If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To whom correspondence should be addressed: Dept. of Anatomy & Cell Biology, The University of Western Ontario, Dental Science Bldg., Rm 00077, London, Ontario N6A 5C1, Canada. Tel.: 519-661-2111 ×86827; Fax: 519-850-2562;
Department of Physiology and Pharmacology, University of Western Ontario, London, Ontario N6A 5C1, CanadaDepartment of Anatomy and Cell Biology, University of Western Ontario, London, Ontario N6A 5C1, Canada
* This work was supported to by the Canadian Institutes of Health Research and Canada Research Chair program (to D. W. L. and D. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
A frameshift mutation generated from a dinucleotide deletion (780-781del) in the GJA1 gene encoding Cx43 results in a frameshift yielding 46 aberrant amino acids after residue 259 and a shortened protein of 305 residues compared with the 382 in wild-type Cx43. This frameshift mutant (fs260) causes oculodentodigital dysplasia (ODDD) that includes the added condition of palmoplantar keratoderma. When expressed in a variety of cell lines, the fs260 mutant was typically localized to the endoplasmic reticulum and other intracellular compartments. The fs260 mutant, but not the G138R ODDD-linked Cx43 mutant or a Cx43 mutant truncated at residue 259 (T259), reduced the number of apparent gap junction plaques formed from endogenous Cx43 in normal rat kidney cells or keratinocytes. Interestingly, mutation of a putative FF endoplasmic reticulum retention motif encoded within the 46 aberrant amino acid domain failed to restore efficient assembly of the fs260 mutant into gap junctions. Dual whole cell patch-clamp recording revealed that fs260-expressing N2A cells exerted severely reduced electrical coupling in comparison to wild-type Cx43 or the T259 mutant, whereas single patch capacitance recordings showed that fs260 could also dominantly inhibit the function of wild-type Cx43. Co-expression studies further revealed that the dominant negative effect of fs260 on wild-type Cx43 was dose-dependent, and at a predicted 1:1 expression ratio the fs260 mutant reduced wild-type Cx43-mediated gap junctional conductance by over 60%. These results suggest that the 46 aberrant amino acid residues associated with the frameshift mutant are, at least in part, responsible for the manifestation of palmoplantar keratoderma symptoms.
Gap junctions are unique intercellular channels that directly connect the cytoplasm of neighboring cells, which are formed by integral membrane proteins called connexins (Cxs)
The abbreviations used are: Cxs, connexins; ANOVA, analysis of variance; ER, endoplasmic reticulum; GFP, green fluorescent protein; PDI, protein-disulfide isomerase; T259, Cx43 mutant truncated at residue 259.
). Each apposing cell contributes a hexameric arrangement of connexins, also known as a connexon, which dock to form a gap junction channel. A connexin polypeptide chain spans the plasma membrane four times, consisting of four transmembrane domains (M1 to M4), two extracellular loops (EL-1 and EL-2) and one intracellular loop (IL), with both the N and C termini (AT and CT, respectively) located in the cytoplasm (
) (Fig. 1). Whereas connexin transmembrane domains and extracellular loops are highly conserved, the intracellular loop and C terminus are known to be the most variable domains across different connexin family members, suggesting that these domains play important roles in the differential functions of connexins.
FIGURE 1Schematic models depicting the topology of human Cx43 and mutants.A, predicted topology of Cx43 depicting four transmembrane domains (M1-M4), two extracellular loops (EL-1 and EL-2), an intracellular loop (IL), an N terminus (AT), and a C terminus (CT). B, T259 represents truncated Cx43 where the remaining 123 amino acids are removed (B-I). fs260 represents a frameshift mutant where residues 260-305 are aberrant amino acid residues which include a “FF” (black) motif (B, II). The CT domain of full-length Cx43 encodes many known or putative phosphorylation sites (yellow) and the binding sites for several known Cx43 binding proteins (B, III).
Connexins exhibit complex tissue distribution patterns that are temporally and spatially regulated during development and differentiation. Among connexin family members (21 in human and 20 in mouse), the GJA1-encoded protein (Cx43) exhibits the most ubiquitous distribution, playing important roles in many tissue types and physiological processes (
). Cx43 channels can be regulated by a pH-sensitive “ball-and-chain” mechanism where the C terminus acts as a gating particle and a region within the intracellular loop serves as a receptor (
). The cytoplasmic acidification renders the “particle” able to bind to the “receptor”, leading to the closure of the Cx43 gap junction channel. This pH gating mechanism was supported by the findings that a truncated mutant of Cx43, where the last 125 amino acid residues were deleted from the CT, became insensitive to intracellular pH change. Furthermore, if the deleted tail segment was co-expressed with the truncated Cx43 the impaired pH gating could be partially restored (
). Channel gating, connexin assembly, and degradation events are also thought to be regulated in part by phosphorylation of one or more of the 14 serine/tyrosine phosphorylation sites encoded-within the C-terminal region of Cx43 (
). The Cx43 C terminus is also able to directly or indirectly interact with a variety of proteins, such as ZO-1, ZO-2, α- and β-tubulin, β-catenin, CIP85, and drebrin (
). Oculodentodigital dysplasia (ODDD), a rare inherited autosomal dominant disorder characterized by developmental abnormalities of the face, eyes, limbs, and dentition, has been linked to mutations in the GJA1 gene (
). Although the Cx43 mutations associated with ODDD are broadly distributed, the majority of the known mutations are located in the N-terminal half of the gene sequence encoding Cx43. Interestingly, the bulk of these mutations are familial with the remaining cases representing sporadic mutations. Also, in all but one case, ODDD-linked GJA1 gene mutations alter a single amino acid, leaving the remainder of the 382 amino acid residues of the Cx43 sequence unaltered. The only mutation localized to the C-terminal region of Cx43 is a frameshift mutation arising from the deletion of two nucleotides (780-781del) (
). The altered sequence translates into 46 aberrant amino acid residues being added after residue 259 until it reaches a premature stop codon after encoding an amino acid residue at the 305 position. The incorporated sequence of 46 aberrant amino acids does not correspond to any known functional polypeptide domain as determined by a BLAST search (
). Thus far, this is the only familial Cx43 ODDD-linked mutation where patients suffer from not only the wide range of syndromes associated with ODDD but also palmoplantar keratoderma. Given that the last native 123 amino acid residues are lost from this frameshift Cx43 mutant it is postulated that the reason for patients suffering from increased disease burden may be related to a critical role played by the Cx43 C terminus in the skin. In fact, mice expressing a Cx43 mutant where the last 125 amino acids of Cx43 are missing die within 1 week after birth because of the loss of skin barrier function (
In the present study, we fully characterized the functional status of the fs260 mutant as it is distinct from other Cx43 ODDD-linked single residue mutations. Importantly, the mechanism responsible for the trafficking defect is encoded within the 46 aberrant amino acids attached to the Cx43 sequence after residue 259. Through the collective use of well understood cell culture models and keratinocytes, our studies shed new insights into why patients harboring this mutant exhibit both ODDD and skin disease.
EXPERIMENTAL PROCEDURES
Cell Culture and Reagents—All cell culture reagents were obtained from Invitrogen or Sigma. The mouse neuroblastoma (N2A) cells, the human cervical carcinoma (HeLa) cells and the normal rat kidney (NRK) cells were purchased from ATCC (American Type Culture Collection). Rat epidermal keratinocytes (REKs) were generously provided by Dr. Vincent C. Hascall (Cleveland Clinic Foundation, Cleveland, OH). Cells were cultured in regular high glucose DMEM supplemented with 10% fetal bovine serum, 2 mm l-glutamine (except N2A cells), 100 units/ml penicillin, and 100 μg/ml streptomycin. All the cell lines were maintained at 37 °C in a moist environment of 95% air and 5% CO2 and subcultured directly into 35 mm tissue culture dishes or plated on glass coverslips as required.
Engineering of Mutant Cx43 cDNAs—The engineering of human Cx43-GFP construct was described previously (
). The human Cx43 mutation-fs260 was constructed with the QuikChange™ Site-directed mutagenesis kit (Stratagene, La Jolla, CA) by using forward 5′-GCTGAGCCCTGCCAAAGACTGGGTCTCAAAAATATGC-3′ and reverse 5′-GCATATTTTTGAGACCCAGTCTTTGGCAGGGCTCAGC-3′ primers and cloned into pcDNA3.1 expression vectors. Sequence analysis confirmed that no other mutation was present. Two nucleotides (780-781TG) were deleted resulting in a frameshift beginning after residue 259 until reaching a premature stop codon after encoding residue 305. The Cx43 mutant, designated fs260 (representing frameshift after encoding residue 259), encoded an irrelevant sequence of 46 aberrant amino acid residues at the C terminus. The fs260-GFP cDNA was generated by PCR using the forward 5′-CGGGGTACCAACATGGGTGACTGGAGC-3′ and reverse 5′-CGGTGGATCCTTGCTTGCTTG-3′ primer by adding KpnI and BamHI restriction sites, then cloning into the pEGFP-N1 vector (BD BioSciences). The resulting construct (fs260-GFP) encoded 6 amino acids linking fs260 to GFP in-frame was sequenced for verification. The T259 mutant was derived from the fs260 construct by deleting the sequence encoding the 46 aberrant amino acids. The primers used to generate the T259 mutant were forward 5′-CGGGGTACCAACATGGGTGACTGGAGC-3′ and reverse 5′-CGGTGGATCCCAGTCTTTGGCAGGGCTC3′. The T259 mutant was then incorporated into the pEGFP-N1 vector by using the KpnI and BamHI restriction sites as described above. Sequencing indicated that truncated Cx43 at residue 259 was linked in-frame by 6 amino acids to GFP. The fs260-GFP construct was further mutated to substitute the putative endoplasmic reticulum FF motif, encoded within the aberrant 46 amino acids, to an AA motif (fs260-FF/AA) by using the forward primer 5′-CGACAGAAACAAGCCGCCTTGCCGCAATTACAACAAGCAAGCAAG, and reverse primer 5′-CTTGCTTGCTTGTTGTAATTGCGGCAAGGCGGCTTGTTTCTGTCG. All GFP-tagged variants were further cloned into the AP2 retroviral vector as described previously (
). Transient Transfection and Retroviral Infection—Cells were transfected with human Cx43-GFP, fs260-GFP, T259-GFP, or fs260-FF/AA-GFP by Lipofectamine 2000 (Invitrogen) as described previously (
). In brief, cells were allowed to reach 50% confluence in 35-mm culture dishes, the plasmids were preincubated with Lipofectamine 2000 reagent for 15 min at room temperature, the complexes were then incubated with the cells for 4-6 h. Cells were washed and incubated in regular medium for 24-48 h before being used for patch-clamp recording, or fixed for immunocytochemistry. For fs260 and free GFP co-transfection experiments, the vectors encoding GFP and fs260 were used at a 1:5 ratio, respectively, to ensure that cells that expressed GFP also expressed fs260. In co-transfection experiments, the cDNAs of fs260-GFP and Cx43-mRFP (Cx43 tagged to monomeric dsRed) were pre-mixed at various ratios prior to being used in transfections.
For cell infection, REKs at 30-40% confluence were subjected to two rounds of retroviral infection with retroviral particles encoding GFP-tagged fs260, T259, G138R or GFP as described previously (
). Infected cells were passed at least twice, evaluated to ensure at least 90% expression efficiency, and subjected to Western blot analysis.
Western Blot Analysis—REKs expressing GFP-tagged fs260, T259, G138R or GFP, were subjected to lysis buffer containing 50 mm Tris-Cl, pH 8.0, 150 mm NaCl, 0.02% sodium azide, 100 μg/ml phenylmethylsulfonyl fluoride, 1% Nonidet P-40, 0.1% SDS, 50 mm NaF, 2 mm EDTA, and a protease inhibitor mixture from Roche Applied Science (Mississauga, ON, Ref.
). The BCA assay reagent kit (Pierce) was used to determine protein concentrations and 20 μg of protein per lane were separated by 10% SDS-PAGE. Proteins were transferred to nitrocellulose membranes and subsequently immunoblotted for Cx43 (antibody to the C-terminal of Cx43; 1:5,000 dilution, Sigma; antibody to the N-terminal of Cx43 1:100 dilution, Fred Hutchinson Cancer Research Center Antibody Development Group, Seattle, WA). The membranes were stripped and re-probed for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) to assess loading levels. Antibody binding was detected using the enhanced chemiluminescence system (Pierce).
Immunocytochemistry—Cell immunofluorescent labeling and imaging were performed as previously described (
). Briefly, cells grown on 12-mm glass coverslips were fixed with 80% methanol/20% acetone (or a 3.7% formaldehyde solution for GFP-expressing cells) at 4 °C for 15 min. HeLa and NRK cells transfected with fs260-GFP or T259-GFP were labeled with a 1:500 dilution of an anti-protein disulfide isomerase (PDI) antibody (Stressgen, Biotechnologies, Victoria, BC), a 1:1000 dilution of an anti-giantin antibody (Babco, Berkeley Antibody Company, Richmond, CA) or a 1:200 dilution of an anti-TGN38 antibody (BD Transduction Laboratories, Lexington, KY). NRK cells expressing fs260-GFP, T259-GFP, or fs260-FF/AA-GFP were also labeled with a 1:50 dilution of anti-Cx43 monoclonal antibody (Fred Hutchinson Cancer Research Center Antibody Development Group). Washed cells were incubated for 1 h in goat anti-mouse or donkey anti-rabbit antibodies conjugated to Texas red or FITC (Jackson ImmunoResearch Laboratories). In some cases, cell nuclei were labeled with Hoechst 33342 (1:1000 dilution; Molecular Probes). As described previously (
), the labeled cells were analyzed and imaged on a Zeiss (Thornwood, NY) LSM 510 META confocal microscope. Digital images were prepared using Zeiss LSM, Adobe Photoshop 7.0 and CorelDraw 12 software. To quantify the incidences of gap junction plaque formation, at least 8 fields (0.09 mm2 each) were randomly picked and over 113 cell-cell interfaces were counted for each experimental condition involving NRK cells and over 39 interfaces were counted between mutant-expressing REKs. The data are presented as mean ± S.D. The one-way ANOVA (Tukey’s Multiple Comparison Test) was used to test statistical significance.
Patch Clamp Electrophysiology—Gap junctional intercellular coupling between isolated N2A cell pairs expressing mutant or wild-type Cx43 was assessed using dual whole cell patch clamp technique as described previously (
). Series resistance was compensated by 80%. In dual color labeling system, only those N2A cell pairs expressing both fs260-GFP and Cx43-mRFP were selected for recording. Wild-type NRK cells or REKs or cells transfected with Cx43 mutants were recorded with single whole cell patch clamp. The cell membrane capacitance was determined and the gap junctional conductance was then calculated. This single cell capacitance recording was developed by de Roos and co-workers (
), which is suited for estimating the gap junctional coupling level between cells within clusters or confluent monolayer cells. Briefly, small cell clusters (5-6 cells) were selected for capacitance recording, and the cell residing in the center of a cluster was voltage-clamped at -60 mV. A depolarization voltage pulse (10 mV, 120 ms) was then applied to induce a transient capacitive current. The current was high-cut filtered at 10 kHz and digitized at 100 kHz. The peak current Ipk and the steady-state current Iss were measured. Data acquisition and analysis were performed via a Digidata 1322A interface and pClamp9 software (Axon Instruments Inc., Union City, CA). The estimated conductance between the patched cell and its surrounding cells, G, was calculated by the following equation: G = Iss × Gser/(Ipk - Iss) where Gser = Ipk/Vp is the series conductance between the patch pipette and the patched cell. Vp is the amplitude of the depolarization voltage pulse. It is noted that the estimated conductance (G)isan underestimation of actual conductance (
). The data are presented as mean ± S.E. The Student’s t test and one-way ANOVA (Tukey’s Multiple Comparison Test) were used to test statistical significance (*, p < 0.05; **, p < 0.01; ***, p < 0.001).
RESULTS
The Cx43 fs260 Mutant Was Localized to the Endoplasmic Reticulum in the Absence or Presence of Wild-type Cx43—Mutations in the GJA1 gene encoding Cx43 (Fig. 1A) underlie the autosomal dominant disorder known as oculodentodigital dysplasia. In the case of the frameshift mutation caused by the deletion of nucleotides 780 and 781, patients would be expected to harbor a truncated variant of Cx43 where 46 aberrant amino acid residues are added after residue 259 (Fig. 1B). Such a Cx43 mutant would lose the majority of its putative or known phosphorylation sites (Fig. 1B, III, yellow residues) and a new double phenylalanine motif would be introduced into the encoded aberrant sequence (Fig. 1B, II, black residues). In addition, all the regulatory properties provided by proteins that bind the last 123 amino acid domain of Cx43 would be lost.
To begin to address the consequences of this frameshift mutation, the fs260-GFP mutant was expressed in a variety of mammalian cells to assess its steady-state cellular distribution. Confocal microscopy revealed that in GJIC-deficient HeLa cells, fs260-GFP rarely formed gap junctional plaque-like structures at the cellcell interface, but rather exhibited a reticular and punctate intracellular localization pattern. Further localization analysis revealed that fs260-GFP co-localized in part with PDI suggesting a population of fs260-GFP was localized in the endoplasmic reticulum (ER) (Fig. 2A). Minimal fs260-GFP was found to localize with the Golgi protein, giantin (Fig. 2B). The ER localization pattern was even more evident in NRK cells where the overlap between the mutant profile and PDI was extensive (Fig. 2C) while the co-localization with the resident Golgi protein, TGN38, was also evident (Fig. 2D). When expressed in NRK cells (see Fig. 4A) or REKs (see Fig. 8A), the intracellular distribution profile of fs260-GFP remained relatively unchanged suggesting that endogenous wild-type Cx43 could not rescue its distribution to the cell surface (Fig. 2, C and D). Immunolabeling for Cx43 in NRK cells revealed the punctate pattern typical of gap junction plaques at cellcell interfaces (Fig. 2E).
FIGURE 2The fs260 mutant rarely forms gap junction plaques and typically co-localizes with a resident protein of the endoplasmic reticulum. When expressed in GJIC-deficient HeLa cells (A and B) or Cx43-positive NRK cells (C and D), fs260-GFP rarely formed identifiable gap junction plaque-like structures at cell-cell interface and was frequently found to co-localize with the resident endoplasmic reticulum protein, PDI, with limited co-localization with giantin or TGN38. Wild-type Cx43 typically assembled into gap junction plaques in NRK cells (E). Hoechst 33342 was used to stain NRK cell nuclei. Bar, 10 μm.
FIGURE 4The fs260 and fs260fs-FF/AA mutants reduced the frequency of endogenous Cx43 gap junction plaques found at cell-cell interfaces in NRK cells. NRK cells expressing fs260-GFP (A), T259-GFP (B), or fs260-FF/AA-GFP (C) were immunolabeled for endogenous Cx43 (red) using an antibody that would only detect endogenous Cx43 and not the co-expressed mutants. At cell-cell interfaces between cells expressing either the fs260-GFP or fs260-FF/AA-GFP mutants there was a loss of gap junction plaques (A, C, open arrows) while plaques were prevalent in cells lacking the mutant (A, C, solid arrows). Hoechst 33342 was used to stain cell nuclei in C. In cells expressing T259-GFP, Cx43 gap junctions were prevalent between both mutant expressing and non-expressing cells (B). Quantification of these results revealed that gap junction plaques were identified in less than 15% of the interfaces between fs260-GFP or fs260-FF/AA-GFP-expressing cells while gap junctions were found in over 80% of the interfaces between cells expressing T259-GFP (D). Bar, 10 μm. A total of 8 fields (0.09 mm2) were randomly picked for assessing the presence or absence of gap junction plaques at cell-cell interfaces (113-126 interfaces were examined for each condition).
FIGURE 8The fs260 mutant inhibits the assembly of Cx43 gap junctions in REKs and appears to reduce the levels of endogenous Cx43. REKs were infected with retroviral particles encoding GFP, fs260-GFP, T259-GFP, or transfected with a cDNA encoding fs260-FF/AA-GFP prior to immunolabeling with an anti-Cx43 directed to the C-terminal domain (A). Note that gap junction plaques were rarely found between fs260 or fs260-FF/AA-expressing cells (green) but prevalent between G138R- or GFP-expressing cells. Hoechst 33342 was used to stain REK cell nuclei. Western blot revealed that the expression of the fs260 mutant reduced the levels of endogenous Cx43 (B). Western blot is representative of three experiments. Bar, 10 μm.
The T259 Mutant Trafficked to the Cell Surface—To assess the role of the 46 additional incorrect amino acid residues in the aberrant localization of the fs260 mutant, a truncated Cx43 mutant was engineered at position 259 (T259, Fig. 1). Unlike the fs260 mutant, when T259-GFP was expressed in HeLa or NRK cells this truncated mutant was transported to the plasma membrane and formed gap junctional plaques at cell-cell interfaces (Fig. 3A-D, open arrows). In T259-GFP-expressing HeLa cells, the localization data also suggested that this truncated mutant was also partially retained in intracellular compartments (Fig. 3, A and B); a situation that was somewhat less pronounced in NRK cells (Fig. 3, C and D).
FIGURE 3The T259 mutant can form gap junctional plaques at cell-cell interfaces in both HeLa and NRK cells. The T259-GFP mutant was transported to the cell surface and formed structures reminiscent of gap junction plaques in both HeLa (A and B) and NRK (C and D) cells (open arrows). In addition, considerable populations of this truncated mutant were found within intracellular compartments of both cell types colocalized in part with resident proteins of the endoplasmic reticulum (PDI, A, C, red) and Golgi apparatus (giantin, B; TGN38, D; red). Hoechst 33342 was used to stain NRK cell nuclei. Bar, 10 μm.
The fs260 Mutant Reduces the Frequency of Gap Junction Plaque-like Structures Formed from Endogenous Cx43—To examine if the expression of the fs260-GFP mutant reduced the prevalence of gap junctions formed from endogenous Cx43, transiently transfected NRK cells were immunolabeled for endogenous untagged Cx43 (i.e. the antibody used is specific for the C-terminal domain and would only recognize endogenous Cx43). As expected, fs260-GFP largely localized to intracellular compartments (Fig. 4A, green) while abundant gap junction plaques were visible as punctate structures between neighboring untransfected cells (Fig. 4A, red, solid arrows). However, in cells expressing the mutant, endogenous Cx43 was restricted to intracellular compartments reminiscent of the ER and there was a drastic reduction in gap junction plaques at the cell-cell interface (Fig. 4A, open arrows). Furthermore, when NRK cells expressed T259-GFP, endogenous Cx43 formed numerous gap junctions at cell-cell interfaces where it co-localized with T259-GFP (Fig. 4B). To assess whether the putative ER-retention FF motif encoded in the aberrant 46 amino acid domain was responsible for the retention of the fs260 mutant within intracellular compartments, the FF motif was mutated to AA (fs260-FF/AA-GFP) and expressed in NRK cells (Fig. 4C). Surprisingly, the fs260-FF/AA-GFP mutant was typically found in intracellular compartments with scattered evidence for its assembly into gap junction plaques (Fig. 4C). Moreover, the fs260-FF/AA-GFP mutant, like the fs260 mutant, reduced the number of gap junction plaques formed from endogenous Cx43 (Fig. 4C). By randomly selecting 8 image fields and examining cell-cell interfaces, gap junctions were detectable on average in less than 15% of the cases where the fs260-GFP or fs260-FF/AA-GFP-expressing cells were adjacent to each other (Fig. 4D). This was in stark contrast to the 81% incidence of detectable gap junction plaques at interfaces of cells expressing the T259-GFP mutant (Fig. 4D) or the 93% incidence of detectable gap junction plaques found between untransfected cells (Fig. 4D). These results strongly suggest that the addition of the 46 aberrant amino acid residues in the frameshift Cx43 mutant fs260, but not strictly the FF motif, are responsible for the impairment of gap junction trafficking of fs260 mutant as well as endogenous Cx43.
N2A Cells Expressing the fs260 Mutant Are Poorly Coupled—To examine whether the fs260 is able to form functional gap junction channels, dual whole cell patch clamp recordings were performed. In N2A cells transfected with fs260-GFP, only a low level of intercellular coupling was exhibited which did not exceed 10% of control cells expressing Cx43-GFP (Fig. 5, A and B). When untagged fs260 mutant was co-expressed in N2A cells with free GFP, a slightly higher level of cell-cell electrical coupling was observed (nearly 20% of the control). There was no significant difference between these two transfectants, as tested by one-way ANOVA Tukey’s test. Unlike the fs260 mutant, when T259-GFP was expressed in N2A cells, the recorded cell pairs were all electrically coupled and the average coupling level was ∼50% of control cells expressing Cx43-GFP (Fig. 5, C and D). Taken together, these results demonstrate that the fs260 mutant has drastically impaired intracellular trafficking and a greatly reduced functional status, while removal of the 46 aberrant amino acid residues could partially restore the impaired trafficking and function.
FIGURE 5Functional gap junction channel formation was greatly reduced for the fs260 mutant but partially restored for the T259 mutant. N2A cell pairs expressing fs260-GFP or untagged fs260 exhibited severely reduced gap junction conductance compared with cell pairs expressing Cx43-GFP, as determined by dual whole cell patch clamp recording (A and B). The difference between tagged and untagged fs260 mutants was not found by one-way ANOVA and Tukey's multiple comparison tests to be statistically significant (A and B). When expressed in GJIC-deficient N2A cells, the T259-GFP mutant formed functional gap junction channels with coupling seen in all cell pairs tested. The average macroscopic conductance exhibited by T259-GFP was nearly 50% of control Cx43-GFP-expressing N2A cell pairs (C and D). The number presented in the brackets represents the n value for each experimental condition.
The fs260 Mutant Acts as a Dominant-Negative on GJIC in Cx43-expressing Cells in a Dose-dependent Manner—To test if the fs260 mutant had a dominant-negative effect on wild-type Cx43, fs260-GFP was transfected into NRK cells and REKs. Small cell clusters of 5-6 cells with obvious fs260-GFP expression (as evidence by the presence of bright green fluorescence) and apposing proximity to each other in NRK or REK cultures were selected to perform single whole cell patch clamp recordings. In response to a 10-mV short pulse, a transient capacitive current was determined, and the gap junctional conductance was calculated. In both NRK cells (Fig. 6, A and B) containing only endogenous Cx43, and REKs (Fig. 6, C and D), which express abundant Cx43, low levels of Cx26 and possibly other connexins (
), fs260-GFP expression reduced intercellular coupling by ∼50% compared with untransfected cells in parallel cultures. As an additional control, we also recorded NRK cell and REK clusters of similar size from the same cultures that were exposed to the transient transfection protocol but showed no evidence of fs260-GFP expression, as assessed by the lack of green fluorescence. These mutant-free NRK cells and REK clusters exhibited robust intercellular electrical coupling that was not significantly different from that of controls as tested in parallel studies (data not shown).
FIGURE 6The fs260 mutant exhibits dominant-negative properties on gap junctional coupling mediated by endogenous Cx43. When Cx43-positive NRK cells (A and B) or REKs (C and D) were engineered to express the fs260-GFP mutant, the estimated gap junctional conductance was reduced to ∼50% of the control level, as revealed by single patch capacitance recording. The numbers presented in brackets represents the n value for each experimental condition.
To assess the potency of the fs260-GFP mutant on inhibiting the function of wild-type Cx43, N2A cells were co-transfected with different ratios of vectors encoding fs260-GFP and monomeric red fluorescent protein (mRFP)-tagged wild-type Cx43 (Cx43-mRFP). Cell pairs that expressed both red and green fluorescence were selected and total gap junction conductance was recorded (Fig. 7, A-C). When both wild-type and mutant Cx43 were predicted to be expressed at equal ratio the total gap junction conductance was reduced by over 60% of controls. When the ratio of mutant to wild-type Cx43 was 2:1, total gap junction conductance was reduced to less than 10% of the control. These findings suggest that the fs260 mutant is acting as a dominant-negative in a dose-dependent manner. Consistently, the incidence of microinjected Lucifer yellow dye transfer from REKs expressing the fs260-GFP mutant was only 22% (n = 45) compared with 100% in REKs expressing the T259-GFP mutant (n = 13).
FIGURE 7The fs260 mutant inhibits Cx43 in a dose-dependent manner. N2A cells were co-transfected with various ratios of cDNA vectors encoding fs260-GFP and Cx43-mRFP. A, cell pairs exhibiting both bright red and green fluorescence were patch-clamped and total gap junction conductance was recorded. B, representative recording traces demonstrated that a Cx43-mRFP-transfected N2A cell-pair exerted robust electrical coupling, while the cell-pair co-transfected with Cx43-mRFP and fs260-GFP at an equal ratio showed dramatically reduced coupling level. C, fs260-GFP dose-dependently inhibited Cx43-mRFP-mediated GJIC. **, p < 0.01; ***, p < 0.001. Bar, 10 μm.
The fs260 Mutant Selectively Impairs Endogenous Cx43 Plaque Formation in Keratinocytes—GFP, fs260-GFP, G138R-GFP, or fs260-FF/AA-GFP were expressed in REKs and cells were immunolabeled for Cx43 using an antibody that would recognize endogenous Cx43 (Fig. 8A). Cx43 gap junction plaque-like structures were rarely observed between fs260 mutant-expressing cells (6.5 ± 6.3% of interfaces) and fs260-FF/AA mutant-expressing REKs (6.9 ± 6.9% of interfaces) while intracellular Cx43 was evident in the fs260-expressing cells but not visible in fs260-FF/AA mutant-expressing REKs. Gap junction plaques were readily detected between control GFP-expressing cells (87.0 ± 9.4% of interfaces) and G138R mutant-expressing REKs (93.8 ± 4.7% of interfaces). Western blot analysis revealed a reduction of endogenous Cx43 in REKs that expressed the fs260 mutant but not in REKs that expressed GFP, T259-GFP, or G138R-GFP (Fig. 8B). These results suggest that, unlike the G138R mutant, the fs260 mutant may reduce the overall cellular levels of endogenous Cx43 possibly by promoting its premature degradation.
DISCUSSION
In recent years, the dysregulation of gap junction function involving many connexin family members has been linked to a variety of disease states in humans, including skin disease, hereditary deafness, cataracts, cancers, developmental abnormalities, and some forms of neuropathies (
). In more recent years, additional mutations were found in the GJA1 gene which often manifest as a pleiomorphic array of syndromes that are characteristic of ODDD (
), the known ODDD Cx43 mutations are not restricted to any particular subdomain of the connexin involved. However, until 2005 no Cx43-ODDD linked mutations were reported in the GJA1 gene encoding the large 148 amino acid C-terminal tail of Cx43, which led to speculations that this domain was either resistance to mutation or mutants in this region were embryonic lethal. Mice lacking Cx43 or expressing a truncated version of Cx43 where the last 125 amino acids are lost, die within the first week after birth suggesting that dysregulation of Cx43 can lead to premature death at least in mouse models (
). In the present study, we characterize a newly reported frameshift mutation in the C terminus of Cx43, which causes both ODDD and skin symptoms in 3 familial patients (
The mechanism by which Cx43 mutations cause ODDD is only beginning to be understood and extends beyond the fact that many of these mutations result in loss of Cx43 function. To date, 11 of the missense ODDD-linked Cx43 mutants have been functionally characterized (
), the Y17S, G21R, A40V, and G138R mutants were transported to the plasma membrane and able to form morphological structures that resembled gap junctional plaques but functional gap junction channels were not detected (
), and whereas the G49K, L90V, I130T, and K134E mutants formed gap junction channels the overall junctional conductance was low compared with wild-type Cx43 (
). Additionally, in a mouse model of human ODDD where a G60S mutation was encoded in the Cx43 gene, junctional coupling in granulosa cells isolated from heterozygous animals exhibited a 90% decrease in junction coupling compared with cells from wild-type littermates (
). This suggests that when the G60S mutant and wild-type Cx43 are expressed at equal levels, the mutant is dominant to wild-type Cx43. Interestingly, mice harboring the autosomal dominant Cx43 mutant have most of the syndromes associated with human ODDD but without obvious skin defects (
) and the fact that other familial missense mutations do not manifest into skin symptoms, it is of interest to determine the mechanism by which this frameshift mutant can cause both ODDD and skin disease. Whereas work by van Steensel et al. (
) alluded to the possibility that the Cx43 protein levels was reduced in the skin, no mechanistic information on how this mutant may underlie palmoplantar keratoderma has been reported.
To dissect the mechanism that links this novel ODDD-causing Cx43 mutant to skin disease, it is first essential to determine its functional characteristics. Data presented in this study revealed that this frameshift mutant is severely impaired in both trafficking and function. More importantly, when expressed in NRK cells with endogenous Cx43, the fs260 mutant exhibited a dominant-negative effect on wild-type Cx43 by both restricting endogenous Cx43 assembly into identifiable gap junction plaques and by impairing the function of endogenous Cx43 by reducing coupling conductance to ∼50% of control level. Hence we propose that the mutant often fails to pass the “quality control” mechanisms associated with the endoplasmic reticulum (
) and is possibly subjected to endoplasmic reticulum associated degradation. Furthermore, the mutant severely reduces the number of gap junction plaques formed by endogenous Cx43 suggesting that it may interact with wild-type Cx43 early in the secretory pathway. The fact that the fs260 mutant can facilitate a low level of GJIC would further suggest that a subpopulation of molecules can escape the “quality control” mechanisms of the endoplasmic reticulum, reach the cell surface and assemble into complete gap junction channels. This low level of Cx43-based GJIC and the fact that it does not completely eliminate endogenous Cx43-based GJIC could explain why ODDD patients harboring this mutation do not display more severe clinical symptoms. To assess the potency of the mutant on wild-type Cx43 function, in co-expression studies where both the fs260 mutant and Cx43 were predicted to be expressed at equal levels, coupling was reduced by slightly more than 60% not dissimilar to our previous studies using cultured granulosa cells from a mouse model of human ODDD (
). At slightly higher mutant to wild-type Cx43 ratios, channel coupling was reduced to near background levels suggesting that this mutant is potent at inhibiting the function of wild-type Cx43. Thus, consistent with many other ODDD-linked Cx43 mutants, the fs260 mutant has a greatly reduced capacity to assemble into functional gap junctions and dominantly inhibits the function of endogenous Cx43 (
Because ODDD patients that express the fs260 mutant have palmoplantar keratodermas, we also assessed the functional status of this mutant in a keratinocyte cell line (REK) that expresses Cx43 and Cx26 at the protein level as well as mRNAs for 7 other connexins (
). Similar to NRK cells, in these undifferentiated keratinocytes the total GJIC measured by coupling conductance was reduced by ∼50% and the incidence of Lucifer yellow dye coupling was reduced by 75%. Because Cx43 appears to be the predominant connexin family member in REKs, these results are consistent with the dominant-negative effect of the fs260 mutant being restricted to Cx43. Unlike the ODDD-linked G138R mutant, the presence of the fs260 mutant reduced the level of endogenous Cx43 found in REKs, suggesting this mutant may cause a skin phenotype in humans by destabilizing endogenous Cx43 in keratinocytes and reducing the overall levels of Cx43-based GJIC. This concept remains to be tested further by expressing a variety of ODDD-linked Cx43 mutants in primary human keratinocytes.
Somewhat to our surprise, the removal of the 46 aberrant amino acids from the frameshift mutant greatly enhanced the functional status of the Cx43 mutant. This finding suggested that the 46 aberrant amino acid extension following the site of the two nucleotide deletion is not benign and changes the phenotype of the mutant. Previously, a cataract-causing Cx46 frameshift mutant, Cx46fs380, containing a sequence of 87 aberrant amino acids starting at residue 380 was characterized and found to be incapable of forming functional gap junction channels or hemichannels in Xenopus oocytes (
). In a subsequent study, this same mutant was found to have impaired trafficking, which was attributed to the expression of a “FF” motif encoded within the aberrant amino acid domain (
). Coincidently, we found a similar “FF” motif in the aberrant sequence of our fs260 mutant (Fig. 1B, II). However, mutation of the “FF” motif to “AA” in the fs260 mutant minimally rescued the formation of gap junction plaques and the assembly of gap junctions composed of endogenous Cx43. These results suggest that sequences other than or including the motif encompassing the two phenylalanines in the aberrant 46 amino acid residues of the 260fs mutant contribute to this trafficking defect.
Several lines of evidence indicate that the mere truncation of Cx43 and the elimination of the large C-terminal domain does not consistently inhibit Cx43 trafficking to the cell surface or the formation of a functional gap junction channel (
). In a previous study, while the T263 truncated mutant was still able to form functional gap junctions, a slightly shorter T231 truncated mutant accumulated within intracellular compartments and reduced the delivery of wild-type Cx43 to the cell surface (
). It has been proposed that the region between amino acid 231 and amino acid 263 is critical for the transport of Cx43 to the cell surface to form functional channels (
). Not unlike our findings where the Cx43 truncated mutant, T259-GFP, had ∼50% wild-type Cx43-GFP macroscopic gap junction conductance levels, N2A cell pairs expressing untagged M257 were coupled with an average macroscopic conductance of 23.4 nS in comparison to 55.3 nS for untagged wild-type Cx43-expressing cell pairs (
). Thus, these studies suggest that the maximum macroscopic junctional coupling that can be achieved is reduced when the C-terminal tail of Cx43 is removed and this finding is not dependent on the presence or absence of the GFP tag.
To date only two patients harboring sporadic K134E or L11P ODDD-linked Cx43 mutation have been identified with any evidence of a skin abnormality (
). In both of these cases, the patient exhibited some degree of hyperkeratosis. Given the fact that only one patient harboring each of these sporadic Cx43 mutations have been described, it is difficult to conclude any generic features that might apply to these particular mutations and the skin phenotype. Nevertheless, it is interesting to note that the intracellular loop region encompassing K134E is believed to interact with the CT of Cx43 in mediating the pH gating (
). However, patients harboring other Cx43 mutants in the same IL region are not known to exhibit skin abnormalities, thus we cannot rule out the possibility that the hyperkeratosis of the palms and soles of the K134E patient might be caused by other factor(s).
In summary, the frameshift Cx43 mutant exhibited loss-of-function and dose-dependent, dominant-negative properties on wild-type Cx43, which may be related to the underlying cause for ODDD. From these findings we speculate that the additional condition of palmoplantar keratoderma not seen in patients harboring familial missense or codon duplication may be related to the loss of most of the phosphorylation sites and interactive sites for Cx43-binding proteins. In particular, Cx43 binding proteins that include ZO-1 interact with other structural junctions allowing cross-talk between gap junctions and both tight and adherens junctions. In the future, it will be of interest to determine if the expression of the fs260 mutant causes dysregulation of the junctional nexus resulting in skin disease.
84095-4&id=gr1.jpg){kind=link}
84095-4&id=gr2.jpg){kind=link}
84095-4&id=gr4.jpg){kind=link}
84095-4&id=gr8.jpg){kind=link}
84095-4&id=gr3.jpg){kind=link}
84095-4&id=gr5.jpg){kind=link}
84095-4&id=gr6.jpg){kind=link}
84095-4&id=gr7.jpg){kind=link}