Mutations within connexin50 (Cx50) have been linked to various cataract phenotypes. To determine the mechanism behind cataract formation we used the paired Xenopus oocyte system in conjunction with transfected HeLa cells and genetically engineered mouse models to examine the functional characteristics of gap junctions in which a cataract-causing mutant of Cx50 (hereafter referred to as Cx50-S50P) is expressed. Channels comprising Cx50-S50P subunits alone failed to induce electrical coupling. However, the mixed expression of Cx50-S50P and wild-type subunits of either Cx50 or Cx46 – to create heteromeric gap junctions – resulted in functional intercellular channels with altered voltage-gating properties compared with homotypic wild-type channels. Additionally, immunofluorescence microscopy showed that channels of Cx50-S50P subunits alone failed to localize to the plasma membrane – unlike channels composed of Cx46 subunits, which concentrated at cell-cell appositions. Cx50-S50P colocalized with wild-type Cx46 in both transfected HeLa cells in vitro and mouse lens sections in vivo. Taken together, these data define the electrophysiological properties and intracellular targeting of gap junctions formed by the heteromeric combination of Cx50 or Cx46 and Cx50-S50P mutant proteins. Additionally, mixed channels displayed significantly altered gating properties, a phenomenon that may contribute to the cataract that is associated with this mutation.
Gap junctions provide a vital pathway for the intercellular transport of the ions and small molecules essential for correct growth and development of multicellular organisms (Bruzzone et al., 1996; Evans and Martin, 2002; Harris, 2001). The structural and, thus, physiological diversity of gap junctions are dependent on the specific connexins expressed in the channel. Currently, the connexin gene family contains over 20 members that are expressed in an overlapping, tissue-dependent pattern (Gerido and White, 2004; Willecke et al., 2002). Each connexin contains a cytosolic N- and C-terminus, as well as four transmembrane domains, linked by a single cytoplasmic loop and the two extracellular loops E1 and E2 (Makowski et al., 1977; Milks et al., 1988). Six connexin proteins oligomerize to form a connexon or hemichannel in the plasma membrane. Hemichannels can be of uniform (homomeric) or varying (heteromeric) connexin composition. Neighboring cells may contribute heteromeric or homomeric connexons to form complete intercellular channels known as homotypic (the combination of two identical uniform connexons), heteromeric (the association of two hemichannels varying in connexin expression), and heterotypic channels (the alignment of two connexons each containing a single unique connexin). Recent studies have shown that altering the connexin composition of gap junctions changes both the permeability and electrophysiological properties of these channels in vitro (Bevans et al., 1998; Cao et al., 1998; Donaldson et al., 1995; Goldberg et al., 1999; Niessen et al., 2000; Valiunas et al., 2002; Veenstra, 1996; White, 2003; White and Bruzzone, 1996).
The vertebrate lens is a functional syncytium connected by gap junctions. In the mammalian lens three connexin family members exhibit distinct but overlapping expression patterns, facilitating the intercellular communication essential to proper organ function. Cx43 is expressed primarily in the epithelium but absent from lens fibers, whereas Cx46 exhibits the opposite expression pattern – as it is found exclusively in lens fiber cells (Gong et al., 1997; Paul et al., 1991). Conversely, Cx50 is present in both, the epithelium and differentiated fibers (Rong et al., 2002; White et al., 1992). This overlapping, yet distinct, pattern of connexin expression contributes to a variation in junctional communication based on the connexin composition in specific cell types (Donaldson et al., 2001; Donaldson et al., 1995; Mathias et al., 1997).
Mutations within the Cx46 and Cx50 gene loci have been linked to various cataract phenotypes in humans and mice (Chang et al., 2002; Graw et al., 2001; Mackay et al., 1999; Runge et al., 1992; Shiels et al., 1998; Steele et al., 1998; Xia et al., 2006b; Xia et al., 2006a; Xu and Ebihara, 1999). In the coding region of Cx50, one such missense mutation caused the substitution of proline for serine at amino acid 50 (S50P) of the protein (Xia et al., 2006b). Mice expressing this mutated protein (hereafter referred to as Cx50-S50P) exhibit a dominant cataract phenotype signified by altered fiber cell formation, dense cataract and posterior capsule rupture (Xia et al., 2006b). Here, we hypothesize that the mutant Cx50-S50P subunits interact with wild-type Cx46 and Cx50 subunits to modulate cataract formation through alterations in the electrophysiological gating properties of heteromeric gap junctions formed between two lens fiber cells. To test this hypothesis we used the dual whole-cell voltage-clamp technique in conjunction with the paired Xenopus oocyte system, and transfected HeLa cells to identify alterations in protein localization and the voltage-gating characteristics of gap junctions comprising wild-type and/or Cx50-S50P mutant subunits.
Connexin expression in Xenopus oocytes
The production of lens fiber connexins in Xenopus oocytes was examined via immunoblot analysis. Oocytes injected with water, wild-type Cx50, Cx46 or Cx50-S50P alone and cells expressing both mutant and wild-type connexins were analyzed using a polyclonal antibody specific for the central cytoplasmic loop of mouse Cx50 (White et al., 1992) or a polyclonal Cx46 antibody raised against the extreme C-terminus (Paul et al., 1991). Immunoblotting revealed a ∼60 kDa band corresponding to Cx50 in oocyte samples injected with wild-type or mutant cRNA transcripts (Fig. 1A). An antibody specific for Cx46 detected two bands of ∼46 kDa in samples injected with either Cx46 cRNA alone or co-injected with Cx46 and Cx50-S50P (Fig. 1B). Expression levels were quantified using band densitometry on replicate immunoblots (n=3). A plot of the mean band-intensity values (normalized to wild-type mean value) (Fig. 1C) showed no significant reduction of Cx50 expression in oocytes injected with wild-type Cx50 or with Cx50-S50P alone, or co-injected samples expressing both wild-type and mutant Cx50 proteins (P>0.05, Student's t-test). Similarly, samples tested for Cx46 expression show equivalent levels of Cx46 production (Fig. 1D) in samples injected with Cx46 alone or in conjunction with Cx50-S50P, as band densitometry showed no significant change in mean band intensity (P>0.05, Student's t-test). Thus, both wild-type and mutant transcripts were synthesized equally and any alteration in channel function cannot be attributed to differences in protein expression.
Cx50-S50P fails to form functional channels in Xenopus oocytes
To test the hypothesis that the S50P mutation alters intercellular communication in the lens we measured gap-junctional conductance Gj in oocyte pairs injected with various combinations of lens fiber connexin cRNAs (Fig. 1E). Control oocyte pairs injected with anti-sense oligonucleotide and water showed negligible junctional conductance. By contrast, homotypic wild-type Cx50 or Cx46 channels displayed mean conductance values of ∼26 μS or ∼20 μS, respectively, a significant increase in Gj and several 100-fold greater than that of the background (P<0.05, Student's t-test). Interestingly, cells injected with both wild-type and mutant Cx50 cRNA transcripts paired to form heteromeric channels, exhibit a mean Gj of ∼20 μS, which was not significantly different compared with homotypic Cx50 channels (P>0.05, Student's t-test). Similarly, oocytes paired to form heteromeric gap junctions comprising Cx46 and Cx50-S50P subunits displayed a mean conductance of 17 μS, a decrease in Gj not significantly different compared with the homotypic Cx46 channel (P>0.05, Student's t-test). Homotypic channels that comprise mutant Cx50-S50P subunits produced a level of conductance significantly lower than that of homotypic wild-type Cx50, indicating that Cx50-S50P proteins alone fail to form functional intercellular channels (P<0.05, Student's t-test). Additionally, the heterotypic pairing of oocytes expressing Cx50 or Cx46 with those containing Cx50-S50P alone failed to significantly couple cells, because levels of junctional conductance were similar to that of the background measured in water-injected pairs. Co-injection of Cx50-S50P with wild-type Cx50 or Cx46 facilitated the formation of functional intercellular channels that displayed slight reductions in mean conductance compared with homotypic wild type. These results show that mutant Cx50 subunits fail to form homotypic or heterotypic gap junctions on their own, and do not significantly inhibit conductance when co-expressed with wild-type lens fiber connexins.
Voltage-gating behavior of wild-type and heteromeric channels
To characterize the electrophysiological properties of homotypic wild-type channels or heteromeric channels expressing both wild-type and mutant connexins, voltage-gating was examined by subjecting oocyte pairs to a series of hyperpolarizing and depolarizing transjunctional voltages (Vj). Fig. 2 shows a comparison of representative traces of junctional currents (Ij) for oocyte pairs expressing homotypic Cx50 (Fig. 2A), and mixed-gap-junction-expressing wild-type and mutant Cx50 proteins (Fig. 2B). Ij decreased symmetrically with time in a voltage-dependent manner for both channel types, and showed a more rapid current decline at greater Vj applications. To analyze channel closure kinetics, the initial 250 mseconds of the decay of current was plotted against time and fit to a mono-exponential function to determine the time constant tau (τ). Representative decays of current were recorded during an application of a +80 mV voltage step for homotypic wild-type (Fig. 2C) and mixed channels that express both wild-type Cx50 and mutant subunits (Fig. 2D). Mixed channels closed 15% faster than wild-type channels, with mean channel closure times of 61 vs 52 mseconds, respectively (P <0.05). At a larger transjunctional potential of 120 mV, both channel configurations closed faster than at 80 mV, exhibiting mean τ values of 34 and 25 mseconds for homotypic wild-type (Fig. 2E) and co-injected (Fig. 2F) channels, respectively (P<0.05, Student's t-test). These slight alterations in channel closure kinetics might indicate an interaction between Cx50-S50P and wild-type Cx50 subunits when the two proteins are co-expressed in cells.
The comparison of homotypic channels comprising Cx46 (Fig. 3A) and heteromeric channels comprising Cx46 and Cx50-S50P subunits (Fig. 3B) revealed that heteromeric channels appear to be more sensitive at larger voltage applications (±80-120 mV) than their homotypic counterparts. Similarly, mixed channels expressing wild-type Cx46 and Cx50 (Fig. 3C) showed that larger voltage potentials result in a more rapid decline towards steady-state when compared with heteromeric gap junctions containing wild-type Cx46 and Cx50-S50P subunits. Comparison of the three representative Ij decays, Cx46 alone (Fig. 3D), gap junctions expressing both Cx46 and Cx50-S50P proteins (Fig. 3E) and channels comprising wild-type Cx46 and Cx50 (Fig. 3F), using one-way analysis of variance (ANOVA) during an 80 mV voltage application reveal that there were statistically significant differences (P<0.05, ANOVA) in the mean channel closure times for all three channel types analyzed. Similarly, when a 120 mV potential was applied to oocyte pairs connected by homotypic Cx46 gap junctions a mean channel closure time of 126 mseconds was recorded (Fig. 3G). Comparison of this τ value with channels comprising wild-type Cx46 and Cx50-S50P (90 mseconds, Fig. 3H), and channels comprising wild-types Cx50 and Cx46 (75 mseconds, Fig. 3I) showed that there were statistically significant differences between the three mean τ values (P<0.05, ANOVA). Thus, the behavior of mixed channels containing wild-type Cx46 and Cx50-S50P was functionally distinct from either homotypic Cx46 channels or heteromeric gap junctions expressing wild-type Cx46 and Cx50 subunits. These data support the hypothesis that Cx50-S50P interacts with wild-type Cx46 to form functional gap junctions with unique voltage gating properties.
The steady-state voltage gating behavior of these channels was examined by plotting Vj against Gj (normalized to the values obtained at ±20 mV) and fitting the data to the Boltzmann equation (Fig. 4, Table 1). Analysis of the equilibrium gating properties showed no drastic changes in the steady-state properties of homotypic Cx50 channels and mixed gap junctions co-expressing wild-type and mutant Cx50 (Fig. 4A). Conversely, mixed channels containing both Cx46 and Cx50-S50P subunits (Fig. 4B, •) displayed a visible shift towards lower transjunctional voltages in the Boltzmann plot when compared with wild-type Cx46 channels (Fig. 4B, □), indicating that heteromeric channels were more voltage sensitive than the homotypic wild-type Cx46 junctions at all tested voltages. Heteromeric channels that comprised wild-type Cx46 and Cx50 (Fig. 4B, ♦) were similar to those containing Cx46 and Cx50-S50P at negative transjunctional potentials, but showed a statistically significant alteration in voltage sensitivity at larger positive potentials (Fig. 4B, *). One-way analysis of variance (ANOVA) revealed that there were statistically significant differences between all three data sets at all positive potentials ≥60 mV (P<0.05).
Taken together, these results provided evidence for the heteromeric interaction between wild-type lens fiber connexins and mutant Cx50-S50P subunits in vitro. Interestingly, the co-expression of wild-type Cx46 and Cx50-S50P presented strong evidence for the functional interaction between these two proteins via, the formation of functional intercellular channels with decreased junctional conductance, visibly altered junctional current decay and significantly altered channel closure kinetics. By contrast, the co-expression of wild-type and mutant Cx50 produced smaller shifts in the electrophysiological properties, although these achieved statistical significance and fully support the interaction of wild-type Cx50 and Cx50-S50P subunits in vitro.
Voltage-gating behavior of heterotypic channels
To determine whether the slight shift in mean channel closure time and the 20% reduction in conductance could be attributed to interaction of the Cx50-S50P mutant with wild-type Cx50, we analyzed the voltage gating characteristics of heterotypic channels where homomeric Cx46 connexons were paired with either Cx50 homomeric hemichannels or heteromeric connexons containing both wild-type and mutant Cx50 subunits. Junctional conductance measurements recorded from oocytes paired to form heterotypic channels (Fig. 5A) showed that heterotypic gap junctions containing only wild-type Cx46 and Cx50 display a mean Gj value of ∼22 μS a significant increase in coupling when compared to the 10.5 μS exhibited by heterotypic channels containing wild-type and mutant Cx50 subunits (P<0.05, Student's t-test). Both channel types displayed a level of conductance significantly higher than that of the water injected negative controls (P<0.05, Student's t-test). Representative current traces for heterotypic oocyte pairs derived from cells expressing only expressing Cx46 and paired with Cx50 alone (Fig. 5B), or oocytes co-expressing both wild-type and mutant Cx50 transcripts (Fig. 5C) exhibit no obvious changes in junctional current decay for all voltage applications assayed. To analyze the gating properties of heterotypic gap junctions, equilibrium gating properties of these channels were analyzed by applying the Boltzmann equation (Fig. 5D). Heterotypic channels expressing both wild-type and mutant Cx50 proteins showed a slight, but statistically significant increase in voltage dependence compared with heterotypic channels expressing wild-type Cx46 and Cx50 proteins alone, as indicated by the downward shift in the Boltzmann curve for positive values of Vj (*P<0.05, Student's t-test, Table 1). Taken together, these data indicate that the reduction in Gj that is seen in mixed channels comprising wild-type and mutant Cx50 proteins, as well as the slight alteration in voltage sensitivity was owing to the heteromeric interaction of CX50-S50P with wild-type Cx50.
Cx50-S50P requires Cx46 to form voltage-sensitive hemichannels
Several studies have suggested that abnormal hemichannel activity plays a role in the development of connexin-related diseases (Gerido et al., 2007; Jiang and Gu, 2005), and Cx46 and Cx50 have both been shown to form hemichannels in vitro (Beahm and Hall, 2002; Ebihara and Steiner, 1993). To test whether Cx50-S50P forms functional hemichannels that might have contributed to the mutant lens phenotypes seen in vivo, whole-cell currents (Im) were recorded from single oocytes expressing wild-type Cx46 or Cx50-S50P proteins alone or in combination (supplementary material Fig. S1). Oocytes expressing wild-type Cx46 alone revealed slowly activated outward currents that increased as voltage applications became larger and more positive. Conversely, CX50-S50P-injected oocytes exhibited negligible current flow at all voltage potentials analyzed. Mixed expression of wild-type Cx46 and Cx50, or wild-type Cx46 and CX50-S50P similarly displayed the presence of intermediate outward currents for both tested conditions. When steady-state currents were plotted against membrane voltage, negligible membrane currents were displayed by cells expressing only Cx50-S50P, whereas cells injected with wild-type Cx46 transcripts exhibited large currents that increased with Vm. Reduced whole-cell currents were observed for cells expressing both wild-type lens fiber connexins as well as cells containing wild-type Cx46 and CX50-S50P proteins. These data suggest that CX50-S50P fails to form hemichannels alone, but can form heteromeric hemichannels with wild-type Cx46.
Aberrant localization of Cx50-S50P in transfected cells and in the lens
It has been previously reported that mutant Cx50 proteins exhibited altered abilities to localize to the cell membrane as well as a reduced capacity to form gap-junctional plaques (Chang et al., 2002; Xia et al., 2006a). To determine whether the reductions in junctional conductance shown by channels of mixed connexin expression was the result of a failure to efficiently localize connexins to the plasma membrane, wild-type Cx46, and/or Cx50-S50P (subcloned into the pIRES2-EGFP and pCS2+ vectors, respectively) were expressed in transiently transfected HeLa cells.
Immunofluorescent images revealed sufficient connexin expression for both wild-type Cx46 (Fig. 6B and 6D) and Cx50-S50P (Fig. 6A and 6C) proteins. Cx50-S50P alone was unable to localize to sites of cell-cell apposition (Fig. 6, arrows), instead accumulating in subcellular compartments surrounding the nucleus (Fig. 6A). Conversely, Cx46 subunits were properly targeted to plasma membrane regions of cell-to-cell interaction in 75% of the cell pairs analyzed (Table 2), a phenomenon consistent with the formation of gap junctions (Fig. 6B, arrowhead). Interestingly, co-transfection of HeLa cells with both wild-type Cx46 and Cx50-S50P cDNA facilitated the colocalization of these proteins in 74% of the cell pairs examined (Fig. 6C and 6D).
Immunostaining of frozen lens sections further confirmed the notion that Cx50-S50P mutant proteins were stabilized in gap junction plaques by Cx46 in vivo. Fluorescent images showed that Cx50-S50P mutant proteins were only detected in the most superficial differentiating fiber cells of Cx50(S50P/S50P)–Cx46(–/–) lenses both lacking endogenous wild-type Cx50 and Cx46 alleles (Fig. 6F). Conversely, substantial staining was extended deeper into differentiating fibers of Cx50(S50P/S50P)–Cx46(+/+) lenses that expressed endogenous wild-type Cx46 (Fig. 6E). Thus, both in vivo and in vitro findings support the hypothesis that the mutant Cx50-S50P protein has the ability to associate with wild-type Cx46 subunits, a phenomenon that might play a role in cataract formation.
In this study, we have used a dual whole-cell voltage clamp method to characterize the electrophysiological properties of gap-junctional channels between paired Xenopus oocytes injected with different combinations of wild-type and mutant connexins. Dual whole-cell voltage clamp has been extensively used to analyze and obtain unique `electrical fingerprints' as a means of identifying functional differences between intercellular channels composed from diverse connexin subunits (Bruzzone et al., 1996; Harris, 2001). Whereas the transjunctional potentials created using this assay may not correlate with the endogenous resting potentials present in the mammalian lens, the distinct functional characteristics revealed by these techniques have enabled the detection of changes in intrinsic channel properties that could only result from the interaction of the Cx50-S50P mutant with wild-type Cx46 and Cx50. These functional interactions have been used to generate a model to explain the physiological pathologies that occur in vivo.
The data presented herein provide in vitro and in vivo evidence for the interaction between wild-type Cx50, wild-type Cx46 and mutant Cx50-S50P subunits. Furthermore, this study provides support for the hypothesis that a unique interaction between mutant and wild-type lens connexins modulates the aberrant lens phenotypes displayed by Cx50-S50P mutant mice (Xia et al., 2006b). Our findings showed that, although Cx50-S50P subunits alone failed to form functional intercellular channels or voltage activated hemichannels in Xenopus oocytes, or target to gap junctions in transfected HeLa cells, upon the mixed expression of wild-type lens fiber connexins and mutant Cx50-S50P subunits they were able to form functional gap junctions, and displayed levels of electrical coupling comparable to those recorded from homotypic wild-type channels. Additionally, heteromeric channels comprising either wild-type Cx46 or Cx50 and S50P proteins exhibited distinct voltage gating properties, and increased the ability of the mutant to localize in junctional plaques. Together, these data imply that unique interactions between connexins in the lens play a crucial role in correct organ function and development.
Recent studies have revealed that connexin diversity plays an integral role in the correct development and function of the mammalian lens, contributing to changes in channel gating and permeation (Gao et al., 2004; Le and Musil, 2001; Martinez-Wittinghan et al., 2003; Rong et al., 2002; White, 2003). In the past, different Cx50 variants have been described in the development of several distinct types of dominant cataract in mice (Chang et al., 2002; Graw et al., 2001; Steele et al., 1998; Xia et al., 2006a). For instance, recently published data revealed that the combination of mutants Cx50-G22R (Xia et al., 2006a) or Cx50-S50P (Xia et al., 2006b) and wild-type lens fiber connexins led to unique aberrations in lens development and function, and the presence of these mutant subunits in gap junction channels led to alterations in channel gating properties and cell-cell communication in vitro. Similarly, mice heterozygous for the S50P mutation showed a unique deficiency in primary fiber cell elongation indicating that an interaction between the wild-type and mutant Cx50 subunits governs primary fiber cell development, whereas homozygous mutant mice exhibited lenses with a unique deficiency in secondary fiber cell development, a finding that suggest involvement of Cx50 in a separate mechanism (Fig. 7) that modulates secondary fiber cell development (Gong et al., 2007; Xia et al., 2006b). Since the S50P mutant alone has been shown to be nonfunctional and failed to properly localize to the membrane (Fig. 1E, Fig. 6), we propose that the mutant must dominantly interact with wild-type subunits to impair lens development in vivo. Our results support these hypotheses and also provide the mechanistic foundation for the unique cataract phenotypes exhibited by the S50P mutant mice.
Taken together these data clearly demonstrate that Cx50 plays a crucial part in lens fiber cell development. Heterozygous Cx50(S50P/+)–Cx46(–/–) mice display a large cystic lumen between the anterior primary fiber cells and the epithelium during embryonic development, whereas postnatal fiber cell development proceeded normally. We propose that this phenomenon is governed by the unique interaction between the wild-type and mutant Cx50 subunits shown here that led to altered intercellular communication mediated by channels of mixed connexin expression. Moreover, we provide support for the hypothesis that Cx50 also plays a separate distinct role in postnatal fiber cell development in vivo. Interestingly, heterozygous Cx50(S50P/–)–Cx46(+/+) mice exhibit normal embryonic development in the presence of wild-type Cx46. However, these animals revealed disrupted secondary fiber development as well as the presence of a dense nuclear cataract and posterior capsule rupture. These findings taken in conjunction with the data shown here indicate an essential role for connexin diversity and Cx50-mediated intercellular communication in postnatal lens development.
In summary, these data indicate that mutated connexin subunits can interact with wild-type lens connexins to modulate both prenatal and postnatal lens development by altering channel function and cell-cell communication. Furthermore, varying the expression of wild-type lens connexins in vivo and in vitro has shown that wild-type Cx50 interacts with mutant Cx50-S50P in a manner that distinctly affects pre-natal development through a slight alteration in channel gating, whereas a separate unique interaction between Cx50-S50P and wild-type Cx46 impairs postnatal lens fiber development through more profound alterations in intercellular communication. The distinctive properties of the heteromeric channels formed by wild-type and mutant lens connexins provide the basic molecular foundation behind a variety of unique cataract phenotypes attributed to various Cx50 mutations (Arora et al., 2006; Graw et al., 2001; Xia et al., 2006a; Xu and Ebihara, 1999). These data indicate that mutated Cx50 proteins act as the molecular mechanism governing altered function of gap junction channels and aberrant intercellular communication in the lens. We believe that a better understanding of these underlying mechanisms may lead to the identification of molecular approaches to cataract prevention.
Materials and Methods
Murine Cx50 was cloned between the XhoI-XbaI sites of the pCS2+ expression vector. Mutant Cx50-S50P cDNA was generated from total mRNAs of homozygous mutant lenses by RT-PCR using a pair of primers (sense 5′-CGGGATCCTAGTGAGCAATGGGCGAC-3′ and anti-sense 5′-GGAATTCGTCATATGGTGAGATCATC-3′). Mutant cDNA was subcloned into pCR-bluntII-TOPO vector (Invitrogen, Carlsbad, CA) and sequenced to confirm the serine for proline amino acid substitution at position 50. Mutant Cx50-S50P cDNA was further subcloned into the pCS2+ and pIRES2-EGFP (Clontech, Palo Alto, CA) expression vectors using the EcoRI restriction sites for expression in Xenopus laevis, sequencing and immunofluorescent staining. Wild-type mouse Cx46 was inserted into the pCS2+ expression vector using the XhoI and SpeI restriction sites.
In-vitro transcription, oocyte microinjection and pairing
The aforementioned wild-type Cx50, Cx46 and mutant Cx50-S50P plasmids were linearized at the NotI restriction site of pCS2+, and transcribed using the SP6 mMessage mMachine (Ambion). Adult Xenopus females were anesthetized, the ovaries were removed and stage V-VI oocytes were collected after the ovarian lobes were de-folliculated in a solution containing 50 mg/ml collagenase B and 50 mg/ml hyaluronidase in modified Barths medium (MB) without Ca2+. Cells were first injected with 10 ng of antisense Xenopus Cx38 oligonucleotide to eliminate coupling caused by endogenous intercellular channels, and cultured overnight in MB medium containing 2 mM of CaCl2. Oligonucleotide-injected oocytes were then injected with either wild-type or mutant Cx50 cRNA transcripts (5 ng per cell) alone or in combination or H2O as a negative control. The vitelline membranes were then removed in a hypertonic solution (200 mM aspartic acid, 10 mM HEPES, 1 mM MgCl2, 10 mM EGTA, and 20 mM KCl at pH 7.4), and the oocytes were manually paired with the vegetal poles apposed in either normal MB medium or MB with elevated Ca2+ (2 mM CaCl2).
Dual whole-cell voltage clamping
Following an overnight incubation, gap-junctional coupling between oocyte pairs was measured using the dual whole-cell voltage clamp technique (Bruzzone et al., 2003; Spray et al., 1981). Current and voltage electrodes (1.2-mm diameter, omega dot; Glass Company of America, Millville, NJ) were pulled to a resistance of 1-2 MΩ with a horizontal puller (Narishige, Tokyo, Japan) and filled with solution containing 3.M KCl, 10 mM EGTA, and 10 mM HEPES pH 7.4. Voltage clamp experiments were performed using two GeneClamp 500 amplifiers (Axon Instruments, Foster City, CA) controlled by a PC-compatible computer through a Digidata 1320A interface (Axon Instruments). The pCLAMP 8.0 software (Axon Instruments) was used to program stimulus and data collection paradigms.
For measurements of junctional conductance, cell pairs were first clamped at –40 mV to eliminate any transjunctional potentials. Following the –40 mV clamp, a single cell was subjected to alternating pulses of ±10-20 mV, whereas the current produced by the change in voltage was recorded in the second cell. The current delivered to the second cell was equal in magnitude to the junctional current. Junctional conductance was calculated by dividing the measured current by the voltage difference, Gj=Ij/(V1-V2).
To determine voltage-gating properties, transjunctional potentials (Vj) of opposite polarity were generated by hyperpolarizing or depolarizing one cell in 20-mV steps (range: ±120 mV) while clamping the second cell at –40 mV. Currents were measured at the end of the voltage pulse, at which time they approached steady-state (Ijss). Macroscopic conductance (Gjss) was calculated by dividing Ijss by Vj, normalized to the values determined at ±20 mV, and plotted against Vj. Data describing the relationship of Gjss as a function of Vj were analyzed using Origin 6.1 (Microcal Software, Northampton, MA) and fit to a Boltzmann relation of the form: Gjss=(Gjmax–Gjmin)/1+ esp[A(Vj–V0)]=Gjmin, where Gjss is the steady-state junctional conductance, Gjmax (normalized to unity) is the maximum conductance, Gjmin is the residual conductance at large values of Vj, and V0 is the transjunctional voltage at which Gjss=(Gjmax –Gjmin)/2. The constant A=nq/kT represents the voltage sensitivity in terms of gating charge as the equivalent number (n) of electron charges (q) moving through the membrane, k is the Boltzmann constant, and T is the absolute temperature.
Preparation of oocyte samples for western blots
Oocytes were collected in 1 ml of buffer containing 5 mM Tris pH 8.0, 5 mM EDTA and protease inhibitors (White et al., 1992) and lysed using a series of mechanical passages through needles of diminishing caliber (20, 22 and 26 Ga). Extracts were centrifuged at 1000 g at 4°C for 5 minutes. The supernatant was then centrifuged at 100,000 g at 4°C for 30 minutes. Membrane pellets were resuspended in SDS sample buffer (2 μl per oocyte), samples were separated on 10% SDS gels and transferred to nitrocellulose membranes. Blots were blocked with 5% BSA in 1×PBS with 0.02% NaN3 for 1 hour and probed with a polyclonal Cx50 antibody, (White et al., 1992) or a polyclonal Cx46 antibody (Paul et al., 1991) followed by incubation with alkaline-phosphatase conjugated anti-rabbit secondary antibody (Jackson laboratories). Band intensities were quantified using Kodak 1D Image Analysis software (Eastman Kodak, Rochester, NY). Values from four independent experiments were normalized to the mean value of band intensity of the wild-type sample.
Electrophysiological recording of hemichannel currents
Macroscopic recordings of hemichannel currents were obtained from single Xenopus oocytes using a GeneClamp 500 amplifier controlled by a Digidata 1320 interface (Axon Instruments, Foster City, CA). pClamp 8.0 software (Axon Instruments) was used to program stimulus and data collection paradigms. To obtain hemichannel I-V curves, cells were initially clamped at –40 mV and subjected to 5-second depolarizing voltage steps ranging from –30 to +60 mV in 10-mV increments. Membrane currents were analyzed by recording hemichannel currents from cells injected with appropriate cRNA transcripts, incubated overnight in MB medium supplemented with 2 mM CaCl2.
Transient transfection and immunocytochemical staining
HeLa cells were plated on 22-mm square coverslips and grown to 50% confluence then transiently transfected with 2 μg of connexin DNA subcloned into the pIRES2-EGFP or pCS2+ vector using Lipofectamine 2000 (Gibco BRL, Gaithersburg, MD). After overnight incubation, cells were fixed with 1% paraformaldehyde in PBS, blocked with 5% BSA in PBS with 0.1% Triton X-100 and 0.02% NaN3. Cells were stained with either a polyclonal anti-Cx50 or anti-Cx46 antibody followed by incubation with an anti-Cy3 goat anti-rabbit secondary antibody, or an anti-Cy2 conjugated fluorescent mouse anti-rabbit secondary antibody (Jackson ImmunoResearch Laboratories). HeLa cells expressing wild-type Cx46 or Cx50-S50P cDNA alone were transfected with the subsequent cDNA using pIRES2+EGFP vector and stained with an anti-Cy3 goat anti-rabbit secondary antibody. For co-transfected cells Cx50-S50P and Cx46 constructs were inserted into the pCS2+ vector using the EcoRI restriction sites. Co-transfected cells were then treated with a polyclonal mouse anti-rabbit Cx50 antibody or a polyclonal goat anti-rabbit Cx46 primary antibody, followed by incubation with a Cy2 mouse anti-rabbit secondary antibody or Cy3 conjugated goat anti-rabbit antibody, to determine co-localization. Cells were viewed and photographed on an Olympus BX51 microscope using an Optronics MagnaFire digital camera. Gap-junctional plaque formation was quantified by immunofluorescent microscopy. Images were photographed at 60× and areas of cell-cell contact were examined for the presence of gap-junctional plaques and counted.
Staining of lens sections
A previously described method was used to prepare frozen lens sections for immunohistochemical analysis (Gong et al., 1997). A rabbit polyclonal antibody against the C-terminal region of Cx50 was generously provided by M. J. Wolosin at Mt Sinai School of Medicine. Sections were counterstained with Rhodamine-phalloidin (Molecular Probe) for detecting F-actin and DAPI (Vector Laboratories) for labeling cell nuclei. Fluorescent images were collected under a Zeiss Axiovert 200 fluorescence microscope with an Axiocam camera.
This work was supported by NIH grants EY13849 (X.G.) and EY13163 (T.W.W.).
Supplementary material available online at http://jcs.biologists.org/cgi/content/full/120/23/4107/DC1
- Accepted September 17, 2007.
- © The Company of Biologists Limited 2007