Control of gap junction formation in early mouse embryos

DEVELOPMENTAL

BIOLOGY

98,155-

164 (1983)

Control of Gap Junction Formation JEANNE Cell Science

Laborato?$es,

R.

Received

STANLEY

MCLACHLIN,

Department

of Zoology, November

The

CAVENEY,

University

10, 1982; accepted

in Early Mouse Embryos AND

of Western in revised

fmm

GERALD Ontario, February

M. KIDDER

London,

Ontario,

N6A

5B7 Canada

18, 1983

Intercellular communication via gap junctions begins in the eight-cell stage in early mouse embryos. We have studied the timing of this event in relation to compaction, and have begun to explore some of the possible control mechanisms underlying it. Gap junction formation was inferred by measuring ionic coupling as well as by observing the intercellular transfer of fluorescent dye. Embryos were obtained early on Day 3 of pregnancy by flushing the oviducts of HA/ICR mice that had been mated with CBGF,/J males. Gap junctions were detected only in those embryos which had achieved the fully compacted state. Inhibition of protein synthesis by cycloheximide treatment beginning as early as the late four-cell stage failed to block compaction or the acquisition of gap junctions, demonstrating that the necessary proteinaceous components are present in advance of these events. In order to test the possibility that gap junctions could be induced to form prematurely, fully compacted, communication-competent eight-cell embryos were aggregated with two- or four-cell embryos. Even after 10 hr of aggregation, no interembryonic gap junctions could be detected. Fully compacted eight-cell embryos when aggregated with each other, however, became ionically coupled within 3-5 hr. The number of interembryonic junctional channels was judged to be effectively small, since the aggregated embryos exhibited obvious ionic coupling but very weak dye coupling. In contrast to gap junction formation within embryos, junction formation between embryos was blocked by cycloheximide. These results demonstrate that gap junction formation in early mouse embryos is under precise temporal control, involving the assembly or mobilisation of preexisting components. This stockpile of components is either unavailable or insufficient to allow the formation of additional gap junctions between aggregated communication-competent embryos without new protein synthesis. INTRODUCTION

The existence of channels which mediate direct intercellular transfer of ions and molecules has been well documented for a wide range of cell types and tissues (reviewed by Loewenstein, 1979). These intercellular membrane channels or gap junctions have been described in the embryos of Triturus, Fund&us, Xenopus, the axolotl, rabbits, and mice (Ito and Loewenstein, 1969; Bennett and Trinkaus, 1970; Sheridan, 1971; DiCaprio et aL, 1974; Hanna et ab, 1980; Ducibella et CAL,1975). Gap junctions are likely to play a role in the coordination of early development, although the exact nature of this role is yet to be determined. The mammalian oocyte and surrounding follicle cells communicate through gap junctions which have been attributed with the role of metabolic cooperativity during oogenesis (Heller and Schultz, 1980; Heller et aL, 1981; Brower and Schultz, 1982). These heterologous junctions disappear just prior to ovulation (Albertini and Anderson, 1974; Anderson and Albertini, 1976; Gilula et aZ., 1978; Moor et aZ.,1981; Dekel et aL, 1981). Early cleavage mouse embryo blastomeres lack communicating junctions but possess large intercellular channels which are capable of passing the tracer, horseradish peroxidase (HRP, M, = 40,000), a molecule too large to cross through gap junctions (Lo and Gilula, 1979a). These cytoplasmic bridges are areas of continuity that persist between sister blastomeres of two- and four-cell mouse

embryos until cytokinesis is completed late in the cell cycle. At this time the blastomeres lose their ability to pass HRP and low-molecular-weight fluorescent dyes, and do not show any evidence of ionic coupling (Lo and Gilula, 1979a). It is not until the eight-cell stage, in association with compaction, that ultrastructural evidence indicates the reappearance of gap junctions (Calarco and Brown, 1969; Ducibella et aL, 1975; Magnuson et ah, 1977). The presence of functional gap junctions at this stage has also been established by the demonstration of ionic coupling and transfer of injected fluorescein (Lo and Gilula, 1979a). More recently, Goodall and Johnson (1982) monitored the passage of fluorescent dye between mouse blastomeres aggregated together from various stages of preimplantation development; their results indicate that the ability to form gap junctions is acquired in the early eight-cell stage and does not require the attainment of the fully compacted state. The stage-specific appearance of gap junctions in the early mouse embryo raises interesting questions about the control mechanisms involved. One possibility is that gap junction components are preformed in the early embryo but are not assembled until a specific chronological age or developmental stage has been reached. This hypothesis would predict that premature assembly of gap junctions will not occur, even if preformed components are present. We have tested this hypothesis by using ionic coupling measurements to look for premature formation of gap junctions in two- and four-cell

155 0012-1606/83 Copyright All rights

$3.00

0 1983 by Academic Press, Inc. of reproduction in any form reserved.

156

DEVELOPMENTAL

BIOLOGY

mouse embryos aggregated with communication-competent eight-cell embryos. No evidence for such early communication was found, despite evidence from inhibition of protein synthesis suggesting the presence of preformed gap junctional components in the four-cell stage of development. A preliminary report of part of this work appeared in abstract form (McLachlin and Kidder, 1981). MATERIALS

AND

1. Collection and Preparation

METHODS

of Embryos

Random-bred HA/ICR female mice (West Seneca Labs, Buffalo, N. Y.) were superovulated by an injection of 4-5 IU pregnant mare’s serum gonadotropin (PMSG, Sigma), followed by a second injection 48 hr later with 5 IU human chorionic gonadotropin (hCG, Sigma). The mice were then paired overnight with CB6/FIJ male mice (Jackson Labs, Bar Harbor, Maine). Successful mating was determined by the presence of a vaginal plug (Day 1) and developmental age was estimated from the time of the hCG injection. Embryos were flushed from the oviducts early on Day 3 of development with flushing Medium I (Spindle, 1980) and cultured at 3’7°C in standard egg culture medium (SECM) (Biggers et al, 1971) as modified by Spindle (1980). The incubator contained an atmosphere of 5% COB in air. In preparation for making embryo aggregates, the zonae pellucidae were removed from small groups of embryos by a brief incubation in acid Tyrode’s solution (pH 2.5). The embryos were then returned to SECM, pushed together, and left undisturbed for a few minutes to facilitate maintenance of close contact before returning the culture dish to the incubator. In successful aggregations, adjacent embryos adhere to one another and establish firm cell-to-cell contact which can withstand vigorous pipetting after less than an hour. 2. Electrophysiology

Ionic coupling was monitored by passing an electrical current through a glass microelectrode inserted in one blastomere while a second electrode in a different blastomere was used to record any change in resting potential. Microelectrodes were made from capillary tubing (W-P Instruments, No. 1BlOOF) using a vertical microelectrode puller (Narashige). Electrodes were filled with 3 M KC1 and had tip resistances of 20-40 MQ in the culture medium (flushing Medium I). All ionic COUpling measurements were carried out at room temperature (25°C) on embryos in this medium. The bathing solution was grounded through a KCl-agar bridge. Microelectrodes were attached to DeFonbrune pneumatic micromanipulators (Beaudouin). The electrotonic potentials and membrane potentials were recorded on a Brush 220 chart recorder. For dye injections, electrode tips were filled with either 50 mM 6-carboxyfluorescein

VOLUME 98.1983

and the remainder of the electrode filled with 3 M KCI, or a solution of 3% Lucifer yellow CH in 0.1 M LiCl. Dye was transmitted into cells iontophoretically with 200 nsec current pulses of 6 X lo-* A (1 pulse per set). The spread of fluorescent dye between blastomeres was observed through a Zeiss IM35 inverted microscope fitted with phase-epifluorescence optics, with narrow-band FITC excitation filter BP 485/20 and barrier filter BP 520-560 in the light path. Our most recent dye injection experiments using Lucifer yellow were carried out with video equipment consisting of an RCA silicon intensifier target (SIT) television camera (Model TC1030 M) coupled through the tine port to the inverted microscope. The video signal from the camera passed through an RCA Video TimeDate Generator (Model TC 1440B) and the combined signals were recorded on a Sony (VO-1600) 3/4-in. videocassette recorder. The video image was later displayed on an RCA 14-in. b/w monitor (TC1214) and photographed. 3. Measurement of Uptake and Incorporation

of

Day 3 embryos were prepared for culture as described. The embryos were first preincubated in SECM containing either 10 or 50 pg/ml cycloheximide (CHX, Sigma) for 2 hr (66-68 hr post-hCG), and then they were transferred to microdrops of the same medium containing 0.2 pML-[35S]methionine (1226.5 Ci/mmole, New England Nuclear, Boston, Mass.) and cultured under paraffin oil (Fisher) for a further 2 hr (68-70 hr posthCG). Following the labelling period embryos were washed through ten 200-~1 drops of phosphate-buffered saline (PBS) containing 0.3% polyvinylpyrrolidone and stored frozen in 10 ~1 PBS. The embryos were thawed and lysed with 200 ~10.5% sodium dodecyl sulfate (SDS). Total methionine uptake was measured by spotting 5~1 samples of the lysate directly onto damp glass fiber filters (Whatman GF/A). The remainder of the lysate was split into two aliquots of 100 ~1 and to each was added 100 pg BSA and 10 ~1 of cold 10 mML-methionine. This preparation was precipitated with 10% trichloroacetic acid (TCA) on ice for at least 4 hr before collecting the acid-insoluble material on a glass fiber filter (Whatman GF/F). Filters were dried thoroughly and then counted in Omnifluor (New England Nuclear). RESULTS

1. Ionic Coupling Is Established in Eight-Cell Compacted Embryos

We began our experiments by verifying the findings of Lo and Gilula (1979a) concerning the acquisition of low-resistance intercellular channels by early mouse embryos. Because these investigators had shown that

MCL~ACHLIN,

CAVENEY,

AND

KIDDER

Gap Junctions

in Early

157

Mouse Embryos

50 mV

50 mV

b

Time

bed?

0 200

0 40I

6,O

t

50 mV

50 mV

d

Time

(set)?

1

20

I

40

FIG. 1. Position of microelectrodes in eight-cell mouse embryos with corresponding chart recordings indicating the presence or absence of ionic coupling. (a) Partially compacted eight-cell embryo and (b) chart recording showing that such embryos do not exhibit ionic coupling. As current is pulsed into one microelectrode there is no corresponding voltage deflection detected by the recording electrode. Arrows indicate membrane potential present at the time of insertion into a single blastomere. The encircled arrow indicates removal of the microelectrode. (c) Fully compacted eight-cell embryo. (d) Chart recording showing the presence of ionic coupling in this embryo. As a current is pulsed into one microelectrode a voltage deflection is detected in the recording microelectrode indicated by the areas designated, c. (a, c) magnification

X880.

sister blastomeres remain connected by cytoplasmic bridges for some time after cytokinesis has occurred, care was taken to avoid this pitfall by inserting the microelectrodes into blastomeres on opposite sides of

the embryo when testing coupling within a single embryo. Whenever possible, successive recordings were taken from two different blastomeres within the same embryo while pulsing into a single blastomere. All eight-

158

DEVELOPMENTAL

BIOLOGY

TABLE 1 EFFECT OF CYCLOHEXIMIDE (CHX) ON INCORPORATION OF [~~SJMETHIONINE IN EARLY CLEAVAGE MOUSE EMBRYOS Uptake experimental/ uptake control

Medium SECM

SECM

+ 10 pg/ml

+ 50 pg/ml

CHX

CHX

Incorporation experimental/ incorporation control

0.80 1.15 1.11

0.046 0.057

0.70 1.00 1.15

0.021 0.041

0.047

0.036

Note. Mouse embryos at the eight-cell stage (Day 3) were preincubated in standard egg culture medium (SECM) containing 10 or 50 gg/ml CHX for 2 hr (66-68 hr post-hCG) followed by a 2 hr labelling period (68-70 hr post-hCG) in 0.2 PM L-pS]methionine. Forty to sixty embryos were incubated in 40-~1 culture drops under oil. Total uptake and total radioactivity incorporated per embryo (1600-2500 cpm/embryo in controls) were determined as described under Materials and Methods. The data represent the results from three separate experiments, each done in duplicate. Because uptake, and hence incorporation, of pS]methionine varied from one experiment to another, the data for each CHX-treated batch are normalized to the respective control batch.

cell compacted embryos tested were found to be ionitally coupled, whereas coupling was not detected between blastomeres of four-cell and early (not fully com-

FIG. 2. Groups of preimplantation (b) Embryos cultured in 50 gg/ml

VOLUME

98, 1983

pacted) eight-cell embryos (Fig. 1). We conclude, in agreement with Lo and Gilula (1979a), that mouse embryo blastomeres become ionically coupled via gap junctions in the eight-cell stage, about 72 hr post-hCG. 2. Gap Junction Formation Does Not Require Concomitant Protein Synthesis Embryos were incubated in cycloheximide (CHX) for various lengths of time prior to the onset of compaction. An effective concentration of inhibitor for early cleavage mouse embryos was determined by examining the effect of 10 and 50 pg/ml CHX on the incorporation of [%]methionine over a period of 2 hr. Both concentrations were found to be effective in substantially reducing methionine incorporation into protein, the higher concentration causing greater than 97% inhibition (Table 1). Embryos were then incubated in 50 pg/ml CHX beginning from 62 to 68 hr post-hCG, and tested for low-resistance intercellular channels. The proportion of embryos reaching the eight-cell stage was about twothirds that of controls and there was no obvious morphological difference between CHX-treated and control eight-cell embryos (Fig. 2), although longer treatments (i.e., those beginning earlier than 62 hr post-hCG) caused most of the embryos to arrest. Inhibition of protein synthesis with cycloheximide for a total period of up to 17 hr beginning from 10 to 4 hr prior to compaction did

mouse embryos at 76 hr post-hCG. (a) Control CHX beginning at 66 hr post-hCG. Magnification

embryos X380.

cultured

in vitro

from

62 hr post-hCG.

MCLUHLIN,

EFFECT

CAVENEY,

AND

Gap Junctions

KIDDER

TABLE 2 OF CYCLOHEXIMIDE ON THE FORMATION OF LOW-RESISTANCE CHANNELS IN COMPACTED EIGHT-CELL MOUSE EMBRYOS

Hours in CHX

Time into CHX (hr post-hCG)I

8 10 12 14 14 17

68 66 68 62 64 62

No. tested

Note. All embryos were incubated in 50 rg/ml CHX and tested the presence of ionic coupling between 76 and 80 hr post-hCG.

Early

159

Mouse Embryos

TABLE 3 TEST FOR LOW-RESISTANCE JUNCTIONAL CHANNELS BETWEEN AGGREGATED EIGHT-CELL MOUSE EMBRYOS

No. coupled

5 5 7 6 7 2

in

5 5 7 6 6 2

Time (hr after

of testing aggregation) 1.5 2 3 5 8 12 14

Time of aggregation (hr post-hCG) 76.5 78 76 74 82 81 78

No. tested 7 8 9 6 10 5 5

No. coupled 0 2 5 6 9 5 5

for

not block compaction nor prevent the establishment of low-resistance junctions in embryos reaching the eightcell stage (Table 2). It would thus appear that gap junction formation in the eight-cell stage depends upon protein synthesis occurring much earlier. Since our longest cycloheximide treatments began prior to the third cleavage, the required protein synthesis could occur in the four-cell stage or even earlier. 3. The Formation of Gap ,Junctions Cannot Be Induced Prematurely

Having demonstrated that gap junction components are formed in advance of the time they are needed, we

sought to test whether functional gap junctions could be induced prematurely by aggregating pre- and postcompaction embryos. We began by demonstrating that aggregated, communication-competent embryos are capable of establishing low-resistance junctions between them. The zonae pellucidae were removed and eight-cell compacted embryos were pushed together. If left undisturbed for a short while cell-cell contact became firmly established and there was progressive adhesion of the cell membranes over a number of hours (Fig. 3). These aggregates were tested for the presence of gap junctions by measuring ionic coupling between the two embryos. Virtually all embryo pairs tested more than 5 hr after aggregation exhibited interembryonic low-

FIG. 3. Formation of compacted eight-cell embryo aggregates. (a) Embryos aggregated for 1.5 hr; there was no evidence of coupling embryos. (b) Embryos aggregated for 3 hr; no evidence of coupling. (c) Embryos aggregated for 5 hr; ionic coupling was detected the two embryos. In a and c, embryos have begun to decompact after several minutes’ pulsing from the microelectrodes. Magnification

between between X960.

160

DEVELOPMENTAL

BIOLOGY

FIG. 4. Test for ionic coupling in aggregates of embryos from different developmental stages. (a) Two-cell/compacted eight-cell pair aggregated for 5 hr prior to testing for ionic coupling; no coupling was detected. The eight-cell embryo of the pair had decompacted during testing. (b) Four-cell/compacted eight-cell pair aggregated for 5 hr prior to testing for ionic coupling; no coupling was detected. Magnification X960.

resistance channels (Table 3). If tested between 1.5 and 2 hr after aggregation, however, only a few of them were ionically coupled, indicating that junctional channels were just becoming established. We therefore took 5 hr as the minimum time necessary for gap junctions to be established between aggregated communicationcompetent embryos. Aggregates were then made from two-cell and four-cell embryos paired with compacted eight-cell embryos (Fig. 4). All pairs were aggregated for at least 5 hr and some for as long as 10 hr in the four-cell/eight-cell pairs. None of the pairs had established low-resistance junctions up to the time of testing (Table 4). We conclude that close contact with the surface of a communication-competent embryo is an insufficient stimulus to promote premature gap junction formation by two- or four-cell mouse embryos, despite the fact that protein constituents of gap junctions are probably present in the four-cell stage.

VOLUME

98, 1983

Some of the experiments just described were also performed using transfer of carboxyfluorescein (Mr = 376) as an indicator of cell coupling, and the results were somewhat different. Aggregates were made by pairing either precompacted eight-cell embryos or compacted eight-cell embryos. After 6-8 hr the dye was injected into a single hlastomere of one of the embryos in each aggregate. In embryo pairs that had not undergone compaction (Fig. 5), injected dye passed only to one other blastomere within the same embryo, presumably its sister blastomere (Lo and Gilula, 1979a). The dye was never observed to spread from one embryo to the other. This is consistent with our observation that aggregates of precompacted eight-cell embryos were never ionically coupled. Surprisingly, however, we could not detect transfer of carboxyfluorescein between aggregated postcompaction embryos either (Fig. 6), despite the fact that the dye spread to other blastomeres within the injected embryo. This same result was reported by Goodall and Johnson (1982); we take it to indicate that the number of gap junctions formed between compacted embryos is too small to allow a detectable amount of carboxyfluorescein to pass between them. As a test of this interpretation we repeated the experiment using Lucifer yellow (M, = 457), which in our experience remains detectable much longer after injection into cells. Aggregates of compacted g-cell embryos were allowed to remain in contact for 8 hr prior to dye injection into a single blastomere (Figs. ?‘a, b); within 17 min the dye had spread throughout the injected embryo (Fig. 7~). One hour later (Fig. 7d) the dye could be seen to have filled the second embryo. This result confirms the fact that the aggregated embryos were in fact coupled by junctional channels large enough to permit the passage of small tracer molecules. -G. The Formation of Gap Junctions between Aggregated Embryos Does Require Concomitant Protein Synthesis Having established that junctions within an embryo can form in the absence of protein synthesis we wanted TABLE 4 TEST FOR LOW-RESISTANCE EMBRYOS AGGREGATED

Stage 2c 3c 4c 4c 4c

+ + + + +

8c 8c 8c 8c 8c

JUNCTIONAL CHANNELS BETWEEN MOUSE AT DIFFERENT DEVELOPMENT STAGES

Time of testing (hr after aggregation)

Time of aggregation’ (hr post-hCG)

No. tested

No. coupled

5 5 5 8 10

49 68 68 69 69

3 1 5 4 3

0 0 0 0 0

a The time of aggregation refers to the age of the younger embryo in each pair; all eight-cell embryos were compacted.

MCL.ACHLIN, CAVENEY, AND KIDDER

Gap

Junctions

in

Early

Mouse Embryos

161

to determine whether the same is true of junctions formed between embryos. If existing junctional components are depleted from externally facing surfaces, then interembryonic junction formation in the eightcell stage might require the synthesis of additional new junctional proteins. Compacted eight-cell embryos were cultured in 50 pg/ml CHX for 3 hr prior to aggregation. They were then paired in the presence of CHX and incubated for a further 8 hr prior to testing. Cycloheximide treatment did not interfere in any way with embryo aggregation. Only about half of the aggregates treated in this way were found to have formed interembryonic low-resistance junctions (Table 5), although ionic coupling was detected within the individual embryos of each pair. When embryos were put into the

FIG. 6. Test for dye coupling (carboxyfluorescein) between compacted &cell embryos aggregated for 8 hours. (a) Bright-field image of embryo aggregate. (b) Fluorescent image of the same embryo aggregate 8 min after injection of carboxyfluorescein into a single blastomere of the lower embryo. The dye has spread to all the blastomeres of the injected embryo but cannot be detected in the upper embryo of the pair. Magnification X969.

same concentration of cycloheximide starting 6 hr prior to aggregation, there was no evidence of coupling between any of the embryo pairs even though coupling was always present within individual members of the pairs (Table 5). Thus there does appear to be a need for new protein synthesis in the establishment of interembryonic gap junctions. FIG. 5. Test for dye coupling between precompacted eight-cell embryos aggregated for 6 hr. (a) Bright field image of embryo aggregate. (b) Fluorescent image of same embryo aggregate 14 min after injection of carboxyfluorescein into a single blastomere of the lower embryo. After this length of time only a small amount of dye has passed into a second blastomere within that embryo. Magnification X990.

DISCUSSION

In this investigation of the control of gap junction formation in early mouse embryos, we have confirmed Lo and Gilula’s (19’79a) finding that blastomeres become ionically coupled coincident with compaction. In all

162

DEVELOPMENTALBIOLOGY

VOLUME 98,1983

FIG. 7. Test for dye coupling (Lucifer yellow) between compacted eight-cell embryos aggregated for 8 hr. (a) Phase-contrast image of embryo aggregate. The microelectrode containing the dye has impaled an outer blastomere of the upper embryo. (b) Fluorescent image of the same embryo at the start of injection of Lucifer yellow. (c) The dye has spread to all blastomeres of the injected embryo 17 min later. (d) The dye has spread into and equilibrated within the lower embryo of the pair one hour and 1’7 min after the initial injection, although the fluorescence intensity is much lower than that in the injected embryo. (a, b) magnification X960; (c, d) magnification X1120.

eight-cell embryos tested, whether single or in aggregates, electrical coupling was never detected if compaction was not fully completed. We confirmed this result by observing the distribution of injected fluorescent dye: the carboxyfluorescein passed only to one other blastomere in precompacted single eight-cell embryos

and in eight-cell precompacted pairs. Rapid transfer to a second blastomere is presumably via the large intercellular channel that persists between sister blastomeres. In contrast to these results, Goodall and Johnson (1982) reported dye transfer between aggregated precompacted eight-cell mouse embryos of a different strain,

MCLACHLIN, CAVENEY, AND KIDDER

Gap

Junctions

in

Early

Mouse

TABLE 5 EFFEGT OF CYCLOHEXIMIDE (CHX) ON THE FORMATION OF LOW-RESISTANCE JUNGTIONAL BETWEEN AGGREGATED EIGHT-CELL COMPACTED EMBRYOS Time into CHX (hr post-hCG) 68 68 68 66

Time of aggregation” (hr post-hCG) 71 71 71 72

Time of testing (hr after aggregation) 8 8 8 10

163

Embryos

CHANNELS

No. coupled

No. not coupled

1 5 6 0

3 2 9 8

Undecided” 0 1 1 0

Note. All control embryo pairs tested 8 hr after aggregation were coupled. A total of 20 aggregates from four separate experiments was tested. Experimental groups were cultured in 50 pg/ml CHX. a Aggregates were scored as undecided in cases where no coupling was detected between embryos, but where it was not possible for technical reasons to measure coupling within individual embryos of the same aggregate.

and concluded that junction formation occurs within 6 hr after division to the eight-cell stage, and prior to the initiation of compaction. The discrepancy between our two sets of results may reflect a strain difference in the timing of junction formation in relation to cleavage and/ or compaction. Even in our embryos, compaction does not appear to be a prerequisite for gap junctions since we frequently observed maintenance of ionic coupling within late eight-cell embryos which had undergone decompaction after prolonged testing. Compaction may, however, serve to maximize cell-cell contact facilitating junction formation. We have further demonstrated that gap junction formation in the eight-cell stage does not require concomitant protein synthesis, the requisite proteinaceous components having been manufactured in the four-cell stage or even earlier. The basis for this conclusion is that a virtually complete blockade of protein synthesis beginning as early as 62 hr post-hCG, when embryos are still in the four-cell stage, failed to prevent compaction or the formation of low-resistance intercellular channels. It is interesting to note that during oogenesis the oocyte plasma membrane had formed gap junctions with the surrounding follicle cells. Although these junctions disappear just prior to ovulation, the intramembrane protein particles which make up the junctional channels may persist for utilization at the time of compaction. This would provide a pool of preexisting junctional components to be used for the rapid formation of junctions in the eight-cell stage. Unfortunately, we were not able to test this hypothesis by extending the cycloheximide treatment into earlier stages because such treatment prevents embryos from reaching the eightcell stage. If gap junction proteins are synthesized by the embryo long in advance of compaction, or if they persist as remnants of former junctional contacts during oogenesis, it is possible that low-resistance junctions might be prematurely induced when two-cell or four-cell embryos are aggregated with postcompaction, communi-

cation-competent embryos. As a prelude to this experiment, we demonstrated that compacted eight-cell embryos do form interembryonic junctions within 3-5 hr after aggregation. The number or at least the total effective cross-sectional area of induced gap junctions was judged to be very small, since they could be detected by ionic coupling measurements but not by monitoring the transfer of carboxyfluorescein. If cells are electrically coupled by even a few gap junctions then ions move readily from one cell to another, whereas transfer of fluorescent dye might not be detected because of the dilution of a small amount of dye passing into a relatively large blastomere (Bennett et aZ., 1981). In order to detect such a small number of junctional channels between aggregated embryos by dye transfer it was necessary to use Lucifer yellow, so that movement of the dye could be monitored over a longer period of time. Several other reports exist involving both insect and mammalian cells where electrical coupling was detected but dye coupling was absent (Rose et CLZ., 1977; FlaggNewton and Loewenstein, 1979; Lo and Gilula, 1979b). It is not clear why coupling between compacted eightcell embryos should be so weak, but it is possible that there are normally few junctional components in the external cell membrane, either because the entire pool is involved in existing gap junctions or because free components are segregated internally. The formation of interembryonic gap junctions was blocked by cycloheximide, indicating a requirement for new protein synthesis. This new protein synthesis could be involved either in the formation of new gap junctional components or in transport processes allowing redistribution of components to the external cell membrane. No communication could be detected between two- or four-cell embryos and eight-cell compacted embryos by ionic coupling measurements even after 10 hr of aggregation, which should have allowed sufficient time for synthesis of extra components by the eight-cell embryo. A similar finding was reported by Goodall and Johnson (1982), who failed to detect dye transfer after 3-hr ag-

164

DEVELOPMENTAL

BIOLOGY

gregations of dissociated blastomeres in phytohaemagglutinin. Our experiments differ in using longer aggregations of intact embryos in the absence of any lectin and, more importantly, in using the sensitive method of measuring ionic coupling to provide a more convincing demonstration of the inability of mouse embryos to form gap junctions prematurely. Thus, these experiments demonstrate that gap junction assembly is under very tight temporal control in the early mouse embryo and support the idea that such junctions may play a role in embryogenesis. We have clearly demonstrated that the formation of gap junctions at compaction involves controlled assembly of preexisting components, but the factors controlling this assembly have yet to be determined. The authors are grateful to Michael Blennerhassett, Richard Safranyos, and Robert Berdan for providing technical advice throughout this work and to Nancy Lewis Parsons for her instructions in making embryo aggregates. We also wish to thank Dr. Janet Rossant for her critical reading of the manuscript. We gratefully acknowledge the members of our laboratory for their,skillful assistance in the collection of embryos and making of culture media. This work was supported by grants (to G.M.K. and SC.) from the Natural Sciences and Engineering Research Council of Canada.

REFERENCES ALBERTINI, D. F., and ANDERSON, E. (1974). The appearance and structure of intercellular connections during the ontogeny of the rabbit ovarian follicle with particular reference to gap junctions. J. Cell Bid 63, 234-250. ANDERSON, E., and ALBERTINI, D. F. (1976). Gap junctions between the oocyte and companion follicle cells in the mammalian ovary. J. Cell BioL 71, 680-686. BENNETT, M. V. L., and TRINKAUS, J. P. (1970). Electrical coupling between embryonic cells by way of ext~acellular space and specialized junctions. J. Cell BioL 44, 592-610. BENNETT, M. V. L., SPRAY, D. C., and HARRIS, A. L. (1981). Electrical coupling in development. Amer. ZooL 21,413-427. BIGGERS, J. D., WHITTEN, W. K., and WHITTINGHAM, D. G. (1971). The culture of mouse embryos in vitro. In “Methods in Mammalian Embryology” (J. C. Daniel, Jr., eds.), pp. 86-116. Freeman, San Francisco. BROWER, P. T., and SCHULTZ, R. M. (1982). Intercellular communication between granulosa cells and mouse oocytes: Existence and possible nutritional role during oocyte growth. Dar. BioL 90, 144-153. CALARCO, P. G., and BROWN, E. H. (1969). An ultrastructural and cytological study of preimplantation development of the mouse. J. Exp. ZooL 171,253-284. DEKEL, N., LAWRENCE, T. S., GILULA, N. B., and BEERS, W. H. (1981).

VOLUME

98, 1983

Modulation of cell-to-cell communication in the cumulus-oocyte complex and the regulation of oocyte maturation by LH. Dev. Biol. 86, 356-362. DICAPRIO, R. A., FRENCH, A. S., and SANDERS, E. J. (19’74). Dynamic properties of electrotonie coupling between cells of early Xenopus embryos. Biophys. J. 14, 387-411. DUCIBELLA, T., ALBERTINI, D. F., ANDERSON, E., and BIGGERS, J. D. (1975). The preimplantation mammalian embryo: Characterization of intercellular junctions and their appearance during development. Dev. BioL 45, 231-250. FLAGG-NEWTON, J., and LOEWENSTEIN, W. R. (1979). Experimental depression of junctional membrane permeability in mammalian cell culture. A study with tracer molecules in the 300 to 800 Dalton range. J. Membr. BioL 50, 65-100. GILULA, N. B., EPSTEIN, M. L., and BEERS, W. H. (1978). Cell-to-cell communication and ovulation. A study of the cumulus-oocyte complex. J. Cell Biol. 78, 58-75. GOODALL, H., and JOHNSON, M. H. (1982). Use of carboxyfluorescein diacetate to study formation of permeable channels between mouse blastomeres. Nature (Lolzdon) 295, 524-526. HANNA, R. B., MODEL, P. G., SPRAY, D. C., BENNETT, M. V. L., and HARRIS, A. L. (1980). Gap junctions in early amphibian embryos. Amer. J. Anat. 158, 111-114. HELLER, D. T., CAHILL, D. M., and SCHULTZ, R. M. (1981). Biochemical studies of mammalian oogenesis: Metabolic cooperativity between granulosa cells and growing mouse oocytes. Den BioL 84,455-464. HELLER, D. T., and SCHULTZ, R. M. (1980). Ribonucleoside metabolism by mouse oocytes: Metabolic cooperativity between the fully grown oocyte and cumulus cells. J. Ezp. ZooL 214, 355-364. ITO, S., and LOEWENSTEIN, W. R. (1969). Ionic communication between early embryonic cells. Dev. BioL 19, 228-243. Lo, C. W., and GILULA, N. B. (1979a). Gap junctional communication in the preimplantation mouse embryo. Cell 18, 399-409. Lo, C. W., and GILULA, N. B. (1979b). Gap junctional communication in the postimplantation mouse embryo. Cell l&411-422. LOEWENSTEIN, W. R. (1979). Junctional intercellular communication and the control of growth. Biochim Biophys. Acta 560, l-65. MAGNUSON, T., DEMSEY, A., and STACKPOLE, C. W. (1977). Characterization of intercellular junctions in the preimplantation mouse embryo by freeze-fracture and thin-section electron microscopy. Dev. BioL 61, 252-261. MCLACHLIN, J. R., and KIDDER, G. M. (1981). Genetic control of intercellular communication in the early mouse embryo. J. Cell BioL 91, 169a. MOOR, R. M., OSBORN, J. C., CRAN, D. G., and WALTERS, D. E. (1981). Selective effect of gonadotrophins on cell coupling, nuclear maturation and protein synthesis in mammalian oocytes. J. EmbrgoL Exp. MorphoL 61, 347-365. ROSE, B., SIMPSON, I., and LOEWENSTEIN, W. R. (1977). Calcium ion produces graded changes in permeability of membrane channels in cell junction. Nature &r&on) 267, 625-627. SHERIDAN, J. D. (1971). Dye movement and low-resistance junctions between reaggregated embryonic cells. Dev. BioL 26, 627-636. SPINDLE, A. I. (1980). An improved culture medium for mouse blastocysts. In Vitro 16, 669-674.