Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/21491957

ZygoticexpressionoftheConnexin43genesuppliessubunitsforgapjunctionassemblyduringmousepreimplantationdevelopment

ARTICLEinMOLECULARREPRODUCTIONANDDEVELOPMENT·SEPTEMBER1991

ImpactFactor:2.53·DOI:10.1002/mrd.1080300103·Source:PubMed

CITATIONS

53

READS

19

5AUTHORS,INCLUDING:

PaulDeSousa

TheUniversityofEdinburgh

74PUBLICATIONS4,083CITATIONS

SEEPROFILE

EricCBeyer

UniversityofChicago

198PUBLICATIONS13,478CITATIONS

SEEPROFILE

DavidLPaul

HarvardMedicalSchool

140PUBLICATIONS18,264CITATIONS

SEEPROFILE

Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,

lettingyouaccessandreadthemimmediately.

Availablefrom:EricCBeyer

Retrievedon:03February2016

MOLECULAR REPRODUCTION AND DEVELOPMENT 3 0 1 S 2 6 (1991)

Zygotic Expression of the Connexin43 Gene Supplies Subunits for Gap Junction Assembly During Mouse Preimplantation Development GUNNAR VALDIMARSSON,' PAUL A. DE SOUSA,l ERIC C. BEYER,2 DAVID L. PAUL: AND

GERALD M. KIDDER' 'Department of Zoology, University of Western Ontario, London, Ontario, Canada; 'Diuision of Hematology, Washington University School of Medicine, St. Louis, Missouri; 3Department of Anatomy and Cellular Biology, Haruard Medical School, Boston, Massachusetts

ABSTRACT De novo assembly of gap junctions begins during compaction in the eight-cell stage of mouse development, and intercellular coupling mediated by gap junctions appears to be required for maintenance of the compacted state. We have begun to explore the expres- sion of the family of genes encoding the connexins, the proteins that form the gap junction channels. We recently reported that a protein with antigenic and size similarity with connexin32, the rat liver gap junction protein, is inherited as an oogenetic product by the mouse zygote, but its gene appears not to be transcribed prior to implantation (Barron et al., Dev Genet 10:318-323, 1989). Here we report that another member of this gene family, connexin43, is transcribed by the embryonic genome from shortly after the time of genomic activation. As revealed by Northern blotting, connexin43 mRNA is absent from ovu- lated oocytes, becomes detectable in the 4-cell stage, and accumulates steadily thereafter to reach a maximum in blastocysts. In contrast, no transcripts of connexin26 could be detected in any preimplantation stage. A protein with antigenic and size similarity with connexin43 from rat heart was found by Western blotting to accumulate from the four-cell stage onward. lmmunofluorescence analysis with embryo whole mounts was used to demonstrate that this protein is incorporated into punctate interblastomeric foci during compaction, consistent with its assembly into gap junction plaques. We conclude that connexin43 is one member of the connexin gene family whose zygotic expression is critical for preimplantation morphogenesis.

Key Words: Gap junction protein, Gene expression, Compaction

INTRODUCTION Gap junctions are collections of intercellular chan-

nels that allow cells to share inorganic ions, small metabolites, and regulatory molecules. The channels are formed when hemichannels, termed connexons, in closely apposed plasma membranes line up end to end (reviewed by Loewenstein, 1987). The connexins are the major protein constituents of isolated gap junctions and are believed to be the subunits from which connex-

ons are assembled (Beyer et al., 1987). Analysis of gap junctions from a variety of rodent organs has revealed at least four distinct, but related, connexins: connexin26 (Nicholson et al., 1987; Zhang and Nicholson, 1989), connexin32 (Kumar and Gilula, 1986; Paul, 1986), connexin43 (Beyer et al., 1987), and connexin46 (Beyer et al., 1988; Kistler e t al., 1988).

Preimplantation mouse embryos acquire functional gap junctions during compaction in the eight-cell stage, accompanied by the critically important morphogenetic processes of cell flattening and blastomere polarization (for review, see Kidder, 1987). The assembly of gap junctions in the eight-cell stage does not depend on cell flattening (Goodall, 1986; Kidder et al., 1987), but functional gap junctions are required for the mainte- nance of the compacted state (Buehr et al., 1987; Lee et al., 1987). Gap junctions are therefore essential for development of the blastocyst, and this is one of the clearest examples of a requirement for intercellular junctional coupling in a developmental process. Identi- fication of the connexin(s) involved is the first step towards determining exactly what role intercellular coupling plays in this process.

The preimplantation embryo also provides a unique opportunity to study de novo assembly of gap junctions, and the connexin genes provide a focus for investiga- tions of the genetic program underlying compaction and cavitation. For these reasons we have begun to characterize the expression pattern of the connexin gene family in preimplantation development. We pre- viously reported that connexin32 is a persistent ooge- netic product present in all preimplantation stages of the mouse,whereas its mRNA cannot be detected at any stage. In addition, we showed that connexin43 mRNA is present by the beginning of cavitation (Barron et al., 1989). Here we report that both mRNA and protein corresponding to connexin43 accumulate through pre- implantation development beginning shortly after ac- tivation of the embryonic genome. This protein is

Received March 18, 1991; accepted April 19, 1991. Address reprint requests to Dr. G.M. Kidder, Department of Zoology, University of Western Ontario, London, Ontario N6A 5B7, Canada.

0 1991 WILEY-LISS, INC.

CONNEXIN43 IN MOUSE EMBRYOS 19

assembled into structures resembling gap junction plaques beginning at compaction. Transcripts of an- other member of the connexin gene family, connexin26, could not be detected.

MATERIALS AND METHODS Embryo Collection

CF1 female mice (Charles River Canada Ltd., St. Constant, Quebec, Canada) were superovulated with pregnant mare serum gonadotropin (PMSG) and hu- man chorionic gonadotropin (hCG) and mated with CB6F,/J males (The Jackson Laboratory, Bar Harbor, ME), and embryos were collected as described previ- ously (Barron et al., 1989). Unfertilized oocytes were collected from females that were superovulated but not mated. Oocytes or embryos to be used for Northern or Western blots were washed by passage through at least three drops of the flushing medium followed by five drops of phosphate-buffered saline containing 3 mg/ml polyvinylpyrrolidone (PBS-PVP). They were then transferred to an Eppendorf tube in 5 pl or less of PBS-PVP and quick frozen in a dry ice-methanol bath. The oocytes or embryos were exposed to culture condi- tions for no more than 1-2 hr prior to freezing. A comparable volume of the last wash from which the embryos had been taken was also frozen. The frozen embryos and companion wash samples were stored at - 70°C until sufficient numbers had been collected for an experiment. Prior to the PBS-PVP washing steps, oocytes were passed through several drops of medium containing 800 U/ml hyaluronidase (Sigma Chemical Co., St. Louis, MO) to remove the cumulus cells. The effectiveness of this treatment was monitored by view- ing a sample under the microscope using Nomarski optics. The staging of oocytes or embryos in relation to hr post-hCG was as follows: oocytes, 16 hr; one-cell zygotes, 18 hr; two-cell, 48 hr; four-cell, 60 hr; eight- cell, 71 hr; compacted morulae, 76-80 hr; late morulae, 82 hr; blastocysts, 94 hr.

Northern Blots Total RNA was extracted from frozen oocytes or

embryos (1,200-1,500 per extraction) using a modifi- cation of the single step method of Chomczynski and Sacchi (1987) as described by Barron et al. (1989). As a positive control, total RNA was also extracted from adult mouse brain, kidney, liver, and heart in the same fashion. The extracted RNAs were separated by elec- trophoresis in 1% denaturing formaldehyde-agarose gels as previously described (Barron et al., 1989) and transferred by capillary action to Hybond-N blotting membrane (Amersham, Oakville, Ontario, Canada) in lox SSC. The blots were probed with cDNAs labeled by the random primer method (Feinberg and Vogelstein, 1983) to a specific activity of 1-2 x lo9 cpdpg. The prehybridization and hybridization solutions consisted of 5x SSPE, 50% formamide, 5x Denhardt's, 10% dextran sulfate, 1% sodium dodecyl sulfate (SDS), and 100 pg/ml sheared denatured salmon sperm DNA. Hy-

bridization was carried out a t 42°C for 40-44 hr at a probe concentration of 1-3 ng/ml. Posthybridization washes were at moderate stringency ( 0 . 5 ~ SSC, 0.5% SDS, 65"C), which is sufficient to prevent cross-hybrid- ization between members of the connexin gene family (Paul, 1986). Autoradiograms were scanned using an LKB Ultroscan XL laser densitometer.

Western Blots For each embryonic stage analyzed, about 600 frozen

embryos were lysed in 20 pl of 2~ Laemmli sample buffer (Laemmli, 1970) containing 1 mM phenylmeth- ylsulfonyl fluoride (PMSF). As a positive control for connexin43, mouse heart plasma membrane prepara- tions enriched for intercalated discs were isolated ac- cording to Green and Severs (1983) except that 1 mM PMSF was added to all solutions. Electrophoresis was carried out in minigels (4% stacking gel, 12% resolving gel) according to Laemmli (19701, and the separated proteins were electrophoretically transferred to nitro- cellulose (Towbin et al., 1979). The blots were blocked for 1-4 hr in 5% skim milk powder in Tris-buffered saline (TBS) (20mM Tris, pH 7.5, 500mM NaCl), washed twice in 0.05% Tween-20 in TBS, and incubated overnight with the primary antiserum in 1% skim milk powder in TTBS. The blot was stripped of unbound primary antiserum and placed in alkaline phosphatase- conjugated secondary antibody (goat antirabbit IgG) in 1% skim milk powder in TTBS for 3-4 hr. The second- ary antibodies and color development reagents (Bio- Rad Laboratories Canada Ltd., Mississauga, Ontario, Canada, or Promega Corp., Madison, WI) were used as suggested by the suppliers.

Immunofluorescence All steps were carried out at 4°C. Embryos were fixed

for at least 1 hr in 1% paraformaldehyde in PHEM buffer (60 mM PIPES, 25 mM HEPES, 10 mM EGTA, and 1 mM MgCI,, pH 6.9) and permeabilized for 20 min with 0.1% Tween-20 in PHEM buffer, then for another 20min in 0.1% Tween-20 in Ca2+-, Mg2+-free PBS (CMF-PBS). The embryos were blocked in CMF-PBS containing 1% bovine serum albumin (BSA) and 0.1% Tween-20 for a t least 30 min and then incubated in the primary antiserum in blocking solution overnight. Following extensive washing in CMF-PBS/Tween to remove the primary antiserum, the embryos were labeled for at least 1 hr with a fluorescein isothiocy- anate-conjugated secondary antibody (goat antirabbit IgG, from ICN Biomedicals Canada Ltd., St. Laurent, Quebec, Canada) in blocking solution also containing 5 pg/ml DAPI stain (Polysciences Inc., Warrington, PA). The secondary antibody solution was removed, and the embryos were mounted in 9:l glycero1:lOx TBS (pH 8.3) and viewed with a Zeiss Photomicroscope I equipped with epifluorescence optics. A few of the embryos were also viewed with a Bio-Rad Lasersharp MRC 500 confocal laser scanning microscope.

20 G. VALDIMARSSON ET AL.

Antisera and cDNAs The two connexin43-specific antisera used in this

study were raised in rabbits immunized with a syn- thetic peptide corresponding to a unique region of the predicted connexin43 polypeptide (amino acids 252- 271; Beyer et al., 1989). Antiserum was used at 1:1,000 dilution for Western blots and either 1500 or 1:11 for immunofluorescence depending on the antiserum. The preimmune serum served as a control for both types of analysis. To test the specificity of the antisera, each was reacted before use with the synthetic peptide. The peptide was dissolved to a final concentration of 0.4 mgl ml in CMF-PBS containing 0.1% Tween-20 and this solution was used to dilute the antiserum to its working concentration. The antiserum-peptide mixture was stirred for 1 hr at 4°C before use.

The two connexin cDNAs used in this study, connexin43 and connexin26, were isolated by screening rat cDNA libraries, and each represents the entire coding sequence of the respective mRNA (Beyer et al., 1987; Zhang and Nicholson, 1989). The connexin43 cDNA was isolated from a heart library and hybridizes to a 3.0 kb RNA present in several adult tissues (Beyer et al., 1987) and in mouse blastocysts (Barron et al., 1989). The connexin26 cDNA was isolated from a liver library and detects a 2.5 kb RNA in preparations from several adult tissues (Zhang and Nicholson, 1989).

As an internal positive control, we also hybridized each Northern blot with a histone H3.2 clone. The thoroughly characterized developmental profile of this message (Giebelhaus et al., 1983; Graves et al., 1985; Barron et al., 1989) and its small size make it ideal for this purpose.

RESULTS Our previous findings that connexin43 mRNA could

be detected in late preimplantation embryos, whereas connexin32 mRNA could not be detected at any preim- plantation stage, led us to postulate that it is the connexin43 gene whose expression is required for the assembly of new connexons as the level of intercellular coupling increases after compaction (McLachlin and Kidder, 1986; Barron et al., 1989). To investigate this further, we began the present study by determining the developmental profile of connexin43 mRNA. A North- ern blot containing total RNA from 1,500 embryos per lane and total RNA from brain, liver, kidney, and heart was probed simultaneously with the connexin43 and histone H3.2 probes. When hybridized separately at moderate stringency, each of these two probes produces a single band (Barron et al., 1989). Figure 1 shows an autoradiogram of this blot exposed for 3 days. The connexin43 cDNA hybridized to a 3 kb RNA in both embryo and organ samples. This RNA is present in the zygote, becomes undetectable at the two-cell stage, and then gradually increases in abundance from the four- cell stage to the blastocyst stage. A subsequent 2 week exposure of this blot failed to reveal a signal at the two-cell stage. The presence of connexin43 mRNA in

Fig. 1. Connexin43 mRNA is present in preimplantation embryos. A Northern blot containing 5 pg total RNA from brain (B), liver (L), kidney (K) and about 2 pg from heart (H) and total RNA from 1,500 embryos each of zygotes (ZY), two-cell (2C), four-cell (4C), eight-cell (8C), late morulae (LM), and early blastocysts (EB) was simulta- neously probed with connexin43 cDNA and the histone H3.2 clone. The connexin43 cDNA hybridized to a single 3.0 kb band (arrow) in brain, kidney, and heart and in all the embryo stages except the two-cell stage. The histone probe hybridized to a single 0.6-0.7 kb band (arrowhead). The autoradiogram was exposed for 3 days.

brain, kidney, and heart but not liver conforms with the known tissue distribution of this message (Beyer et al., 1987). To determine if the signal in zygotes could have resulted from residual cumulus cell contamination, RNA was extracted from 1,200 oocytes that were subjected to several successive hyaluronidase treat- ments and then verified microscopically to be free from the cumulus cells, which were collected separately. As is shown in Figure 2, removal of the cumulus cells resulted in loss of the 3 kb band from the oocytes, although some of the histone signal remained. The isolated cumulus cells, as expected, produced a very strong connexin43 signal. Densitometric scans of var- ious exposures of the blot shown in Figure 1 revealed that connexin43 and histone transcripts increase at about the same rate following activation of the embry- onic genome, although the histone transcripts are much more abundant. Our measurement of a 16-fold increase in histone H3.2 mRNA on a per-embryo basis between the late two-cell and blastocyst stages is in reasonable agreement with published data from quan- titative dot blotting (Graves et al., 1985) indicating a ninefold increase. Since the ratio of hybridization sig- nals provided by the connexin43 and histone probes

CONNEXIN43 IN MOUSE EMBRYOS 21

Fig. 2. Connexin43 mRNA is not present in oocytes after complete removal of cumulus cells. This Northern blot contained total RNA from 1,200 unfertilized oocytes (0) and from the cumulus cells (C) that had been removed from them by extensive hyaluronidase treatments. This resulted in loss of the connexin43 band (arrow) from the oocytes, while a portion of the histone signal (arrowhead) remained. The cumulus cells retained all of the connexin43 mRNA and a large portion of the histone mRNA. The autoradiogram was exposed for 3 days. A second exposure for 2 weeks failed to reveal a 3 kb band in the oocyte lane.

Fig. 3. Connexin26 mRNA was not detected in preimplantation embryos. The same Northern blot shown in Figure 1 was reprobed simultaneously with the connexin26 and histone H3.2 probes. The connexin26 cDNA hybridized to a single 2.5 kb band (arrow) in brain, liver, and kidney RNA, which was not detected in heart or embryo RNA. The histone profile (arrowhead) from Figure 1 was reproduced, indicating retention of the RNA. The autoradiogram was exposed for 14 days.

debris. In general, the level of this doublet is high in the zygote (but see Discussion), decreases sharply at the two-cell stage, and then increases gradually t o the blastocyst stage. The ratio of intensities of the two bands of the doublet was variable. The high-molecular- weight band (M, > 84,000) present in all the embryo stages is due to endogenous alkaline phosphatases (see Barron et al., 1989). To ascertain the specificity of the antiserum, an embryo lysate from 800-900 blastocysts was probed with the preimmune serum. No signal other than the alkaline phosphatase band was detected (data

does not change appreciably between the four-cell and blastocyst stages, we can use the data of Graves et al. (1985) to estimate that connexin43 transcripts undergo a sevenfold increase per embryo during this period. On a per-cell basis, however, the connexin43 transcript level may decline somewhat between the late morula and early blastocyst stages.

The Northern blot in Figure 1 was reprobed with connexin26 cDNA along with the histone DNA after the previous radioactivity had decayed. No connexin26 signal was detected in the embryo RNAs, even on autoradiograms exposed for 2 weeks, whereas the pre- vious histone mRNA profile was reproduced (Fig. 3). The presence of a 2.5 kb connexin26 mRNA in brain, liver, and kidney is in agreement with published data (Zhang and Nicholson, 1989).

In a second set of experiments, we used Western blot analysis to investigate connexin43 levels in preimplan- tation embryos. Figure 4 shows one such blot where anticonnexin43 antiserum was used to probe embryo lysates from various stages of preimplantation devel- opment. A doublet of polypeptides of M, 43,000-48,OOO was detected in all of the embryo samples. It was not detected in any of the last washes, verifying that it could not have been derived from reproductive tract

not shown). We used the same antiserum to localize connexin43

in embryo whole mounts. In four-cell and eight-cell uncompacted embryos, only patchy cytoplasmic stain- ing was observed (Fig. 5A), but, beginning with com- paction, punctate interblastomeric staining could be seen (Fig. 5C). No such immunoreactivity was detected using the preimmune serum (Fig. 5E). Preincubation of the antiserum with the synthetic peptide against which it was raised reduced the level of staining to that of the preimmune serum (not shown). No immunoreactivity was detected in the zona. Of a total of 39 precompaction embryos examined after reaction with this antiserum, all showed cytoplasmic staining, but not one showed interblastomeric staining. In contrast, 92% of the 38 postcompaction embryos examined showed cytoplasmic

22 G. VALDIMARSSON ET AL.

Fig. 4. Connexin43 is present throughout preimplantation devel- opment. A Western blot containing zygote (ZY), two-cell (ZC), four-cell ( 4 0 , eight-cell ( 8 0 , late morula (LM), and early blastocyst (EB) lysates (600 embryos each) and an intercalated disc-enriched mem- brane preparation from mouse heart (H) was probed with a connexin43 specific antiserum. A doublet with M, 43,00048,000 was

staining and 66% showed punctate interblastomeric staining. A few of these embryos were examined with a confocal microscope, which confirmed the distinction between patchy cytoplasmic staining and punctate interblastomeric staining in postcompaction embryos (Fig. 6). In addition, it became apparent that some of the cytoplasmic immunoreactivity is arranged in a perinuclear array. The possible significance of this is considered in the Discussion. The punctate interblas- tomeric staining is temporally and spatially consistent with connexin43’s being assembled into gap junction plaques at a time when, by both morphological (Duci- bella et al., 1975) and physiological criteria (Lo and Gilula, 1979; McLachlin et al., 1983), gap junctions first form in the embryo.

As a further test of the specificity of the immunocy- tochemistry, an experiment was performed using a second antiserum raised against the same peptide and affinity purified with it. In this case the embryo groups were scored blind; i.e., their identities were revealed to the experimenter only after scoring had been completed. Of 31 postcompaction embryos (76-80 hr post-hCG) treated with this antiserum, all showed cytoplasmic staining and 61% showed punctate inter- blastomeric staining. None of the 30 embryos treated with that antiserum after it had been absorbed with the peptide showed either staining pattern.

DISCUSSION The results presented here demonstrate that a pro-

tein with antigenic and structural similarity with connexin43, a known gap junction protein from rat organs, accumulates during preimplantation develop- ment of the mouse and is incorporated into what appear

detected in all but the four-cell stage while a single band was detected in the heart preparation. The panel on the extreme right shows a four-cell lane from another Western blot probed with the same antiserum. It was included to show that faint connexin43 bands (arrow) could be detected in some embryo batches at that stage.

to be gap junction plaques concomitant with compac- tion. This protein is present along with another, which, according to the same criteria, is a homolog of rat connexin32 (Barron et al., 1989). However, our North- ern blot results taken together with our previous study, in which we failed to detect any connexin32 mRNA in preimplantation embryos (Barron et al., 19891, strongly suggest that the connexin43 gene is the predominant, if not the only, known member of the rodent connexin gene family transcribed during this early period of embryogenesis (preliminary results indicate that connexin46 mRNA is not present either). We therefore consider connexin43 to be one of the first products of the zygotic genome; the connexin32 present in the same embryos, on the other hand, must be an oogenetic product (Barron et al., 1989). While we believe we are detecting the mouse equivalents of rat connexin43 and its mRNA, the possibility that we are detecting the products of other related genes cannot be totally dis- counted. This seems unlikely given the established specificity of the antisera (Beyer et al., 19891, the stringency of the washes for the Northern blots (see Paul, 1986; Beyer et al., 1987), and the fact that the sizes of the bands on both Northern and Western blots conformed closely with those found in rat organs.

Our finding of a doublet of polypeptides reacting with the connexin43 antiserum is intriguing, especially since only a single polypeptide was recognized in heart. One possible explanation is that connexin43 undergoes posttranslational modification, such as proteolytic cleavage or phosphorylation, during preimplantation development. Multiple polypeptides reactive with connexin43 antibodies have been noted in other cell types, and in some cases this has been demonstrated to

CONNEXINIS IN MOUSE EMBRYOS 23

Fig. 5. Detection of connexin43 in embryo whole mounts by indirect immunofluorescence. Specific immunoreactivity was detected in both four-cell (A) and eight-cell embryos (C), although punctate staining at the interface of apposed blastomeres, suggesting connexin43 assembly into gap junction plaques (arrows in C), was seen only in embryos undergoing or having completed compaction. Staining in embryos treated with preimmune serum (E) was confined to low levels of uniformly distributed fluorescence. The cell numbers in A, C, and E were determined by counting DAPI-stained nuclei as shown in B, D, and F, respectively. Bar = 15 pm.

24 G. VALDIMARSSON ET AL.

The presence of connexin43 and its mRNA in four- cell embryos and their accumulation thereafter are consistent with data from transcriptional and transla- tional inhibition experiments, which showed that in the four-cell stage there is already a sufficient stockpile of gap junctional precursors to establish coupling in the eight-cell stage (McLachlin et al., 1983; McLachlin and Kidder, 1986) and with the increase in size and number of gap junction plaques that occurs between the eight- cell and morula stages (Magnuson et al., 1977). The developmental profile of connexin43 mRNA suggests that transcription of the connexin43 gene is initiated as part of the general activation of the zygotic genome that takes place in the two-cell stage (reviewed by Schultz, 1986). Moreover, the profile of accumulation of this message is similar to that of actin, histone, and Na',K'-ATPase a-subunit mRNAs as well as most of the unidentified transcripts that first appear following genomic activation (Giebelhaus et al., 1983; Watson et al. 1990; Taylor and Piko, 1987). The fact that we did not detect connexin43 transcripts in two-cell embryos may simply mean that the abundance of the message had not yet reached the detection limit of the Northern blot assay. Such low abundance could be due either to a low transcription rate or to a high turnover rate for this message.

Polypeptides bearing the epitope recognized by our antisera become incorporated into plaque-like foci in opposed plasma membrane regions during compaction, indicating that at least some of the immunoreactive material is gap junction precursor. Prior to this, cloud- like foci of immunoreactivity can be seen in the cyto- plasm; some of these persist after gap junction assem- bly has begun. The fact that preincubation with the synthetic peptide abolished both the punctate inter- blastomeric staining and the patchy cytoplasmic stain- ing indicates that both types of immunoreactivity are specific to the synthetic peptide epitope and therefore accurately represent the distribution of connexin43. This conclusion is further supported by our recent finding that a third affinity-purified polyclonal anti- body, raised against a synthetic peptide corresponding to a different segment of the C-terminal cytoplasmic tail of connexin43, produces the same staining patterns (De Sousa and Kidder, in preparation). Although we are convinced that the foci of immunoreactivity in opposed plasma membranes represent sites of gap junction assembly, discovering the identity of the patches of immunoreactivity in the cytoplasm will require further experimentation. One possibility is that they are sites of processing or storage of nascent connexin43 prior to its being inserted into plasma membranes. The perinuclear arrangement of some of the sites, as revealed by confocal microscopy, suggests that they may be Golgi complexes; indeed, our whole mount images with the conventional microscope bear a striking resemblance to those obtained using an anti- body against a Golgi antigen (Maro et al., 1985).

Do both connexin32 and connexin43 contribute t o the establishment of gap junctional coupling at the time of

Fig. 6. Analysis of the distribution of connexin43 in a compacting eight-cell embryo by indirect immunofluorescence using confocal microscopy. Most of the immunoreactivity was contained in numerous cytoplasmic foci, including a fairly regular array surrounding each nucleus (small arrows). In addition, putative gap junction plaques could be identified as discrete, brightly staining interblastomeric foci (large arrow). Bar = 25 pm.

result from posttranslational phosphorylation (Crow et al., 1990; Musil et al., 1990). The variable ratio between upper and lower bands on our blots could be explained by the action of phosphatases during the cell lysis procedure, or it might be developmentally signif- icant. We plan to make this the subject of future investigations.

At present, it is difficult to interpret our detection of connexin43 in one-cell zygotes and two-cell embryos by immunoblotting. The connexin43 transcript in zygotes is clearly a contaminant contributed by residual cumu- lus cells, which are rich in connexin43 junctions (Beyer et al., 1989; Risek et al., 1990). If the connexin43 gene is transcribed during oogenesis, its transcripts must decay by the time of ovulation. The protein itself, however, may be a product of the oocytes that persists beyond fertilization. On the other hand, even carefully denuded oocytes or zygotes might be expected to retain connexin43 of cumulus cell origin since the gap junc- tions which couple the oocyte to the cumulus prior to ovulation are located on the tips of cumulus cell pro- jections that penetrate through the zona and into the oocyte (reviewed by Larsen and Wert, 1988). It is not known if these junctions are retained by the oocyte or withdrawn during cumulus expansion. Electron micro- scopic analysis will be needed to identify the connexins that constitutes the oocyte-cumulus gap junctions.

CONNEXIN43 IN MOUSE EMBRYOS 25

compaction? This is certainly a possibility, since colo- calization of two different connexins in the same gap junction plaque has been demonstrated in hepatocytes (Nicholson et al., 1987). Our results make it very likely that connexin43 plays a role in gap junction assembly, and two other recent reports have been interpreted as demonstrating a functional role for connexin32 as well. Lee et al. (1987) injected affinity-purified polyclonal antibodies, raised against connexin32 from rat liver, into eight-cell compacted mouse blastomeres. The an- tibodies caused the injected blastomeres to uncouple from their neighbors and eventually to decompact. The same antibodies stained the surface membranes of morula-stage embryos in a punctate pattern. The most straightforward explanation for these results is that the antibodies were binding to connexin32 localized in functional gap junctions. On the other hand, the anti- bodies used by Lee et al. (1987) are known to recognize at least one conformational epitope that may be con- served among the connexins (Milks et al., 1988), and an examination of their reactivity towards gap junctions from connexin43-containing organs such as heart has not been published. In the second report, Bevilacqua et al. (1989) described microinjection experiments with connexin32 antisense RNA, which gave results very similar to those of Lee et al. (1987); i.e., the antisense RNA caused uncoupling and decompaction. The appar- ent absence of connexin32 mRNA in preimplantation embryos makes it likely, however, that the antisense RNA in that experiment was cross-hybridizing with connexin43 message in the low-stringency environ- ment of the cytoplasm (see Paul, 1986; Beyer et al., 1987). In our view, the involvement of connexin32 in the establishment of coupling during compaction has not yet been conclusively demonstrated. Our current hypothesis is that connexin32 is an oogenetic product that may be left over from oocyte-cumulus cell gap junctions and that the gap junctions formed de novo during compaction are composed primarily, if not ex- clusively, of connexin43.

ACKNOWLEDGMENTS We thank Cynthia Pape and Andrew Watson for help

with the embryo collections, Linda Musil for supplying one of the antisera, and Bio-Rad Laboratories (Canada) Ltd. for providing access to a confocal microscope. The work was supported by grants from NSERC Canada (to G.M.K.) and the NIH (GM37751 to D.L.P. and EY08368 and HL45466 to E.C.B.).

REFERENCES Barron DJ, Valdimarsson G, Paul DL, Kidder GM (1989):

Connexin32, a gap junction protein, is a persistent oogenetic product through preimplantation development of the mouse. Dev Genet 10:318-323.

Bevilacqua A, Loch-Caruso R, Erickson RP (1989): Abnormal devel- opment and dye coupling produced by antisense RNA to gap junction protein in mouse preimplantation embryos. Proc Natl Acad Sci USA 86:544&5448.

Beyer EC, Goodenough DA, Paul DL (1988): The connexins, a family

of related gap junction proteins. In EL Hertzberg and RG Johnson (eds): “Gap Junctions.” New York: Alan R. Liss, Inc., pp 167-175.

Beyer EC, Kistler J , Paul DL, Goodenough DA (1989): Antisera directed against connexin43 peptides react with a 43-kD protein localized to gap junctions in myocardium and other tissues. J Cell Biol 108595-605.

Beyer EC, Paul DL, Goodenough DA (1987): Connexin43: A Protein from rat heart homologous to a gap junction protein from liver. J Cell Biol 105:2621-2629.

Buehr M, Lee S, McLaren A, Warner A (1987): Reduced gap junctional communication is associated with the lethal condition characteristic of DDK mouse eggs fertilized by foreign sperm. Development 101:449-459.

Chomczynski P, Sacchi N (1987): Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162:156-159.

Crow DS, Beyer EC, Paul DL, Kobe SS, Lau AF (1990): Phosphory- lation of connexin43 gap junction protein in uninfected and Rous sarcoma virus-transformed mammalian fibroblasts. Mol Cell Biol 10: 1754-1763.

Ducibella T, Albertini DF, Anderson E, Biggers JD (1975): The preimplantation mammalian embryo: characterization of intercel- M a r junctions and their appearance during development. Dev Biol 45:231-250.

Feinberg A, Vogelstein B (1983): A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal Biochem 137:6-13.

Giebelhaus DH, Heikkila JJ, Schultz GA (1983): Changes in the quantity of histone and actin messenger RNA during the develop- ment of preimplantation mouse embryos. Dev Biol 98:148-154.

Goodall H (1986): Manipulation of gap junctional communication during compaction of the mouse early embryo. J Embryol Exp Morphol91:283-296.

Graves RA, Marzluff WF, Giebelhaus DH, Schultz GA (1985): Quan- titative and qualitative changes in histone gene expression during early mouse embryo development. Proc Natl Acad Sci USA 82:5685-5689.

Green CR, Severs NJ (1983): A simplified method for the rapid isolation of cardiac intercalated discs. Tissue Cell 15: 17-26.

Kidder GM (1987): Intercellular communication during mouse em- bryogenesis. In BD Bavister (ed) “The Mammalian Preimplantation Embryo: Regulation of Growth and Differentiation In Vitro.” New York: Plenum Publishing Corporation, pp 43-64.

Kidder GM, Rains J , McKeon J (1987): Gap junction assembly in the preimplantation mouse conceptus is independent of microtubules, microfilaments, cell flattening, and cytokinesis. Proc Natl Acad Sci USA 84:3718-3722.

Kistler J, Christie D, Bullivant S (1988): Homologies between gap junction proteins in lens, heart and liver. Nature 331:721-723.

Kumar NM, Gilula NB (1986): Cloning and characterization of human and rat liver cDNAs coding for a gap junction protein. J Cell Biol 103:767-776.

Laemmli UK (1970): Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685.

Larsen WJ, Wert SE (1988): Roles of cell junctions in gametogenesis and early embryonic development. Tissue Cell 202309-848.

Lee S, Gilula NB, Warner AE (1987): Gap junctional communication and compaction during preimplantation stages of mouse develop- ment. Cell 51:851-860.

Lo CW, Gilula NB (1979): Gap junctional communication in the preimplantation mouse embryo. Cell 18:399-409.

Loewenstein WR (1987): The cell-to-cell channel of gap junctions. Cell 48:725-726.

Magnuson T, Demsey A, Stackpole CW (1977): Characterization of intercellular junctions in the preimplantation mouse embryo by freeze-fracture and thin-section electron microscopy. Dev Biol 61:252-261.

Mar0 B, Johnson MH, Pickering SJ, Louvard D (1985): Changes in the distribution of membranous organelles during mouse early devel- opment. J Embryol Exp Morphol 90:287-309.

McLachlin JR, Caveney S, Kidder GM (1983): Control of gap junction formation in early mouse embryos. Dev Biol 98:155-164.

McLachlin JR, Kidder GM (1986): Intercellular junctional coupling in

26 G. VALDIMARSSON ET AL.

preimplantation mouse embryos: Effect of blocking transcription or translation. Dev Biol 117:146-155.

Milks LC, Kumar NM, Houghten R, Unwin N, Gilula NB (1988): Topology of the 32-kd liver gap junction protein determined by site-directed antibody localizations. EMBO J 7:2967-2975.

Musil LS, Beyer EC, Goodenough DA (1990): Expression of the gap junction protein connexin43 in embryonic chick lens: Molecular cloning, ultrastructural localization, and post-translational phos- phorylation. J Membrane Biol 116:163-175.

Nicholson B, Dermietzel R, Teplow D, Traub 0, Willecke K, Revel J P (1987): Two homologous protein components of hepatic gap junc- tions. Nature 329:732-734.

Paul DL (1986): Molecular cloning of cDNA for rat liver gap junction protein. J Cell Biol 103:123-134.

Risek B, Guthrie S, Kumar N, Gilula NB (1990): Modulation of gap junction transcript and protein expression during pregnancy in the rat. J Cell Biol 110:269-282.

Schultz GA (1986): Utilization of genetic information in the preim- plantation mouse embryo. In J Rossant and RA Pedersen (eds): “Experimental Approaches to Mammalian Embryonic Develop- ment.” New York: Cambridge University Press, pp 239-265.

Taylor KD, Piko L (1987): Patterns of mRNA prevalence and expres- sion of B1 and B2 transcripts in early mouse embryos. Development 101:877-892.

Towbin H, Staehelin T, Gordon J (1979): Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Proce- dure and some applications. Proc Natl Acad Sci USA 76:4350-4354.

Watson AJ, Pape C, Emmanuel JR, Levinson R, Kidder GM (1990): Expression of Na,K-ATPase a and p subunit genes during preim- plantation development of the mouse. Dev Genet 11:41-48.

Zhang JT, Nicholson B (1989): Sequence and tissue distribution of a second protein of hepatic gap junctions, Cx26, as deduced from its cDNA. J Cell Biol 109:3391-3401.