.

The antiserum to the 100kD protein associated with the acid organelle/lysosome fraction gave a diffuse granular pattern of reaction (Fig. 2i–1), but, as for Golgi-antigen, the granular foci were larger at 8- and 16-cell stages (Fig. 2k, 1). Moreover, in some late 8-cell embryos and most 16-cell embryos, there was a concentration of 100kD antigen in clumps at the centre of the embryo (Fig. 2k,1).

The use of the whole embryo mounts allows a general assessment of changing temporal and spatial patterns of antigen distribution, and suggests that for clathrin, Golgi and 100kD antigens changes occur at the 8-cell stage and later. In order to visualize these changes more clearly, we used pairs of 8-cell or 16-cell blastomeres. This approach reduces background interference from fluorescent emission outside the plane of focus and also permits a more accurate temporal staging of blastomeres within the fourth and fifth developmental cell cycles.

Preparations of isolated 4-cell and 8-cell blastomeres were made, cultured and examined at hourly intervals for evidence of division to yield two 0 h 8-cell or 16cell blastomeres (a 2/8 or 2/16 pair). Pairs were then cultured for up to 11 h before being examined for their surface phenotype (assessed by binding of TMRTC-succinyl-Con A) and the distribution of organelles. Surface phenotype in 2/8 pairs was categorized as being apolar if Con A was uniformly bound and polar if Con A binding was restricted to the apical region of the cell (compare Fig. 4b apolar with Fig. 4k, polar). For 2/16 pairs, three surface phenotypes were defined, namely polar in which Con A binding was restricted to a limited area of membrane (e.g.upper cell Fig. 7d), bright apolar, in which the surface is brightly labelled over all or most of its surface (e.g. upper cell Fig. 7b) and dull apolar, in which a uniform weak labelling was observed (e.g. lower cell Fig. 7b). We have shown previously (Johnson & Ziomek, 1981) that the phenotype of a 2/16 couplet depends upon the way in which the polarized 1/8 cell divides. Thus if the cleavage plane is oriented perpendicular to the axis of polarity one bright and one dim apolar cell result, and the bright area ‘shrinks’ over a 1h period to form a discrete pole (e.g. Fig. 7b converts to Fig. 7d). If the division plane is oriented along the axis of polarity of the 1/8 blastomere, two polar cells result (e.g. Fig. 7f,h). Five patterns of organelle distribution were observed, and these are indicated schematically in Fig. 3.

The changing distribution of intracellular clathrin in relation to surface polarity at the 8-cell stage is summarized in Table 1. Three points emerge from these data. First, the incidence of surface polarity increased with time (Table 1, column 9); this result confirms previous observations (Ziomek & Johnson, 1980). Second, the distribution of intracellular clathrin was mainly zonal soon after division (Table 1, line 1, columns 3 and 4), but thereafter the proportion of cells homogeneous or zonal for clathrin declined whilst the proportion polar for clathrin increased (summarized in Table 1, column 10). Third, cells were detectably polarized for intracellular clathrin before showing evidence of surface polarity (Table 1, compare columns 9 and 10). Fig. 4 shows examples of cells that were apolar at their surface but homogeneous (Fig. 4c,d,e), zonal (Fig. 4a,b) or polar (lower cell Fig. 4f,g and Fig. 4h,i lower blastomere) for clathrin distribution. In addition, when the surface pole was present, it invariably overlay the pole of clathrin (Fig. 4j,k and upper cell Fig. 4h,i) and in almost all cases examined this dual pole was opposite to the point of contact with the partner cell (e.g. Fig. 4j,k). Moreover, clathrin concentration in the polar region was not confined to the cytoplasm, the membrane overlying the pole of clathrin-positive vesicles also staining clearly (e.g. Fig. 4h,j). Enhanced membrane staining for clathrin was also observed in regions of cell apposition.

The redistribution of clathrin from homogeneous to polar pattern, and its relationship to the surface pole of Con A binding, are not artefacts of the in vitro culture of 2/8 pairs, as was shown by the staining patterns of blastomeres isolated from precompact 8-cell and from compacted late 8-cell embryos (Fig. 41,m), in which a homogeneous clathrin distribution predominated in the former and a polar pattern in the latter.

The analysis of intracellular clathrin distribution in 2/16 pairs is complicated by the cell heterogeneity at the 16-cell stage. Thus, two populations of cells exist that differ in surface phenotype (bright or polar; dull and apolar) as well as in properties, developmental fate and lability (reviewed Johnson, 1985). One such property is the tendency after 4 to 8 h in culture of the polar cells to envelope the apolar cells (e.g. see Fig. 8e–h), this being a reflection of the role of polar cells in situ as precursors of the trophectoderm of the 32-cell blastocyst stage (Ziomek & Johnson, 1982). Envelopment makes scoring of surface and intracellular phenotype more difficult. In most cases therefore we incubated the 2/16 couplets in the presence of a monoclonal antibody to cadherin (see Materials & Methods), a surface homotypic, Ca²⁺-dependent, adhesion molecule; this antibody prevented cells from flattening on each other and blocked the process of envelopment (c.f. Fig. 8e–h with Fig. 8i–1). It did not interfere, under the conditions used here, with polarization of surface or intracellular organelles at the 8-cell stage (compare lines 6 and 7 in Table 1; also Johnson, 1985).

The data summarizing the distribution of intracellular clathrin in 2/16 pairs is summarized for each cell subpopulation in Table 2. Four points emerge from these data. First, immediately after division, most cells regardless of surface phenotype showed a non-polar distribution of clathrin (Table 2, lines 1 and 5; Fig. 7a,b). Second, at later stages most cells with a polar surface phenotype also manifested a polar distribution of clathrin (Table 2, columns 4 and 8, lines 2–4). Third, in these cells the clathrin and surface poles were almost always scored as coincident (Fig. 7c–f); in cells cultured in the presence of the antibody to cadherin the clathrin tended to cluster between the nucleus and the surface pole, but when flattening and envelopment occurred the clathrin-positive staining was displaced laterally as the nucleus became located closer to the surface membrane. The surface and clathrin poles showed no consistent relationship to the contact point with the other cell (Table 2, columns 9 and 10, lines 2–4; this latter result confirms and extends a previous report; Johnson & Ziomek, 1981). Fourth, at later stages, most cells with an apolar surface phenotype did not have a polar distribution of clathrin (Table 2, columns 5–8, lines 6–8).

The data on the distribution of the Golgi antigen with time and in relation to surface polarity are summarized in Table 3 and illustrated in Fig. 5. No clear trend towards a polar organization of the Golgi antigen is evident. Immediately after division the Golgi antigen appeared in many cells to concentrate in a single polar aggregate in association with the spindle pole (Table 3, line 1; Fig. 5a,b). However, thereafter the Golgi antigen was dispersed throughout the cell in multiple aggregates of varying size and distribution (Fig. 5c–h).

Data for the distribution of Golgi antigen in the two cell subpopulations identifiable at the 16-cell stage are summarized in Table 4, from which four points emerge. First, with time Golgi antigen changed from a dispersed into an increasingly aggregated organization and ultimately into a single aggregate (designated polar in Table 4; columns 8 and 9, Fig. 7g,h). Second, this concentration of the Golgi antigen occurred regardless of whether cells had a polar or an apolar surface phenotype. Third, the polar aggregate was not obviously dr consistently related to the contact point with the other cells (Table 4; column 11). Fourth, in cells that had surface poles, the Golgi antigen was more often on the axis of polarity than off it (Table 4, column 10). Moreover at 5–6 h, and especially in pairs in which flattening and envelopment occurred, the Golgi was mainly located coincident with the surface pole (e.g. Fig. 7g, upper cell) whereas at the later time point a more basal location opposite to the pole was more frequent, (e.g. Fig. 7g, lower cell).

In general the endoplasmic reticulum antigen showed an apolar distribution at all time points and in all cell types examined (Table 5). Only two deviations from this pattern were observed. First, as cells flattened on each other at the 8-cell or 16cell stage, a zone free of endoplasmic reticulum developed adjacent to the zone of contact in some polar cells (Table 5, columns 5 and 10; Fig. 6a–d; Fig. 8a,b). When flattening was reduced in the presence of the monoclonal antibody to cadherin, zonal clearance of antigen was reduced (Table 5, column 10 compare lines 3 and 4, also compare Fig. 8a and 8c). Second, in some polar cells the apical cytoplasmic zone also appeared to be relatively deficient in ER antigen (e.g. Fig. 6c, upper cell).

The distribution of the 100kD antigen is recorded in Table 6. As for the Golgi antigen, little evidence of redistribution to a focal, polar state was evident during the 8-cell stage, most cells showing a homogeneous granular pattern (Fig. 6e,f). Only rarely was a polar localization of antigen observed (Fig. 6g,h).

The distributions of 100kD antigen in the two cell subpopulations identifiable at the 16-cell stage are summarized in Table 7. Four points emerge from these data. First, with time the 100kD antigen concentrated into a single aggregate (designated polar in Table 7, column 9; Fig. 8g-l). Second, this concentration occurred regardless of the cell surface phenotype. Third, the polar aggregate was not obviously or consistently related to the point of contact with the other cell (Table 7; column 11). Fourth, in cells that had surface poles the 100kD antigen tended to lie along the axis of polarity (Table 7; column 10). In the 5–6 h group there was a particularly high incidence of flattening and envelopment and the lysosomes tended to concentrate in the ‘arms’ of the outer cell processes that extend round the apolar cell (see Fig. 8e). In pairs cultured in the presence of the antiserum to cadherin the lysosomal antigen was concentrated initially between the nucleus and the pole, but at the later time point had shifted to the opposite or basal side of the nucleus in many polar cells scored (Table 7, lines 3 and 4, column 10; Fig. 8i,e,k).

The process of de novo polarization of blastomeres in the mouse early embryo is of central importance to the generation of cell diversity in the blastocyst (Johnson, 1985) and of considerable interest as a cell biological phenomenon. The polarization process is oriented by contact signals from other cells (Ziomek & Johnson, 1980), and is initiated at a characteristic stage of development. The acquisition of polarized features by the cell occurs progressively, new polar features being acquired, and established polar features being elaborated, at successive 8-, 16- and 32-cell stages (see Johnson, 1985; Fleming etal. 1984; Fleming & Pickering, 1985). In this paper we have examined the changing distribution with time of four membranous organelles, as inferred from antigenic distribution, and have detected three distinctive patterns of change.

The levels of the endoplasmic reticulum antigen detected appear to increase after the 1-cell stage but otherwise the antigen was distributed uniformly throughout the cytoplasm at all stages examined except adjacent to contact zones with other cells. A similar distribution is observed in differentiated cells (Louvard et al. 1982). Exclusion of actin (Johnson & Maro, 1984) and myosin (Sobel, 1983) from contact zones has also been described previously, and in this study intracellular clathrin was likewise excluded from contact zones.

Cytoplasmic clathrin also showed a dispersed distribution (other than in contact zones) except in two situations. First’ in all cells that were polarized (or polarizing) at their surface, clathrin also accumulated in a focal aggregate or pole that was located immediately underneath the surface pole. Moreover, this polarity of clathrin preceded by several hours the occurrence of detectable polarity at the surface. The redistribution of intracellular clathrin coincided with that reported for endosomes (Fleming & Pickering, 1985) and for filamentous, cytoplasmic actin (Johnson & Maro, 1984), both of which also polarized in advance of the cell surface and which also colocalized with clathrin. It seems probable that these three events are linked either causally to each other or via some underlying mechanism that affects each. The reorganization of clathrin, endosomes and actin thus provides an early indication of cell polarization at the 8-cell stage. However, clathrin polarity is not as stable as the later developing surface pole. Thus, at mitotic (or meiotic) division, the clathrin redistributed from its polar location to the spindle; an association between microtubules or tubulin and coated vesicles or clathrin has also been observed in other mitotic and interphase cell types (Imhof et al. 1983; Kelly et al. 1983; Louvard & Reggio, 1981; Louvard et al. 1983; Pfeffer, Drubin & Kelly, 1983). The association between the spindle and clathrin presumably ensures that the latter is distributed to each daughter cell in a 2/16 couplet, in which it is then relocated in a polar distribution only in the progeny cells that had surface poles. Thus, at the 16-cell stage the surface pole appears to act as, or be associated with, an organizing focus for cytoplasmic polarity. In most nonpolar cells of 2/16 couplets, clathrin remained distributed throughout the cell with no obvious polar cluster. In a few cells, a polar cluster did form after 5–9 h at the 16-cell stage but only in cells treated with antibody to cadherin. Under such conditions, envelopment of the apolar cell is prevented, and it is known from previous work that non-enveloped apolar 1/16 cells will start to develop elements of polarity at about this time (Ziomek & Johnson, 1982). The polar clustering of clathrin could therefore represent the earliest manifestations of this regulative polarization.

The distributions of the 100kD (acidic organelle/lysosome) and the Golgi antigens are similar to each other but differ from that of clathrin, an observation in striking contrast to the situation in fully differentiated cells in which clathrin and Golgi antigen tend to colocalize (Louvard & Reggio, 1981; Louvard et al. 1983), and in which the acid vesicle antigen can be detected in endosomes and coated vesicles (Reggio et al. 1984). These differences presumably relate to the relative immaturity of processing pathways in blastomeres (Fleming & Pickering, 1985). Both the 100kD and Golgi antigens were dispersed up until the 16-cell stage, although during the 8-cell and early 16-cell stage the antigens were increasingly aggregated into fewer, larger clumps. During mitosis, the antigens distributed to both poles of the spindle, thereby presumably ensuring transmission of each organelle to both progeny. During the 16-cell stage, each antigen became organized into a single (polar) clump, but did so regardless of the surface phenotype of the cell. Thus, unlike the focal and polar distribution of clathrin, that of the Golgi/lysosomal antigens did not appear to be related to the development of polarity but more to the maturation of endocytic and secretory function in the cells. Analysis of the maturing endocytic pathway at this time supports the view that major changes in the organization of and capacity for lysosomal processing occur during the 16-cell stage (Fleming & Pickering, 1985), with the first appearance of secondary lysosomes. However, although the concentration of Golgi/ lysosomal antigens into a single aggregated focus represents a maturational change in both polar and apolar cell types at the 16-cell stage, the location at which the foci of antigen developed in cells with polar surface phenotypes may be related to the axis of polarity of each cell in this subpopulation. Thus, in antibody-treated polar cells there was the suggestion of a shift from an initially mainly apical to a later mainly basal location. In non-antibody-treated pairs there is a suggestion that, unlike clathrin, the lysosomal and Golgi antigens localize in the enveloping arms of the polar cells away from the nuclei. In intact embryos, the aggregates of antigen appear to locate basally. Independent cytochemical evidence also suggests that at the late 16- and early 32-cell stages lysosomal-like bodies and the Golgi apparatus locate basally (Fleming & Pickering, 1985) or in the enveloping arms of polar, trophectoderm cells (Fleming et al. 1984).

Alignment of the Golgi apparatus along the axis of polarization through the cell has been observed in differentiated cells, in which an association with the microtubule organizing centre is also reported (Carpen, Virtanen & Saksela, 1982; Kupfer, Dennert & Singer, 1983). However, in late 8-cell mouse blastomeres the MTOC is located apically (unpublished observations by B. Maro and S. J. Pickering), again stressing that early embryonic cells may not have established the range of interorganelle associations characteristic of more mature cells. It is also noteworthy, that in embryonic chick corneal epithelium the Golgi apparatus shifts from an apical to basal position during two periods each correlating in time with the appearance of an acellular collagenous matrix beneath the epithelium (Trelstad, 1970). Significantly, in this regard, in mouse embryos, all three polypeptide subunits of laminin are first synthesized and secreted basolaterally from the 16-cell stage (Leivo, Vahari, Timpl & Wartiovaara, 1980; Cooper & MacQueen, 1983).

Thus, a temporal sequence for the development of polarity in mouse early blastomeres may be emerging. Actin, endosomal and clathrin redistribution are evident early in the 8-cell stage. The surface polarization becomes evident later in the 8-cell stage. Golgi and lysosomal bodies align on the polar axis during the 16cell stage at the same time as endocytic processing pathways mature. In the accompanying paper (Johnson & Maro, 1985) we describe experiments to disrupt selectively elements of this sequence of maturation, and as a result we propose a model for the process of polarization.

We wish to thank our research colleagues and Professor S. J. Singer for their critical advice and reading of the manuscript, Gin Flach, Ian Edgar and Sheena Glenister for their technical assistance, and Dr M. Takeichi for supplying the antibody to cadherin. The work was supported by grants to M. H. Johnson and P. R. Braude from the Medical Research Council and the Cancer Research Campaign. B. Maro is a visiting EMBO Research Fellow.