Figure 1.
Identification of the Gja8R205G mutation in Nm2249 mice.
(A) Photos of eyes and lenses of Nm2249 heterozygous (Nm/+) and Nm2249 homozygous (Nm/Nm) mutant mice in the CWXS/Agl strain background at the age of 3 weeks. Scale bars, 1 mm. (B) Genome-wide screening data of the phenotypes and genotypes of 105 backcrossed mice, generated between mutant Nm2249 and wild-type CAST/Ei mice. The Nm2249 mutation was mapped to a region near D3Mit11 and D3Mit122 on chromosome 3. (C) The Gja8 gene is near the region where the Nm2249 mutation was mapped. (D) DNA sequencing data confirmed a missense mutation (C to G) in the Gja8 gene of Nm2249 mutant (Nm-Gja8), which resulted in the arginine at codon 205 of wild-type Gja8 (WT-Gja8) being replaced by a glycine (R205G). (E) Homozygous Gja8Nm2249/Nm2249 mutant mice were bred with the Gja8−/− knockout mice for a Gja8 allelic test. Phenotypic comparison of eyes and lenses of Gja8Nm2249/Nm2249, Gja8Nm2249/− and Gja8−/− mice at the age of 3 weeks was shown. The similarity of small lens size in these mutant mice further indicated that Gja8R205G is the causative mutation in Nm2249 mice. Scale bars, 1 mm.
Figure 2.
Endogenous Cx46 influences cataract formation caused by the Gja8R205G mutation.
(A) Lens photos of wild-type (Gja8+/+ Gja3+/+), Gja8R205G heterozygous (Gja8R205G/+Gja3+/+), Gja8R205G homozygous (Gja8R205G/R205G Gja3+/+), and Gja8R205G Gja3 double mutant (Gja8R205G/R205G Gja3−/−) mice at the age of 3 weeks. These mutant lines were bred into the C57BL/6J strain background. While the Gja8R205G homozygous (Gja8R205G/R205G Gja3+/+) lens revealed vacuole-like defects in the lens periphery (indicated by a white arrow), the Gja8R205G/R205G Gja3−/− double mutant lens showed a clear lens periphery (indicated by a white arrow). Scale bar, 1 mm. (B) Histological sections of Gja8R205G homozygous mutant (Gja8R205G/R205G Gja3+/+) and Gja8R205G Gja3 double mutant (Gja8R205G/R205G Gja3−/−) lenses from 3-week-old littermates. The Gja8R205G/R205G Gja3+/+ lens section showed disorganized peripheral fibers with vacuoles or enlarged extracellular spaces (labeled with asterisk) near the lens bow region (indicated by an arrowhead), while the double mutant lens section shows organized and elongated periphery fiber cells at the bow region (indicated by an arrowhead). Scale bar, 50 µm.
Figure 3.
Biochemical and immunological characterization of Cx50 and Cx46 proteins.
(A) Total lens homogenates, prepared from Gja8R205G/R205G (a), Gja8R205G/− (b) and Gja8−/− (c) littermates at one week of age, were examined by a Coomassie-blue stained gel (left panel) and by western blot using an anti-Cx50 antibody and an anti-Cx46 antibody (right panels). Arrowheads indicate phosphorylated proteins while the arrow indicates non-phosphorylated proteins. Compared to Gja8−/− lenses, mutant Gja8R205G/R205G lenses showed decreased phosphorylated Cx46 proteins. (B) Immunostaining results of lens frozen sections showed typical punctate signals of Cx50 (red) and Cx46 (green) connexin proteins in wild-type (Gja8+/+Gja3+/+) sample, but only diffuse tiny spots of Cx50 and Cx46 could be seen in Gja8R205G homozygous mutant (Gja8R205G/R205G Gja3+/+) sample. Diffuse tiny spots of Cx50 were observed in Gja8R205G Gja3 double mutant (Gja8R205G/R205G Gja3−/−) sample. These lens samples were prepared from 1-week-old mice. Scale bar, 20 µm.
Figure 4.
Confocal imaging of GFP-positive (GFP+) live lenses from wild-type (WT), Gja8 knockout (Gja8−/−), Gja8R205G homozygous mutant (Gja8R205G/R205G) and Gja8R205G Gja3 double mutant (Gja8R205G/R205G Gja3−/−) mice at the age of 2
weeks. (A) Confocal images showed the typical mosaic pattern of GFP in lens central epithelial cells in all lenses examined. Aberrant epithelial cells appeared only in the Gja8R205G/R205G lens. (B) Images displayed the anterior Y-suture (delineated by the white lines in the WT lens) where the ends of underlying fiber cells come into contact. These elongated fiber cells in the WT lens had normal uniform GFP signal. There was a partial delay in Y-suture closure in the Gja8−/− lens. In the Gja8R205G/R205G mutant lens, aberrant GFP+ cells were observed in the open suture region where the opposing fiber cells failed to elongate fully to close the anterior Y-suture. The Gja8R205G/R205G Gja3−/− double mutant lens also had an open suture (black area) due to insufficient fiber cell elongation. (C) An anterior (top) to posterior (bottom) sectional view from three-dimensional z-stack reconstructions of GFP+ lenses showing the lens equator (indicated by an arrow) and peripheral fiber cells. Only Gja8R205G/R205G homozygous lenses displayed disorganized fiber cells. Scale bars, 50 μm. (D) A cross sectional view from three-dimensional reconstructions of z-stacks of GFP+ lenses at the lens equator. Lens epithelial cells are on the left. In the WT lens, peripheral fiber cells were precisely organized and displayed a mosaic GFP expression pattern while inner fiber cells showed uniform GFP signal. In the Gja8−/− lens, peripheral fiber cells remained organized. In the Gja8R205G/R205G homozygous mutant lenses, peripheral fiber cells were completely disorganized with vacuole-like structures or enlarged extracellular spaces (indicated by an asterisk). However, in Gja8R205G/R205G Gja3−/− double mutant lenses, peripheral fiber cells were mainly organized.
Figure 5.
Electrophysiological assays of wild-type Cx50 and Cx46 connexins and mutant Cx50-R205G subunits in paired Xenopus oocytes.
(A and B) Western blot analyses of oocytes showed equivalent levels of wild-type and mutant Cx50 when expressed alone or in co-injected cells (A). Western blot also demonstrated similar levels of wild-type Cx46 in both conditions assayed (B). (C and D) Band densitometry quantitatively confirmed that mean protein expression was not significantly changed (P>0.05). (E) Junctional conductance measurements recorded from Xenopus oocyte pairs injected with wild-type Gja8, wild-type Gja3 or mutant Gja8R205G transcripts alone or in combination. Cell pairs expressing wild-type Cx50 or Cx46 subunits alone formed functional gap junctions with mean conductance values of approximately 30µS and 17µS, respectively. Oocytes co-injected with both wild-type Gja8 and mutant Gja8R205G transcripts failed to form functional gap junction channels and exhibited a level of coupling not significantly higher than that of the water-injected control (P>0.05). Conversely, the co-expression of Cx46 and Cx50-R205G subunits did not significantly alter junctional conductance (P>0.05) as these channels displayed a mean Gj of 16 µS. Heterotypic channels failed to form functional channels with levels of conductance significantly higher than that of the water-injected controls (P>0.05). Cx50-R205G subunits alone failed to produce functional intercellular channels.
Table 1.
Boltzmann parameters for homomeric Cx46 and heteromeric Cx46/Cx50-R205G channels.
Figure 6.
Voltage-gating properties of Cx46 and mixed Cx46/Cx50-R205G channels.
The decay in junctional current (Ij) induced by transjunctional voltage (Vj) was plotted as a function of time for gap junctions comprised of Cx46 (A) and mixed Cx46/Cx50-R205G (B). Voltage was stepped to ±120 mV in 20 mV increments. At all voltage applications >±20 mV, mixed channels showed a more rapid current decay toward steady state. Analysis of channel closure kinetics was based on representative traces displaying the initial 250 ms of current decay recorded after application of a +120 mV transjunctional voltage, for gap junctions comprised of homomeric wild-type Cx46 (n = 5) (C) and heteromeric channels containing both Cx46 and Cx50-R205G (n = 5) (D). Current traces were fit to monoexponential decay to determine the mean time constant, τ. Heteromeric Cx46 and Cx50-R205G channels closed ∼25% faster than homomeric Cx46 channels, displaying a significant increase in mean channel closure time (p<0.05). (E) Comparison of equilibrium conductance. Steady state conductance was measured when current decay reached equilibrium, normalized to the values at ±20 mV and plotted as a function of Vj. The steady state reduction in conductance for heteromeric Cx46 and Cx50-R205G channels (n = 8) was greater than the reduction for homomeric wild-type Cx46 channels (n = 5).