Characterization of a Mouse Cx50 Mutation Associated with the No2 Mouse Cataract

Abstract

purpose. Recently, a missense mutation in the mouse connexin 50 (Cx50) gene has been associated with the nuclear opacity 2 (No2) mouse cataract. This missense mutation (D47A) resulted in an aspartate-to-alanine substitution at amino acid position 47 in the first extracellular domain of Cx50. To better understand the role of Cx50 in the pathogenesis of congenital cataract, the functional consequences of the D47A mutation in the Xenopus oocyte expression system were studied.

methods. D47A was constructed using polymerase chain reaction (PCR) mutagenesis. Xenopus oocytes were injected with in vitro transcribed cRNA encoding wild-type mouse Cx50 (Cx50wt), wild-type rat Cx46 (Cx46wt), D47A, or combinations of wild-type and mutant connexins. The oocytes were then devitellinized and paired. Gap junctional conductance (G _j ) was measured using a dual two-microelectrode voltage-clamp technique.

results. Homotypic oocyte pairs expressing wild-type Cx50 or Cx46 were well coupled. In contrast, oocytes injected with D47A cRNA did not form gap junctional channels when paired homotypically. To test whether the D47A mutation could interact with wild-type connexins in a dominant negative manner, oocytes were injected with equal amounts of mutant and wild-type connexin cRNA, mimicking the heterozygous condition. Expression of D47A did not inhibit the development of junctional conductance in paired oocytes induced by wild-type Cx50 or Cx46.

conclusions. These results indicate that the D47A mutation acts as a loss-of-function mutation without strong dominant inhibition. In No2 mice, the mutation would be predicted to result in a reduction in intercellular communication, leading to cataractogenesis. It may also cause other qualitative changes such as a change in permeability for small molecules.

Gap junction channels are intercellular pathways between adjacent cells for the exchange of ions and metabolites smaller than 1 kDa. ¹ The gap junction channel is made of two hemichannels, each contributed separately by two adjoining cells. The hemichannels are composed of six subunits called connexins. The connexins belong to a multigene family composed of at least 14 members. ¹ Gap junctional channels are defined as homotypic when they contain a single type of connexin and heterotypic when each hemichannel of the pair is composed of a different connexin subtype. Heteromeric channels are those in which the hemichannels are composed of more than one connexin type. 

The lens is an avascular organ that is highly dependent on intercellular communication for volume regulation and metabolic homeostasis. ² Three connexins have been identified in the rodent lens: Cx43, Cx46, and Cx50. ^{3 4 5} Mouse connexin 50 (mCx50) is expressed only in the lens, where it forms gap junctional channels between fiber cells. ⁴ Connexin 46 is also found in lens fiber–fiber gap junctions, whereas connexin 43 is expressed in lens epithelial cells. ³  

Mutations in connexins have been linked to several genetic diseases including X-linked Charcot–Marie–Tooth disease (CMTX), a demyelinating peripheral neuropathy that is associated with mutations in Cx32 ^{6 7 8} ; hereditary nonsyndromic deafness, which is associated with mutations in Cx26 ⁹ ; and visceroatrial heterotaxia syndromes, which are associated with mutations in Cx43. ¹⁰ Recently, a missense mutation in the mouse connexin 50 gene (Gja8) has been associated with the nuclear opacity 2 (No2) mouse cataract, a congenital hereditary bilateral cataract that is inherited in a semidominant manner. ^{11 12} This missense mutation results in a substitution of aspartic acid-to-alanine at amino acid position 47 in the first putative extracellular domain of Cx50. To understand better the role of Cx50 in the pathogenesis of congenital cataract, we studied the functional consequences of the Cx50D47A mutation by testing its ability to induce gap junctional coupling between paired oocytes. 

Materials and Methods

Mutagenesis

cDNAs encoding mouse Cx50 and rat Cx46 in the pSP64TII vector ¹³ were provided by Thomas W. White and Daniel A. Goodenough (Harvard University, Boston, MA). To generate Cx50D47A, two primers corresponding to adjacent regions in the E1 domain of Cx50 were synthesized. The sense primer, 5′-CTGAGCAATCTGATTTTGTATGCAACACC-3′, corresponding to nucleotides from 406 to 434, contained nucleotides encoding amino acid 47 to 56 of Cx50, with nucleotide 406 changed from A to C, resulting in the conversion of aspartic acid 47 to alanine. The antisense primer, 5′-CGCCCCACACAAACTCCGCT-3′, corresponding to nucleotides from 405 to 386, corresponds to amino acids 40 to 47 of Cx50. Cx50 in the SP64T vector was amplified using a commercial kit (LA PCR; Takara Shuzo, Ostsu, Japan) according to the manufacturer’s protocol. The polymerase chain reaction (PCR) conditions were as follows: 1 minute at 94°C; 25 cycles at 98°C for 20 seconds and 68°C for 10 minutes; and 1 cycle at 72°C for 10 minutes. The PCR-amplified product was digested with the restriction enzyme, DpnI, to select against the parental, nonreplicated DNA. ¹⁴ Subsequently, the PCR-amplified product was purified with a PCR purification kit (QIAquick; Qiagen, Chatsworth, CA), polished with a PCR polishing kit (Stratagene, La Jolla, CA), and ligated to itself to generate Cx50D47A SP64T. The mutant construct was sequenced to ensure that PCR amplification did not introduce any new mutations (DNA Sequencing Facility, Iowa State University, Ames, IA). The recombinant plasmid DNA was linearized with the restriction enzyme, SalI. cRNAs were in vitro transcribed with SP6 polymerase (mMessage mMachine kit; Ambion, Austin, TX) following the manufacturer’s protocol. The transcripts were purified on a G-50 Sephadex column (Boehringer Mannheim, Indianapolis, IN) to remove unincorporated rNTPs, precipitated with isopropanol, and resuspended in diethyl pyrocarbonate–treated water. The cRNA was quantitated by measuring the absorbance at 260 nm and stored as 3-μl aliquots at– 80°C. 

Preparation of Xenopus Oocytes

Female Xenopus laevis was anesthetized, and a partial ovariectomy was performed. The frogs were maintained and treated in accordance with National Institutes of Health guidelines and with the ARVO Statement for Use of Animals in Ophthalmic and Vision Research. The oocytes were treated with 10 mg/ml collagenase type IA (Sigma, St. Louis, MO) for 20 minutes, manually defolliculated, and injected with an oligonucleotide antisense to endogenous Cx38, as previously described. ¹⁵ The oocytes were then injected with 3 to 4 ng cRNA for mouse Cx50, mouse Cx50D47A, or rat Cx46, either alone or in combination, and allowed to incubate for an additional 6 to 48 hours. Then the oocytes were devitellinized and paired as previously described. ¹⁶ Electrophysiological measurements were performed 6 to 18 hours after pairing. 

Western Blot Analysis of Connexin Proteins

Plasma membrane–enriched preparations of Xenopus oocytes were prepared as previously described. ^{3 17} The proteins were resolved on a sodium dodecyl sulfate–containing 9% polyacrylamide gel and transferred to nitrocellulose. The western blots were probed with the anti-Cx50 monoclonal antibody 6-4-B2-C6 (kindly provided by Viviana Berthoud and Eric Beyer, University of Chicago, IL). ¹⁸ The primary antibody was detected with alkaline phosphatase–conjugated goat anti-mouse Ig (Boehringer–Mannheim, Indianapolis, IN). 

Electrophysiological Measurements and Analysis

Dual two-microelectrode voltage-clamp recordings of gap junctional channels were performed (Axoclamp 2A and a Geneclamp 500 amplifier; Axon Instruments, Foster City, CA). The current and voltage electrodes were filled with 3 M KCl and had resistances of 0.1 to 0.5 MΩ. The tips of the electrodes were back filled with 1% agar in 3 M KCl to prevent KCl from leaking out of the electrodes and damaging the oocytes. Data acquisition and analysis were as performed (Pentium computer equipped with a TL-1 labmaster board and Pclamp6 software; Axon Instruments, Austin, Texas). Currents were filtered at 50 Hz using a four-pole Bessel filter. All experiments were performed at room temperature (22°C–24°C). For simple measurement of gap junctional conductance, both cells of the pair were initially voltage clamped to –40 mV and a 5- to 10-mV pulse was applied to one cell. Under these conditions, the change in current recorded in the second cell would be equal in magnitude and opposite in polarity to the current flowing through the gap junction and could be divided by change in transjunctional voltage to determine junctional conductance, G _j . To evaluate the transjunctional voltage dependence of the gap junctions, transjunctional voltage-clamp steps were applied between ±70 mV in 10-mV increments from a holding potential of –40 mV. The initial and steady state junctional currents were measured at 40 msec and 24 seconds, respectively, after application of the voltage-clamp step. The normalized steady state junctional conductance (G _j∞) versus transjunctional voltage (V _j ) relation was determined by normalizing the steady state conductance values to the values at ±10 mV. The G _j∞–V _j relation was fit to a Boltzmann equation: G _j∞= G _jmin + (G _jmax− G _jmin)/{1+exp[A*(V _j − V₀)]}, where G _j∞ is the steady state conductance, G _jmin is the minimum conductance, G _jmax is the maximum conductance, A is the cooperativity constant, and V ₀ is the voltage at which the decrease in G _j∞ is half maximal. Oocyte pairs with resting membrane potentials more negative than –15 mV were selected for analysis. 

Results

Expression of Cx50D47A in Oocyte Pairs

We compared the functional properties of the D47A mutation with those of the wild-type protein by testing their ability to form gap junctional channels in the paired oocyte system. The results of these experiments are summarized in Table 1 . Oocytes injected with cRNA for D47A did not induce the formation of gap junctional channels when paired homotypically. In contrast, homotypic oocyte pairs expressing wild-type Cx50 were well coupled. Figure 1 A shows typical junctional currents from a homotypic pair expressing wild-type Cx50. The Cx50 gap junctional current rapidly inactivated to a new steady state level on application of transjunctional voltage-clamp steps to potentials greater than ±10 mV. The time course of inactivation became progressively faster at larger transjunctional potentials. Figure 1B shows a plot of the normalized steady state junctional conductance (G _j∞) versus transjunctional voltage (V _j ) relation. G _j∞ declined symmetrically at positive and negative V _j values. The mean G _j∞–V _j curve (n = 4) could be described by a Boltzmann function with G _jmax = 1.04; G _jmin = 0.16, A = 0.25 and V ₀ = 22.76 mV (Table 2) . These values are similar to those previously reported by White et al. ¹⁷  

To test whether the D47A mutation could interact with wild-type Cx50 in a dominant negative manner, we coinjected oocytes with equal amounts of mutant and wild-type Cx50 cRNA, mimicking the heterozygous situation. Expression of D47A did not inhibit the development of junctional conductance in paired oocytes induced by wild-type Cx50 (Table 1) . Moreover, the time course of inactivation and the G _j∞–V _j curve were not altered by coexpression of D47A with wild-type Cx50 (Table 2) . These results indicate that the D47A mutation acted as a loss-of-function mutation without having a dominant negative effect. 

To investigate further the mechanisms underlying the behavior of the D47A mutation, immunoblot analysis of membrane-enriched preparations of oocytes was performed. Oocytes injected with wild-type or mutant Cx50 cRNA synthesized a protein of approximately 70 kDa that was recognized by the anti-Cx50 monoclonal antibody 6-4-B2-C6 (Fig. 2) . The amount of wild-type and mutant Cx50 protein was similar. No major proteins were detected in antisense-injected control oocytes. These results indicate that the loss of the function without dominant inhibition exhibited by the D47A mutant was not caused by the failure of the mutant protein to reach the plasma membrane. 

Effect of Coexpression of Cx50D47A with Wild-type Cx46

Previous biochemical studies have shown that Cx50 forms heteromeric gap junctional channels with Cx46 in lens fiber cells. ^{19 20} Thus, we were interested in determining whether coexpression of the D47A mutant with wild-type rat Cx46 would inhibit gap junctional coupling. Oocyte pairs injected with wild-type rat Cx46, either alone or in combination with wild-type mouse Cx50 cRNA, efficiently made gap junctional channels (Table 3) . Oocytes coinjected with Cx46 and D47A mutant cRNA were also well coupled, showing that the D47A mutant did not significantly inhibit the ability of wild-type Cx46 to form gap junctional channels. Figure 3 shows representative junctional current traces and plots of normalized G _j∞ versus V _j for homotypic Cx46 and heteromeric (Cx46+Cx50) and (Cx46+Cx50D47A) pairs. The junctional currents recorded for Cx46 pairs displayed voltage sensitivity at larger transjunctional voltages. The G _j∞–V _j relation decreased symmetrically for V _j s of opposite polarities with a V ₀ of 52.86 mV (Table 4) . Similar findings for Cx46 gap junctional channels have been reported by White et al. ¹⁷ The junctional currents recorded from oocyte pairs expressing (Cx46+Cx50D47A) also decayed in a time- and voltage-dependent manner at V _j s of ±30 mV or more. The mean G _j∞-–V _j relation could be described by a Boltzmann function with a V ₀ of 38.09 mV. In contrast to both Cx46 and (Cx46+Cx50D47A) pairs, (Cx46+Cx50) pairs displayed a much greater sensitivity to voltage. The Boltzmann parameters for the mean G _j∞–V _j relation were similar to the values for homotypic pairs expressing wild-type Cx50 with a V ₀ of 25.91 mV. 

Discussion

This study shows that mouse Cx50D47A acts as a loss-of-function mutation without dominant inhibition. The observation that the D47A mutation does not have a strong dominant negative effect on wild-type Cx50 or Cx46 suggests that it is unable to coassemble with the wild-type connexins. Alternatively, it is possible that the D47A mutation can coassemble with wild-type connexins and that the function of these heteromeric channels depends on the number and placement of the mutant subunits. Further studies are needed to distinguish between these two possibilities. 

The finding that the D47A mutation leads to loss of function is not surprising. The E1 loop of the connexin protein is a highly conserved and functionally important domain in gap junctional coupling and gating. The invariance of the aspartic acid at amino acid position 47 is suggestive of its importance. Mutations at this position would potentially alter the structure of the E1 loop and perturb its ability to dock with an opposing connexon. 

The No2 mouse mutation has been described as semidominant because heterozygous mice have a milder form of cataract than do homozygous mice. These observations are consistent with the notion that D47A acts as a loss-of-function mutation without dominant inhibition. Consequently, cataract formation occurs when the amount of wild-type Cx50 is reduced below a critical level, and the severity of the cataract depends on the amount of reduction. In addition, the No2 mice exhibit a reduction in total ocular mass of approximately 30% compared with wild-type, suggesting that the Cx50 gap junctional channels are also involved in the regulation of growth. ¹¹ A similar reduction in ocular size and diffuse nuclear opacities has been observed in homozygous Cx50 knockout mice. ²¹ However, no phenotype was observed in heterozygotes suggesting that the effect of the D47A mutation cannot be completely reproduced by knocking out one allele. 

Recently, a human congenital zonular pulverulent cataract has been linked to a missense mutation in human Cx50 converting proline 88 to serine. ²² Unlike the D47A mutation, expression of the P88S mutant with wild-type Cx50 in Xenopus oocyte pairs results in a profound inhibition of intercellular coupling, indicating that it acts as a loss-of-function mutant with dominant inhibition. ²³ It would be interesting to determine whether expression of the P88S mutation in mice by homologous recombination would result in a more severe form of cataract than does the D47A mutation. 

 Reprint requests: Lisa Ebihara, Department of Physiology and Biophysics, FUHS/The Chicago Medical School, 3333 Green Bay Road, North Chicago, IL 60064.

 Supported by Grant EY10589 from the National Institutes of Health.

 Submitted for publication December 15, 1998; accepted February 9, 1999.

 Proprietary interest category: N.

 

Table 1.

 

Table 1.

 

Conductance of Xenopus Oocyte Pairs Injected with Mouse Cx50 Wild-type and/or Cx50D47A Mutant cRNA

Cell Injected			Mean Conductance ± SEM*	Number of Pairs	Expression Time, †
Cell 1	Cell 2
Control	Control	0.033 ± 0.012	20	2 days
Control	Control	0.032 ± 0.008	12	o/n
Cx50wt	Cx50wt	29.1 ± 5.77	14	2 days
Cx50wt	Cx50wt	9.43 ± 5.10	4	o/n
D47A	D47A	0.017 ± 0.003	5	2 days
D47A	D47A	0.047 ± 0.016	11	o/n
Cx50wt+D47A	Cx50wt+D47A	25.2 ± 12.8	5	2 days
Cx50wt+D47A	Cx50wt+D47A	6.12 ± 2.68	3	o/n

 The total amount of cRNA injected into each oocyte was held constant. Therefore, oocytes coinjected with mutant and wild-type Cx50 cRNA at a 1:1 ratio would be expected to have junctional conductances that were reduced by 50% compared with conductance in oocytes injected with wild-type C×50 alone, if the D47A mutation does not have a dominant negative effect.

*  Data are expressed in microsiemens.

†  Oocytes were injected with wild-type or mutant C×50 cRNA and allowed to incubate at 18°C, either overnight or for 2 days before pairing.

Figure 1.

 

Voltage dependence of Cx50wt homotypic pairs, gap junctional current traces (A) and plot of mean normalized steady state gap junctional conductance versus transjunctional voltage (B). In (A), both cells of the pair were held at a constant holding potential and 24-second voltage-clamp steps were applied between ±70 mV in 10-mV increments. In (B), the mean normalized steady state gap junctional conductance was plotted as a function of transjunctional voltage (n = 4). The results are expressed as mean ± SEM. The solid line is the best fit of the experimental data to a Boltzmann equation with G _jmax = 1.04, G _jmin = 0.16, A = 0.25, and V ₀ = 22.76 mV (Table 2) .

Table 2.

 

Table 2.

 

Boltzmann Parameters of Normalized Mean Steady State Conductance of Gap Junctional Channels Expressed in Xenopus Oocyte Pairs

Cell Injected			Gjmax	Gjmin	A	V0 , *	Number of Pairs
Cell 1	Cell 2
Cx50wt	Cx50wt	1.04	0.16	0.25	22.76	4
Cx50wt+D47A	Cx50wt+D47A	1.10	0.18	0.17	20.23	6

 Data were fit to a Boltzmann equation: G _j∞ = G _jmin + (G _jmax − G _jmin)/{1+exp[A

*  (V _j −V ₀)]}.

*  Data are in millivolts.

Figure 2.

 

Immunoblot analysis of oocytes injected with cRNAs for mouse Cx50wt (lane 1), Cx50D47A (lane 2), and (Cx50wt+Cx50D47A) (lane 3). The plasma membrane–enriched proteins were separated on a 9% polyacrylamide gel, and then transferred to nitrocellulose. The blots were probed with the anti-Cx50 monoclonal antibody 6-4-B2-C6. All three groups of oocytes produced a 70-kDa protein band. No band was detected in antisense-treated control oocytes (data not shown).

Table 3.

 

Table 3.

 

Conductance of Xenopus Oocyte Pairs Coinjected with Rat Cx46 Wild-type and Mouse Cx50 Wild-type or D47A cRNA

Cell Injected			Mean Conductance ± SEM*	Number of Pairs	Expression Time
Cell 1	Cell 2
Control	Control	0.046 ± 0.016	5	2 days
Control	Control	0.026 ± 0.010	10	o/n
Cx46wt+Cx50wt	Cx46wt+Cx50wt	16.8 ± 4.73	4	2 days
Cx46wt+Cx50wt	Cx46wt+Cx50wt	14.5 ± 7.16	9	o/n
Cx46wt+D47A	Cx46wt+D47A	3.70 ± 3.64	3	2 days
Cx46wt+D47A	Cx46wt+D47A	4.06 ± 1.37	6	o/n
Cx46wt	Cx46wt	9.03 ± 3.22	5	o/n

*  Data are in microsiemens.

Figure 3.

 

Typical junctional current traces (left) and mean G _j∞ –V _j curves (right) for oocyte pairs injected with (A) Cx46wt, n = 3; (B) (Cx46wt+Cx50D47A), n = 4; or (C) (Cx46wt+Cx50wt), n = 1 cRNA. The experimental protocol was the same as that described in Figure 1 . The solid lines are the best fit of experimental data to a Boltzmann equation whose parameters are shown in Table 4 .

Table 4.

 

Table 4.

 

Boltzmann Parameters of Normalized Mean Steady State Conductance of Gap Junction Channels Expressed in Xenopus Oocyte Pairs

Cell Injected			Gjmax	Gjmin	A	V0 , *	Number of Pairs
Cell 1	Cell 2
Cx46wt	Cx46wt	1.02	0.25	0.09	52.86	3
Cx46wt+Cx50wt	Cx46wt+Cx50wt	1.02	0.21	0.24	25.91	1
Cx46wt+D47A	Cx46wt+D47A	1.08	0.32	0.09	38.09	4

 Data were fit to a Boltzmann equation G _j∞ = G _jmin + (G _jmax − G _jmin)/{1+exp[A

*  (V _j −V ₀)]}.

*  Data are expressed in millivolts.

The authors thank Jay Pal for reviewing the manuscript and Xiaoqin Liu for technical assistance.