Cataracts Are Caused by Alterations of a Critical N-Terminal Positive Charge in Connexin50

Abstract

purpose. To elucidate the basis of the autosomal dominant congenital nuclear cataracts caused by the connexin50 mutant, CX50R23T, by determining its cellular distribution and functional behavior and the consequences of substituting other amino acids for arginine-23.

methods. Connexin50 (CX50) mutants were generated by PCR and transfected into HeLa or N2a cells. Expressed CX50 protein was detected by immunoblot analysis and localized by immunofluorescence. Intercellular communication was assessed by microinjection of neurobiotin or by double whole-cell patch-clamp recording.

results. HeLa cells stably transfected with CX50R23T or wild-type CX50 produced immunoreactive CX50 bands of identical electrophoretic mobility. Whereas HeLa cells stably expressing CX50 contained abundant gap junction plaques, CX50R23T localized predominantly in the cytoplasm. HeLa cells expressing wild-type CX50 showed large gap junctional conductances and extensive transfer of neurobiotin, but those expressing CX50R23T did not show significant intercellular communication by either assay. Moreover, CX50R23T inhibited the function of coexpressed wild-type CX50. Three CX50R23 substitution mutants (CX50R23K, CX50R23L, and CX50R23W) formed gap junction plaques, whereas two mutant substitutions with negatively charged residues (CX50R23D, CX50R23E) did not form detectable plaques. Only the mutant with a positive charge substitution (CX50R23K) allowed neurobiotin transfer at levels similar to those of wild-type CX50; none of the other mutants induced transfer.

conclusions. These results suggest that replacement of amino acid 23 in CX50 by any residue that is not positively charged would lead to cataract formation.

Cataracts are a major cause of impaired vision and blindness throughout the world. ¹ Congenital cataracts are responsible for 10% of childhood blindness and are a common treatable cause of childhood visual impairment. ² A substantial fraction of isolated (nonsyndromic) childhood cataracts are inherited. Inheritance is most commonly autosomal dominant, although autosomal recessive and X-linked forms have also been reported. Hereditary congenital cataracts have been linked to mutations of several different genes, including the gap junction proteins connexin46 (CX46) and connexin50 (CX50). ^{3 4 5 6 7 8 9 10 11 12 13 14 15 16 17}  

Maintenance of a transparent lens depends on gap junction-mediated communication between its cells. Gap junctions are plasma membrane specializations containing clusters of intercellular channels. These channels allow intercellular transfer of ions and molecules of ≤1000 Da. Within the lens, gap junction channels form a network of cell-to-cell channels that facilitate exchange of water, ions, nutrients, and metabolites, especially between fiber cells. ^{18 19}  

A gap junction channel contains 12 subunit proteins (connexins) distributed among two coaxially aligned hexameric hemichannels located in the plasma membranes of the apposing cells. Lens fiber cells primarily express CX50 ²⁰ and CX46, ²¹ whereas the lens epithelial cells mainly express connexin43 (CX43). ^{22 23} Large numbers of gap junction channels may cluster together within the appositional membranes to form plaques. 

Genetic data from rodents have also emphasized the importance of alterations of the lens fiber connexins in cataractogenesis. Several strains of mice or rats that develop cataracts harbor mutations of the Gja3 (CX46) or Gja8 (CX50) genes. ^{24 25 26 27 28} Targeted ablation of the murine Gja3 (CX46) ²⁹ or Gja8 (CX50) ^{30 31} genes also causes cataracts. In fact, not only loss of CX50, but also its overexpression, can result in cataracts. ³²  

Investigators in several previous studies have examined the behavior of cataract-associated mutant connexins in expression systems, to elucidate cellular and physiological abnormalities that may lead to cataract formation. Some lens connexin mutants exhibit loss of function—for example, small or no gap junction conductances were detected when they were expressed in paired Xenopus oocytes. ^{17 28 33 34} When coexpressed with wild-type connexins, some mutants exert a dominant-negative effect abolishing conductance, ^{15 35 36} whereas some other mutants do not eliminate coupling, but may affect channel gating properties. ³⁷ Several mutants do not form gap junction plaques. ^{15 36} In one of these cases a unique sequence in the mutant protein (generated by a frame shift) impairs trafficking to the plasma membrane. ³⁸  

The human mutation, CX50R23T (which leads to the substitution of a threonine for arginine-23), was first described in members of an Iranian family with autosomal dominant, bilateral, congenital nuclear cataract that progressed to cause visual impairment by the second decade of life. ⁹ The mechanisms responsible for the development of this cataract are not understood. However, this mutation affects a critical position within the connexin molecule at the predicted interface between the N-terminal cytoplasmic domain and the first transmembrane domain (Fig. 1) . Structural models suggest the importance of this residue for maintaining associations of transmembrane domains. ³⁹ Substitutions of this amino acid have been identified in the mutant forms of at least three other connexins associated with other diseases. ^{40 41 42 43 44} The present experiments were designed to elucidate the mechanism of cataract formation in patients carrying the mutation CX50R23T by examining the cellular distribution and physiological function of the CX50R23T mutant and of site-directed mutants containing other amino acid substitutions at this position. 

Materials and Methods

Generation of Connexin Mutants

CX50 mutants were generated by PCR using DNA polymerase (Phusion High-Fidelity polymerase; New England Biolabs, Ipswich, MA) and human CX50 wild-type DNA in pcDNA3.1(+) or pcDNA3.1/Hygro(+) (Invitrogen, Carlsbad, CA) as a template. The primers were designed and mutagenesis was performed according to the strategy used previously by Minogue et al. ³⁸ Primers facing opposite directions and spanning the DNA region encoding amino acids 14 to 35 were designed to amplify the sequence of the full construct (including the vector sequence); the plasmid was regenerated by re-ligation of the PCR product. All mutations were produced using the same sense primer (encoding amino acids 25-35): 5′-TGGCTCACCGTGCTTTTCATCTTCCGGATCCTC-3′. 

The antisense primers encoded amino acids 14 to 24 and contained the mutated codon 23 (designated in bold): 5′-GACTGTGCCGATGACGGTGGAGTGCTCATTCAC-3′ (CX50R23T); 5′-GACGTCGCCGATGACGGTGGAGTGCTCATTCAC-3′ (CX50R23D); 5′-GACTTCGCCGATGACGGTGGAGTGCTCATTCAC-3′ (CX50R23E); 5′-GACTTTGCCGATGACGGTGGAGTGCTCATTCAC-3′ (CX50R23K); 5′-GACTAAGCCGATGACGGTGGAGTGCTCATTCAC-3′ (CX50R23L); 5′-GACCCAGCCGATGACGGTGGAGTGCTCATTCAC-3′ (CX50R23W). 

The DNA sequences coding for the CX50 mutants were fully sequenced at the Cancer Research Center DNA Sequencing Facility of the University of Chicago, to ensure that PCR amplification did not introduce additional unwanted mutations. 

Chemicals

All chemicals were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise specified. 

Cell Culture and Transfections

All cell culture media and supplements were obtained from Invitrogen, unless otherwise noted. HeLa and Neuro-2a (N2a) cells were grown in MEM supplemented with 0.1 mM nonessential amino acids, 10% fetal bovine serum (US Bio-Technologies Inc., Pottstown, PA), 2 mM glutamine, 10 U/mL penicillin G, and 10 μg/mL streptomycin sulfate. Transfections were performed with transfection reagents (Lipofectin and PLUS Reagent) used according to the manufacturer’s instructions. For stable transfections of CX50R23T in pcDNA3.1(+), clones were selected by their resistance to geneticin (1 mg/mL). Generation of HeLa cells stably transfected with human CX50 (HeLaCx50) has been described. ³⁶  

Antibodies

Rabbit polyclonal anti-CX50 antibodies were characterized previously. ³⁶ Mouse monoclonal anti-CX50 IgM antibodies were purchased from Invitrogen. Cy3-conjugated goat anti-rabbit IgG, Cy2-conjugated goat anti-mouse IgG, and horseradish peroxidase-conjugated goat anti-rabbit IgG antibodies were obtained from Jackson ImmunoResearch (West Grove, PA). 

Immunoblot Analysis

Cells were grown to 90% to 100% confluence in 100-mm dishes and harvested in PBS supplemented with EDTA-free protease inhibitors (Complete Mini; Roche, Indianapolis, IN), 1 mM EDTA, and 2 mM PMSF. Harvested cells were sonicated and cell homogenates were stored at −20°C until subjected to electrophoresis and immunoblot analysis. Aliquots from cell homogenates were mixed with loading dye and boiled. The proteins were resolved on 10% SDS-containing polyacrylamide gels and transferred to membranes (Immobilon P; Millipore, Bedford, MA). The membranes were blocked in 5% nonfat dry milk in Tris-buffered saline (TBS; pH 7.4) for at least 30 minutes at room temperature and incubated in rabbit polyclonal anti-CX50 antibodies overnight at 4°C. After they were rinsed in TBS and TBS containing 0.05% Tween 20, the membranes were incubated in HRP-conjugated goat anti-rabbit IgG antibodies for 30 minutes at room temperature, followed by more washes in TBS. The detection of bound HRP-conjugated antibodies was performed with enhanced chemiluminescence (GE Healthcare, Piscataway, NJ) and exposure of blots to blue biofilm (Denville Scientific, South Plainfield, NJ). 

Immunofluorescence

Cells plated on four-well chamber slides (Lab Tek; Nalge Nunc International, Naperville, IL) or glass coverslips were allowed to reach 80% to 90% confluence and subjected to immunofluorescence staining, as described previously. ³⁸ Briefly, the cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS; pH 7.4), blocked and permeabilized in 4% normal goat serum with 1% Triton X-100 in PBS and incubated in primary antibodies, either overnight at 4°C or for 2 hours at room temperature. The cells were then rinsed with PBS and incubated in Cy2- and/or Cy3-conjugated secondary antibodies for 30 to 45 minutes at room temperature. After the cells were rinsed with PBS, the coverslips were mounted using 2% n-propylgallate in PBS/glycerol (1:1). 

Specimens were studied by microscopy with a ×40 objective (Plan Apochromat in an Axioplan 2 microscope; (Carl Zeiss Meditec, Munich, Germany). The images were captured with a digital camera (Axiocam; Carl Zeiss Meditec) using Zeiss AxioVision software. Overlap images were generated with one of two image-analysis software programs (Image J (http://rsb.info.nih.gov/ij/ developed by Wayne Rasband and provided in the public domain by the National Institutes of Health, Bethesda MD; or Adobe Photoshop; Adobe Systems Inc., San Jose, CA). 

Microinjection of Gap Junction Tracers

Cells cultured on glass coverslips (80%–90% confluence) were transferred to F-12 medium (Invitrogen) buffered with 15 mM HEPES. Individual cells were impaled and injected for 3 minutes with a picospritzer (model PLI-188; Nikon, Tokyo, Japan) and a micropipette filled with 5% Lucifer yellow (charge = −2, M_r = 457; Molecular Probes-Invitrogen, Eugene, OR) and 4% neurobiotin (charge = +1, M_r = 322.8; Vector Laboratories, Burlingame, CA). 

After injection, the cells were fixed in 4% paraformaldehyde for 15 to 30 minutes and then permeabilized with methanol/acetone (1:1) for 2 minutes at room temperature. The neurobiotin tracer was detected by staining the cells with Cy3-streptavidin conjugate (Sigma-Aldrich) for 30 minutes at room temperature. The extent of intercellular transfer of both tracers was determined by counting the number of adjacent cells containing the tracer. Statistical analysis was performed with Student’s t-test. 

For injections of transiently transfected cells, the cells were cotransfected with a eukaryotic expression vector that produces GFP (GTX-EMD-pcDNA3.1neo, courteously provided by John Kyle, University of Chicago), and clusters of transfected cells were identified by their green fluorescence. 

Electrophysiological Methods

Experiments were performed on cell pairs. Junctional currents from cell pairs were recorded using the double whole-cell, patch-clamp technique, as described previously. ^{45 46} Briefly, glass coverslips with adherent cells were transferred to an experimental chamber mounted on the stage of an inverted microscope (model IX71; Olympus, Lake Success, NY). The chamber was perfused with bath solution containing 150 mM NaCl, 10 mM KCl, 2 mM CaCl₂, 5 mM HEPES (pH 7.4), 5 mM glucose, 2 mM CsCl, and 2 mM BaCl₂. Patch pipettes were pulled from glass capillaries (code GC150F-10; Harvard Apparatus, Holliston, MA) with a horizontal puller (DMZ-Universal, Zeitz-Instrumente, Munich, Germany). The patch pipettes were filled with solution containing 120 mM Kaspartate, 10 mM NaCl, 3 mM MgATP, 5 mM HEPES (pH 7.2), and 10 mM EGTA (pCa ∼8) that had been filtered through 0.22-μm pores. When filled, the resistance of the pipettes measured 1 to 5 MΩ. 

Results

Predominantly Cytoplasmic Distribution of CX50R23T

To examine the cellular behavior of CX50R23T, HeLa cells were stably transfected with the mutant (HeLaCX50R23T) and compared with HeLa cells that stably expressed wild-type CX50 (HeLaCX50). 

The presence of CX50 and CX50R23T protein in the transfected cells was verified by immunoblot analysis (Fig. 2A) . A band of the expected size (M_r ∼62 kDa) was detected in both cell lines. CX50 immunoreactivity was absent from homogenates of untransfected cells (not shown). 

The cellular distribution pattern of CX50 and CX50R23T in stable transfectants was determined by immunofluorescence. Wild-type CX50 immunoreactivity was detected at appositional membranes and in a perinuclear cytoplasmic location in HeLaCX50 cells (Fig. 2B) . Gap junction plaques were frequent, and many large plaques were observed. In contrast, the mutant CX50R23T was predominantly located in the cytoplasm (Figs. 2C 2D) . No gap junction plaques were present between most cells (Fig. 2C) . However, in rare cases, CX50R23T was found at appositional membranes; the very few plaques that could be observed were small (Fig. 2D) . 

Formation of Functional Gap Junction Channels

To test the ability of CX50R23T to form functional gap junction channels, we microinjected individual cells within monolayers of HeLaCX50 or HeLaCX50R23T cells with the gap junction tracers Lucifer yellow and neurobiotin and quantified intercellular communication by counting the number of tracer-containing neighbors (Fig. 3A) . Neither cell line allowed significant transfer of Lucifer yellow (Fig. 3B) , consistent with the lack of permeability of CX50 channels to this dye. ⁴⁷ The HeLaCX50 cells showed extensive intercellular transfer of neurobiotin (19.54 ± 3.40 coupled cells; n = 39); in contrast, the cells expressing CX50R23T showed almost no transfer of neurobiotin (0.37 ± 0.10 coupled cells, n = 38; Fig. 3B ). 

To examine the effects of the mutant on conductance of current-carrying ions, gap junctional coupling in pairs of HeLaCX50 or HeLaCX50R23T cells was also studied using the double whole-cell, patch-clamp technique (Table 1) . The HeLaCX50 cells were well coupled, with 100% of cell pairs studied showing coupling. These cell pairs had a mean total junctional conductance of 18.71 nS. In contrast, most pairs of HeLaCX50R23T cells were not coupled; moreover, the few cell pairs that were coupled had a mean total junctional conductance of only 0.50 nS. 

Representative examples of gap junction currents (I _j) recorded between HeLaCX50 and HeLaCX50R23T cell pairs are shown in Figure 4A(left and right, respectively). Currents elicited from HeLaCX50 cells using a symmetrical bipolar pulse protocol (5s, from ±10 to ±110 mV) showed voltage and time-dependent current deactivation. In the cases in which HeLaCX50R23T cells were coupled, junctional currents were of much lower amplitudes (as exemplified in Fig 4A , right); they resembled currents recorded from cells expressing CX45 (i.e., they were consistent with endogenous channels present in HeLa cells). 

The summary plot of normalized steady state conductances (g _ss) versus transjunctional voltage (V _j) from five HeLaCX50 cell pairs is shown in Figure 4B . The data were best fitted to the Boltzman equation with the following parameters: V _j,0 = −53 mV/+55 mV and normalized g _jmin = 0.26/0.22 for negative/positive V _j, respectively. 

In the HeLaCX50 and N2aCX50cells, channels were observed with unitary conductances of ∼200 pS, consistent with the single-channel sizes previously observed for CX50 channels. ^{47 48} Figure 4Cillustrates single-channel currents obtained from a poorly coupled N2aCX50 cell pair. Voltage steps of V _j = ±50 mV revealed associated unitary channel conductances of 194 and 200 pS for negative and positive V _j, respectively. In contrast, the few channels detected in coupled pairs of HeLa CX50R23T cells had single channel conductances of <40pS, consistent with endogenous channels that have been detected in HeLa cells ⁴⁹ (data not shown). 

To confirm that CX50R23T did not form low-conductance channels, CX50 and CX50R23T were transiently transfected into N2a cells, a cell line devoid of endogenous connexin channels. All the pairs of N2a cells expressing CX50 exhibited junctional conductances that were similar to those determined in the HeLaCX50 cell pairs (i.e., 16.2 nS). In contrast, none of the pairs of cells expressing CX50R23T were coupled (Table 1) . 

Because the cataracts are inherited as an autosomal dominant trait in the family where CX50R23T was discovered, ⁹ we also examined the consequences of coexpression of CX50R23T with its wild-type counterpart by transiently transfecting the HeLaCX50 cells with CX50R23T. We found very low gap junctional conductances in many of these cell pairs (indistinguishable from those in HeLaCX50R23T cells), and we found that some (6/18) of these cell pairs were totally uncoupled (Table 1) . 

Taken together, these data indicate that CX50R23T does not form functional channels allowing passage of either gap junction-permeant tracers (neurobiotin) or current-carrying ions, and it acts as a dominant-negative inhibitor of wild-type CX50. 

Effect of Substitution of Other Amino Acid Residues for R23

To study the importance of R23 in determining the ability of CX50 to produce gap junction plaques and to form functional channels, we generated several other CX50R23 substitution mutants. The mutants were chosen to replace amino acids that differed in charge, size, and/or polarity. R23 in CX50 was replaced with the negatively charged glutamate (E) or aspartate (D), positively charged lysine (K), bulky and nonpolar tryptophan (W), or nonpolar leucine (L). All these mutants, as well as wild-type CX50 and CX50R23T, were transiently transfected into HeLa cells and analyzed for plaque formation by immunofluorescence. As expected, the wild-type CX50 formed plaques (Fig. 5A) , whereas CX50R23T had a cytoplasmic distribution (Fig. 5B) . Detectable plaques failed to form in the negatively charged substitution mutants CX50R23D and CX50R23E (Figs. 5C 5D) . All other mutants (CX50R23K, CX50R23L, and CX50R23W) showed plaque formation (Figs. 5E 5F 5G) . 

To test whether functional gap junction channels would form in these mutants, the constructs were transiently transfected into HeLa cells, and intercellular communication was assessed after microinjection of the gap junction tracers Lucifer yellow and neurobiotin. Similar to the results obtained with the stably transfected cells, we found no evidence of Lucifer yellow transfer among cells expressing wild-type CX50. (None of the R23 substitution mutants allowed Lucifer yellow transfer, either.) As in the stably transfected cells, cells transiently transfected with wild-type CX50 showed extensive intercellular transfer of neurobiotin (Fig. 6) . Among the plaque-forming mutants, only the positively charged substitution CX50R23K exhibited neurobiotin transfer at levels comparable to those in cells expressing wild-type CX50 (Fig. 6) ; the number of coupled cells for CX50R23K (12.05 ± 2.97, n = 20) was not significantly different from those observed for CX50 (13.05 ± 2.85, n = 19). None of the other mutants, including the plaque-forming substitution mutants CX50R23L and CX50R23W, produced levels of neurobiotin transfer that were significantly different from those observed in untransfected (parental) HeLa cells (Fig. 6) . 

Because CX50R23L and CX50R23W formed gap junction plaques but did not allow transfer of neurobiotin, they were also tested for electrical coupling by double whole-cell, patch-clamp recording after transient transfection. When expressed by themselves in N2a cells, we never observed coupling (n = 15 and 18 for CX50R23L and CX50R23W, respectively). Moreover, each of these mutants acted as a dominant-negative inhibitor of wild-type CX50, abolishing conductances when transfected into HeLaCX50 cells. For HeLaCX50 transfected with CX50R23L, 11 of 11 cell pairs tested were uncoupled. For HeLaCX50 transfected with CX50R23W, 10 of 11 cell pairs were uncoupled, whereas one cell pair had a conductance of 2.8 nS. Control HeLaCX50 cells studied in parallel had a mean conductance of 27.5 nS (n = 4). 

Discussion

The results suggest two abnormalities of the cataract-associated human mutant CX50R23T: (1) It is severely impaired in trafficking and formation of gap junction plaques, and (2) it is nonfunctional. The loss of function (and dominant-negative inhibition of the function of coexpressed wild-type CX50) may explain the development of cataract in affected patients, since cataract also develops in mice with deletion of CX50. ³⁰  

The mutation R23T is located at the boundary between the cytoplasm and the predicted beginning of the first transmembrane domain of CX50. It replaces a positively charged amino acid with an uncharged, polar amino acid. All connexins contain a positively charged amino acid (either arginine [R] or lysine [K]) at the corresponding position. Thus, it may not be surprising that we found no cellular or functional consequence of substituting lysine for arginine in CX50R23K. 

Positively charged amino acids in the hydrophilic intracellular domains of membrane proteins are thought to play an important role in the orientation and insertion of hydrophobic transmembrane elements. ⁵⁰ Of the six substitutions for amino acid R23 in CX50 that we tested, three readily formed gap junction plaques when transiently transfected into HeLa cells (R23K, R23L, R23W); only the two that contained negatively charged amino acids (R23D and R23E) completely failed to form plaques. Thus, a complete reversal of the positive charge may lead to removal of a membrane anchor disrupting connexin membrane topology (leading to the observed absence of gap junction plaques). 

The importance of this positively charged residue is further emphasized by the identification of mutations of this residue in other connexins associated with other hereditary diseases including X-linked Charcot-Marie-Tooth disease (CX32R22G, CX32R22P, CX32R22E, CX32R22Q), ^{40 41 44} erythrokeratodermia variabilis (CX30.3R22H), ⁴² and oculodentodigital dysplasia (CX43K23T). ⁴³ Substitutions for R23 may have additional structural consequences, since Fleishman et al. ³⁹ have hypothesized that this amino acid may play a critical role in stabilizing the interactions of the connexin transmembrane helices by forming a salt bridge with a negatively charged amino acid at the cytoplasmic end of the fourth transmembrane domain. 

The N-terminal connexin domain appears to be critical for several aspects of connexin function. Site-directed mutagenesis studies have shown that residues in this region influence various physiological properties including transjunctional voltage-dependent gating, unitary conductance, sensitivity to regulation by polyamines, and permeability. ^{51 52 53 54 55 56 57} In addition to R23, many of the other N-terminal amino acids (especially charged or polar ones) are altered in connexin-associated diseases such as sensorineural deafness (CX26, CX30, and CX31 http://davinci.crg.es/deafness/) or X-linked Charcot-Marie-Tooth disease (CX32 http://www.molgen.ua.ac.be/CMTMutations/). In addition, a mutation of the neighboring amino acid residue, G22, in the murine CX50, CX50G22R, causes semidominant cataract in the Lop10 mouse; CX50G22R fails to form plaques and functional gap junction channels by itself. ²⁶  

Our results demonstrate that plaque formation by a connexin variant does not ensure gap junction function. Among our CX50R23 mutants, three (R23K, R23L, and R23W) exhibited proper plasma membrane trafficking and gap junction plaque formation, but only one, CX50R23K, produced functional channels. A discrepancy between plaque formation and function has also been observed for disease-causing mutations within the N-terminal regions of CX26 and CX43. CX26S19T forms plaques, but does not allow intercellular transfer of Lucifer yellow. ⁵⁸ CX43Y17R and CX43G21R localize within gap junction plaques but do not induce junctional conductances in transfected N2a cells. ⁵⁹  

Individuals carrying the CX50R23T mutation have congenital nuclear cataracts that progress with age. ⁹ We have shown that CX50R23T is nonfunctional and can completely inhibit the function of wild-type CX50. Thus, even though affected individuals are heterozygous, the expression of the mutant allele would render them effectively null for CX50 function. Therefore, we hypothesize that this mutant leads to cataracts through loss of CX50 mediated intercellular communication. Mice that are homozygous for the targeted disruption of CX50 develop irregular, nuclear pulverulent lens opacities ^{30 31} ; the lenses of these mice exhibit morphologic changes as early as postnatal day 7 and show particulate precipitates within the nucleus at 6 months of age. ³⁰ Thus, the phenotype and progression of cataracts in humans carrying the CX50R23T mutation seem reasonably similar to those in CX50-null mice. 

In summary, replacement of a positively charged arginine by an uncharged polar threonine at position 23 in CX50 leads to impaired gap junction plaque formation and loss of function. Our current data demonstrate that the positive charge at position 23 is critical for normal CX50 function. We hypothesize that the replacement of amino acid residue 23 by amino acids that are not positively charged is likely to lead to formation of cataract. 

²  Present affiliation: Department of Ophthalmology, University of Heidelberg, Heidelberg, Germany.

 Supported by a fellowship within the Postdoctoral Program of the German Academic Exchange Service (BCT) and by National Eye Institute Grant EY08368 (ECB) and National Heart, Lung, and Blood Institute Grant HL59199 (ECB, PRB).

 Submitted for publication December 27, 2007; revised February 20, 2008; accepted April 16, 2008.

 Disclosure: B.C. Thomas, None; P.J. Minogue, None; V. Valiunas, None; G. Kanaporis, None; P.R. Brink, None; V.M. Berthoud, None; E.C. Beyer, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Diagram of the membrane topology of human CX50 showing the position of the R23T substitution.

Figure 2.

 

Expression levels and immunolocalization of wild-type CX50 and CX50R23T proteins in stably transfected HeLa cells. (A) Immunoblot analysis of HeLaCX50 and HeLaCX50R23T with polyclonal anti-CX50 antibodies revealed similar levels of immunoreactive CX50 protein of identical electrophoretic mobility. Left: migration positions of the molecular mass standards. (B–D) Immunofluorescence localization of wild-type CX50 (B) and CX50R23T (C, D). The wild-type CX50 formed abundant gap junction plaques (B, arrows). CX50R23T localized predominantly in the cytoplasm (C), but it also showed very rare formation of small gap junction plaques (D, arrows). Bar: (B) 20 μm; (C) 16 μm; (D) 12 μm.

Figure 3.

 

Intercellular transfer of microinjected gap junction permeant tracers in stably transfected HeLa cells. (A) Photomicrographs obtained after microinjection of a solution containing Lucifer yellow (LY) and neurobiotin (NB) into HeLa cells expressing CX50 (wt) or CX50R23T (R23T). Neither cell line showed significant intercellular transfer of Lucifer yellow; this yellow dye allows identification of the injected cell. Cells expressing wild-type CX50 displayed extensive transfer of neurobiotin, whereas those expressing CX50R23T did not allow significant intercellular passage of this tracer. Bar: (top) 40 μm; 50 μm (bottom). (B) The number of HeLaCX50 (wt) or HeLaCX50R23T (R23T) cells receiving Lucifer yellow (▪) or neurobiotin ( ) from the injected cell. Bars represent the mean ± SEM; n = 39 (wt), n = 38 (R23T); *P < 0.0001, Student’s t-test.

Table 1.

 

Gap Junction Conductances Induced by Wild-Type CX50 or CX50R23T in Transfected HeLa and N2a Cells

Cell Type	Connexin	Total Cell Pairs (n)	Coupled Cell Pairs (n)	Uncoupled Cell Pairs (n)	Conductance (Mean nS ± SEM)
HeLa	CX50	10	10	0	18.71 ± 2.91
	CX50R23T	19	5	14	0.50 ± 0.21
	CX50+CX50R23T	18	12	6	0.20 ± 0.25
N2a	CX50	12	12	0	16.2 ± 4.5
	CX50R23T	19	0	19	0.0 ± 0.0

Figure 4.

 

Macroscopic and single-channel properties of CX50 gap junctions. (A) Gap junction currents (I _j) elicited by a bipolar pulse protocol from HeLa cell pairs transfected with wild-type CX50 (left) and CX50R23T (right). (B) Summary plot of normalized steady state g _j versus V _j from five HeLaCX50 cell pairs. The continuous line represents the best fit of the data to the Boltzman equation with V _j,0 = −53 mV/+55 mV and g _j,min= 0.26/0.22, for negative/positive V _j, respectively. (C) Pulse protocol (V ₁ and V ₂) and associated single-channel current (I ₂) recorded from a N2aCX50 cell pair during maintained V _j of ± 50 mV. Dashed line: zero current level. The all-points current histograms revealed conductances of 194 and 200 pS for negative and positive V _j, respectively.

Figure 5.

 

Cellular localization of CX50R23 substitution mutants. HeLa cells were transiently transfected with wild-type CX50 (A), CX50R23T (B), CX50R23D (C), CX50R23E (D), CX50R23K (E), CX50R23L (F), or CX50R23W (G), and CX50 immunoreactivity was detected by immunofluorescence microscopy. Only CX50R23K, CX50R23L, and CX50R23W formed gap junction plaques (arrows). CX50R23D and CX50R23E failed to form plaques. Bar: (A, B, E, G) 20 μm; (C) 17 μm; (D) 13 μm.

Figure 6.

 

Intercellular transfer of neurobiotin in HeLa cells expressing CX50R23 substitution mutants. Shown are the number of coupled cells that received the tracer from the injected cell. Only CX50R23K, the mutant containing the positive charge substitution, allowed neurobiotin transfer at levels similar to wild-type CX50. None of the other mutants induced intercellular transfer of neurobiotin at levels that differed significantly from those observed in nontransfected (parental) HeLa cells. Bars represent the mean ± SEM; *P < 0.05 compared with untransfected HeLa cells, Student’s t-test.