The E368Q Mutant Allele of GJA8 is Associated with Congenital Cataracts with Intrafamilial Variation in a South Indian Family.

PURPOSE: To determine the basis of the autosomal dominant congenital cataracts in a three generation south Indian pedigree. METHODS: The proband and several family members underwent a complete ophthalmic examination. The coding regions of eight candidate genes (CRYAA, CRYBB2, CRYGC, CRYGD, GJA3, GJA8, AQP0, and PITX3) were amplified by PCR and directly sequenced. Wild type and mutant connexin50 (Cx50) were expressed by stable transfection of HeLa cells. Their cellular distributions and function were examined by immunofluorescence microscopy and by microinjection of gap junction permeant tracers, respectively. RESULTS: Congenital cataracts (with some variations in phenotype) segregated as an autosomal dominant trait within a three generation pedigree. Three affected individuals (proband, sibling and mother) showed a sequence variation in the candidate gene GJA8 encoding Cx50: a c.1102G>C transversion encoding a substitution of glutamate for glutamine at position 368 (E368Q). This substitution was absent from an unaffected family member (paternal aunt) and 100 healthy controls of the same ethnicity. In transfected HeLa cells, both wild type Cx50 and E368Q localized to gap junction plaques, and supported similar levels of intercellular transfer of Neurobiotin. CONCLUSIONS: The E368Q mutant allele of GJA8 is associated with autosomal dominant congenital cataracts with phenotypic variability. E368Q forms gap junction plaques and functional channels in transfected HeLa cells.

A novel connexin50 mutation associated with congenital nuclear pulverulent cataracts.

PURPOSE: To screen for mutations of connexin50 (Cx50)/GJA8 in a panel of patients with inherited cataract and to determine the cellular and functional consequences of the identified mutation. METHODS: All patients in the study underwent a full clinical examination and leucocyte DNA was extracted from venous blood. The GJA8 gene was sequenced directly. Connexin function and cellular trafficking were examined by expression in Xenopus oocytes and HeLa cells. RESULTS: Screening of the GJA8 gene identified a 139 G to A transition that resulted in the replacement of aspartic acid by asparagine (D47N) in the coding region of Cx50. This change co-segregated with cataract among affected members of a family with autosomal dominant nuclear pulverulent cataracts. While pairs of Xenopus oocytes injected with wild type Cx50 RNA formed functional gap junction channels, pairs of oocytes injected with Cx50D47N showed no detectable intercellular conductance. Co-expression of Cx50D47N did not inhibit gap junctional conductance of wild type Cx50. In transiently transfected HeLa cells, wild type Cx50 localised to appositional membranes and within the perinuclear region, but Cx50D47N showed no immunostaining at appositional membranes with immunoreactivity confined to the cytoplasm. Incubation of HeLa cells transfected with Cx50D47N at 27 degrees C resulted in formation of gap junctional plaques. CONCLUSIONS: The pulverulent cataracts present in members of this family are associated with a novel GJA8 mutation, Cx50D47N, that acts as a loss-of-function mutation. The consequent decrease in lens intercellular communication and changes associated with intracellular retention of the mutant connexin may contribute to cataract formation.

A novel GJA8 mutation is associated with autosomal dominant lamellar pulverulent cataract: further evidence for gap junction dysfunction in human cataract.

PURPOSE: To identify the gene responsible for autosomal dominant lamellar pulverulent cataract in a four-generation British family and characterise the functional and cellular consequences of the mutation. METHODS: Linkage analysis was used to identify the disease locus. The GJA8 gene was sequenced directly. Functional behaviour and cellular trafficking of connexins were examined by expression in Xenopus oocytes and HeLa cells. RESULTS: A 262C>A transition that resulted in the replacement of proline by glutamine (P88Q) in the coding region of connexin50 (Cx50) was identified. hCx50P88Q did not induce intercellular conductance and significantly inhibited gap junctional activity of co-expressed wild type hCx50 RNA in paired Xenopus oocytes. In transfected cells, immunoreactive hCx50P88Q was confined to the cytoplasm but showed a temperature sensitive localisation at gap junctional plaques. CONCLUSIONS: The pulverulent cataract described in this family is associated with a novel GJA8 mutation and has a different clinical phenotype from previously described GJA8 mutants. The cataract likely results from lack of gap junction function. The lack of function was associated with improper targeting to the plasma membrane, most probably due to protein misfolding.

Functional role of the carboxyl terminal domain of human connexin 50 in gap junctional channels.

Gap junction channels formed by connexin 50 (Cx50) are critical for maintenance of lens transparency. Because the C-terminus of Cx50 can be cleaved post-translationally, we hypothesized that channels formed by the truncated Cx50 exhibit altered properties or regulation. We used the dual whole-cell patch-clamp technique to investigate the macroscopic and single-channel properties of gap junctional channels formed by wild-type human Cx50 and a truncation mutant (Cx50A294stop) after transfection of N2A cells. Our results show that wild-type Cx50 formed functional gap junctional channels. The macroscopic Gjss-Vj relationship was well described by a Boltzmann equation with A of 0.10, V0 of 43.8 mV and Gjmin of 0.23. The single-channel conductance was 212 +/- 5 pS. Multiple long-lasting substates were observed with conductances ranging between 31 and 80 pS. Wild-type Cx50 gap junctional channels were reversibly blocked when pHi was reduced to 6.3. Truncating the C-terminus at amino acid 294 caused a loss of pHi sensitivity, but there were no significant changes in single-channel current amplitude or Gjss-Vj relationship. These results suggest that the C-terminus of human Cx50 is involved in pHi sensitivity, but has little influence over single-channel conductance, voltage dependence, or gating kinetics.

Heteromeric connexons formed by the lens connexins, connexin43 and connexin56.

In the eye lens, three connexins have been detected in epithelial cells and bow region/differentiating fiber cells, suggesting the possible formation of heteromeric gap junction channels. To study possible interactions between Cx56 and Cx43, we stably transfected a normal rat kidney cell line (NRK) that expresses Cx43 with Cx56 (NRK-Cx56). Similar to the lens, several bands of Cx56 corresponding to phosphorylated forms were detected by immunoblotting in NRK-Cx56 cells. Immunofluorescence studies showed co-localization of Cx56 with Cx43 in the perinuclear region and at appositional membranes. Connexin hexamers in NRK-Cx56 cells contained both Cx43 and Cx56 as demonstrated by sedimentation through sucrose gradients. Immunoprecipitation of Cx56 from sucrose gradient fractions resulted in co-precipitation of Cx43 from NRK-Cx56 cells suggesting the presence of relatively stable interactions between the two connexins. Double whole-cell patch-clamp experiments showed that the voltage-dependence of Gmin in NRK-Cx56 cells differed from that in NRK cells. Moreover, stable interactions between Cx43 and Cx56 were also demonstrated in the embryonic chicken lens by co-precipitation of Cx43 in Cx56 immunoprecipitates. These data suggest that Cx43 and Cx56 form heteromeric connexons in NRK-Cx56 cells as well as in the lens in vivo leading to differences in channel properties which might contribute to the variations in gap junctional intercellular communication observed in different regions of the lens.

Heteromeric mixing of connexins: compatibility of partners and functional consequences.

Cx43 is widely expressed in many different cell types, and many of these cells also express other connexins. If these connexins are capable of mixing, the functional properties of channels containing heteromeric connexons may substantially influence intercellular communication between such cells. We used biochemical strategies (sedimentation through sucrose gradients, co-immunoprecipitation, or co-purification by Ni-NTA chromatography) to examine heteromeric mixing of Cx43 with other connexins (including Cx26, Cx37, Cx40, Cx45, and Cx56) in transfected cells. These analyses showed that all of the tested connexins except Cx26 formed heteromeric connexons with Cx43. We used the double whole-cell patch-camp technique to analyze the electrophysiological properties of gap junction channels in pairs of co-expressing cells. Cx37 and Cx45 made a large variety of functional heteromeric combinations with Cx43 based on detection of many different single channel conductances. Most of the channel event sizes observed in cells co-expressing Cx40 and Cx43 were similar to those of homomeric Cx43 or Cx40 hemichannels in homo- or hetero-typic configurations. Our data suggest several different possible consequences of connexin co-expression: (1) some combinations of connexins may form heteromeric connexons with novel proeprties; (2) some connexins may form heteromeric channels that do not have unique properties, and (3) some connexins may be incompatible for heteromeric mixing.

Connexin46 mutations linked to congenital cataract show loss of gap junction channel function.

Human connexin46 (hCx46) forms gap junctional channels interconnecting lens fiber cells and appears to be critical for normal lens function, because hCx46 mutations have been linked to congenital cataracts. We studied two hCx46 mutants, N63S, a missense mutation in the first extracellular domain, and fs380, a frame-shift mutation that shifts the translational reading frame at amino acid residue 380. We expressed wild-type Cx46 and the two mutants in Xenopus oocytes. Production of the expressed proteins was verified by SDS-PAGE after metabolic labeling with [(35)S]methionine or by immunoblotting. Dual two-microelectrode voltage-clamp studies showed that hCx46 formed both gap junctional channels in paired Xenopus oocytes and hemi-gap junctional channels in single oocytes. In contrast, neither of the two cataract-associated hCx46 mutants could form intercellular channels in paired Xenopus oocytes. The hCx46 mutants were also impaired in their ability to form hemi-gap-junctional channels. When N63S or fs380 was coexpressed with wild-type connexins, both mutations acted like "loss of function" rather than "dominant negative" mutations, because they did not affect the gap junctional conductance induced by either wild-type hCx46 or wild-type hCx50.

Functional expression and biochemical characterization of an epitope-tagged connexin37.

To study the gap junction protein connexin37 (Cx37), we stably transfected cell lines with constructs of human Cx37 containing the epitope tag FLAG (DYKDDDDK). A Cx37 construct containing the FLAG moiety at the carboxyl terminus (Cx37F) was expressed in BWEM cells, and did not substantially alter the levels of endogenous Cx43 in these cells. Immunostaining showed that Cx37F colocalized with Cx43 at cell-cell contacts. Pulse-chase metabolic labeling and immunoprecipitation with anti-FLAG antibodies indicated that Cx37F was synthesized as a protein that ran at 35.9 +/- 0.9 kDa on reducing SDS-PAGE but chased into a slower migrating band at 38.0 +/- 1.0 kDa. This shift in mobility was due to phosphorylation on serine residues, based on [(32)P]-metabolic labeling, immunoprecipitation, and phosphoamino acid analyses. The transition to the phosphoCx37F correlated with a loss of solubility in 1% Triton X-100. Based on the [(35)S]-methionine pulse-chase experiments, the half-life of the labeled Cx37F was approximately 3 h, which is within the range reported for other connexins. Analysis of dye injection experiments indicated that dye transfer was reduced in Cx37-transfected cells in comparison to parental BWEM cells, suggesting that formation of heteromeric Cx37-Cx43 channels reduced the molecular permeability of communication between these cells. Moreover, the similarities of previously demonstrated kinetic details and modification of Cx43 to our new data regarding Cx37 provide evidence for a commonality in processing and assembly steps of these two connexins.

Gap junctions in the chicken pineal gland.

The chicken pineal gland, which contains a heterogeneous cell population, sustains a circadian rhythm of activity. Synchronization of cellular activity of heterogeneous cells might be facilitated by gap junctional intercellular channels which are permeable to ions and second messengers. To test this possibility, we looked for morphologically identifiable gap junctions between the different pineal cells, used antibodies and cDNA probes to screen for the presence of connexins, and tested for functional intercellular coupling. By transmission electron microscopy and immunocytochemistry, gap junctions and connexins were observed between pinealocyte cell bodies, stromal cells, astrocytes, and astrocyte and pinealocyte processes. Two gap junctional proteins, connexin43 and connexin45, were detected by immunocytochemistry, immunoblotting and RNA blot analysis. Functional intercellular coupling was observed in the gland by transfer of low molecular weight dyes. Dye transferred between homologous and heterologous cells. These data suggest that homologous and heterologous gap junctions may provide a mechanism for coordination of the cellular responses of the elements of the biological clock which are induced by lighting cues to produce the circadian rhythm of pineal activity.

PKC isoenzymes in the chicken lens and TPA-induced effects on intercellular communication.

PURPOSE: Because lens connexins are phosphoproteins and intercellular communication between lens cells may be modulated by connexin phosphorylation, experiments were designed to characterize the expression of protein kinase C (PKC) isoenzymes in the chicken lens and in lentoid-containing cultures and to study the effects of 12-O-tetradecanoylphorbol-13-acetate (TPA) treatment on the distribution of PKC isoenzymes and intercellular communication. METHODS: The presence and distribution of PKC isoenzymes were studied by immunoblot analysis and immunofluorescence in chicken lens sections and in cell cultures under control conditions and after treatment with TPA. Intercellular communication was assessed by transfer of microinjected Lucifer yellow. RESULTS: PKC alpha, gamma, iota, epsilon, and mu were detected in lens homogenates by immunoblot analysis. The levels of PKC alpha, gamma, iota, and mu decreased between the 7th and the 18th embryonic days. Levels of PKC epsilon remained relatively constant during the period of study. Similarly, lens cells in culture expressed isoenzymes alpha, gamma, epsilon, iota, and mu. PKC beta was not detected in lens or culture homogenates. In lens sections, all PKC isoenzymes analyzed were present in epithelial cells, in the annular pad region, and in the posterior aspect of fiber cells. The anti-PKC gamma antibody also stained fiber cell membranes. Analysis of lentoid cultures by immunofluorescence revealed that PKC gamma, epsilon, and iota and minimal amounts of PKC alpha were present in lentoid cells. Treatment with 200 nM TPA for 15 to 30 minutes induced translocation of PKC gamma to the plasma membrane of lentoid cells and significantly reduced the transfer of microinjected Lucifer yellow. CONCLUSIONS: Several PKC isoenzymes are expressed by lens cells in situ and in culture. The gamma isoenzyme, present in lens fibers, was activated in lentoid cells by TPA, a known activator of PKC. We have previously demonstrated TPA-induced phosphorylation of the gap junction protein connexin56 (Cx56). The new data presented in the current study demonstrate that TPA treatment also decreased intercellular communication. Taken together, the results suggest that differential phosphorylation of Cx56 by PKCgamma may induce a conformational change in the protein which, in turn, might lead to channel closure.

Connexin and gap junction degradation.

Many of the subunit proteins (connexins) that form gap junctions are rather dynamic, with half-lives of only a few hours. Thus, alterations in connexin turnover and degradation may represent significant mechanisms for the regulation of intercellular communication. We describe a pharmacological approach to determining pathways of connexin degradation. Cell cultures are left untreated or treated with inhibitors of lysosomal or proteasomal proteolysis. Changes in connexin levels, localization, or decay curves (derived from pulse-chase experiments) are assessed by immunoblotting, immunofluorescence, and immunoprecipitation, respectively. Such experiments have provided evidence that connexin43 degradation involves both the lysosome and the proteasome.

Molecular mechanism underlying a Cx50-linked congenital cataract.

Mutations in gap junctional channels have been linked to certain forms of inherited congenital cataract (D. Mackay, A. Ionides, V. Berry, A. Moore, S. Bhattacharya, and A. Shiels. Am. J. Hum. Genet. 60: 1474-1478, 1997; A. Shiels, D. Mackay, A. Ionides, V. Berry, A. Moore, and S. Bhattacharya. Am. J. Hum. Genet. 62: 526-532, 1998). We used the Xenopus oocyte pair system to investigate the functional properties of a missense mutation in the human connexin 50 gene (P88S) associated with zonular pulverulent cataract. The associated phenotype for the mutation is transmitted in an autosomal dominant fashion. Xenopus oocytes injected with wild-type connexin 50 cRNA developed gap junctional conductances of approximately 5 microS 4-7 h after pairing. In contrast, the P88S mutant connexin failed to form functional gap junctional channels when paired homotypically. Moreover, the P88S mutant functioned in a dominant negative manner as an inhibitor of human connexin 50 gap junctional channels when coinjected with wild-type connexin 50 cRNA. Cells injected with 1:5 and 1:11 ratios of P88S mutant to wild-type cRNA exhibited gap junctional coupling of approximately 8% and 39% of wild-type coupling, respectively. Based on these findings, we conclude that only one P88S mutant subunit is necessary per gap junctional channel to abolish channel function.

Cultured chicken embryo lens cells resemble differentiating fiber cells in vivo and contain two kinetic pools of connexin56.

The lens is an avascular organ in which gap junctions play a pivotal role for cell physiology and transparency. Here we evaluate a lens culture system as a model for studies of lens gap junction dynamics. In culture, chicken embryo lens cells initially form a monolayer of epithelial cells. Subsequently, the epithelial cells differentiate into lentoids, birefringent multicellular structures composed of fiber-like cells. We examined the cultures for the expression of cellular markers and lens fiber specific proteins using immunofluorescence and immunoblot analysis. We also determined the half-life of connexin56 (Cx56), a fiber-specific gap junction protein. All lens cells in culture expressed actin, endoplasmic reticulum proteins and N-cadherin. Only lentoid cells expressed the lens fiber connexins, Cx45.6 and Cx56. Cx56 localized at appositional membranes and did not co-localize with endoplasmic reticulum proteins or N-cadherin. Two pools of Cx56 were detected in these cultures, one with a half-life of a few hours and the other with a half-life of days. The two pools contained phosphorylated forms of Cx56 of different apparent molecular weights. These results suggest that lens cells in culture can be used as a model for the study of lens biology. They also suggest that phosphorylation of Cx56 might be regulating the stability of the protein.

Co-expression of lens fiber connexins modifies hemi-gap-junctional channel behavior.

Lens fiber cells contain two gap junction proteins (Cx56 and Cx45.6 in the chicken). Biochemical studies have suggested that these two proteins can form heteromeric connexons. To investigate the biophysical properties of heteromeric lens connexons, Cx56 was co-expressed with Cx45.6 (or its mouse counterpart, Cx50) in Xenopus oocytes. Whole-cell and single-channel currents were measured in single oocytes by conventional two-microelectrode voltage-clamp and patch clamp techniques, respectively. Injection of Cx56 cRNA induced a slowly activating, nonselective cation current that activated on depolarization to potentials higher than -10 mV. In contrast, little or no hemichannel current was induced by injection of Cx50 or Cx45.6 cRNA. Co-expression of Cx56 with Cx45.6 or Cx50 led to a shift in the threshold for activation to -40 or -70 mV, respectively. It also slowed the rate of deactivation of the hemichannel currents. Moreover, an increase in the unitary conductance, steady state probability of hemichannel opening and mean open times at negative potentials, was observed in (Cx56 + Cx45.6) cRNA-injected oocytes compared with Cx56 cRNA-injected oocytes. These results indicate that co-expression of lens fiber connexins gives rise to novel channels that may be explained by the formation of heteromeric hemichannels that contain both connexins.

The gap-junction protein connexin 56 is phosphorylated in the intracellular loop and the carboxy-terminal region.

The lens gap-junction protein, connexin 56, is modified by phosphorylation. Two-dimensional mapping of tryptic phosphopeptides of 32P-labeled connexin 56 from primary chicken-lens cultures showed that treatment with 12-O-tetradecanoylphorbol 13-acetate (TPA) induced an increase in phosphorylation of connexin 56 at specific constitutively phosphorylated sites. Treatment with 8-Br-cAMP or forskolin did not induce substantial changes in connexin 56 phosphorylation. Two phosphorylation sites within connexin 56, S493 and S118, were identified after HPLC purification and peptide sequencing of tryptic phosphopeptides from bacterially expressed connexin 56 fusion proteins phosphorylated by protein kinase C or protein kinase A in vitro. Comparisons of the two-dimensional maps of tryptic phosphopeptides from in vitro phosphorylated connexin 56 fusion proteins and in vivo phosphorylated connexin 56 showed that S493 and S118 were constitutively phosphorylated in lentoid-containing cultures, and that treatment with TPA induced an increase in phosphorylation of the peptides containing S118. It is suggested that phosphorylation of connexin 56 at S118 is involved in the TPA-induced decrease in intercellular communication and acceleration of connexin 56 degradation.

Distinct behavior of connexin56 and connexin46 gap junctional channels can be predicted from the behavior of their hemi-gap-junctional channels.

The gap-junctional protein rat connexin46 (Cx46) has the unusual ability to form voltage-gated channels in the nonjunctional plasma membrane of Xenopus oocytes (Paul et al., 1991; Ebihara and Steiner, 1993). These have been suggested to be gap-junctional hemichannels or connexons. The Xenopus oocyte system was used to characterize the functional properties of a closely related lens gap-junctional protein, chicken connexin56 (Cx56) (Rup et al., 1993) and to contrast them to those of rat Cx46. Single oocytes injected with either Cx56 or Cx46 cRNA developed time-dependent, outward currents that activated on depolarization. The currents induced by Cx56 and Cx46 showed differences in steady-state voltage dependence and in their degree of rectification. Furthermore, the voltage-dependent properties of the nonjunctional channels induced by the connexin cRNAs in external solutions containing low concentrations of calcium ions could account remarkably well for the behavior of the intercellular channels formed by Cx56 and Cx46 in paired oocytes. These results suggest that many of the voltage-dependent properties of the hemi-gap-junctional channels are retained by the intercellular channels.

Characterization of the gap junction protein connexin56 in the chicken lens by immunofluorescence and immunoblotting.

PURPOSE: To characterize the chicken lens gap junction protein, connexin56 (Cx56). METHODS: The methods used were immunoblotting, immunofluorescence, alkaline phosphatase treatment, in vitro translation, and primary tissue culture. RESULTS: Connexin56 translated in vitro showed a single band with an electrophoretic mobility of approximately 66 kd. Multiple bands of 67 to 90 kd were detected in immunoblots of chicken lens homogenates using antibodies raised against a peptide from Cx56. Most, if not all, of these bands represented different phosphorylated forms of Cx56, because immunoreactive Cx56 was detected as a doublet of 65 to 67 kd after treatment of lens homogenates with alkaline phosphatase. Indirect immunofluorescence demonstrated that Cx56 was localized at membrane appositions between lens fibers and bow region cells. Levels of Cx56 increased from embryonic days 4 to 15; thereafter, levels remained fairly constant until hatching, after which they declined. Before embryonic day 9, the slowest migrating bands were not as abundant as they were at later ages. After embryonic day 20, less Cx56 was observed by immunofluorescence in the nucleus than in the cortex; however, both regions had similar levels of Cx56 as measured by immunoblotting. The pattern of bands differed between the two lens regions, suggesting differential protein modification. Immunoreactive Cx56 bands of 35 to 38 kd were detected unless homogenates were prepared in the presence of ethylenediaminetetraacetic acid (EDTA). Cx56 was also detected in lentoid-containing primary cultures derived from chicken lens. CONCLUSIONS: Cx56 is a phosphoprotein. Its appearance and modification by phosphorylation, as detected by immunoblotting, correlate with lens fiber differentiation.

Bovine connexin44, a lens gap junction protein: molecular cloning, immunologic characterization, and functional expression.

PURPOSE: To identify, clone molecularly, characterize immunochemically, and express functionally a bovine lens gap junction protein (connexin). METHODS: The methods used were polymerase chain reaction, genomic cloning, RNA and DNA blotting, bacterial expression of a fusion protein, immunoblotting, alkaline phosphatase treatment, Xenopus oocyte expression, and voltage clamp technique. RESULTS: A bovine genomic clone encoding a polypeptide of 44,424 d, termed connexin44 (Cx44), was isolated. Cx44 was most closely related to the lens connexins rat Cx46 and chicken Cx56. The Cx44 DNA hybridized to a 2.5 kb mRNA detected only in lens RNA. The carboxyl terminal 161 amino acids from Cx44 were expressed in bacteria fused to maltose binding protein (MBP). The Cx44/MBP fusion protein reacted in immunoblots with anti-rat Cx46(411 to 416) antibodies and with the monoclonal antibody 5H1, but not with a monoclonal antibody to MP70 nor with antibodies to other connexins. Cx44 translated in vitro from the cloned DNA showed a single band with an apparent electrophoretic mobility of approximately 50 kd on polyacrylamide gels containing sodium dodecyl sulfate. Multiple bands of 53 to 57 kd were detected by immunoblotting in homogenates of bovine lens; these bands were reduced to a broad band of approximately 50 kd by alkaline phosphatase treatment, suggesting that they represented phosphorylated forms of Cx44. Cx44 RNA injected in single oocytes induced a large and characteristic time- and voltage-dependent current. Overexpression of Cx44 produced depolarization and cell lysis. Junctional currents that could be regulated by transjunctional voltage were induced between paired oocytes injected with Cx44 RNA. Observations in paired oocytes suggested the assembly of hemichannels into junctional channels. CONCLUSIONS: Cx44 is a phosphoprotein component of bovine lens fiber gap junctions. Although it has a relatively distinct sequence, it shares sequence similarity, immunologic cross-reactivity, and electrophysiological properties with rat Cx46. These data suggest that Cx44 is the protein previously identified in several immunohistochemical studies of bovine lens gap junctions that used anti-rat Cx46 antibodies. They also suggest that the formation of intercellular channels by pairing of hemichannels might prevent the cell lysis induced by the opening of unpaired hemichannels.