Search

A previously isolated endocytic trafficking mutant (TRF1) isolated from HuH-7 cells is defective in the distribution of subpopulations of cell-surface receptors for asialoorosomucoid (asialoglycoprotein receptor (ASGR)), transferrin, and mannose-terminating glycoproteins. The pleiotropic phenotype of TRF1 also includes an increased sensitivity to Pseudomonas toxin and deficient assembly and function of gap junctions. HuH-7xTRF1 hybrids exhibited a normal subcellular distribution of ASGR, consistent with the TRF1 mutation being recessive. A cDNA expression library derived from HuH-7 mRNA was transfected into TRF1 cells, which were subsequently selected for resistance to Pseudomonas toxin. Sequence analysis of a recovered cDNA revealed a unique isoform of casein kinase 2 (CK2), CK2alpha". Western blot analysis of TRF1 proteins revealed a 60% reduction in total CK2alpha expression. Consistent with this finding, the hybrids HuH-7xHuH-7 and HuH-7xTRF1 expressed equivalent amounts of total CK2alpha. Immunoblots using antibodies against peptides unique to the previously described CK2 isoforms CK2alpha and CK2alpha' and the novel CK2alpha" isoform showed that, although TRF1 and parental HuH-7 cells expressed comparable amounts of CK2alpha and CK2alpha', the mutant did not express CK2alpha". Based on the genomic DNA sequence, RNA transcripts encoding CK2alpha" apparently originate from alternative splicing of a primary transcript. Protein overexpression following transfection of TRF1 cells with cDNAs encoding either CK2alpha or the newly cloned CK2alpha" restored the parental HuH-7 phenotype, including Pseudomonas toxin resistance, cell-surface ASGR binding activity, phosphorylation, and the assembly of gap junctions. This study suggests that HuH-7 cells express at least three CK2alpha isoforms and that the pleiotropic TRF1 phenotype is a consequence of a reduction in total CK2 expression.

Vertebrate tissues use multiple junctional types to establish and maintain tissue architecture, including gap junctions for cytoplasmic connectivity and tight junctions (TJs) for paracellular and/or cell polarity barriers. The integral membrane proteins of gap junctions are connexins, whereas TJs are a complex between occludin and members of a recently characterized multigene family, the claudins. In normal brain, astrocytes are coupled by gap junctions composed primarily of connexin43 (Cx43), whereas TJs have not been detected in these cells. We now show that treatment of primary human astrocytes with the cytokine interleukin-1beta (IL-1beta) causes rapid induction of claudin-1, with an expression pattern reciprocal to loss of Cx43. Treatment also led to protracted downregulation of occludin but no change in expression of zonula occludens proteins ZO-1 and -2. Immunofluorescence staining localized claudin-1 to cell membranes in IL-1beta-treated astrocytes, whereas freeze-fracture replicas showed strand-like arrays of intramembranous particles in treated cells resembling rudimentary TJ assemblies. We conclude that in human astrocytes, IL-1beta regulates expression of the claudin multigene family and that gap and tight junction proteins are inversely regulated by this proinflammatory cytokine. We suggest that in pathological conditions of the human CNS, elevated IL-1beta expression fundamentally alters astrocyte-to-astrocyte connectivity.

Small gap junction plaques are associated with tight junction strands in some cell types including hepatocytes and it is thought that they may be closely related to tight junctions and the establishment of cell polarity. In order to examine roles of gap junctions in regulating expression and structure of tight junctions, we transfected human Cx32 cDNA into immortalized mouse hepatocytes (CHST8 cells) which lack endogenous Cx32 and Cx26. Immunocytochemistry revealed that endogenous integral tight junction protein occludin was strongly localized and was colocalized with Cx32 at cell borders in transfectants, whereas neither was detected in parental cells. In Northern blots, mRNAs encoding occludin and the other integral tight junction proteins, claudin-1 and -2, were induced in the transfectants compared to parental cells. In Western blots, occludin protein was increased in the transfectants compared to parental cells, and binding of occludin to Cx32 protein was demonstrated by immunoprecipitation. In freeze fracture of the transfectants, tight junction strands were more numerous and complex compared to parental cells, and small gap junction plaques appeared within induced tight junction strands. Nevertheless, no change in barrier function of tight junctions was observed. These results indicate that in hepatocytes, gap junction, and tight junction expression are closely coordinated, and that Cx32 may play a role in regulating occludin expression., (Copyright 1999 Academic Press.)

The Trf1 cell line, selected from the human hepatoma cell line HuH-7, manifests altered trafficking of various plasma membrane proteins. In particular, there is a striking loss of State 2 asialoglycoprotein receptors. This cell line is shown here to also manifest defects in function and assembly of gap junctions comprising connexin43 (Cx43). No alteration of Cx43 expression or phosphorylation was apparent. Nevertheless, immunostaining of Cx43 revealed that fewer and smaller gap junctions were present at appositional membrane areas in Trf1 cells as compared with parental HuH-7. This correlated with a significant attenuation in gap junction-mediated communication between Trf1 cells as demonstrated by markedly decreased dye transfer and their reduced ability to propagate mechanically evoked Ca(2+) waves. Isoelectric focusing (IEF) of Cx43 in HuH-7 cells indicated that the pIs of this protein were significantly lower than that predicted from its amino acid sequence; no differences in pI were evident in Cx43 from Trf1 cells and the HuH-7 cell line. The effects of the Trf1 mutation on assembly and function of gap junctions indicate that this mutation influences trafficking of Cx43. Connexins differ in several respects from other membrane proteins thus far analyzed in Trf1 mutants: gap junctions localize exclusively to the lateral cell surface; they are not glycoproteins; and they do not play a role in endocytic pathways. The disruption of trafficking of Cx43 by this mutation suggests that the Trf1 phenotype is a defect at a common point along the trafficking pathway of cell-surface proteins, irrespective of their ultimate destination on the cell surface or their glycosylation profile.

We have cloned a protein from bovine brain, brain-derived voltage-dependent anion channel 1 (BR1-VDAC), that is identical to a recently sequenced plasmalemmal-bound porin from human lymphocytes. mRNA hybridization indicates that BR1-VDAC is widely distributed throughout nervous and nonnervous tissues. In situ localization substantiated that the BR1-VDAC is associated with the plasmalemma of astrocytes. A monoclonal antibody that recognizes the N terminus of the BR1-VDAC protein completely blocks an astrocytic high-conductance anion channel that has electrophysiological similarities with the mitochondrial VDAC. Since the high-conductance anion channel in astrocytes has been shown to respond to hypoosmotic solutions, its molecular identification provides the basis for a better understanding of volume regulation in brain tissue.

Despite sparse autonomic innervation, the smooth muscle cells of the corpus cavernosum relax and contract synchronously to achieve penile erection and flaccidity. As with other smooth muscle cell types, the excitation process in the corpora is presumably propagated through gap junctions to allow the diffusion of current-carrying ions and second messenger molecules from cell to cell. Using both molecular and immunocytochemical techniques, we have identified gap junctions between human corporal smooth muscle cells in situ and in culture. Northern analyses demonstrated that corporal smooth muscle cells express the gap junction protein connexin43, but not connexin26 mRNA. Immunoblots showed the presence of connexin43 isoforms, whereas connexin32 was not detected. Immunocytochemical studies in cultured cells identified prominent connexin43 immunoreactive puncta between cells, as well as within the cytoplasm. In addition, gap junction membranes both in situ and in culture were labelled in thin section by anti-connexin43 antibodies using the immunogold technique. We conclude that the presence and distribution of gap junctions in this sparsely innervated tissue may provide an important mechanism of intercellular communication among the smooth muscle cells, and thus play a major role in coordinating tissue contraction and relaxation.

Distinct patterns of expression of gap junction, or connexin, mRNAs were observed in periportal vs. pericentral hepatocytes. The two cellular fractions (isolated from rat livers by perfusion) were more than 90% parenchymal, as determined by flow cytometry for a hepatocyte-specific marker. The periportal and pericentral fractions were identifiable due to enrichment in enzymatic activities previously shown to be differentially expressed in the respective regions of liver. Northern blot analyses revealed that mRNA encoding connexin 26 was 2.8 times more abundant in the periportal than in the pericentral cells, while connexin 32 mRNA was equally distributed. Messenger RNA from each fraction was radiolabeled in order to compare the relative abundance of the connexin mRNAs in each fraction. The ratio of connexin 26 to connexin 32 mRNA in the portal fraction was about 0.085, and in the central fraction about 0.038. Connexin 26 mRNA was transcribed, however, at a faster rate than connexin 32 mRNA by nuclei isolated from both cellular fractions. Connexin 26 mRNA was transcribed at 3.9 times the rate in nuclei from the periportal than from the pericentral cells. These data suggest that while the zonation of connexin 26 mRNA synthesis in liver appears to be controlled transcriptionally, posttranscriptional regulatory mechanisms determine the relative abundance of the connexin mRNAs.

Gap junction channels provide a pathway for exchange of ions and small molecules between coupled cells, and this exchange is believed to be critical for normal tissue growth and development. As a test for a role of gap junction-mediated intercellular communication in control of cell growth, we have compared growth rates of communication-deficient human tumor cells (SKHep1) with clones stably transfected with cDNA encoding the rat liver gap junction protein connexin 32. In culture, growth rates for parental and transfected clones were similar. However, when sizes of tumors were evaluated following injection of these clones into athymic nude mice, growth rates for two well-coupled clones were significantly lower than for communication-deficient or poorly coupled clones. This study demonstrates that growth rate of these tumor cells in situ is negatively correlated with strength of intercellular communication.

Pairs of SKHep1 cells, which are derived from a highly metastatic human hepatoma, were studied using the whole cell voltage clamp technique with patch-type electrodes containing CsCl as the major ionic species. In 12 of 81 cell pairs, current flow through junctional membranes was detectable; in the remaining 69 cell pairs, junctional conductance was less than the noise limit of our recording apparatus (worst case: 10 pS). Macroscopic junctional conductance (gj) in the small percentage of pairs where it was detectable ranged from 100 to 600 pS. Unitary junctional conductance (gamma j) determined in the lowest conductance pairs or after reducing conductance with a short exposure to the uncoupling agent halothane was 25-35 pS. To study properties of gap junction channels formed of connexin32, the parental SKHep1 cell line was stably transfected with a plasmid containing cDNA that encodes connexin32, the major gap junction protein of rat liver cells. In 85 of 98 pairs of voltage clamped connexin32-transfected SKHep1 cells, macroscopic gj was greater than 1 nS; gj increased with time after dissociation (from 1.8 +/- 0.6 [mean +/- SE; n = 7] nS at 2 h after plating to 9.3 +/- 2.2 [n = 9] nS, the maximal value, at 24 h). Unitary conductance of gap junction channels between pairs of transfected SKHep1 cells was measured in low conductance pairs and after reducing gj by exposure to halothane or heptanol. Histograms of gamma j values in transfected cells, in 10 experiments where greater than 100 transitions were measurable, displayed two peaks; 120-130 pS and 25-35 pS. The smaller size corresponded to channels that were occasionally detected in the parental cells. We therefore conclude that connexin32 forms gap junctions channels of the 120-130 pS size class.

Communication-deficient cells (the SKHep1 cell line) were stably transfected with a plasmid containing cDNA which encodes the major gap junction protein of rat liver, connexin32. Application of the dual whole-cell voltage clamp technique with patch electrodes to pairs of transfected SKHep1 cells revealed strong sensitivity of junctional conductance (gj) to transjunctional voltages (Vjs) of either polarity, with the ratio of minimal to maximal gj (gmin/gmax) being approximately 0.1 at the highest Vjs. Steady-state gj values as a function of voltages of either polarity were well fit by the Boltzmann equation. V0, the voltage at which gj was reduced by 50%, was approximately 25-30 mV; A, the Boltzmann parameter describing voltage dependence, was approximately 0.06 (corresponding to an energy difference between states of approximately 1 kCal/mol and to approximately 2 gating charges moving through the field). The kinetics of the transjunctional voltage dependence were slow (tau greater than 5 s at 20-40 mV, tau = 2 s at and beyond 70 mV). Voltage sensitivity of the opening rate constant (alpha) was approximately 30% lower than that of the closing rate constant (beta) over the Vj range 0-70 mV; at higher voltages, voltage sensitivity of alpha and beta saturated. The kinetic response of gj to a paradigm in which gj was first rendered low by a prepulse of opposite polarity indicated that the voltage sensors are likely to be arranged in series. Transitions between open and closed states in response to transjunctional voltages of either polarity are single order processes; transitions from one closed state to the other involve passage through the open state.

The connexins form a family of membrane spanning proteins that assemble into gap junction channels. The biophysical properties of these channels are dependent upon the constituent connexin isoform. To begin identifying the molecular basis for gap junction channel behavior in the human heart, a tissue that expresses connexin43, we used site-directed mutagenesis to generate mutant cDNAs of human connexin43 with shortened cytoplasmic tail domains. Premature stop codons were inserted, resulting in proteins corresponding in length to the mammalian isoforms connexin32 and connexin26, which are expressed primarily in liver. All constructs restore intercellular coupling when they are transfected into SKHep1 cells, a human hepatoma line that is communication deficient. Whereas wild-type connexin43 transfectants display two distinct unitary conductance values of about 60 and 100 pS, transfectants expressing the mutant proteins, from which 80 and 138 amino acids have been deleted, exhibit markedly different single-channel properties, with unitary conductance values of about 160 and 50 pS, respectively. Junctional conductance of channels composed of wild-type connexin43 is less voltage-sensitive compared with transfectants expressing wild-type connexin32. However, neither of the connexin43 truncation mutants alters this relative voltage insensitivity. These results suggest that the cytoplasmic tail domain is an important determinant of the unitary conductance event of gap junction channels but not their voltage dependence. Furthermore, since the mutant connexins are missing several consensus phosphorylation sites, modification of these particular sites may not be required for membrane insertion or assembly of human connexin43 into functional channels.

The properties of astroglial gap junction channels and the protein that constitutes the channels were characterized by immunocytochemical, molecular biological, and physiological techniques. Comparative immunocytochemical labeling utilizing different antibodies specific for liver connexin 32 and connexin 26 and antibodies to peptides corresponding to carboxy-terminal sequences of the heart gap junction protein (connexin 43) indicates that the predominant gap junction protein in astrocytes is connexin 43. The expression of this connexin in cultured astrocytes was also established by Western and Northern blot analyses. Cultured astrocytes expressed connexin 43 mRNA and did not contain detectable levels of the mRNAs encoding connexin 32 or connexin 26. Further, the cells contained the same primary connexin 43 translation product and the same phosphorylated forms as heart. Finally, electrophysiological recordings under voltage-clamp conditions revealed that astrocyte cell pairs were moderately well coupled, with an average junctional conductance of about 13 nS. Single-channel recordings indicated a unitary junctional conductance of about 50-60 pS, which is of the same order as that found in cultured rat cardiac myocytes, where the channel properties of connexin 43 were first described. Thus, physiological properties of gap junction channels appear to be determined by the connexin expressed, independent of the tissue type.

We report here experiments undertaken in pairs of hepatocytes that demonstrate a marked voltage sensivity of junctional conductance and, thus, contradict earlier findings reported by this laboratory (Spray, D.C., R.D.ginzberg, E.A., E. A. Morales, Z. Gatmaitan and I.M. Arias, 1986, J. Cell Biol. 101:135-144; Spray C.D. R.L. White, A.C. Campos de Carvalho, and M.V.L. Bennett. 1984. Biophys. J. 45:219-230) and by others (Dahl, G., T. Moller, D. Paul, R. Voellmy, and R. Werner. 1987. Science [Wash. DC] 236:1290-1293; Riverdin, E.C., and R. Weingart. 1988. Am. J. Physiol. 254:C226-C234). Expression in exogenous systems, lipid bilayers in which fragments of isolated gap junction membranes were incorporated (Young, J.D.-E., Z. Cohn, and N.B. Gilula. 1987. Cell. 48:733-743.) and noncommunicating cells transfected with connexin32 cDNA (Eghbali, B., J.A. Kessler, and D.C. Spray. 1990. Proc. Natl. Acad. Sci. USA. 87:1328-1331), support these findings and indicate that the voltage-dependent channel is composed of connexin32, the major gap junction protein of rat liver (Paul, D. 1986. J. Cell Biol. 103:123-134).

Connexin43 is the predominant gap junction protein expressed in the heart. To determine the relation between cardiac maturation and gap junction gene expression, the developmental profiles of connexin43 mRNA and protein were examined in the rat heart. Connexin43 mRNA levels accumulate progressively (eightfold) during embryonic and early neonatal stages, accompanied by a parallel, but temporally delayed, accumulation of connexin43 protein (15-fold). As the heart matures further, both mRNA and protein levels subsequently decline, to about 50% and 30% of their maximum levels, respectively. These observations suggest that increases in intercellular coupling that characterize cardiac development do not depend solely on modulation of connexin43 gene expression, but rather are likely to involve organization of gap junction channels into the intercalated disc.

1. Channels from isolated liver junctional membranes were incorporated into lipid bilayers and studied under voltage clamp conditions. Detergent treatment of junctional membrane fragments greatly increased the incidence of channel incorporation but did not noticeably alter the properties of the incorporated channels. Incorporation resulted in channel activity displaying an approximately symmetric voltage dependence in which conductance was decreased with imposed transmembrane voltages exceeding +/- 20 mV. A residual voltage-independent conductance was also detected in membranes in which liver junctional membranes were incorporated. The magnitude of this voltage-insensitive component varied from less than 20% to more than 75% of the total conductance. 2. These results are generally similar to those described by Young, Cohn and Gilula (Cell, 48: 733-743, 1987) in incorporation experiments following detergent treatment of isolated gap junction membranes. However, we interpret these data as indicating the existence of distinct channel populations in the incorporated membrane fractions. Our results suggest that a population of larger conductance channels (greater than or equal to 150 pS) contributes the voltage-dependent component of the membrane conductance, while smaller channels (unitary conductance about 50-150 pS) contribute the voltage-independent component. The biophysical properties of the larger channel are comparable to those seen in communication-deficient cells transfected with connexin32, confirming a report describing conductance of bilayers in which electroeluted 27-kDa liver gap junction protein was inserted. 3. These findings indicate that connexin32 comprises the larger, voltage-dependent channels seen in the bilayer experiments in which liver junctional membranes are incorporated.

Phosphorylation of connexin 32, the major liver gap-junction protein, was studied in purified liver gap junctions and in hepatocytes. In isolated gap junctions, connexin 32 was phosphorylated by cAMP-dependent protein kinase (cAMP-PK), by protein kinase C (PKC) and by Ca2+/calmodulin-dependent protein kinase II (Ca2+/CaM-PK II). Connexin 26 was not phosphorylated by these three protein kinases. Phosphopeptide mapping of connexin 32 demonstrated that cAMP-PK and PKC primarily phosphorylated a seryl residue in a peptide termed peptide 1. PKC also phosphorylated seryl residues in additional peptides. CA2+/CaM-PK II phosphorylated serine and to a lesser extent, threonine, at sites different from those phosphorylated by the other two protein kinases. A synthetic peptide PSRKGSGFGHRL-amine (residues 228-239 based on the deduced amino acid sequence of rat connexin 32) was phosphorylated by cAMP-PK and by PKC, with kinetic properties being similar to those for other physiological substrates phosphorylated by these enzymes. Ca2+/CaM-PK II did not phosphorylate the peptide. Phosphopeptide mapping and amino acid sequencing of the phosphorylated synthetic peptide indicated that Ser233 of connexin 32 was present in peptide 1 and was phosphorylated by cAMP-PK or by PKC. In hepatocytes labeled with [32P]orthophosphoric acid, treatment with forskolin or 20-deoxy-20-oxophorbol 12,13-dibutyrate (PDBt) resulted in increased 32P-incorporation into connexin 32. Phosphopeptide mapping and phosphoamino acid analysis showed that a seryl residue in peptide 1 was most prominently phosphorylated under basal conditions. Treatment with forskolin or PDBt stimulated the phosphorylation of peptide 1. PDBt treatment also increased the phosphorylation of seryl residues in several other peptides. PDBt did not affect the cAMP-PK activity in hepatocytes. It has previously been shown that phorbol ester reduces dye coupling in several cell types, however in rat hepatocytes, dye coupling was not reduced by treatment with PDBt. Thus, activation of PKC may have differential effects on junctional permeability in different cell types; one source of this variability may be differences in the sites of phosphorylation in different gap-junction proteins.

Gap junctions permit the passage of ions and chemical mediators from cell to cell. To identify the molecular genetic basis for this coupling in the human heart, we have isolated clones from a human fetal cardiac cDNA library which encode the full-length human cardiac gap junction (HCGJ) mRNA. The predicted amino acid sequence is homologous to the rat cardiac gap junction protein, connexin43 (Beyer, E. D., D. Paul, and D. A. Goodenough. 1987. J. Cell Biol. 105:2621-2629), differing by 9 of 382 amino acids. HCGJ mRNA is detected as early as fetal week 15 and persists in adult human cardiac samples. Genomic DNA analysis suggests the presence of two highly homologous HCGJ loci, only one of which is functional. Stable transfection of the HCGJ cDNA into SKHep1 cells, a human hepatoma line which is communication deficient, leads to the formation of functional channels. Junctional conductance in pairs of transfectants containing 10 copies of the HCGJ sequence is high (approximately 20 nS). Single channel currents are detectable in this expression system and correspond to conductances of approximately 60 pS. These first measurements of the HCGJ channel are similar to the junctional conductance recorded between pairs of rat or guinea pig cardiocytes.

The gene family encoding gap junction proteins (connexins) consists of several known members, and multiple connexins are frequently coexpressed by coupled cells. To characterize the channel properties of the major rat liver gap junction protein (connexin 32) in isolation from other gap junction proteins, we have introduced the cDNA encoding it into a human hepatoma cell line (SKHep1) in which we have identified a gap junction deficiency. In this cell line, dye coupling was absent and junctional conductance was near zero. Connexins and connexin 32 mRNA were not detectable by immunocytochemistry and Northern blot analysis. After transfection and selection, cells were strongly coupled with regard to dye and electrical current, and connexin 32 mRNA and punctate connexin 32-immunoreactive membrane contacts were abundant. Functional gap junction channels were still expressed after 19 passages of the cells, indicating stable transfection. When junctional conductance was rendered reversibly low by exposing the cells to agents that uncouple other cell types, currents through single gap junction channels could be observed. The unitary conductance of these expressed channels was about 120-150 pS, a value that is distinctly larger than in heart cells, which express a different gap junction protein.

Gap junctions are aggregates of transmembranous channels which bypass the extracellular space by transporting messenger molecules and ions from one cytoplasmic source to an adjacent cytoplasmic interior. The channels join the plasma membranes of adjacent cells by bridging the extracellular space between them. Thereby, cellular "compartments" which were once considered to be individual units are, in actuality, interconnected by a system of pathways which form a functional cellular syncytium. The evolutionary importance of a generalized intercellular communication system can be appreciated when one considers the widespread prevalence of gap junctions within animals of all multicellular phyla, and within almost all tissues of vertebrates. Only a few population of cells such as skeletal muscle cells (which are fused to form functional syncytia) and circulating blood cells are not equipped with gap junctions. This paper provides a brief review of the diverse structural, molecular and functional aspects of gap junctions as revealed by current research.

By using antibodies directed against gap junction proteins of liver (connexins 26 and 32) and heart (connexin 43), we have localized immunoreactivity to specific cell types in frozen sections of adult rodent brains. Connexin 32 reactivity was found in oligodendrocytes and also in a few neurons, whereas reactivity to connexins 26 and 43 was localized to leptomeningeal cells, ependymal cells, and pineal gland. Immunoreactivity with antibodies to connexin 43 also occurred in astrocytes. Furthermore, during embryonic and postnatal maturation of brain tissues, gap junction proteins were differentially expressed. Connexins 43 and 26 predominated in the neuroepithelium of embryonic brains, whereas connexin 32 was virtually absent. Between 3 and 6 weeks after birth, connexin 26 largely disappeared from immature brain; this time course corresponded to the increased expression of connexin 32. Expression of connexin 43 remained high throughout embryonic and postnatal development. These findings demonstrate that gap junction expression in the brain is diverse, with specific cell types expressing different connexins; this cell-specific distribution may imply differences in the function of these intercellular channels in different loci and developmental stages.

Membrane-permeant cAMP derivatives (dibutyryl- and 8-bromo-cAMP) increase gap-junctional conductance within minutes when applied to voltage-clamped pairs of rat hepatocytes. Glucagon also increases junctional conductances, but the response has a more rapid onset and is more rapidly reversible. The glucagon effect can be prevented by intracellular injection of the protein inhibitor of the cAMP-dependent protein kinase (Walsh inhibitor), indicating that the catalytic subunit of cAMP-dependent protein kinase is directly involved. The 27-kDa major gap junction polypeptide is phosphorylated when liver cells dissociated into small groups are incubated with 32P. Addition of 8-bromo-cAMP to cells increases the incorporation of 32P into the 27-kDa junctional protein. Serine is the amino acid residue that is phosphorylated. When isolated liver gap junctions are incubated in the presence of catalytic subunit of the cAMP-dependent protein kinase, the 27-kDa gap junction polypeptide is phosphorylated with low stoichiometry on serine. The rapid increases in gap junctional conductance caused by agents that elevate cAMP and phosphorylation of the gap junction protein by cAMP-dependent protein kinase suggest that cAMP-dependent phosphorylation of the gap junction channel modulates the conductance of liver gap junctions.

Antibody raised against isolated rat liver gap junctions was microinjected into coupled cells in culture to assess its influence on gap junctional conductance. A rapid inhibition of fluorescent dye transfer and electrical coupling was produced in pairs of freshly dissociated adult rat hepatocytes and myocardial cells as well as in pairs of superior cervical ganglion neurons from neonatal rats cultured under conditions in which electrotonic synapses form. The antibodies have been shown by indirect immunofluorescence to bind to punctate regions of the plasma membrane in liver. By immunoreplica analysis of rat liver homogenates, plasma membranes, and isolated gap junctions resolved on NaDodSO4/polyacrylamide gels, binding was shown to be specific for the 27-kDa major polypeptide of gap junctions. This and similar antibodies should provide a tool for further investigation of the role of cell-cell communication mediated by gap junctions and indicate that immunologically similar polypeptides comprise gap junctions in adult mammalian cells derived from all three germ layers.

Intercellular communication via gap junctions, as measured by dye and electrical coupling, disappears within 12 h in primary rat hepatocytes cultured in serum-supplemented media or within 24 h in cells in a serum-free, hormonally defined medium (HDM) designed for hepatocytes. Glucagon and linoleic acid/BSA were the primary factors in the HDM responsible for the extended life span of the electrical coupling. After 24 h of culture, no hormone or growth factor tested could restore the expression of gap junctions. After 4-5 d of culture, the incidence of coupling was undetectable in a serum-supplemented medium and was only 4-5% in HDM alone. However, treatment with glycosaminoglycans or proteoglycans of 24-h cultures, having no detectable gap junction protein, resulted in synthesis of gap junction protein and of reexpression of electrical and dye coupling within 48 h. Most glycosaminoglycans were inactive (heparan sulfates, chondroitin-6 sulfates) or only weakly active (dermatan sulfates, chondroitin 4-sulfates, hyaluronates), the weakly active group increasing the incidence of coupling to 10-30% with the addition of 50-100 micrograms/ml of the factor. Treatment of the cells with 50-100 micrograms/ml of heparins derived from lung or intestine resulted in cells with intermediate levels of coupling (30-50%). By contrast, 10-20 micrograms/ml of chondroitin sulfate proteoglycan, dermatan sulfate proteoglycan, or liver-derived heparin resulted in dye coupling in 80-100% of the cells, with numerous cells showing dye spread from a single injected cell. Sulfated polysaccharides of glucose (dextran sulfates) or of galactose (carrageenans) were inactive or only weakly active except for lambda-carrageenan, which induced up to 70% coupling (albeit no multiple coupling in the cultures). The abundance of mRNA (Northern blots) encoding gap junction protein and the amounts of the 27-kD gap junction polypeptide (Western blots) correlated with the degree of electrical and dye coupling indicating that the active glycosaminoglycans and proteoglycans are inducing synthesis and expression of gap junctions. Thus, proteoglycans and glycosaminoglycans, especially those found in abundance in the extracellular matrix of liver cells, are important in the regulation of expression of gap junctions and, thereby, in the regulation of intercellular communication in the liver. The relative potencies of heparins from different tissue sources at inducing gap junction expression are suggestive of functional tissue specificity for these glycosaminoglycans.