Phosphorylation of a Gap Junction Protein ... - Semantic Scholar

1 downloads 0 Views 1MB Size Report
we examined whether there is a popula- tion of tyrosine-phosphorylated connexin43 molecules present in the iysate that was not recognized by the C43 antibod.
Vol.

1, 661-668,

December

1990

Cell

Phosphorylation Correlates with Inhibition Communication

Tyrosine

.

.

Adele J. Filson,2 Roobik Werner R. Loewenstein, Howard University Philadelphia, Physiology

Hughes

Medical

Azarnia, Eric C. Beyer, and Joan S. Brugge3

Institute

of Pennsylvania Pennsylvania and Biophysics,

of a Gap Junction of Cell-to-Cell

and

Department

of Microbiology,

School of Medicine, 19104 [A. J. F., J. S. B.]; Department of University of Miami School of Medicine,

Miami, Florida 33101 [R. A., W. R. L.]; and Departments and Medicine, Washington University School of Medicine, St. Louis, Missouri 631 10 [E. C. B.]

of Pediatrics

Abstract

Cell-to-cell communication is achieved by passage small molecules through gap junction membrane channels. The expression from Rous sarcoma virus,

of

of the transforming gene v-src, induces a rapid and

dramatic reduction in cell-to-cell communication in cultured cells. To determine whether connexin43, a major gap junction protein expressed in fibroblasts, is a target for the v-src protein tyrosine kinase activity, we examined the phosphorylation state of connexin43 in cells expressing variants of src. Using an antipeptide serum that recognizes connexin43, we demonstrate that this protein is phosphorylated on serine and tyrosine residues in avian and mammalian cells expressing activated src proteins. Connexin43 from control cells and cells expressing nonactivated variants of the src protein

was

phosphorylated

solely

on serine

residues. In Iysates from v-src-transformed cells, all phosphorylated connexin43 molecules were cleared from the lysate by sequential immunoprecipitations using the phosphotyrosine antibodies, suggesting that each molecule of phosphorylated connexin43 contains both phosphoserine and phosphotyrosine. We have also examined junctional permeability in cells expressing src variants and find that loss of cell-to-cell communication correlates with tyrosine phosphorylation

of connexin43.

Introduction The 16-20-A-wide membrane are the conduits for intercellular gans and tissue (i). This direct munication appears to play control of cellular growth and communication is dramatically formed with the v-src gene protein tyrosine kinase (12-14).

channels in gap junctions communication in ormode of cell-to-cell coman important role in the differentiation (2-7). This impaired in cells trans(8-li), which encodes a The v-src gene product,

& Differentiation

Protein

as well as other PTKs,4 can regulate cell-to-cell communication. Cultured cells which express a temperaturesensitive v-src gene product display a reduction in cornmunication within minutes after shifting from the nonpermissive temperature to the permissive temperature (9, 1 5). This reduction in communication correlates with an increase in the tynosine-specific protein kinase activity of pp6Oc as well as with the induction of cellular transformation (1 2, 16). Cell-to-cell communication is also reduced in cells transformed with polyomavirus middle T antigen (17). Middle T binds to, and activates, several cellular PTKs, including the cellular snc protein (18-23), suggesting that the activated cellular snc protein bound to middle T antigen has an effect on communication similar to that of the viral src protein. Finally, activation of two receptor PTKs, the receptors for epidermal growth factor and platelet-derived growth factor, also causes a loss in cellto-cell communication (24), implicating two other PTKs in junctional regulation. The cell-to-cell channels are bipartite structures made of six subunits (25, 26) that are tilted and arranged symmetrically around a central aqueous passageway (27, 28). The cDNAs encoding the channel proteins from a variety of tissues have been sequenced, revealing a gap junction protein family with high homology and several conserved structural motifs (29-36). A cDNA recently cloned and sequenced from a rat heart cDNA library gives a predicted molecular weight of 43,000 (31). This protein, termed connexin43 (31) or a (36), is expressed in a variety of tissues, including fibroblasts, ovarian granulosa cells, conneal epithelial cells, and Xenopus embryos (36, 37). Connexin43 is also expressed in embryonic chicken lens cells, and the primary amino acid sequence ofthis protein is highly homologous to the rat counterpart (38). The mechanism whereby PTKs regulate cell-to-cell communication has not yet been elucidated. The v-src gene product is localized at the plasma membrane, including regions of cell-to-cell contact (39) where the gap junction proteins might conceivably be directly phosphorylated by the v-src gene product. Using an antiserum raised against connexin43, we examined the phosphorylation of connexin43 in fibroblasts transformed with vsrc. We also addressed whether tyrosine phosphorylation of connexin43 correlates with src-induced inhibition of cell-to-cell communication by examining cells expressing variants

Received 8/28/90. I This work was funded by Research Grants CA-27951 0. 5. B.) and CA14464 (W. R. L.) from the NIH. 2 A. J. F. is presently on leave from the Cellular and Developmental Biology Program, State University of New Vork, Stony Brook, NV 11790. 3 To whom requests for reprints should be addressed.

Growth

of the

src protein.

The abbreviations used are: PTK, protein tyrosine kinase; cDNA, plementary DNA; SDS, sodium dodecyl sulfate; RSV, Rout sarcoma CEF, chicken embryo fibroblast; DMEM, Dulbecco’s modified medium.

4

comvirus; Eagle’s

661

662

Tyrosine

Phosphorylation

of a Gap

Junction

Protein

C

A

iI’ .

1

2

3

ti

P-Ser

-

P-Thr

-

P-Tyr-

123

Fig. 1. Analysis of connexin43 phosphorylation in normal BALB/c3T3 cells and SRD3T3 cells. Cultures of BALB/c3T3 cells (Lane 1 ) and SRD3T3 cells (Lanes 2 and 3) were labeled with 32P, for 2 h. A, proteins were immunoprecipitated with the following antibodies: antipeptide serum directed against connexin43 (Lanes 1 and 2); no antibodies (Lane 3). B, SDS-polyacrylamide gel treated with i N KOH for 2 h, at 55’C. C, phosphoamino acid analysis of connexin43 from SRD3T3 cells. Lines, migration of ninhydrin-stained markers.

Results Phosphorylation of Connexin43 in Normal and v-srctransformed 3T3 Cells. To examine the phosphorylation of connexin43 in normal and v-src-transformed mouse BALB/c3T3 cells (SRD3T3), we immunoprecipitated this protein from lysates of 32P-iabeled cells. The connexin43 protein immunoprecipitated from both normal and transformed cells was radiolabeled with 32P, and migrated as a diffuse M, 43,000-45,000 doublet on SDS-polyacryiamide gels (Fig. lA). Connexin43 immunoprecipitates from SRD3T3 cells displayed greater 32P incorporation than that from normal cells, despite similar amounts of connexin43 protein present in each cell line (as determined by [35S]methionine labeling; data not shown). To determine the alkali stability of the radiolabeled phosphate on connexin43, the same SDS-poiyacryiamide gel was incubated in hot alkali, which selectively hydrolyzes phosphosenine and phosphothreonine residues (Fig. iB) (40). The 32P,-labeled connexin43 from the normal 3T3 cells (Fig. 1 B, Lane 1 ) was completely alkali labile, whereas the 32P,-iabeled connexin43 from the v-src-transformed 3T3 cells (Fig. lB. Lane 2) was partially alkali stable, suggesting that this protein contained phosphotyrosine. Phosphoamino acid analysis, shown in Fig. 1C, confirmed that connexin43 from v-src-transformed 3T3 cells contained phosphotyrosine and indicated that phosphotyrosine represented approximately 10% of total radioiabeied phosphoamino acids. Phosphoamino acid analysis performed on the upper and lower M, 43,000 doublet revealed that both forms of connexin43 from the v-src-transformed cells contained similar amounts of

phosphotyrosine (data not shown). Connexin43 was predominantly phosphorylated on serine residues in both the normal (data not shown) and the v-src-transfonmed cells, with only negligible amounts of phosphothreonine detected in the protein from v-src-transformed cells. No phosphotyrosine was detected on connexin43 from the normal 3T3 cells. These results indicate that connexin43 is phosphorylated on tyrosine in RSV-transformed mouse cells. in addition, there appears to be an increase in senine phosphorylation on connexin43 induced by the expression of the viral src protein. immunoprecipitations and immunoblots performed with antibody against connexin32 indicated that there was no detectable connexin32 present in these cells (data not shown). Immunoprecipitation of a Mr 43,00045,000 Doublet with Phosphotyrosine Antibodies. Using an antibody directed against phosphotyrosine, we observed a Mr 43,000-45,000 doublet that was precipitated from lysates made from 32P,-iabeied SRD3T3 cells (Fig. 2A). This band, which comigrated with the protein precipitated with the connexin43 peptide antisera (Fig. 2A, Lane 5), was not detected in the normal BALB/c3T3 cells (Fig. 2A, Lane 2). Partial proteolytic peptide map analysis revealed that both the upper and the lower bands of the Mr 43,000 45,000 doublet immunoprecipitated with the antibody

to

phosphotyrosine

were

identical

to

the

proteins

immunoprecipitated with the antibodies to C43 (connexin43) (Fig. 2B). Several different monoclonal and polycional antibodies directed against phosphotyrosine were able to immunoreact with connexin43 from v-src-transformed 3T3 cells (data not shown). The ability of the antibodies directed against phosphotyrosine to immunoreact with tyrosine-phosphorylated connexin43 molecules allowed us to examine what portion of the phosphorylated connexin43 population from the v-src-transformed 3T3 cells is phosphorylated on tyrosine. To address this question, a lysate prepared from 32P,-iabeied v-src-transformed 3T3 cells was pre-cleaned of proteins which immunoreact with the phosphotyrosine antibody by performing three consecutive immunoprecipitations with the phosphotyrosine antibody. This preadsonbed lysate was then immunoprecipitated with the C43 antibodies. Little or no connexin43 was immunoprecipitated following this procedure (Fig. 3, Lane 7), suggesting that a majority, if not all, of the phosphorylated connexin43 recognized by the C43 antibodies contains phosphotyrosine. As a control, an identical lysate was subjected to three consecutive pneadsorptions with rabbit anti-mouse antibody and then immunoprecipitated with the C43 antibodies. There was only a minor reduction in the amount of connexin43 remaining in this lysate compared to the total connexin43 immunoprecipitated from a lysate that had not been pre-cleared with the rabbit anti-mouse antibody (Fig. 3, Lanes 6 and 3), indicating that the manipulations involved with this procedure did not cause a severe reduction in the amount of connexin43. in addition, we examined whether there is a population of tyrosine-phosphorylated connexin43 molecules present in the iysate that was not recognized by the C43 antibod ies. Three consecutive immunopreci pitations were performed with the C43 antibodies. This precleared lysate was subsequently immunoprecipitated with a phosphotyrosine antibody. Only a small propontion of connexin43 was precipitated from this pre-cleaned

Cell

Balb/c3T3

A

SRD3T3

II

B

Upper

I

Last IP:

II

cr

0.

Lower

II

a:

>-

>-

I-

I-

Q

Q -

cQ

Cl)

00.00 I

I

I

I

C

C

C

C

C

C

c

(B

Ce

C

(B

(B

First

IP:

RaM

antiPTYR

antiC43

Growth

& Differentiation

anti-

anti-

anti-

anti-

PTYR

PTYR

C43

C43

4

+

663

44

3X

3X

3X

3X

RaM

antiC43

RaM

antiPTYR

1’. -

S-43K

43K

...

p

1 Fig. 2.

2

345

1234

Immunoprecipitation of a M, 43,000-45,000 antibodies. A, immunoprecipitates (Lanes 1 and 2) and SRD3T3 (Lanes 3-5)

photyrosine

c3T3

doublet from

cultures

lysates

with from

labeled

phosBALB/ with 32P

for 2 h. Antibodies used were as follows: preimmune antipeptide serum (Lanes 1 and 3), antibodies to phosphotyrosine (anti-PTYR) (Lanes 2 and 4), antipeptide serum directed against connexin43 (anti-C43) (Lane 5). B, the upper (Lanes 1 and 2) and lower (Lanes 3 and 4) M, 43,000-45,000 bands immunoprecipitated with anti-C43 (Lanes 1 and 3) and phosphotyrosine antibodies (Lanes 2 and 4) were excised from the gel and subjected to limited proteolysis with 8 protease as described in ‘Materials and Methods.’

lysate (Fig. 3, Lane 5). in a control lysate that was preadsorbed with rabbit anti-mouse antibody, there was little on no reduction in the amount of connexin43 protein imrnunoprecipitated with the phosphotyrosine antibody (Fig. 3, Lane 4) compared to the amount of connexin43 irnmunoprecipitated from a lysate that was not precleaned (Fig. 3, Lane 2). These results confirm that the Mr 43,000-45,000 proteins immunoprecipitated with the

phosphotyrosine antibody and the C43 antibodies are identical and demonstrate that only a minor population of tynosmne-phosphorylated connexin43 protein is not recognized by the C43 antibodies. Taken

together,

the

results

each molecule of connexin43, molecules (see “Discussion”), senine is also phosphorylated

shown

in Fig. 3 suggest

that

or complex of connexin43 that is phosphonylated on on tynosine. Since the phos-

1

234567

Fig. 3. Analysis of connexin43 population phosphorylated on tyrosine. Cultures of SRD3T3 cells were labeled with 32P, for 2 h. Lanes 1 -3 contain immunoprecipitates not subjected to preadsorption with antibodies. Antibodies used were as follows: rabbit anti-mouse immunoglobulin (RaM) (Lane 1 ), phosphotyrosine antibodies plus RaM (Lane 2), and anti-C43 (Lane 3). In Lanes 4-7, lysates were preadsorbed 3 times with the following antibodies: RaM (Lanes 4 and 6), anti-C43 (Lane 5), and phosphotyrosine antibodies (Lane 7). Following preadsorption, lysates were immunoprecipitated again with the following antibodies: (Lanes 4 and 5), and anti-C43 (Lanes 6 and

phoamino

demonstrated

acid

analysis

phosphotyrosine

antibodies

7).

of connexin43

shown

a high ratio of phosphosenine

in Fig.

1C

to phospho-

tyrosine, these data suggest that each connexin43 molecule contains multiple senine phosphorylation sites, as well as at least one tyrosine phosphorylation site. Relationship between Tyrosine Phosphorylation of

Connexin43

and Reduction

in Junctional

Communication

in CEFs. Does tynosine phosphorylation of connexin43 correlate with src-induced reduction in junctional cornmunication? To address this point, we examined connexin43 phosphorylation and junctional communi-

664

Tyrosine

Phosphorylaiion

of a Gap

40 0’

A0

Junction

Protein

/O

L,0

B00’

5’-5’

at

.i#{216} -43K

--

.q

-43K

0 U S

0 C 0

123

4

5

S

12345

S

E 0 0.

P-Sercontrol P-Thr

RSV v-src

RSV v-src

2A

RSV

c-src

Fig. 5. Junctional permeability in src-expressing CEFs. Mean incidence of permeable interfaces, normalized to the mean incidence in CEF controls (100%) ± SE (bars). The absolute mean incidence in CEF was 0.65. Statistical significance levels of the differences: RSVv-src or RSVvsrc2A vs. CEF, P 0.000003; RSVv-src2A vs. RSVv-src, P = 0.000003; RSVc-src vs. CEF, P = 0.2 (t test with Bonferroni’s correction for multiple comparisons). Number of independent trials: CEF controls, 34; RSVv-src, 33; RSVv-src2A, 64; RSVc-src, 45.

-

P-Tyr-

v-src

v-snc2A

c-src

0

Fig. 4. Analysis of connexin43 phosphorylation in uninfected CEFs or CEFs expressing variants of the src protein. Uninfected (Lanes 1 and 2) or CEFs infected with RSVc-src (Lane 3), RSVv-src2A (Lane 4), or RSVvsrc (Lane 5) were labeled with 32P for 2 h. Proteins were immunoprecipitated from cell lysates using the following antibodies: no antibodies (Lane 1), and anti-C43 (Lanes 2-5). A, SDS-polyacrylamide gel. B, SDS-polyacrylamide gel treated with 1 N KOH for 2 h at 55’C. C, phosphoamino acid analysis was performed on connexin43 immunoprecipitated from each culture. Equal counts per mm were spotted in each lane. Lines indicate the migration of ninhydrin-stained markers.

cation in CEFs infected with an RSV-derived vector carrying either c-src, wild-type v-src, or v-src2A, a mutant variant of v-src that generates a nontransforming product which contains an alanine substitution at glycine 2 (41, 42). The v-src2A gene product cannot be myristylated and therefore does not associate with the plasma membrane (43). The connexin43 protein expressed in CEFs is 90% homologous to rodent connexin43 and is necognized by the peptide antiserum (38). Figure 4A illustrates that the connexin43 immunoprecipitated from 32P,-iabeled CEFs was labeled in each cell type, with the highest levels of 32P, incorporation detected in the v-src-transformed cells (Fig. 4A, Lane 5). The differences in 32P, labeling of connexin43 in each culture were not attnibutable to different amounts of protein present (as determined by [35S]methionine labeling; data not shown). As shown in Fig. 4B, the phosphorylated connexin43 protein from the v-src-transformed cells was partially alkali stable, suggesting that the connexin43 from these cells contamed phosphotyrosine (Fig. 4B, Lane 5). Trace amounts of alkali-stable connexin43 were detected in CEFs infected with the v-src2A-carrying virus (Fig. 4B, Lane 4), whereas the phosphorylated connexin43 from both the uninfected and c-src virus-infected CEFs was completely alkali labile (Fig. 4B, Lanes 2 and 3).

Phosphoamino acid analysis of connexin43 from v-srctransformed CEFs indicated that this protein is phosphorylated predominantly on serine residues, with phosphotyrosine representing approximately 1 0% of total radiolabeled phosphoamino acids (Fig. 4C). The connexin43 immunoprecipitated from the v-src2A-infected cells contains both phosphoserine and phosphotyrosine, with phosphotyrosine representing only about 1 -2% of total nadiolabeled phosphoamino acids. The connexin43 from both uninfected CEFs and c-src-infected CEFs displayed no detectable phosphotynosine, indicating that the expression of c-src at levels 10-15fold higher than RSVc-src-infected

endogenous CEFs (44)]

levels [as is insufficient

ble tyrosine phosphorylation In parallel experiments, was examined by measuring cent

dye

Lucifer

summarizes tions in the in

the

Yellow

of connexin43. cell-to-cell communication the transfer of the fluores-

across

the

the results obtained various CEF cultures.

v-src-transformed

compared

to

the

is the case for to induce sta-

cell

junctions.

Fig.

in steady-state The junctional

CEFs

was

nontransformed

markedly cell

reduced

counterparts.

reduction in junctional transfer was also observed v-src2A-expressing cells, but this reduction was erably less than cells. junctional not

statistically As in other

brane

different from that of control cell types (45), nonjunctional was

of oncogenes. loss from

Methods”) times

was the

Thus,

duced

by v-src

tions

in junctional

not

the and

detectably

for

over in

v-src2A

expression

permeability.

of

junctional

reductions

by

the

of the rates of “Materials and

periods

which

was

CEFs. cell mem-

altered

Measurements these cells (see

negligible

duration

scored.

A in the consid-

that seen in wild-type v-src-transformed transfer in CEFs overexpressing c-src

permeability

expression fluorescence

5

conditransfer

10

mm,

transfer

junctional

transfer

represent

30 was in-

reduc-

Cell

Discussion

In this report, we have examined the phosphorylation of the major cell-to-cell channel protein, connexin43 or a, in fibroblasts expressing variants of the src gene. Our studies have focused on src-expressing cells because earlier work has shown that cells transformed with v-src exhibit a dramatic reduction in cell-to-cell communication. Using an antipeptide serum directed against connexin43, we have found that tyrosine phosphorylation of connexin43 correlates with the src-induced inhibition in junctional communication. While these studies were in progress, Crow et a!. (46) also reported that connexin43 is phosphorylated on tyrosine in v-src-transformed vole cells. These results provide evidence that tyrosine phosphorylation of connexin43 may be instrumental in regulating junctional permeability in v-src-tnansformed cells. We have found a correlation between tyrosine phosphorylation of connexin43 and loss of junctional cornmunication in chicken cells expressing variants of the src gene. Connexin43 was also phosphorylated on tyrosine in v-src-transformed mouse cells; this phosphorylation correlates with a dramatic reduction in junctional communication [previously reported by Azannia and Loewenstein (9)]. The evidence that this correlation is found in two phyletically distant cell types suggests that tynosine phosphonylation

may

be a general

mode

of cell-to-cell

channel regulation. Down-regulation of junctional permeability and tyrosine phosphorylation of connexin43 occurred in only the RSVv-srcand RSVv-src2A-infected cells which express variants of the src protein with activated kinase activity; no changes in junctional permeability were observed in cells overexpressing c-src. The expression of c-src at levels 10-15-fold higher than endogenous levels was insufficient to cause either detectable tyrosine phosphorylation of connexin43 or a detectable reduction in junctional permeability in CEFs. NIH3T3 cells overexpressing avian c-src exhibit a reduction in permeability (47). it is possible that avian and mammalian cells respond differently to the overexpression of avian c-src. We chose CEFs to study the effect of overexpression of c-src on junctional communication and connexin43 phosphorylation because the levels of connexin43 are very low in NIH3T3 cells (data not shown) (46). The extent of tyrosine phosphorylation of connexin43 correlated with the extent of inhibition of cell-to-cell communication. The connexin43 protein from CEFs expressing pp60 contained at least 5-fold higher levels of tyrosine phosphorylation than cells expressing the nonmynistylated variant, pp6ovsC2A, and displayed significantly greater reduction in junctional permeability than the v-src2A-expressing cells. The exact role of tyrosine phosphorylation

of

connexin43

in

the

v-src-induced

down-regulation of cell-to-cell communication, as well as the role of other phosphorylations of connexin43, cannot be determined at this point. In both mouse and chicken cell transformed by wildtype v-src, there was an increase in the total level of phosphate incorporation in connexin43 that could not be accounted for by the increase in phosphotyrosmne. Since the connexin43 protein from the v-src-tnansformed cells contained only phosphosenine and phosphotyrosine, our results indicate that v-src also induces an in-

Growth

& Differentiation

665

crease in senine phosphorylation of connexin43. The proposed cytoplasmic portion of connexin43 contains 23 senine residues (31) which could potentially serve as phosphoacceptors. Expression of v-src has been shown to activate several senine kinases (including 56 kinase, MAP2 kinase, and potentially the c-raf protein) (48-50) which could contribute to the increased level of senine phosphorylation on connexin43. Phosphopeptide mapping is required to determine the number of additional phosphonylated

senine

residues

and

their

location,

as well

as the sites of tyrosine phosphorylation. With this information, it will be possible to mutate these phosphoacceptor sites to assess the functional significance of these phosphonylation events. The data from the pre-cleaning experiment in Fig. 3 showed that almost the entire population of phosphorylated connexin43 molecules can be cleaned from v-srctransformed 3T3 cell lysates with antibodies to phosphotyrosine. These results suggest that each molecule of phosphate-labeled connexin43 is phosphorylated on tynosine. Alternatively, if connexin43 remains associated in a complex with other connexin molecules in radioimmunoprecipitation assay solubilization buffer, these resuits indicate that at least one connexin43 molecule in the complex is phosphorylated on tyrosine. Although these experiments are complicated by the fact that we used an antipeptide serum (the effectiveness of which could be influenced by any number of posttransiational modifications, including phosphonylation of connexin43), the reciprocal pneadsonption experiment with the C43 antibodies indicated that the C43 antipeptide serum used here

was

able

to

recognize

the

vast

majority

of

connexin43 molecules phosphorylated on tynosine. Funthen evidence that these antibodies recognize the same population of proteins comes from experiments in which immunoblots of connexin43 were probed with antibodies to phosphotynosmne. Connexin43 protein immunoprecipitated from SRD3T3 cell lysates with antibodies to either connexin43 or phosphotyrosmne showed the identical levels of immunoreactivity, suggesting that the connexin43 molecules precipitated by both antibodies contained similar levels of phosphotyrosine (data not shown). We have also found that connexin43, immunoprecipitated with antibodies to connexin43 or phosphotynosine, contained equal amounts of alkali-stable phosphate label (data not shown). Although these data mdicate that the molar amount of phosphotynosine present on the connexin43 isolated with the two different antibodies is the same, we are unable to determine the exact stoichiometry of tynosine phosphorylation, since the pencentage of unphosphonylated connexin43 molecules could not be established in these experiments. The identification of connexin43 as a direct on indirect target of v-src kinase activity raises the question of whether tyrosine phosphorylation of connexin43, and its associated inhibition of junctional communication, is essential for v-src-induced oncogenic transformation. It has been proposed that cell-to-cell communication regulates the intercellular transfer of small growth-regulatory cytoplasmic molecules (2, 51). This suggestion has drawn support from several studies showing a correlation between

the

inhibition

of

junctional

communication

and

the induction of cell proliferation (4, 45, 47, 52). Moreover, treatment with growth factors, such as epidermal and platelet-derived growth factors, causes a rapid ne-

666

Tyrosine

Phosphorylation

of a Gap

Junction

Protein

duction in communication in various cell types (24). The present studies provide evidence for a correlation between the induction of oncogenic transformation, tyrosine phosphorylation of cell-to-cell channel protein, and the reduction in communication. Connexin43 was phosphorylated on tyrosine in both mouse and chicken cells transformed by v-src; in both cell types, v-src induced a dramatic reduction in junctional communication. We have also found that connexin43 is phosphorylated on tyrosine in chicken cells transformed by another oncogene encoding the src-nelated PTK, pp6rs (data not shown). In the experiments reported here, the overexpression of the c-src gene product-which can cause subtle changes in morphology and growth properties of CEFs but does not allow efficient growth of cells in soft agan on the induction of tumors in vivo (53)-did not cause a reduction in junctional communication. In cells expressing the nontransforming v-src variant, pp602”, there was a minor reduction in junctional communication associated with a small amount of tynosine phosphorylation of connexin43. These minimal effects on junctional channels can be attributed to the small population of v-src2A protein that is able to associate with cellular membranes. We have found that approximately 20-30% of this nonmynistylated

mutant

protein

fractionates

with

cellular

membranes under conditions in which cells are disrupted in hypotonic lysis buffers (data not shown). Further analysis of connexin43 phosphorylation in cells expressing conditional mutants of the src protein will be helpful in drawing definitive conclusions about the importance of connexin43 phosphonylation in transformation.

directed ICN.

against

and Methods

Cells, Plasmids, Viruses, and Antibodies. Normal mouse BALB/c3T3 cells and 3T3 cells stably transformed with the Schmidt-Ruppin D strain of RSV (SRD3T3s) (54) were grown in DMEM high glucose (GIBCO) plus 10% calf serum (CIBCO). Chicken embryo fibroblasts were prepared from 1 1 -day-old gs-negative embryos (SPAFAS, Inc., Norwich, CT) and grown at 41 #{176}C in high glucose DMEM plus 1% chicken serum (GIBCO)-5% calf serum (J. R. Scientific). An avian netnovinus vector, RCAN [obtamed from S. Hughes (55)], was modified to carry either the wild-type v-src gene from the SR-A strain of RSV (pRSVv-snc), the avian c-src gene obtained from H. Hanafusa (pRSVc-snc), or a mutant variant of the v-src gene (pRSVv-src2A). pRSVv-src2A was constructed by using a 44-base oligonucleotide provided by R. Clark and F. McCormick (CETUS Corp.) from v-src containing a twobase pair change which altered the second codon from CCC to GCT. in each pRSV construct, the p0! gene from the Bryan strain of RSV was substituted for the SchmidtRuppin p0! gene.5 Experiments using CEFs were performed 2-3 days postinfection, when more than 90% of the cells were infected, as determined by the extent of transformation of the pRSVv-src-infected cells. A polyclonal antipeptide antibody that recognizes amino acids 252-271 of connexin43 was prepared by E. Beyen (37). A monoclonal antibody against connexin32 was a gift of D. Paul (56). A monoclonal antibody, PY2O,

S

L. Fox,

unpublished

results.

was

obtained

from

In Vivo Labeling, Immunoprecipitation, Partial Peptide Mapping, and Phosphoamino Acid Analysis. To assay connexin43 phosphonylation, 3T3 cells and CEFs were labeled with 1 mCi of carrier-free 32P (ICN Pharmaceuticals, Inc., Irvine, CA)/ml in phosphate-free DMEM for 2-3 h. Cells were rinsed twice with STE (150 mti NaC110 mM Tnis, pH 7.2-5 mM EDTA) and lysed in nadioimmunoprecipitation assay buffer (150 mi NaCI-5 mi EDTA-iO mM Tnis, pH 7.2-0.i% SDS-i% sodium deoxycholate-i% Triton X-iOO). Immunoprecipitation using either 1 zi of the connexin antipeptide antisera on 1 zg of PY2O was performed on lysates (normalized by levels of tnichloroacetic acid-precipitable counts) for 2 h. FormaIm-fixed Staphy!ococcus aureus (Pansorbin) was used as an

immunoabsonbent

rabbit PY2O

anti-mouse was used as a second antibody for the immunoprecipitations, as described previously

(57).

One-half

of each

as described

previously

immunoprecipitate

was

(57);

analyzed

on a 10% SDS-polyacrylamide gel (58), destained in 10% methanol-iO% acetic acid, and dried. The other half of the immunoprecipitate was analyzed on an SDS-polyacrylamide gel which was treated with 1 N KOH for 2 h at 55#{176}C and incubated in 10% methanol-iO% acetic acid (40). SDS-polyacrylamide gels were dried and exposed on XAR-5 (Kodak) film with a Cronex Lightning Pius intensifying screen. For partial proteolytic peptide analysis, radiolabeled polyacrylamide

proteins were gel and subjected

excised from an to electrophonesis

SDSin

the presence of 50 ng of Staphy!ococcus V8 protease, as described by Cleveland et a!. (59). Phosphoamino acid analysis

Materials

phosphotynosmne

was

performed

on

32P-Iabeled

connexin43

iso-

lated on an SDS-polyacnylamide gel, as described previously (60). Spots on autonadiograms obtained for phosphoamino acid analysis were quantitated by densitometric scanning with a Molecular Dynamics computing densitometer. Measurement of Cell-to-Cell Communication. Junctional permeability was probed with Lucifer Yellow CH (Molecular Probes, Inc.; 10% aqueous solution). The fluorescent dye was microinjected into the cells with the aid of a micropipet by pneumatic pressure pulses controlled by a solenoid valve. The intracellular and cell-tocell flow of the fluorescence was videoneconded for analysis. cidence

The junctional ol permeab!e

transfer interlaces,

was

indexed

by the

in-

that is, the ratio of firstorder fluorescent neighbors to total first-order neighbors, at 20 5, for each injection (15). The permeability probings were carried out under strictly parallel conditions and with the same batch of serum for each cell type. Thus, the data in Fig. 5 can be directly compared with one another. The cultures were at room temperature during the testing of cell-to-cell transfer. The testing was limited to periods of 30 mm, in order to avoid the effects of the lack of CO2 on junctional permeability. in control cultures, the incidence of permeable interfaces was not affected by this condition over a period of 50 mm. Nonjunctional membrane permeability was measured to ascertain that the changes in this parameter were not responsible for the oncogene-mnduced effects on junctional transfer. Cells without neighbors (sparse cultures) were injected with Lucifer Yellow, and the rate of fluorescent loss was measured by means of a photodiode

Cell

onto which the image of a part of a cell was projected through the microscope (61). Fluorescence was excited with 430-nm light pulses of 5-s duration, 1 pulse/mm.

v-src2A Dunn

gene, Mike for critically

References 1 . Loewenstein, to-cell membrane 2. Loewenstein, growth. Biochim.

W. R. Junctional intercellular communication. channel. Physiol. Rev., 61: 829-913, 1981. W. R. Junctional communication Biophys. Acta, 560: 1-65, 1979.

3. Warner, A. E., Guthrie, S. C., and Gilula, junctional protein selectively disrupt junctional early amphibian embryo. Nature (Lond.), 311:

and

The control

N.

B. Antibodies communication 127-131, 1984.

P. P., Bertram, J. S., and Loewenstein, W. R. Growth of transformed cells correlates with their junctional communication normal cells. Cell, 44: 187-196, 1986. Fraser,

disruption process

S. E., Green,

C.

R.,

Bode,

H.

R., and

of gap junctional communication in hydra. Science (Wash. DC), 237:

6. Lee, S., Gilula,

N. B., and Warner, and compaction during preimplantation Cell, 51: 851-860, 1987. 7. Guihrie, development.

S. C., and Gilula, Trends Neurosci.,

N.

inhibition with

B. Selective

B. Gap junctional 12; 12-16, 1989.

M. M., Menko, A. S., Johnson, J. D. Rapid and reversible reduction infected with a temperature-sensitive J. Cell. Biol., 91: 573-578, 1981.

communication

W. R. Intercellular communication and ofjunctional permeability by the src gene: mutant Rout sarcoma virus. J. Membr.

10. Azarnia, R., and Loewenstein, control of growth. XI. Alteration in a revertant cell with normal 212, 1984.

W. R. Intercellular communication and of junctional permeability by the src gene cytoskeleton. I. Membr. Biol., 82: 207-

12. with

Collett, M. S., and Erikson, the avian sarcoma virus sit

75: 2021-2024, 1 3.

Hunter,

R. L. Protein gene product.

D., and Matsumura, of gap junctional Sci. USA, 82: 5360-

kinase activity associated Proc. Nail. Acad. Sci. USA,

1978. Sefton,

B. M. The

sarcoma virus phosphorylaies 1311-1315, 1980.

T., and

tyrosine.

transforming Proc.

gene NatI.

product

Acad.

Sci.

of Rout USA,

77:

14. Levinson, A. D., Oppermann, H., Levintow, L., Varmus, H. E., and Bishop, J. M. Evidence that the transforming gene of avian sarcoma virus encodes a protein kinase associated with a phosphoprotein. Cell, 15: 561-572, 1978. 1 5. Rose, B., Vada, T., and Loewenstein, W. R. Downregulation of cellto-cell communication by the viral sic gene is blocked by TMB-8 and recovery of communication is blocked by vanadate. J. Membr. Biol., 94: 129-142,

1986.

16. Sefton, B. M., Hunter, T., Beemon, K., and Eckart, the phosphorylation of tyrosine is essential for cellular Rout Sarcoma virus. Cell, 20: 807-816, 1980. 1 7. Azarnia, downregulates

946-950,

R., and Loewenstein, junctional cell-to-cell

W.

R. Polyomavirus communication.

W. Evidence transformation

that by

middle T antigen Mol. Cell. Biol., 7:

1987.

1 8. Courtneidge, S. A., and Smith, protein associates with the product (Lond.), 303: 435-439, 1983.

A. E. Polyoma virus of the c-src cellular

transforming gene. Nature

19. Bolen, J. B., Thiele, C. J., Israel, M. A., Vonemoto, W., Lipsich, L. A., and Brugge, J. S. Enhancement of cellular src gene product associated tyrosyl kinase activity following polyoma virus infection and transformation. Cell, 38: 757-777, 1984. 20. Kornbluth, virus middle 1987.

S., Sudol, M., and Hanafusa, H. Association T antigen with c-yes protein. Nature (Lond.),

Caspar,

D. L., Goodenough,

26. Baker, T. S., Sosinsky, ough, D. A. Gap junction obtained 98, 1985.

of polyoma325: 171-173,

3845-3855,

1988.

D. A.,

W. R. Growth factors J. Membr. Biol., 106:

Makowski,

L. A.,

with

cationic

G. E., Caspar, D. L. D., Gall, structures. VII. Analysis of

and

anionic

negative

and

Phillips,

microscopy C., and connexin

J. Mol.

stains.

Unwin,

P. N. T., and

Paul,

D.

Ennis,

evidence 1983.

L. Molecular

J. Cell.

Biol.,

103:

the

cloning 123-134,

of

low

184:

30. Kumar, N. M., and Gilula, N. human and rat liver cDNAs coding Biol., 103: 767-776, 1986.

for

B. Cloning for a gap

31. Beyer, E., Paul., D. L., and Goodenough, from rat heart homologous to a gap junction Biol.,

105: 2621-2629,

32. Beyer, E. C., family of related Johnson 1988.

(edt.),

35.

Gap

Nature

Gimlich,

opmental Xenopus.

liver

Junctions,

(Lond.),

pp.

732-734,

N. M.,

for 1988.

N. M., and Gilula, junction mRNAs

Musil, 1. 5., Beyer, E. C., and gap junction protein connexin43 cloning, ultrasiructural localization tion. J. Membr. Biol., 1 16: 163-175,

40. Cooper, J. A., and induce phosphorylation Biol.,

42. Kamps, glycine of mation.

NatI.

between gap 331: 721-723, 0., Willecke, of hepatic

K., gap

N. B. Sequence a gap

junction

of develprotein

in

N. B. Differential regulation in Xenopus embryos. J. Cell.

L., and Goodenough, D. A. Antisera react with a 43kd protein localized other tissues. J. Cell. Biol., 108: 595-

of the ASV src cells by electron 1979.

1981. E. A.,

Pellman,

p60” myristylation Mol. Cell. Biol.,

Acad.

R. List,

Goodenough, D. A. Expression of the in embryonic chick lent: molecular and post-translational phosphoryla1990.

M. P., Buss, i. E., and pp60): prevents fatty

Proc.

Alan

Hunter, T. Four different classes of retroviruses of tyrosines present in similar cellular proteins.

1: 394-407,

F. R., Garber,

sequence in the cell transformation.

a protein J. Cell.

connexins: a and R. G.

Vork:

39. Willingham, M. C., Jay, G., and Pastan, I. Localization gene product to the plasma membrane of transformed microscopic immunocytochemistry. Cell, 18: 125-134,

Cell.

of J. Cell.

1987.

of mRNA coding 107: 1065-1073,

38.

New

D., Traub, components

Gilula,

37. Beyer, E. C., Kistler, J., Paul, D. directed against connexin43 peptides to gap junctions in myocardium and 605, 1989.

Cross,

junction

liver.

D. A. The E. L. Hertzberg

S. Homologies Nature (Lond.),

and

36. Gimlich, R. L., Kumar, of the levels of three gap Biol., I 10: 597-605, 1990.

41.

gap

and characterization junction protein.

167-175.

R., Teplow, protein

329:

R. 1., Kumar, expression J. Cell. Biol.,

Mol.

I. Cell.

1987.

Nicholson, B. J., Dermietzel, Revel, J-P. Two homologous

junctions.

rat

-

in gap

pattern.

D. A. Connexin43: protein from

Paul, D. L., and Goodenough, gap junction proteins. In:

33. Kistler, i., Christie, D., and Bullivant, junction proteins in lens, heart and liver. 1988. 34. and

X-ray

81

between

changes

angle

cDNA 1986.

x-ray

Goodenimages

Biol.,

P. D. Calcium-mediated

from

W.

and

27. Unwin, P. N. T., and Zampighi, G. Structure of the junction communicatingcells. Nature (Lond.), 283: 545-549, 1980.

protein.

R., and Loewenstein, control of growth. X. Alteration a study with temperature-sensitive Biol., 82: 191-205, 1984.

J., 7:

EMBO

C. Gap junction structures. I. Correlated electron diffraction. J. Cell. Biol., 74: 605-628, 1977.

29.

and

virus.

24. Maldonado, P. E., Rose, B., and Loewenstein, modulate junctional cell-to-cell communication. 203-210, 1988.

junction structure: Biol., 97: 1459-1466,

R. G., Sheppard, J. R., and of junctional permeability mutant of avian sarcoma

11. Chang, C. C., Trosko, J. E., Kung, H-J., Bombick, F. Potential role of the src gene product in inhibition communication on NIH3T3 cells. Proc. NatI. Acad. 5364, 1985.

of polyoma

Horak, I. D., Kawakami, T., Gregory, F., Robbins, K. C., and Bolen, J. B. Association of p60” with middle tumor antigen in murine polyomavirus-transformed rat cells. 1. Virol., 63: 2343-2347, 1989.

28.

patterning

A. E. Gap junctional communication stages of mouse development.

Sheridan, in cells

9. Azarnia,

N.

interferes with 49-55, 1987.

8. Atkinson,

virus.

Gilula,

of

to gap in the

4. Mehta,

5.

cell-

T antigen

23.

25.

the

667

22. Cheng, S. H., Harvey, R., Espino, P. C., Semba, K., Vamamoto, T., Smith, T. K., and Smith, A. E. Peptide antibodies to the human c-Iyn gene product demonstrate pp58”” is capable of complex formation with the middle

of the Kenneth

& Differentiation

21 . Kypta, R. M., Hemming, A., and Courtneidge, S. A. Identification and characterization of p59” (a src-Iike protein tyrosine kinase) in normal and polyoma virus transformed cells. EMBO J., 7: 3837-3844, 1988.

Acknowledgments We thank Donna Fossaceca for construction DeMarco for technical assistance, and D. reading the manuscript.

Growth

Sci.

USA,

D.,

and

and membrane 4: 1834-1842,

Hanafusa,

H.

association 1984.

A short and

for

Sefton, B. M. Mutation of NH2-ierminal acylation and morphological transfor82: 4625-4628,

1985.

43. Buss, J. E., Kamps, M. P., and Sefton, B. M. Myristic acid is attached to the transforming protein of Rout sarcoma virus during or immediately after synthesis and is present in both soluble and membrane-bound forms of the protein. Mol. Cell. Biol., 4: 2697-2704, 1984. 44. Nemeth, S. P., Fox, within the amino-terminal

L. G., DeMarco, half of the

M., and c-src gene

Brugge, product

J. S. Deletions that alter the

668

Tyrosine

Phosphorylation

functional 45.

activity

Azarnia,

Junctional cogenes

of a Gap

of the

R.,

Mitcho,

M.,

intercellular in transformation.

46. Crow, D. Phosphorylation RSv-transformed 1990.

S.,

protein.

Junction

Mol. Shalloway,

communication Oncogene,

Protein

Cell.

Biol., D.,

9: 1 109-1119,

1989.

Loewenstein,

W.

and

is cooperatively 4: 1161-1168,

inhibited 1989.

A. J., Payne,

D. M.,

Weber,

M.

J.,

and

Evidence that pp42, a major tyrosine kinase target protein, activated serine/threonine protein kinase. Proc. Nail. Acad. 6940-6943, 1989.

T. W.

is a mitogenSci. USA, 86:

50. Morrison, D. K., Kaplan, D. R., Rapp, U., and Roberts, T. M. Signal transduction from membrane to cytoplasm: growth factors and membrane-bound oncogene products increase Raf-1 phosphorylation and associated protein kinase activity. Proc. Nati. Acad. Sci. USA, 85: 88558859, 1988. 51. Acad.

Loewenstein, Sci.,

137:

W.

R. Permeability

441 -472,

of membrane

junctions.

Ann.

NV

1966.

52. Mehta, P. P., Bertram, J. S., and Loewenstein, W. retinoids on cellular growth correlate with their actions communication. J. Cell. Biol., 108: 1053-1065, 1989.

R. The on gap

actions junctional

54. Brugge, J. S., and Darrow, D. Analysis of phosphotransferase activity of two avian sarcoma teins. J. Biol. Chem., 259: 4550-4557, 1984.

the catalytic domain virus transforming

of pro-

J. J., Petropoulos, C. J., and Sutgrave, P. insertion of foreign DNA into helperJ. Virol., 61: 3004-3012, 1987.

56. Goodenough, D. A., Paul, D. L., and Jesaitis, L. A. Topological distribution of two connexin32 antigenic sites in intact and split rodent hepatocyte gap junctions. J. Cell. Biol., 107: 1817-1824, 1988.

56 kinase or phorbol

Sturgill,

53. Iba, H., Takeya, T., Cross, F., Hanafusa, T., and Hanafusa, H. Rout sarcoma virus variants which carry the cellular src gene instead of the viral src gene cannot transform chicken embryo fibroblasts. Proc. Nail. Acad. Sci. USA, 81: 4424-4428, 1984.

55. Hughes, S. H., Greenhouse, Adaptor plasmids simplify the independent retroviral vectors.

S., Kmiecik, T. E., Shalloway, D., and Loewensrc gene product regulates junctional cell-to-cell (Wash. DC), 239: 398-401, 1988.

48. Blenis, J., and Erikson, R. Regulation of a ribosomal protein activity by the Rout sarcoma virus transforming protein, serum, ester. Proc. Nail. Acad. Sci. USA, 82: 7621-7625, 1986. Rossomando,

by on-

Beyer, E. C., Paul, D. L., Kobe, S. S., and Lau, A. of connexin43 gap junction protein in uninfecied and mammalian fibroblasts. Mol. Cell. Biol., 10: 1754-1763,

47. Azarnia, R., Reddy, stein, W. R. The cellular communication. Science

49.

R.

of

57. Lipsich, L. A., Lewis, A. J., and Brugge, antibodies that recognize the transforming viruses. J. virol., 48: 352-360, 1983. 58. Laemmli, of the head

U. K. Cleavage of bacteriophage

59. Cleveland, D. W., Fischer, K. Peptide mapping by limited analysis by gel electrophoresis. 60. Brugge, I. S., and Darrow, ylation of a 50,000 molecular 250-253, 1982. 61 . Flagg-Newton, and cyclic AMP.

J. L., DahI, I. Upregulation

of structural T4. Nature

J. S. Isolation proteins of

of monoclonal avian sarcoma

proteins during the (Lond.), 227: 680-685,

S. G., Kirschner, M. proteolysis in sodium I. Biol. Chem., 252:

W., and Laemmli, U. dodecyl sulfate and 1 102-1 106, 1977.

D. Rout sarcoma virus-induced weight cellular protein. Nature G.,

and Loewenstein, W. of junctional membrane

and junctional membrane particles by administration or phosphodiesterase inhibitor. J. Membr. Biol., 63:

assembly 1970.

phosphor(Lond.), 295: R. Cell junction permeability

of cyclic 105-121,

nucleotide 1981.