At least one polar body, almost invariably the second, persists intact to the early blastocyst stage in nearly two-thirds of mouse conceptuses of the PO strain. The distribution in early blastocysts of these surviving polar bodies was highly non-random. Most not only lay in the midregion of the embryonic-abembryonic axis but, on discovering that early blastocysts are bilaterally rather than radially symmetrical about this axis, were found to align with the bilateral axis. Cell marking experiments failed to detect movement of polar bodies relative to the surface of the conceptus during either cleavage or blastulation. That the distribution of degenerating polar bodies and their presumed debris was similar to intact ones also argued against their motility, as did the finding that at all stages second polar bodies were attached to conceptuses by a thin, extensible, weakly elastic ‘tether’. Although the transfer of small fluorochromes between them was rarely observed beyond second cleavage, the second polar body and conceptus could remain coupled ionically up to the blastocyst stage. It is concluded that the second polar body normally remains attached to the conceptus through persistence of the intercellular bridge formed during its abstriction, and therefore provides an enduring marker of the animal pole of the zygote. Hence, according to the distribution of polar bodies, the axis of bilateral symmetry of the early blastocysts is normally aligned with the animalvegetal axis of the zygote and its embryonic-abembryonic axis is orthogonal to it. Such relationships suggest that, at least in undisturbed development, specification of the axes of the blastocyst depends on spatial patterning of the zygote.

Spatial patterning of the egg has been implicated in early development in a sufficient variety of invertebrates and vertebrates to suggest that it is the norm among metazoa. Mammals are regarded as exceptions because the dependence of their initial cell differentiation on cytoplasmic ‘determinants’ in the egg (Dalcq, 1957) is held to be untenable (Wolstenholme and O’Connor, 1965). However, in species in which egg organization is important, it seems to be concerned more with the establishment of axes and global patterning than differentiation of specific types of cells (Wilson, 1928; Davidson, 1986). That egg organization might be involved in specifying axes in mammals has not been considered, largely because of the impressive ability of the early conceptus to regulate its development following removal, addition or rearrangement of cells (Gardner, 1996). However, even in highly regulative embryos axes may be specified at the onset of cleavage during normal development (Cameron et al., 1987). It has also been argued that organization of the embryo proper begins too late in development to depend on spatial information in the egg in mammals (Tarkowski and Wroblewska, 1967). This objection is undermined by evidence that the mouse conceptus is bilaterally symmetrical by the late blastocyst stage (Smith, 1980, 1985; Gardner, 1990), and that the bilateral axis of the embryo is aligned with that of the conceptus (Gardner et al., 1992).

Determining whether egg organization is involved in spatial patterning in mammals is complicated by the apparent absence of candidate markers of such organization that are discernible in the living state. The cytoplasmic territories identified by Dalcq and his colleagues as defining bilateral symmetry in the egg (Jones-Seaton, 1950; summarized in Dalcq, 1957) can only be distinguished readily in fixed material that has been stained histochemically and, even then, only before the first cleavage in the mouse. The possibility that polar bodies (Pbs) might serve as enduring markers of the orientation and polarity of the animal-vegetal (AV) axis of the egg has been dismissed on two grounds. Their survival through cleavage is said to be low (Tarkowski and Wroblewska, 1967), and their motility incompatible with retention of their ancestral position (e.g. Lewis and Wright, 1935; Borghese and Cassini, 1963).

The present investigation was prompted by chance observations, which suggested that intact Pbs were common, and their distribution strikingly non-random, in early blastocysts of the PO (Pathology, Oxford) strain.

Oocytes and conceptuses

Closed-bred PO mice were used throughout this investigation. All conceptuses were obtained following natural mating of females that were selected for oestrus by external inspection (Champlin et al., 1973). Hepesbuffered MTF medium (Gardner and Sakkas, 1993) was used for recovery, holding at room temperature, and manipulation of oocytes and conceptuses, and bicarbonate-buffered MTF was used for both their short- and long-term culture. Minimal exposure of conceptuses to acidified Tyrode’s saline (AT), prepared as described by Pratt (1987), was used to remove the zona pellucida (zona). For studying patterns of chromatin distribution in Pbs, oocytes and conceptuses were incubated for 5 minutes at 37°C in 5 μg/ml Hoechst 33342 (Sigma) in MTF-Hepes and rinsed in fresh medium before examination by fluorescence microscopy.

All manipulations were done in hanging drops of medium in Puliv (Leitz) chambers charged with heavy paraffin oil (BDH, UK). Oocytes and conceptuses were immobilized by applying suction to a heat-polished holding pipette via a microinjector (Narishige IM-6, Japan), as described elsewhere (Gardner, 1978). Siliconized (Repelcote, BDH, UK), solid glass microneedles whose tips had been rounded with a De Fonbrune microforge (Beaudouin, France) were used to relocate Pbs through the intact zona and investigate their attachment to the conceptus.

Single blastomeres were labelled with horseradish peroxidase (HRP) or DiI (Molecular Probes, USA). HRP (Grade 1, Boehringer, Germany) was dissolved in 0.05 M KCl that also contained 1%(w/v) lysinated fluorescein dextran or tetramethylrhodamine dextran (both 40,000 MW: Molecular Probes, USA) for monitoring injections by fluorescence microscopy. For labelling at the late 2-/early 4-cell stage the final HRP concentration was approx. 10%(w/v), whereas for later labelling it was 6%.The enzyme was injected using microelectrodes pulled from borosilicate, filament-containing, 1 mm capillary tubing (either GC100TF-15 or GC100F-15, Clark Electromedical Instruments, UK), with a micropipette puller (model P-87, Sutter, USA). The holding pipette was inserted into an electrode holder with a side port so it also served as the indifferent electrode (Beddington and Lawson, 1990). Injection was controlled with a Neurolog System (Digitimer, Welwyn Garden City, UK) coupled to an oscilloscope (Type 546B, Tetronix, Guernsey) for monitoring transmembrane potential. Microelectrode penetration was effected by briefly depressing the negative capacitance button on the preamplifier unit, and a depolarizing (+) current of 4-8 nA was then applied, either continuously or in variably spaced pulses of 500 ?seconds, until the intensity of fluorescence of the co-injected dextran conjugate was judged satisfactory. Alternatively, the high frequency electronic oscillation caused by depressing the negative capacitance button was used to induce bulk flow of fluid from the tip of the microelectrode (Warner and Bate, 1987). Fluorescein complexon (FC, Kodak, UK), made up at 100-300 mg/ml in deionized water, was injected similarly.

Individual cells were labelled with DiI, essentially as reported recently (McCain and McClay, 1994), by injection with a small drop of a saturated solution of the dye in corn oil (Mazola, UK), which was microfuged at 13,000 rpm before use. Pipettes pulled from siliconized Leitz capillary tubing with an aperture of less than 1 μm were back-filled with a small volume of the solution followed by corn oil alone, and then connected to a De Fonbrune microinjector (Beaudouin, France) containing heavy paraffin oil. The DiI usually spread rapidly from the drop to label both the internal and plasma membranes of injected cells.

A complication in labelling individual cells in the preimplantation conceptus is that bridges connecting sister cells can remain patent throughout, and even beyond, the succeeding cell cycle (see e.g. Goodall and Johnson, 1984; Goodall, 1986). Failure of restriction of label to the injected cell was encountered equally frequently with HRP and DiI. Following injection of 1/2 blastomeres, this either resulted in uniform labelling of the entire conceptus or, more commonly, weaker labelling of one half than the other. Fully labelled conceptuses were rejected unless there was an obvious difference in intensity of staining between the two halves. The problem could largely be avoided by postponing labelling until shortly after the second cleavage, when injection of a 1/4 blastomere usually resulted in uniform staining that was entirely confined to sister 2/4 blastomeres. Conceptuses labelled with DiI were examined live by fluorescence microscopy. Those labelled with HRP were fixed in 0.25% glutaraldehyde in 0.1 M phosphate buffer, pH 7.2, for 6 minutes before being stained for enzyme activity for up to 15 minutes, as described elsewhere (Beddington and Lawson, 1990). Labelling was evaluated by placing conceptuses in hanging drops of a Puliv chamber so that they could be rotated with a blunt glass probe attached to a Leitz micromanipulator. DiI proved less satisfactory for longer-term labelling experiments than HRP because it became punctate and perinuclear with time, thus failing to delimit cell boundaries.

For investigating ionic coupling between the conceptus and Pb(s), the procedure of Goodall and Johnson (1984) was adapted for use in Puliv chambers where conceptuses could be visualized much more clearly. This entailed making the Agar/KCl bridge from a micropipette with a tip diameter of approx. 100 μm, and including a standard holding pipette so that conceptuses could be immobilized securely. The microelectrodes had a resistance of between 60 and 90 MΩ when filled with 0.2 M KCl. Regular checks were made to see if there was any shift in potential of the recording microelectrode on applying current pulses to the current injection microelectrode when both were extracellular. This was to guard against possible artefactual coupling due to increase in resistance of the small Agar/KCl bridge through evaporation of fluid. Recording is relatively simple from the late 8-cell stage since all cells are in communication via gap junctions, except briefly during mitosis (Goodall and Maro, 1986). Initially, the current injection microelectrode was inserted into a cell on one side of the conceptus and 1 second pulses of a 1 nA hyperpolarizing current was applied to it repeatedly at 1 second intervals. The recording microlectrode was then placed in a cell near the Pb on the opposite side of the conceptus. If coupling was observed, the recording electrode was moved into the Pb. If this also showed coupling, care was taken to verify that the electrode tip was in the Pb, and that coupling ceased once the Pb had deteriorated or the microelectrode was extracellular. Finally, to check that the current injection microelectrode remained in place in a cell that was still coupled with the rest of the conceptus, the recording microelectrode was once more inserted into a cell.

For zygotes, the coupling tests before and after penetrating the Pb were done with both microelectrodes in the vitellus. 2- and 4-cell stages posed the greatest difficulty, since coupling between blastomeres depends on the variable retention of patency of intercellular bridges, and it is seldom possible to decide to which blastomere the Pb is attached.

For examining attachment sites of Pbs by transmission electron microscopy (TEM), blastocysts were placed individually in small hanging drops of MTF-Hepes, each of which was adjacent to a much larger drop of half-strength Karnovsky’s fixative (Karnovsky, 1965). In turn, blastocysts were then oriented with a blunt glass probe so that the Pb was directed towards the tip of a holding pipette with an aperture of approx. 30 μm. Sufficient suction was then applied to the pipette to draw in part of the zona and pull the Pb slightly away from the surface of the trophectoderm. The glass probe was then used to merge the drop containing the immobilized blastocyst with that containing fixative. Blastocysts were exposed thus to fixative for 6 minutes before being released from the holding pipette and transferred to a much larger volume of fresh fixative at 4°C overnight. Following fixation, blastocysts were osmicated and processed for examination in a Philips 400 microscope.

Only early and expanding blastocysts with intact Pb(s), whose movement was restricted within the perivitelline space, were selected for examination by SEM. Blastocysts were first immobilized by suction and the Pb(s) moved by at least their own diameter by applying a glass probe to the intact zona. Some were then transferred immediately to half-strength Karnovsky’s fixative via a PBS rinse, and others were cultured for 1-2 hours before fixation. After 10-25 minutes in fixative at room temperature, the blastocysts were again rinsed briefly in PBS before being incubated in 0.5% (w/v) Pronase (grade B, Calbiochem, USA) in PBS at room temperature. Once denuded of their taraldehyde in 0.1 M cacodylate, where they remained at 4°C for between 5 and 20 hours. They were then transferred via 1% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) to 13 mm diameter circular No. 1 coverslips coated with poly-L-lysine, as described by Pratt (1987), before being immersed in 0.1 M cacodylate buffer. Thereafter, the blastocysts were processed for viewing in a Philips 515 SEM.

Axes

Conventionally, the animal pole of the egg or zygote is where the Pbs are formed (e.g. Balinsky, 1970). A complication in the mouse is that the 1st Pb does not usually lie directly above the second meiotic spindle in mature oocytes (Zamboni, 1970: Calarco, 1995). Since this is explained more readily by movement of the 1st Pb than the spindle, the animal pole was assumed to be aligned with the latter before fertilization and with the 2nd Pb thereafter. The vegetal pole was taken as the point on the surface diametrically opposite the animal pole, and the animal-vegetal (AV) axis as the line connecting the two (see Fig. 1A). The conceptus is assumed to be spherically symmetrical until the blastocyst stage, when a new axis of polarity is defined by the formation of an eccentric blastocoelic cavity. This axis, referred to as the embryonic-abembryonic (EmAb) axis, extends from the centre of the polar to the centre of the mural trophectoderm (Fig. 1B). The general view that symmetry is radial about this axis until the primitive streak forms is contradicted by observations reported here and elsewhere (Smith, 1980, 1985; Gardner, 1990; Gardner et al., 1992).

Fig. 1.

Diagrams illustrating polarity of the zygote (A) and blastocyst (B) in which the animal-vegetal (AV) axis of the zygote and the embryonic-abembryonic (EmAb) axis of the blastocyst are indicated by dashed lines. The horizontal line in (B) represents the polar-mural junction (p/mj), the boundary between the polar region of the trophectoderm overlying the inner cell mass (ICM) and the mural region embracing the blastocoelic cavity.

Fig. 1.

Diagrams illustrating polarity of the zygote (A) and blastocyst (B) in which the animal-vegetal (AV) axis of the zygote and the embryonic-abembryonic (EmAb) axis of the blastocyst are indicated by dashed lines. The horizontal line in (B) represents the polar-mural junction (p/mj), the boundary between the polar region of the trophectoderm overlying the inner cell mass (ICM) and the mural region embracing the blastocoelic cavity.

Staging

Up to compaction, conceptuses are referred to by the number of blastomeres they contain, except for the 1-cell stage for which the term zygote is reserved. Individual cells of 2-, 4- and 8-cell stages are denoted as 1/2, 1/4 or 1/8 blastomeres. The term morula (M) is applied to compacted conceptuses recovered on the 3rd day postcoitum (p.c.), and very late morula (VLM) to compacted conceptuses recovered on the morning of the 4th day in which no extracellular fluid was discernible in the dissecting microscope. The former would be expected to be composed of between 8 and 16 cells and the latter of nearer 32 than 16 cells. Conceptuses with obvious extracellular fluid but which lacked a clearly demarcated blastocoelic cavity were classified as nascent blastocysts. Blastocysts with an intact zona and unitary blastocoele were assigned to one of four categories according to the size of the blastocoele relative to the ICM, as judged by visual assessment of the M/P ratio (see Fig. 1). The four categories studied, together with their abbreviations, are as follows:

Survival of Pbs during preimplantation development

Throughout cleavage and at the early blastocyst stage conceptuses could readily be scored for the presence or absence of Pbs without removing the zona. This was not possible once expansion of the blastocoele had begun because the Pbs became increasingly flattened between the zona and the trophectoderm. However, Pbs were usually obvious in denuded expanding blastocysts and, providing the zona was removed carefully, they remained attached to the trophectoderm there-after. In some more advanced blastocysts, Pbs appeared to be in the process of phagocytosis by the trophectoderm. At all stages, the great majority of conceptuses without intact Pbs had either lysed ones or foci of debris, which were almost invariably attached to the conceptus rather than free in the perivitelline space.

The proportion of conceptuses with intact Pbs at successive stages of preimplantation development from the mid-zygote to the expanded blastocyst is shown in Fig. 2. This gradually decreases during cleavage from 100%in zygotes to about two-thirds in VLMs. Little further decline occurs during blastulaion, so that 64% of EBs had at least one intact Pb. A further decrease to below 50% is evident once blastocoelic expansion begins. Regardless of stage, most conceptuses with intact Pbs had only one such body (see legend to Fig. 2).

Fig. 2.

Histogram showing percentage of conceptuses with intact Pbs at successive stages of preimplantation development in vivo. The total number of conceptuses examined at each stage is shown in brackets. The occurrence of more than one intact Pb was recorded for all zygotes and for the majority of positive conceptuses at other stages, as follows: Zyg, 27/130 (21%); 2- to 4-cell, 58/333 (17%); 8- cell, 10/100 (10%); M, 36/209 (17%); VLM, 9/93 (10%); VEB, 4/23 (17%); EB, 27/280 (10%); ExB, 13/108 (12%); ExdB, 1/14 (7%).

Fig. 2.

Histogram showing percentage of conceptuses with intact Pbs at successive stages of preimplantation development in vivo. The total number of conceptuses examined at each stage is shown in brackets. The occurrence of more than one intact Pb was recorded for all zygotes and for the majority of positive conceptuses at other stages, as follows: Zyg, 27/130 (21%); 2- to 4-cell, 58/333 (17%); 8- cell, 10/100 (10%); M, 36/209 (17%); VLM, 9/93 (10%); VEB, 4/23 (17%); EB, 27/280 (10%); ExB, 13/108 (12%); ExdB, 1/14 (7%).

Where the internal details of Pbs could be visualized clearly by Nomarski optics, singletons were typically nucleated, as was one member of the pairs and triplets. However, because many Pbs could not be scored confidently for the presence or absence of a nucleus in this way, additional early blastocysts, 2-4-cell conceptuses and oocytes with intact Pbs, were examined by fluorescence microscopy after staining with Hoechst 33342. In oocytes, 1st Pbs always had a dispersed rather than a nucleus-like pattern of fluorescence (Fig. 3A), regardless of whether they remained single or had divided, and this was typically retained by those surviving beyond first cleavage (Fig. 3B). Whilst only a minority of 1st Pbs divided, many were obviously ‘dumb-bell’ shaped. In conceptuses with a single Pb, this body almost invariably exhibited a nucleuslike pattern of fluorescence (125/128=98%), arguing that it was the 2nd rather than the 1st (Fig. 3D,E). Consistent with this is the finding that in the great majority of conceptuses with two or three Pbs, only one Pb had such a nucleus-like pattern (54/61=89%: Fig. 3B,C). However, because this pattern was observed in more than one Pb in the remaining seven (=11%), it is not a completely reliable way of discriminating between 1st and 2nd Pbs. Finally, among the total of 263 Pbs examined by fluorescence microscopy, seven gave no signal above back-ground. These may have been products of deutoplasmolysis rather than Pbs (Dalcq, 1957). Such occasional abstricted, membrane-bound, cytoplasmic masses were usually readily distinguishable from Pbs by their larger size and marked granularity, but might be misclassified when small.

Fig. 3.

Chromatin patterns in Pbs stained with Hoechst 33342. (A) Oocyte showing diffuse staining in the 1st Pb (arrow) which, as often the case, is remote from the second meiotic metaphase (Zamboni, 1970; Calarco, 1995). (B) Diffuse staining in the 1st Pb of a 2-cell stage compared with the condensed, nucleus-like staining in the 2nd Pb of the same conceptus (C). (D,E) Differential interference contrast and fluorescence microscopic images of two 2nd Pbs stained after their detachment from EBs. Bars: 20 μm (A-C); 10μm (D,E).

Fig. 3.

Chromatin patterns in Pbs stained with Hoechst 33342. (A) Oocyte showing diffuse staining in the 1st Pb (arrow) which, as often the case, is remote from the second meiotic metaphase (Zamboni, 1970; Calarco, 1995). (B) Diffuse staining in the 1st Pb of a 2-cell stage compared with the condensed, nucleus-like staining in the 2nd Pb of the same conceptus (C). (D,E) Differential interference contrast and fluorescence microscopic images of two 2nd Pbs stained after their detachment from EBs. Bars: 20 μm (A-C); 10μm (D,E).

These findings show that at least one Pb, almost invariably the 2nd, survives intact to the early blastocyst stage in almost two-thirds of PO conceptuses developing in vivo. As discussed later, comparable survival of Pbs was found in early blastocysts that developed in vitro from the late 2-/early 4-cell stage.

Location of Pbs on the surface of early blastocysts

The initial chance observation on early blastocysts was that surviving Pbs seemed to occur more frequently at the junctional region between polar and mural trophectoderm (p/mj; see Fig.1) than elsewhere. To investigate whether their distribution was indeed non-random, a series of carefully staged EBs were photographed either before or after removing the zona so that the location of the visual centre of all intact attached Pbs could be recorded as a percentage of the distance along the EmAb axis from the embryonic pole. The criterion for including blastocysts in this series was that their p/mj lay between 34% and 54% along this axis. The lower limit was set to avoid distortion of the distribution of Pbs due to the onset of a decline in proliferation of the mural relative to polar trophectoderm (Copp, 1978), and the upper limit to exclude blastocysts in which blastocoele formation was too early for the p/mj to be clearly defined.

73 out of 116 early blastocysts freshly recovered from the uterus had intact Pbs. In six of the 73 the location of the Pbs could not be ascertained because they became detached during removal of the zona. 59 of the 67 scorable specimens had a single Pb, and each of the remaining eight had two Pbs, which were adjacent in four cases, somewhat apart in a further three cases, and almost at opposite sides of the blastocyst in the 8th case. The distribution of all 75 Pbs in relation to the polar (P), central (C) and mural (M) thirds of the EmAb axis is shown in Fig. 4A. Since EBs are approximately spherical, each third has about the same surface area and would therefore be expected to have a similar number of Pbs, if the distribution of the latter were random (see legend to Fig. 4). In fact, while the number of Pbs in the P and M thirds is similar, it is considerably less in both of these than in the C third, the observed distribution being highly significantly different from random (Fig. 4, legend). As shown in Fig. 4B, when the frequency of Pbs is plotted at 10% intervals along the EmAb axis, it approximates a normal distribution centred on 50%.

Fig. 4.

Distribution of intact and degenerate Pbs relative to the EmAb axis of EBs/ExBs that developed in vivo versus in vitro. (A) Axial distribution of Pbs in relation to polar (P), central (C) and mural (M) thirds of the blastocyst’s surface. Assuming the blastocysts are essentially spherical, the area of the outer convex surface of each third should be the same (=2π?rh). Hence, if the distribution of Pbs was random, equal numbers would be expected to occupy each third of the surface. The observed distribution of intact Pbs is very significantly different from expected for both the blastocysts developing in vivo (??=31.9: P<<0.001) and in vitro (??=16.4: P<<0.001). Whilst the distribution of degenerate Pbs did not differ significantly from random in the in vivo series (??=5.87? P<0.05), it did in the in vitro series (?2=8.63: P<0.02). The summed distribution of intact and degenerate Pbs was also significantly non-random in both series (in vivo, ??=28.9: P<<0.001; in vitro, ?2=23.6: P<<0.001). (B,C) Plot at 10% intervals of the axial distribution of intact Pbs in blastocysts developing in vivo (B) and in vitro (C).

Fig. 4.

Distribution of intact and degenerate Pbs relative to the EmAb axis of EBs/ExBs that developed in vivo versus in vitro. (A) Axial distribution of Pbs in relation to polar (P), central (C) and mural (M) thirds of the blastocyst’s surface. Assuming the blastocysts are essentially spherical, the area of the outer convex surface of each third should be the same (=2π?rh). Hence, if the distribution of Pbs was random, equal numbers would be expected to occupy each third of the surface. The observed distribution of intact Pbs is very significantly different from expected for both the blastocysts developing in vivo (??=31.9: P<<0.001) and in vitro (??=16.4: P<<0.001). Whilst the distribution of degenerate Pbs did not differ significantly from random in the in vivo series (??=5.87? P<0.05), it did in the in vitro series (?2=8.63: P<0.02). The summed distribution of intact and degenerate Pbs was also significantly non-random in both series (in vivo, ??=28.9: P<<0.001; in vitro, ?2=23.6: P<<0.001). (B,C) Plot at 10% intervals of the axial distribution of intact Pbs in blastocysts developing in vivo (B) and in vitro (C).

To determine whether the markedly non-random distribution of Pbs depended on development in vivo, the above analysis was repeated on a series of EBs that had developed in vitro from the late 2- to early 4-cell stage. Here, of a total of 113 blastocysts examined, 67 had intact Pbs (=59%), and 60 of the latter were suitable for measuring the axial location of Pbs. Four of the remaining seven with intact Pbs were too-well cavitated, two contracted spontaneously before photography, and the last was morphologically aberrant with excluded blastomeres. The results for the 60 blastocysts, 57 with single intact Pbs and three with pairs, are shown in Fig. 4A, C. Here too, the distribution of Pbs differed significantly from random (Fig. 4, legend).

In both the in vivo and in vitro series, most of the blastocysts without intact Pbs had either lysed ones or foci of debris attached to their surface. In the in vivo series, only 10% of blastocysts had neither intact Pbs nor readily identifiable debris, the corresponding value for the in vitro series being just 6%. The axial distribution of degenerate Pbs and foci of debris, as assessed visually rather than by measurement, is also presented for both series in Fig. 4A. Whilst this was not significantly nonrandom for the in vivo blastocysts, it was for both the in vitro blastocysts alone and for the two series combined. When the summed distribution of both intact and degenerated Pbs is considered, it also shows a significant surplus in the C region relative to the M and P regions, in both the in vivo and the in vitro series (Fig. 4, legend).

Shape of early blastocysts about the EmAb axis

To obtain suitable photographs of early blastocysts for determining the axial level of their Pbs, it was necessary to orient them with their EmAb axis horizontal and the Pb(s) to one side. Frequently, this required little or no rotation of the blastocysts about this axis. Furthermore, as blastocysts were rotated from their preferred orientation, their diameter seemed to decrease, suggesting they were oval rather than round. This impression was confirmed by measuring the diameter of 37 early blastocysts, first in the orientation in which they settled naturally, and then after rotation through approximately 90°. For the greater diameter (GD), the mean was found to be 81.0 μm (s.e.m. ± 1.2), and for the lesser diameter (LD) 74.9 (s.e.m. ± 1.3) μm, a difference of 8%.When measurements were repeated on 11 of the blastocysts after they had been left standing at approx. 26°C for about 20 minutes after removal of the zona, the difference between the GD and LD was maintained.

A much larger series of EBs and ExBs, together with a smaller sample of ExdBs, were then classified according to whether their shape orthogonal to the EmAb axis was obviously oval rather than round. Scoring was done visually, either by rotating the blastocysts about their EmAb axis or immobilizing them by suction with this axis vertical (see Fig. 5A-F). In oval blastocysts with intact Pb(s), whether or not these bodies were aligned with the GD was recorded in all cases except where they were too close to one or the other pole to enable this to be decided. The data are presented in Table 1, from which it is evident that the great majority of early blastocysts are oval, the proportion among those with intact Pbs being the same as in the entire sample. Both the incidence of an oval shape and the presence of intact Pbs declines markedly with expansion of the blastocoele (Table 1). Most interestingly, however, Pbs were judged to be aligned with the GD in the great majority of oval early and more advanced blastocysts that were scorable (Table 1; Fig. 5A-C,F).

Table 1.

Frequency of oval shape and alignment of Pb(s) with the bilateral axis (= GD) according to stage of blastocyst expansion

Frequency of oval shape and alignment of Pb(s) with the bilateral axis (= GD) according to stage of blastocyst expansion
Frequency of oval shape and alignment of Pb(s) with the bilateral axis (= GD) according to stage of blastocyst expansion
Fig. 5.

(A,C) The same EB photographed in its preferred orientation (A) and following rotation through approx. 90° about its EmAb axis, with focus on the mid-plane (B) and on the upper surface (C) to show the Pb, which is clearly aligned with its greater diameter. (D,E) Another EB showing an oval shape about its EmAb axis and alignment of the Pb with the greater diameter (D) rather than the lesser diameter (E). (F) EB oriented with embryonic pole uppermost, again showing oval shape about its EmAb axis and alignment of Pb with the GD. Arrows indicate Pbs in A, C, D and F. Bar, 20 μm.

Fig. 5.

(A,C) The same EB photographed in its preferred orientation (A) and following rotation through approx. 90° about its EmAb axis, with focus on the mid-plane (B) and on the upper surface (C) to show the Pb, which is clearly aligned with its greater diameter. (D,E) Another EB showing an oval shape about its EmAb axis and alignment of the Pb with the greater diameter (D) rather than the lesser diameter (E). (F) EB oriented with embryonic pole uppermost, again showing oval shape about its EmAb axis and alignment of Pb with the GD. Arrows indicate Pbs in A, C, D and F. Bar, 20 μm.

Corresponding data for early blastocysts that had developed in vitro from the late 2-/early 4-cell stage are also presented in Table 1. While these blastocysts were oval as frequently as those that had developed in vivo, they tended to be less conspicuously so. This may account for the somewhat lower frequency with which surviving Pbs were scored as being aligned with the GD in blastocysts grown in vitro. Degenerating Pbs or debris were also aligned with the GD in the majority of both in vivo- and in vitro-developed blastocysts in which their location was recorded (see footnote to Table 1).

Clearly, early blastocysts are typically bilaterally rather than radially symmetrical about their EmAb axis, and Pbs are distributed non-randomly on their surface in relation to both this axis and that of bilateral symmetry.

Are Pbs relocated during blastulation?

Whether the restricted distribution of Pbs in EBs could be accounted for by their relocation during blastulation was investigated by injecting a small drop of DiI in corn oil into a blastomere immediately underlying the Pb in VLMs that were then cultured in vitro to the EB/ExB stage. VLMs were used to ensure that the loss of resolution of marking through division of the injected cell was minimal. Altogether, the locus of the Pb was marked in 107 VLMs with intact Pbs and a further 21 with degenerating Pbs. It is clear from the results presented in Table 2 that Pbs or their debris abutted or overlapped DiI-labelled trophectoderm in the great majority of the resulting blastocysts (Fig. 6A and B). However, the task of assessing the significance of these findings was complicated by the variability in size and shape of the patch of DiIlabelled cells in the trophectoderm which, notwithstanding marking at the VLM stage, could consist of four or more cells. Hence, the single intact Pb in each of a further series of VLMs was relocated immediately after an adjacent blastomere had been labelled by applying a blunt probe to the intact zona pellucida. The Pb was moved haphazardly, first in one direction and then, after rotating the morula, in another until its location relative to the marked blastomere, which the DiI-containing oil drop was too small to identify, could no longer be recollected. As shown in Table 2, the Pb or its putative debris was remote from the labelled trophectoderm cells in most of the resulting blastocysts. Nevertheless, like their undisturbed counterparts, relocated Pbs occurred more frequently in the C than in the P or M regions of the EmAb axis.

Table 2.

Location of Pbs relative to labelled trophectoderm in blastocysts following injection of DiI into an adjacant blastomere at VLM stage

Location of Pbs relative to labelled trophectoderm in blastocysts following injection of DiI into an adjacant blastomere at VLM stage
Location of Pbs relative to labelled trophectoderm in blastocysts following injection of DiI into an adjacant blastomere at VLM stage
Fig. 6.

(A,B) Combined brightfield/fluorescence photographs of EBs developing in vitro from VLMs in which a single cell adjacent to the Pb was labelled with DiI. In both cases the Pb has remained associated with the clonal descendants of the labelled cell throughout blastulation. (C-F) Four EBs stained for HRP activity in which the Pb overlaps or abuts the boundary between the clonal descendants of the labelled and unlabelled 1/2 blastomeres. The Pb is indicated by a large arrow in C and E, and is delimited by small arrows in D and F. Bars: 20 μm.

Fig. 6.

(A,B) Combined brightfield/fluorescence photographs of EBs developing in vitro from VLMs in which a single cell adjacent to the Pb was labelled with DiI. In both cases the Pb has remained associated with the clonal descendants of the labelled cell throughout blastulation. (C-F) Four EBs stained for HRP activity in which the Pb overlaps or abuts the boundary between the clonal descendants of the labelled and unlabelled 1/2 blastomeres. The Pb is indicated by a large arrow in C and E, and is delimited by small arrows in D and F. Bars: 20 μm.

Clearly, displacement of Pbs relative to the surface of the conceptus is exceptional during the transition from morula to blastocyst, regardless of whether these bodies remain intact throughout the process or degenerate before or during it. Hence, if Pbs are motile, their movement must normally cease before blastulation.

Initially, externally applied particles including both carboxylated or noncarboxylated latex beads were tested as possible markers of the locus of Pbs (Fleming and George, 1987), as were cortically injected oil drops (Wilson et al., 1972).

However, none proved satisfactory when the ability of two or more foci of label to retain their relative position through to blastulation was assessed. Therefore the labelling of one of the blastomeres in contact with the Pb that was used to ascertain whether Pbs move during blastulation was applied to earlier cleavage stages. If Pbs retain their ancestral position throughout cleavage they should remain at the boundary formed by the clonal descendants of the sister 2-cell blastomeres through to the EB stage, providing two conditions are upheld. One is that the first cleavage is meridional with respect to the A-V axis of the egg so that Pbs lie in its plane (see Fig. 1). The other is that coherent clonal growth obtains up to the early blastocyst stage so external blastomeres bear the same spatial relationship to each other as to the parts of the surface of the zygote from which they arise.

The results of such an analysis using HRP or DiI as labels are presented in Table 3. They show that, regardless of the stage of labelling or scoring, most surviving Pbs either abutted or overlapped the clonal boundaries (Fig. 6C-F). The lower frequency of on-boundary Pbs recorded with DiI compared to HRP may reflect the fact that at longer intervals between labelling and analysis, DiI tended not to delimit cell boundaries as clearly as HRP. Clones varied considerably in size and shape between conceptuses that were matched in stage for both labelling and scoring. Both because of this variability and that of the diameter of Pbs, it proved difficult to use any group of conceptuses presented in Table 3 to obtain an accurate estimate of the mean proportion of the surface to which Pbs that respected boundaries were thereby constrained. Particular interest attaches to the results on blastocysts labelled at the 2-cell stage because the scope for movement of Pbs is greatest in this series. By the same token, however, the proportion of the surface occupied by 1/2 clonal boundaries also tended to be the largest and hence most likely to be overlaid by Pbs by chance. Relocation experiments could not be used to assess the significance of the restricted distribution of Pbs following the labelling at the 2-cell stage because, regardless of where Pbs are moved, they invariably end up in the interblastomeric groove. Therefore, instead, the frequency of ‘on-boundary’ Pbs observed following labelling at the 2-cell stage was compared with that expected from an estimate of the proportion of the surface of the blastocyst they might occupy. The task of estimating the lengths of clonal boundaries was easier with 2-cell than with later labelling because their shape was usually less complex. For 1/2 clones, on-boundary Pbs were estimated to occupy no more than 50% of the blastocyst’s surface, as described in the footnote to Table 3. As also shown in this footnote, the observed proportion of Pbs that were onboundary was significantly greater than 50%. The results of lineage-labelling of 1/4-, 1/8-, or 1/16-blastomeres are similar to those for 1/2 blastomeres (Table 3) and would seem to be even less readily attributable to chance since, as noted above, the boundaries of later clones will occupy a progressively smaller proportion of the surface of the conceptus.

Table 3.

Clonal analysis of Pb distribution during cleavage

Clonal analysis of Pb distribution during cleavage
Clonal analysis of Pb distribution during cleavage

The data recorded in Table 3 relate only to conceptuses with intact Pbs. However, both degenerating Pbs and debris also tended to lie at or near the clonal boundaries. This was true in 3/3 cases in the HRP 2-4-cell/EB series and in 8/11 cases in the corresponding DiI series.

Further evidence of whether Pbs move during cleavage was sought by examining the spatial relationship between such bodies in conceptuses with two or three intact ones. The presence of three Pbs was assumed to be due to division of the 1st, since two of the bodies were invariably intimately attached to each other, often to an extent that made it impossible to decide if they were separate entities or a single cell arrested in cytokinesis. However, regardless of whether the 1st Pb was double or single, in no case was it found to be attached directly to the 2nd. Therefore, if either or both Pbs were able to move, they would be expected to do so independently. As can be seen from the data in Table 4, the frequency with which 1st and 2nd Pbs were separate rather than adjacent was no higher in morulae and blastocysts than at the onset of cleavage. A possible explanation for the abrupt and significant transient increase in incidence of their separation between the late 2- and early 4-cell stage is presented in the Discussion.

Table 4.

Frequency with which multiple Pbs were adjacent versus apart, according to stage of development in vivo

Frequency with which multiple Pbs were adjacent versus apart, according to stage of development in vivo
Frequency with which multiple Pbs were adjacent versus apart, according to stage of development in vivo

Attachment of Pbs to conceptuses

Providing conceptuses were exposed minimally to AT, and were pipetted carefully once dissolution of the zona was complete, Pbs remained attached to their surface thereafter in the overwhelming majority of cases. Attachment of Pbs was not an artefact of exposure to AT, since it could clearly be demonstrated by applying a blunt glass probe to the zona of untreated blastocysts, and was also evident after the zona had been removed microsurgically. In many cases where a Pb detached during the handling of denuded conceptuses, it was one of two such bodies, and was almost invariably classified as a 1st Pb when stained with Hoechst 33342.

Despite varying greatly in overall shape, Pbs tend to make fairly extensive contact with the surface of the conceptus at the zygote stage and during early cleavage. Between compaction and the EB stage they are usually more rounded, and their contact with the conceptus is typically focal. When pushed in any direction in VLMs or EBs by applying a blunt probe to the intact zona, they normally return repeatedly to their initial location. With blastocyst expansion, Pbs become progressively flattened between the trophectoderm and zona and are thus harder to identify prior to AT treatment. Thereafter, they round up to a variable extent, thereby usually revealing limited contact with the underlying trophectoderm.

The attachment of Pbs was investigated more closely using either a fine pipette or a blunt glass probe to withdraw them gently from the surface of immobilized denuded conceptuses. In EBs, Pbs could readily be rocked from side to side, as expected from their focal attachment. More interestingly, after being pulled well clear of the conceptus and then released, Pbs could return to their initial location on its surface within a few seconds. One had to be separated by more than 90 ?m before detaching rather than returning to the surface of the blastocyst. More often, however, the distance they could be withdrawn before detachment was between 15 and 40 ?m. Blastocysts could readily be suspended via their Pbs without obviously extending the attachment between them. The site of attachment was commonly to one side of the Pb rather than at the centre of its surface facing the trophectoderm.

As in EBs, the attachment sites of Pbs also behaved like very extensible, weakly elastic, ‘tethers’ in Ms and VLMs. At earlier stages, although Pbs could not normally be rocked initially, this was possible after their more extensive contact with the surface of the conceptus had been disrupted with a pair of blunt glass probes. Providing this loosening was done carefully, a focal contact between the Pb and vitellus usually persisted, which showed similar properties to that found in morulae and blastocysts.

The more extensive attachment of Pbs to earlier than later conceptuses was also evident from the relative ease with which they could be relocated by applying a blunt glass probe to the intact zona. The proportion of cases where they could not be moved without risk of damaging either them or adjacent blastomeres was highest at the late-2-/early 4-cell stage and declined progressively thereafter (Table 5).

Table 5.

Ease of relocating Pb(s) according to stage of development

Ease of relocating Pb(s) according to stage of development
Ease of relocating Pb(s) according to stage of development

Pbs that had been relocated in Ms, VLMs or in EBs were also usually found to be attached focally to the trophectoderm. However, these attachments differed from those of undisturbed Pbs in showing little or no extensibility, typically breaking as soon as the Pb was no longer in contact with the surface of the conceptus.

Pb ‘tethers’ could not be visualized clearly by light microscopy using either phase or differential interference contrast optics, or dark field illumination. Occasionally, when Pbs were held just clear of the surface of denuded conceptuses, a bridging thread was just discernible, particularly with Heine phase contrast microscopy. However, in no case was the image clear enough to yield a convincing photographic record. Therefore, further attempts were made to resolve such ‘tethers’ in VLMs and EBs by fluorescence microscopy after labelling Pbs or, occasionally, adjacent blastomeres with DiI, and also by TEM and SEM.

A very thin, cable-like structure could be seen by epi-fluorescence microscopy to extend over the trophectoderm from a minority of Pbs that had been injected with DiI in oil, particularly after the labelled Pbs had been moved slightly in conceptuses with the zona intact, or drawn just clear of the surface of those that had been denuded. Apart from often lacking a clear focal bulge, it resembled the fluorescent threads which extended between occasional nonadjacent pairs of presumed sister trophectoderm cells that became labelled following injection of DiI into a single cell. Representative examples of this structure, whose appearance was indistinguishable from that seen following labelling of the Pb with DiI in zygotes, are shown in Fig. 7A and B.

Fig. 7.

(A,B) Fluorescence microscopic images of 2nd Pbs of EBs that were labelled in situ with DiI. In both cases the label has clearly spread into a very thin process. (C) TEM image of the region between the Pb (p) and trophectoderm (t) of an EB to show part of a putative tether. (D-F). SEM images of the same EB at low (D), medium (E) and high magnification (F), showing a flattened Pb with a cable-like ‘tether’ (arrow in E) extending over the trophectoderm. (G) SEM of Pb of another EB in which the putative ‘tether’ (arrow) appears less extended. (H) Combined brightfield/fluorescence image of a VEB in which DiI injected into the Pb clearly spread to an adjacent trophectoderm cell. (1)Similar image of a VLM in which FC injected into the Pb failed to spread to the trophectoderm. Bars: 20 μm (A,B,H,I); 1 μm (C,F); 10 μm (D); 5 μm (E,G).

Fig. 7.

(A,B) Fluorescence microscopic images of 2nd Pbs of EBs that were labelled in situ with DiI. In both cases the label has clearly spread into a very thin process. (C) TEM image of the region between the Pb (p) and trophectoderm (t) of an EB to show part of a putative tether. (D-F). SEM images of the same EB at low (D), medium (E) and high magnification (F), showing a flattened Pb with a cable-like ‘tether’ (arrow in E) extending over the trophectoderm. (G) SEM of Pb of another EB in which the putative ‘tether’ (arrow) appears less extended. (H) Combined brightfield/fluorescence image of a VEB in which DiI injected into the Pb clearly spread to an adjacent trophectoderm cell. (1)Similar image of a VLM in which FC injected into the Pb failed to spread to the trophectoderm. Bars: 20 μm (A,B,H,I); 1 μm (C,F); 10 μm (D); 5 μm (E,G).

In no case was direct continuity between the Pb and the surface of the EB seen in thin sections examined by TEM. Nevertheless, images consistent with focal tethering of Pbs to the trophectoderm were obtained from several specimens (e.g. Fig. 7C). Candidate sites of focal attachment of Pbs to the trophectoderm were also identified by SEM. While most appeared to be in an unextended state (Fig. 7G), one was clearly elongated and essentially cable-like, in conformity with the fluorescent images (Fig. 7D-F).

The nature of the focal attachment between the Pb and conceptus was investigated further using FC and DiI as probes to determine whether cytoplasmic or membrane continuity persisted between them. This was prompted by finding two cases of unequivocal staining of Pbs for HRP activity following injection of the enzyme into 1/2 blastomeres in the labelling experiments described in Section 5. Conceptuses between the late zygotic and EB stage were injected intracellularly with FC in one series of experiments and with DiI in a second. In some cases, label was injected into the Pb rather than the conceptus. Care was taken to ensure that the blastomere to which each Pb was attached was labelled by injecting all possible candidates in cases of uncertainty. While both fluorescent probes routinely yielded strong labelling of Pbs following injection into zygotes, this was not the case when late 2-cell or more advanced conceptuses were injected. Unequivocal labelling of Pbs through spread of label from the later conceptus was never seen with either FC or DiI, although a single very brightly and uniformly labelled trophectoderm cell was observed in two instances following injection of DiI into the Pb of EBs (Fig. 7H and I). In blastocysts, punctate staining of one or more trophectoderm cells adjacent to Pbs was relatively common following injection of the latter with DiI but, being peculiar to specimens that were cultured before examination, was attributed to the onset of phagocytosis of the Pbs. Nevertheless, as noted earlier, at all stages DiI injection into either the conceptus or Pb often revealed a thread extending between them. Furthermore, although DiI never labelled the Pb throughout following injection into conceptus, it sometimes did so focally at its point of tethering.

Evidence of ionic coupling between the conceptus and Pb was also sought, since this is a much more sensitive way of detecting intercellular communication than dye transfer. A difficulty with this approach was that Pbs tolerated microelectrode penetration less well than normal cells, particularly from the 2-cell stage onwards. Many Pbs began to pale and show a decline in transmembrane potential immediately on penetration. Even in those in which a potential difference of up to 15-20 mV persisted thereafter, it was seldom maintained for more than 10-30 seconds. Ionic coupling between the Pb and conceptus was detected consistently prior to first cleavage, when the proportion of Pbs surviving penetration was greatest. Thereafter, there was a marked decline in rate of their survival. However, a substantial minority of Pbs that withstood penetration showed unequivocal coupling, even in VLMs and EBs (Table 6; Fig. 8).

Fig. 8.

Oscilloscope traces of ionic coupling between a 2nd Pb and zygote (A), and within an EB (B) versus between the EB and its 2nd Pb (C). In each case the upper and middle traces are from the current injection and recording microelectrodes, respectively. The lower trace shows the pulses of depolarizing current whose duration and spacing were both 1 second. Vertical bar, 10 mV.

Fig. 8.

Oscilloscope traces of ionic coupling between a 2nd Pb and zygote (A), and within an EB (B) versus between the EB and its 2nd Pb (C). In each case the upper and middle traces are from the current injection and recording microelectrodes, respectively. The lower trace shows the pulses of depolarizing current whose duration and spacing were both 1 second. Vertical bar, 10 mV.

Table 6.

Incidence of ionic coupling between conceptus and Pb at different stages of development

Incidence of ionic coupling between conceptus and Pb at different stages of development
Incidence of ionic coupling between conceptus and Pb at different stages of development

At least one Pb, typically the 2nd, remains intact until the EB stage in nearly two-thirds of conceptuses of the PO strain. Furthermore, the distribution of surviving Pbs in EBs is clearly non-random, the majority lying at or near the middle of the EmAb axis. With the discovery that EBs are bilaterally rather than radially symmetrical about this axis, the distribution of Pbs proved even more closely circumscribed, since the great majority were also aligned with the axis of bilateral symmetry (Table 1; Fig. 5). One explanation for this finding is that Pbs survive differentially according to their location; another is that they are redistributed actively or passively during blastulation. Differential survival appears untenable for two reasons. First, because the frequency of intact Pbs declines very little between the VLM and EB stage, it would have to operate before blastulation when the conceptus is devoid of obvious regional surface differentiation that might account for it. Second, degenerate Pbs and focal debris did not show the reciprocal distribution to intact Pbs that regional differences in survival would lead one to expect.

The possibility that Pbs become redistributed actively or, in the case of degenerate ones, passively, during blastulation was not supported by the results of experiments in which an adjacent blastomere was labelled at the VLM stage. Both intact and degenerate Pbs remained by or close to the labelled cells at the EB or ExB stage in the great majority of cases, whereas such juxtaposition was exceptional where Pbs had been relocated randomly immediately after blastomere labelling. Nevertheless, relocated Pbs were observed more often in the C than in P or M regions of blastocysts, suggesting that any Pbs becoming detached during cleavage might accumulate preferentially in the C region. Because conceptuses suffer considerable compression on passing through the oviduct (Nichols and Gardner, 1989), detachment is more likely in those developing in vivo than in vitro. Whether it occurs often enough to explain the higher ratio of C/P+M Pbs in the in vivo than the in vitro series of blastocysts is doubtful in view of the obvious extensibility of the attachment site of most Pbs.

In extending this analysis to see if Pbs move at any stage before blastulation, it proved necessary to resort to clonal analysis. The rationale was that if Pbs were free to move, their distribution on the surface of VLMs and EBs would not be expected to bear any consistent relationship to the clonal boundaries resulting from injection of a 1/2 blastomere or an adjacent blastomere at any later stage in cleavage. In practice, the great majority of Pbs were found to abut or overlap such clonal boundaries following labelling at all stages. Furthermore, in the 1/2 injections the frequency of on-boundary Pbs was significantly higher than expected from the estimated proportion of the surface they occupy. This is not what one would predict if the plane of first cleavage bears no consistent relationship to the A-V axis of the zygote, as Evsikov et al. (1994) claim. However, while first cleavage is unquestionably normally meridional in the PO strain, occasional marked departures from it are seen (R.L.G., unpublished observations). Hence, not all Pbs would be expected to be on-boundary even if they did not move. The possibility that Pbs move before the end of the 2-cell stage cannot be discounted by the clonal analysis since such movement would be expected to be confined to the single, deep, continuous interblastomeric groove and therefore be limited to the boundary between the two 1/2 blastomeres. However, evidence against this is provided by the difficulty experienced in relocating them in the zygote and 2-cell stage (see Table 5) and, as discussed later, by the finding that they are tethered to the surface of the conceptus.

Even if Pbs were only loosely tethered or even unattached, how freely could they move during cleavage? Although variable in size, they are typically squeezed in the perivitelline space, which presumably explains why from the 2- to early 8-cell stage they invariably lie in the grooves between blastomeres. Hence, if they did move either actively or passively their movement would be confined to these grooves. While, as noted earlier, there is scope for their complete circumferential movement at the 2-cell stage, this will be constrained progressively by the intersection of successive interblastomeric grooves until the process of compaction is completed. Thereafter, except during the relatively asynchronous mitoses of 4th and 5th cleavage, the surface of the conceptus is smooth and should therefore permit unrestricted movement of Pbs. However, labelling of an adjacent blastomere provided no evidence that Pbs normally move in compacted conceptuses, notwithstanding the greater precision with which their location could be marked later in cleavage.

That the frequency with which Pbs were apart rather than together in the minority of conceptuses with two or three such bodies was no higher at the end than at the beginning of cleavage also argues against their being motile. The roughly 90° rotation of one blastomere relative to the other, which Gulyas (1975) holds to be a general feature of second cleavage in eutherian mammals, could explain why this frequency shows a temporary but significant increase between the late 2- and early 4-cell stages, providing the plane of first cleavage regularly passes between the Pbs. Counter-rotation of the products of the rotated blastomere during subsequent cleavages would be required to account for the fact that this sudden increase in the frequency with which Pbs are apart is not sustained.

The failure to find any evidence that Pbs move during cleavage is clearly at variance with the conclusion drawn from recordings of living conceptuses by time-lapse cinemicrography. However, while such recordings demonstrate that Pbs actively change shape, proof that they actually move relative to the surface of the conceptus is lacking, as Borghese and Cassini (1963) acknowledge. Of particular relevance is the finding that Pbs do not simply adhere to the conceptus but are almost invariably attached to it focally by an extensible, weakly elastic, ‘tether’. The most obvious way that such enduring tethering of Pbs could be accomplished is through persistence of the intercellular bridge formed during their abstriction. That this bridge persists routinely to the late zygotic stage is evident from the efficiency with which FC and DiI are transferred to Pbs following their injection into the vitellus and from the data on ionic coupling (Table 6). That it can remain patent thereafter was first suggested by two cases where Pbs showed staining for HRP activity after injection of the enzyme into 1/2 blastomeres. Although transfer of fluorochromes was detected only very exceptionally beyond first cleavage, ionic coupling between conceptus and Pb was recorded sporadically even in VLMs and EBs. Whilst the absence of coupling shows that intercellular bridges are no longer patent, it does not constitute proof that they no longer exist. Failure to detect dye transfer or electrical coupling might be due to loss of the continuity of the plasma membrane, which has been seen in broken intercellular bridges (Mullins and Biesele, 1977), providing this can occur before they break. Whether or not this is so, the close similarity in both the properties and dimensions of the tether in the blastocyst and zygote is very difficult to explain other than by persistence of the same structure. This implies that the intercellular bridge between the conceptus and 2nd Pb can survive routinely for at least 3.5 days. While this may seem improbable, it is evident from the pattern of spread of HRP that intercellular bridges between blastomeres can remain patent for two or more cleavage cell cycles in the mouse (Goodall and Johnson, 1984; Goodall, 1986). Bearing in mind the earlier caveat regarding patency, the period for which they persist could be longer. Because the 2nd Pb is poorly endowed with organelles (Longo, 1987) and has only a partially replicated genome (Howlett and Bolton, 1985), the intercellular bridge linking it with the conceptus might be expected to survive even longer than those between blastomeres, particularly if disruption of such bridges requires tension between the products of cell division (Mullins and Biesele, 1977).

Among the other features of tethers were their extensibility, mechanical strength, and the difficulty with which they could be resolved by light microscopy. Candidate intercellular bridges of up to 17 ?m in length were observed by SEM in the ectoderm of the chick embryo (Bellairs and Bancroft, 1975). The fact that these could extend between cells that were separated by as many as 5-6 intervening ones argues that they can be strong as well as extensible. According to Mullins and Biesele (1977), intercellular bridges can be as little as 0.2 ?m in diameter, so that it is not surprising that they were seldom discernible by light microscopy. Unequivocal intercellular bridges that could be seen by FM extending between nonadjacent DiI-labelled trophectoderm cells or across the blastocoele from a labelled trophectoderm to an ICM cell also proved difficult to visualize by other types of light microscopy. Structures of the appropriate size connecting the Pb to the EB were observed by SEM. The only way in which tethers differed from typical intercellular bridges was that those of 2-cell or later stage conceptuses tended to lack a central mid-body (Buck, 1963). Conceivably, the intercellular bridge between the conceptus and 2nd Pb becomes a hemi-bridge through integration of the mid-body into one or other structure. Occasional integration of the mid-body into the Pb might explain why a fluorescent focus was sometimes visible on Pbs whose tether became labelled after injecting DiI into the conceptus.

Collectively, the evidence obtained in this investigation argues that the 2nd Pb normally remains connected to the surface of the conceptus at its site of production throughout, and even beyond, the period for which it remains intact. It can therefore serve as an enduring marker for the animal pole of the zygote. Hence, the implication of the strikingly nonrandom distribution of Pbs in Ebs is that the axes of the blastocyst and zygote are related. Specifically, the axis of bilateral symmetry of the blastocyst is normally coincident with, and its EmAb axis orthogonal to, the A-V axis of the zygote (Fig. 9). This finding, together with others, suggests that organization of the egg may play a role in patterning of the conceptus during normal development in the mouse. Thus, evidence has been provided that the mouse conceptus is bilaterally rather than radially symmetrical from the late blastocyst stage (Smith, 1980, 1985; Gardner, 1990), and that the antero-posterior axis of the definitive embryo and conceptus share a common orientation (Gardner et al., 1992). The present investigation has revealed that the EB is also bilaterally symmetrical even when it has developed in vitro from the late 2-cell stage. What has yet to be established is whether the bilateral axis of the EB is conserved through to gastrulation. Finally, it is notable that no regularity was discernible in the way clones resulting from the vital labelling of 1/2 blastomeres mapped onto the EmAb axis of the blastocyst. Hence, if information specifiying this axis also resides in the zygote, the plane of the meridional first cleavage evidently bears no fixed relationship to it (Fig. 9).

Fig. 9.

Diagrams illustrating how, according to present findings, the axes of the EB (thick outline) map onto the structure of the zygote (thin outline). (A) View from the embryonic pole of blastocyst showing alignment of its axis of bilateral symmetry (greater diameter), which is not discernibly polarized, with the AV axis of the zygote (thick line with arrowhead at animal pole). (B) View from the animal pole of the zygote showing how the EmAb axis of the blastocyst (continuous thick line) is not only orthogonal to its AV axis but, evidently, bears no fixed relationship to the plane of 1st cleavage (dashed thin line).

Fig. 9.

Diagrams illustrating how, according to present findings, the axes of the EB (thick outline) map onto the structure of the zygote (thin outline). (A) View from the embryonic pole of blastocyst showing alignment of its axis of bilateral symmetry (greater diameter), which is not discernibly polarized, with the AV axis of the zygote (thick line with arrowhead at animal pole). (B) View from the animal pole of the zygote showing how the EmAb axis of the blastocyst (continuous thick line) is not only orthogonal to its AV axis but, evidently, bears no fixed relationship to the plane of 1st cleavage (dashed thin line).

I thank Barbara Luke for her expert assistance with electron microscopy, and Frances Brook, Tim Davies, Wendy Gardner and Jo Williamson for help in preparing the manuscript. The work was supported by the Royal Society and the ICRF.

Balinsky
,
B. I.
(
1970
).
An Introduction to Embryology
. 3rd edn.
Philadelphia
:
W. B. Saunders Company
.
Beddington
,
R. S. P.
and
Lawson
,
K.
(
1990
).
Clonal analysis of cell lineages
.
In Postimplantation Mammalian Embryos, A Practical Approach
(ed.
A. J.
Copp
and
D. L.
Cockroft
), pp.
267
292
.
Oxford
:
IRL Press
.
Bellairs
,
R.
and
Bancroft
,
M.
(
1975
).
Mid-bodies and beaded threads
.
Am J. Anat
.
143
,
393
398
.
Borghese
,
E.
and
Cassini
,
A.
(
1963
).
Cleavage of the mouse egg
.
In Cinemicrography in Cell Biology
(ed.
G. G.
Rose
), pp.
263
277
.
New York
:
Academic Press
.
Buck
,
R. C.
(
1963
).
The central spindle and the cleavage furrow
.
In The Cell in Mitosis
(ed.
L.
Levine
), pp.
55
63
.
New York
:
Academic Press
.
Calarco
,
P. G.
(
1995
).
Polarization of mitochondria in the unfertilized mouse oocyte
.
Dev. Genet
.
16
,
36
43
.
Cameron
,
R. A.
,
Hough-Evans
,
B. R.
,
Britten
,
R. J.
and
Davidson
,
E. H.
(
1987
).
Lineage and fate of each blastomere of the eight-cell sea urchin embryo
.
Genes Dev
.
1
,
75
84
.
Champlin
,
A. K.
,
Dorr
,
D. L.
and
Gates
,
A. H.
(
1973
).
Determining the stage of the estrous cycle in the mouse by the appearance of the vagina
.
Biol. Reprod
.
8
,
491
494
.
Copp
.
A. J
. (
1978
).
Interaction between inner cell mass and trophectoderm of the mouse blastocyst. I. A study of cellular proliferation
.
J. Embryol. Exp. Morph
.
48
,
109
125
.
Dalcq
,
A. M.
(
1957
).
Introduction to General Embryology
.
Oxford
:
University Press
.
Davidson
,
E. H.
(
1986
).
Gene Activity in Early Development
. 3rd edn.
Orlando
:
Academic Press
.
Evsikov
,
S. V.
,
Morozova
,
L. M.
and
Solomko
,
A. P.
(
1994
).
Role of ooplasmic segregation in mammalian development
.
Roux’s Arch. Dev. Biol
.
203
,
199
204
.
Fleming
,
T. P.
and
George
,
M. A.
(
1987
).
Fluorescent latex microparticles: a non-invasive short-term cell lineage marker suitable for use in the mouse early embryo
.
Roux’s Arch. Dev. Biol
.
196
,
1
11
.
Gardner
,
D. K.
and
Sakkas
,
D.
(
1993
).
Mouse embryo cleavage, metabolism and viability: role of medium composition
.
Hum. Reprod
.
8
,
288
295
.
Gardner
,
R. L.
(
1978
).
Production of chimeras by injecting cells or tissue into the blastocyst
.
In Methods in Mammalian Reproduction
(ed.
J. C.
Daniel
), pp.
137
165
.
New York
:
Academic Press
.
Gardner
,
R. L.
(
1990
).
Location and orientation of implantation
.
In Establishing a Successful Human Pregnancy, Serono Symposia Publications
, vol.
66
(ed.
R. G.
Edwards
), pp.
225
238
.
New York
:
Raven Press
.
Gardner
,
R. L.
(
1996
).
Can developmentally significant spatial patterning of the egg be discounted in mammals?
Hum. Reprod. Update
2
,
3
27
.
Gardner
.
R. L.
,
Meredith
,
M. R.
and
Altman
,
D. G.
(
1992
).
Is the anterior-posterior axis of the fetus specified before implantation in the mouse?
J. Exp. Zool
.
264
,
437
443
.
Goodall
,
H.
(
1986
).
Manipulation of gap junctional communication during compaction of the mouse early embryo
.
J. Embryol. Exp. Morph
.
91
,
283
296
.
Goodall
,
H.
and
Johnson
,
M. H.
(
1984
).
The nature of intercellular coupling within the preimplantation mouse embryo
.
J. Embryol. Exp. Morph
.
79
,
53
76
.
Goodall
,
H.
and
Maro
,
B.
(
1986
).
Major loss of junctional coupling during mitosis in early mouse embryos
.
J. Cell Biol
.
102
,
568
574
.
Gulyas
,
B.
(
1975
).
A reexamination of cleavage patterns in eutherian mammalian eggs: rotation of blastomere pairs during second cleavage in the rabbit
.
J. Exp. Zool
.
193
,
235
248
.
Howlett
,
S. K.
and
Bolton
,
V. N.
(
1985
).
Sequence and regulation of morphological and molecular events during the first cycle of mouse embryogenesis
.
J. Embryol. Exp. Morph
.
87
,
175
206
.
Jones-Seaton
,
A.
(
1950
).
Etude de l’organisation cytoplasmic de l’oeuf de Rongeurs principalement quant a la basophile ribonucleique
.
Arch. Biol
.
61
,
291
444
.
Karnovsky
,
M. J.
(
1965
).
A formaldehyde-glutaraldehyde fixative of high osmolarity for use in electron microscopy
.
J. Cell Biol
.
27
,
137A
.
Lewis
,
W. H.
and
Wright
,
E. S.
(
1935
).
On the early development of the mouse egg
.
Cont. Embry. Carnegie Inst. Wash
.
25
,
113
143
.
Longo
,
F.
(
1987
).
Fertilization
.
London
:
Chapman and Hall
.
McCain
,
E. R.
and
McClay
,
D. R.
(
1994
).
The establishment of bilateral symmetry in sea urchin embryos
.
Development
120
,
395
404
.
Mullins
,
J. M.
and
Biesele
,
J. J.
(
1977
).
Terminal phase of cytokinesis in D-985 cells
.
J. Cell Biol
.
73
,
672
684
.
Nichols
,
J.
and
Gardner
,
R. L.
(
1989
).
Effect of damage to the zona pellucida on development of preimplantation embryos in the mouse
.
Hum. Reprod
.
4
,
180
187
.
Pratt
,
H. M. P.
(
1987
).
Isolation, culture and manipulation of pre-implantation mouse embryos
.
In Mammalian Development: A Practical Approach
(ed.
M.
Monk
), pp.
13
42
.
Oxford
:
IRL Press
.
Smith
,
L. J.
(
1980
).
Embryonic axis orientation in the mouse and its correlation with blastocyst relationships to the uterus. I. Relationships between 82 hours and 4 1/4 days
.
J. Embryol. Exp. Morph
.
55
,
257
277
.
Smith
,
L. J.
(
1985
).
Embryonic axis orientation in the mouse and its correlation with blastocyst relationships to the uterus. II. Relationships from 4 1/4 to 9 1/2 days
.
J. Embryol. Exp. Morph
.
89
,
15
35
.
Tarkowski
,
A. K.
and
Wroblewska
,
J.
(
1967
).
Development of blastomeres of mouse eggs isolated at the 4- and 8-cell stage
.
J. Embryol. Exp. Morph
.
18
,
155
180
.
Warner
,
A. E.
and
Bate
,
C. M.
(
1987
).
Techniques for dye injection and cell labelling
.
In Microelectrode Techniques: The Plymouth Workshop Handbook
(ed.
N. B.
Standen
,
P. T. A.
Gray
and
M. J.
Whitaker
), pp.
169
185
.
The Company of Biologists
,
Cambridge
.
Wilson
,
E. B.
(
1928
).
The Cell in Development and Heredity
, 3rd edn.
New York
:
Macmillan
.
Wilson
,
I. B.
,
Bolton
,
E.
and
Cuttler
,
R. H.
(
1972
).
Preimplantation differentiation in the mouse egg as revealed by microinjection of vital markers
.
J. Embryol. Exp. Morph
.
27
,
467
479
.
Wolstenholme
,
G. E. W.
and
O’Connor
,
M.
(eds) (
1965
).
Preimplantation Stages of Pregnancy: A Ciba Foundation Symposium
.
London
:
Churchill
.
Zamboni
,
L.
(
1970
).
Ultrastructure of mammalian oocytes and ova
.
Biol. Reprod. Suppl
.
2
,
44
63
.

It has been reported recently that the expression of Brachyury during gastrulation and A-P patterning of the fetus thereafter are perturbed following brief exposure of 2- or 8-cell mouse conceptuses to lithium (Rogers, I. and Varmuza, S. (1996). Epigenetic alterations brought about by lithium treatment disrupt mouse embryo development. Molecular Reproduction and Development 45, 163-170). Lithium is presumed by the authors to cause a heritable alteration in the competence of cells to respond to positional cues that are established later in development. However, an alternative possibility raised by the present investigation is that it directly affects a spatial patterning system which is already present during cleavage.