mopdf.com

J.-J. Tong et al.: The Connexin46 Mutant14718 C in L-15 (GIBCO-Invitrogen, Carlsbad, CA) containing 2 mM CaCl2 prior to performing electrophysiological experiments.protein, Zaza green, or human Cx46-EGFP or humanT19M-EGFP using Happyfect (Mayflower Bioscience, St.Louis, MO). Cells were tested for dye uptake 1 day later bya 20-min exposure to sodium Ringer’s solution (150 mMsodium gluconate or NaCl, 4.7 mM KCl, 5 mM glucose,and 5 mM HEPES, pH 7.4) with or without divalent cations and to which 10 lM 40 ,6-diamino-2-phenylindoledihydrochloride (DAPI) was added. Then, cells werewashed in the sodium Ringer’s solution containing1 mM Ca2? and 1 mM Mg2? without dyes and examinedby epifluorescence using a Nikon Eclipse inverted microscope equipped with a CCD camera (Photometrics, Tucson,AZ) associated with image analysis software (NIS Elements AR 3.0, Nikon). Uptake of DAPI was quantified byplacing a region of interest (ROI) over the nuclei of cellsexpressing the reporter protein and measuring the meanintensity of DAPI after correction for background fluorescence measured in a cell-free region. Two to five 209fields of view were analyzed for each treatment to generatehistograms of the intensity of DAPI in reporter proteinexpressing cells. Dead cells identified by morphologicalcriteria were excluded from the data analysis.Time-lapse experiments using rat Cx46 or T19M wereperformed using a closed bath insert (RC-37FC; WarnerInstruments) as previously described (Ebihara et al. 2011).Four lM DAPI was added to all of the perfusion solutions.The rate of DAPI uptake under control conditions, following removal of divalent cations, and after addition ofLa3? was determined by fitting the time course of DAPIuptake to a linear regression.For measurement of gap junctional coupling, connexincRNA-injected oocytes were devitellinized and paired aspreviously described (Ebihara 1992). Double two-microelectrode voltage clamp experiments were performed usingGeneClamp 500 (Molecular Devices, Sunnyvale, CA) anda TEV-200A (Dagan Corporation, MN) as previouslydescribed (Xu and Ebihara 1999). Pulse generation anddata acquisition were performed using a PC computerequipped with PCLAMP9 software and a Digidata 1322Adata acquisition system (Molecular Devices). All experiments were performed at room temperature (20–22 C).Hemi-channel currents were recorded from singletransfected HeLa cells using the whole cell variant of thepatch clamp technique. The resistance of the patch pipetteswas 2–4 MX when filled with standard internal solution.After rupturing the membrane patch, the series resistancewas usually \10 MX and was therefore not compensated.The internal solution contained: 140 mM CsCl, 10 mMEGTA, 2 mM MgATP, 3 mM Na2ATP, 10 mM HEPES,pH 7.4. The standard external bath solution contained150 mM Na gluconate, 4.7 mM KCl, 1 mM MgCl2, 1 mMCaCl2, 5 mM glucose, 5 mM HEPES, pH 7.4. The membrane potentials were not corrected for liquid-junctionpotentials.Expression of Connexins in Xenopus OocytesResultsConnexin cRNAs were synthesized using the mMessagemMachine in vitro transcription kit (Ambion, Austin, TX)according to the manufacturer’s instructions. The amountof cRNA was quantitated by measuring the absorbance at260 nm.Adult female Xenopus laevis frogs were anesthetizedwith tricaine and a partial ovariectomy was performed inaccordance with protocols approved by the Animal Careand Use Committee at Rosalind Franklin University inNorth Chicago, IL. The oocytes were manually defolliculated after treating them with collagenase IA (Worthington Biochemical Corporation, Lakewood, NJ).Stage V and VI oocytes were selected and pressureinjected using a Nanoject variable microinjection apparatus (model No. 3-000-203, Drummond Scientific,Broomal, PA) with 36.8 nl of 0.5–600 ng/ll of connexincRNA and 5 ng/36.8 nl of oligonucleotides antisense tomRNA for Xenopus Cx38 as previously described(Ebihara 1996). The oocytes were incubated overnight atWe initially tested whether T19M was able to induce gapjunctional (intercellular) conductances in paired Xenopusoocytes using the double two-electrode voltage clamptechnique. Because rodent Cx46 induces higher levels ofgap junctional currents than human Cx46 in this system,experiments were initially performed after injectingcRNAs encoding either wild-type or mutant rat Cx46.Homotypic oocyte pairs expressing wild-type rat Cx46were well coupled, but oocyte pairs injected with rat T19McRNA showed no coupling above control levels (Fig. 1a).Furthermore, oocytes expressing rat T19M failed to inducesignificant coupling when paired heterotypically withoocytes expressing wild-type Cx46 (Fig. 1a).To determine if co-expression of T19M affected thejunctional conductance produced by wild-type lens connexins, we tested pairs of oocytes co-injected with equalamounts of cRNAs encoding rat T19M and wild-type ratCx46 or mouse Cx50. The conductances induced in oocytepairs co-expressing T19M with wild-type Cx46 did notElectrophysiological Measurements123

148Fig. 1 T19M does not induce gap junctional coupling whenexpressed by itself, and it acts as a loss-of-function mutation withoutdominant-negative inhibition when co-expressed with wild-type lensconnexins. Bar graphs show mean gap junctional conductances inpairs of oocytes expressing different combinations of wild-type andmutant lens connexins as determined using the double two-electrodevoltage clamp technique. a Rat Cx46 or T19M were expressed aloneor in combination with each other. b Mouse Cx50 was expressedalone or in combination with either rat Cx46 or T19M. AS indicatesoocytes that were injected with no cRNA (i.e., Xenopus Cx38antisense oligonucleotide alone). The number of pairs tested isindicated within parentheses. *p \ 0.001 (Student’s t test comparedwith Cx46-injected oocyte pairs); **p \ 0.001 (Student’s t testcompared with Cx50-injected oocyte pairs)differ significantly from those expressing only wild-typeCx46 (Fig. 1a). Similarly, the conductances induced inoocyte pairs co-expressing T19M with Cx50 did not differsignificantly from those expressing Cx50 alone or those coexpressing wild-type Cx50 and wild-type Cx46 (Fig. 1b).These results suggest that T19M acts as a loss-of-functionmutation without dominant-negative inhibition of intercellular communication.To determine the ability of T19M to form gap junctionplaques, we performed immunofluorescence staining ontransiently transfected HeLa cells. Cells transfected withwild-type rat Cx46 showed strong staining at gap junctionplaques and some staining in the perinuclear region, likelycorresponding to the Golgi apparatus (Fig. 2a). Surprisingly, in multiple independent experiments we detectedvery few cells containing immunoreactive Cx46 aftertransfection with rat T19M at our usual time of observation(48 h). Eighteen hours following transfection, mostCx46T19M-expressing cells showed a normal morphologywith strong perinuclear staining and no gap junctional123J.-J. Tong et al.: The Connexin46 Mutantplaques (Fig. 2b). By 24 h after transfection, cells hadstarted to round up and the abundance of cells expressingthe Cx46 mutant was decreased (Fig. 2c). After 48–72 h,cells showed low intensity of Cx46T19M immunoreactivity, lacked detectable gap junctions, and were rounded up(Fig. 2d). The striking decline in numbers of T19Mexpressing cells and their rounded morphology suggestedthat the mutated protein had toxic effects.To assess the potential toxicity of T19M, we quantifiedthe proportion of rounded up Zaza-green positive cells(expressed with the connexin using the bidirectional promoter vector, PBI-CMV3) after transfection. For theseexperiments, we used a different transfection reagent(Happyfect), because it has been reported to be less toxic tocultured cells. Seventy-two hours after transfection, 89 %(149/167) of the T19M transfected cells were rounded upas compared to only 14 % (21/126) of the wild-type Cx46and 16 % (51/318) of vector-transfected cells. The morphology of the cells expressing T19M was strikingly different from that of the cells expressing wild-type Cx46 asillustrated in Fig. S1. In addition to rounding up, many ofthe T19M expressing cells showed extensive membraneblebbing which is often seen in dying cells.To test the hypothesis that the apparent cell toxicity inT19M-expressing cells was a consequence of aberrantT19M hemi-channel activity, we examined uptake of theconnexon-permeant dye, DAPI, in transfected HeLa cells.Cells that expressed rat Cx46 or T19M were identified bythe fluorescence of the reporter protein, Zaza green(Fig. 3). When incubated in the absence of divalent cationsor in the presence of 1 mM external Ca2? concentration([Ca2?]o and 1 mM [Mg2?]o), cells expressing T19Mshowed significantly increased DAPI uptake compared tocells expressing wild-type Cx46 or control (Zaza-greennegative) cells. The T19M-induced dye uptake was mostlyblocked by increasing the [Ca?2]o to 5 mM.In some experiments, we used propidium iodide (PI) andDAPI as the connexon-permeant tracers (data not shown).These experiments showed a linear relationship betweenthe rates of DAPI and PI uptake. However, the rate of PIuptake was much smaller than the rate of DAPI uptake.To examine the effect of wild-type and mutant Cx46 onhemi-channel activity in greater detail, we measured theuptake of DAPI by time-lapse recordings in HeLa cellsexpressing wild-type rat Cx46, T19M, or vector alone(Fig. 4). In the presence of control external solution (containing 1 mM Ca2?and 1 mM Mg2?), cells expressingT19M showed significantly higher rates of DAPI uptakethan those expressing wild-type Cx46 or vector alone.After changing the bathing medium to a solution free ofdivalent cations, the rate of DAPI uptake increased byapproximately twofold in the T19M cells and by up to19-fold in the Cx46 cells. Little or no increase in dye

J.-J. Tong et al.: The Connexin46 Mutant149Fig. 2 T19M is inefficient at forming gap junction plaques. Photomicrographs show the distribution of wild-type rat Cx46 (a) and T19M (b–d) at the indicated times following transient transfection of HeLa cells. Bar 30 lMFig. 3 T19M causes increaseduptake of connexon-permeantdyes. Photomicrographs showexamples of HeLa cells thatwere transfected with wild-typerat Cx46 (a–c) or T19M (d–i)(using the vector PBI-CMV3which also drives expression ofZaza green) and incubated a daylater with DAPI in Na gluconateRinger’s solution containing0 mM Ca2? (a–f) or 5 mMCa?2 (g–i) for 20 min. Phasecontrast images (a, d, g). Zazagreen fluorescence (b, e, h).DAPI fluorescence (c, f, i).After a 20-min incubation incontrol solution containing0 mM Ca2?, cells expressingT19M showed DAPI uptake (e,f) that was mostly inhibited by5 mM Ca2? (h, i). Bar graphsummarizes the quantificationof the DAPI uptake data (j).Data are graphed asmean SEM. The number ofcells tested is indicated withinparentheses. *p \ 0.001(Mann–Whitney rank sum testcompared with wild-type ratCx46-transfected cells)123

150J.-J. Tong et al.: The Connexin46 MutantFig. 4 The rate of DAPI uptake is increased by lowering divalentcations and inhibited by La3?. Average time course of DAPI uptakeby transfected HeLa cells in control solution (1 mM Ca2?,1 mM Mg2?), in external solutions with no added divalent cationsand in control solution plus 200 lM La3?. a Wild-type rat Cx46(closed circles); T19M (open triangles). To measure changes in therate of dye uptake over time, the mean DAPI fluorescence intensityper pixel from ROI’s located in the nuclei of Zaza-green positive cellswere normalized to mean DAPI fluorescence intensity of the ROI’s at60 min, averaged and plotted as a function of time. The cells wereinitially bathed in control solution (containing 1 mM Ca2?,1 mM Mg2?). Then, the cells were exposed to a solution containingno added divalent cations followed by reperfusion with controlsolution containing 200 lM La3?. All the solutions contained 4 lMDAPI. b Bar graph shows the rates of DAPI uptake in cellsexpressing wild-type Cx46, T19M, or vector alone in the presence of1 mM Ca2?, 1 mM Mg2? (gray bar); or 1 mM Ca2?, 1 mM Mg2?,0.2 mM La3? (black bar). Data are presented as the mean SEM.*p \ 0.002 (Mann–Whitney rank sum test compared with T19Mtransfected cells); ?p \ 0.001 (Mann–Whitney rank sum test compared with vector-transfected cells). The number of cells analyzed isindicated within parenthesesuptake was observed in control cells expressing vectoralone (data not shown). Dye uptake was completelyblocked by application of 200 lM lanthanum, a nonspecific connexin hemi-channel blocker. Dye uptake was also*90 % blocked in Cx46 cells and 77 % blocked in T19Mcells by application of 500 lM carbenoxolone. Theseresults suggest that the T19M channels are less sensitive toblockade by divalent cations than wild-type Cx46 hemichannels. Interestingly, the effect of divalent cation-freeconditions on the wild-type Cx46 cells was slow, requiringgreater than 15 min to reach a new steady-state. In contrast,the effect of divalent cation-free conditions on the T19Mcells was relatively fast, reaching steady-state in \2 min.To determine the effects of T19M on plasma membraneconductance of HeLa cells, whole cell patch clampexperiments were performed with cesium in the pipette andsodium in the bath. Figure 5 compares representativefamilies of current traces (and averaged steady-state I–Vrelationships) recorded from HeLa cells expressing Cx46,T19M, or Zaza green alone in the presence of 1 mM[Ca?2]o and 1 mM [Mg?2]o. Both Cx46- and T19Mexpressing cells exhibited currents that were mostly closedat -60 mV and activated in response to depolarizingvoltage clamp steps in a time- and voltage-dependentmanner. These currents were not observed in either nontransfected cells or Zaza-green-transfected cells, indicatingthat they could be attributed to Cx46 and T19M hemichannels. Both wild-type and mutant currents could bereadily observed in cells transfected with similar amountsof DNA even when the bath solution contained 1 mM Ca2?(and 1 mM Mg2?). However, the T19M mutation alteredthe kinetics of channel gating. T19M currents activatedmore rapidly and had a more pronounced and prolongedinactivation phase than the wild-type Cx46 currents at largepositive potentials as illustrated in Fig. 6a and b. In addition, the time course of deactivation of the T19M currentsat negative potentials was slower than that of wild-typeCx46. This effect was quantified by measuring the timerequired for the tail current to decay to 50 % of its peakvalue (t1/2). Over the transmembrane voltage (Vm) rangebetween -80 and -40 mV, the average t1/2 values weresignificantly longer for T19M than for wild-type Cx46(Fig. 6d). This effect did not appear to correlate withchanges in current density since the amplitudes of theT19M tail currents used in the data analysis were smallerthan those of the WT tail currents (Fig. 6c).To determine if the small, persistent inward currentobserved at negative potentials in T19M-expressing cellswas due to hemi-channels, we used the nonspecific hemichannel blocker, La?3 (John et al. 1999; Contreras et al.2002). Application of 200 lM La3? blocked most of thetime- and voltage-dependent component of the current(Fig. 7). It also reduced the inward current at the holdingpotential and decreased the cell input conductance to values comparable to those observed in vector-transfectedcontrol cells. Similar results were obtained for wild-typeCx46. The Cx46 hemi-channel currents could also bepartially blocked by 200 lM carbenoxolone. These findings suggest that a small number of Cx46 hemi-channelsremain open even in the presence of divalent cations andthat these channels can account for the dye uptake observedin the presence of divalent cations.123

J.-J. Tong et al.: The Connexin46 Mutanta151bCx46cT19MVector100 pA100 pA100 pA.2 s.2 s.2 sd 500Current (pA)400300T19MrCx46Vector2001000-100-60 -40 -20 0 20 40 60 80Voltage (mV)Fig. 5 Representative families of current traces recorded from singleHeLa cells transfected with vector alone (a), wild-type rat Cx46 (b),or T19M (c). Families of current traces were recorded in response to aseries of voltage clamp steps between -60 and 50 mV in incrementsof 10 mV from a holding potential of -60 mV. Dashed line indicateszero current level. d Average steady-state I–V relationships for vectoralone (open squares, n 5), wild-type (solid circles, n 4), andT19M (open triangles, n 3)Because the T19M was originally identified in a humanfamily (Santhiya et al. 2010), we investigated whetherhuman wild-type Cx46 and T19M behaved similarly.Indeed, wild-type human Cx46 localized extensively to gapjunction plaques with some localization in the cytoplasm intransfected HeLa cells (Fig. 8a). In contrast, human T19Mimmunoreactivity was detected mainly in intracellularcompartments; on very rare occasions a faint staining atappositional membranes was detected (Fig. 8b).Similar to the results obtained with the rat constructs,HeLa cells expressing human T19M-EGFP showed muchhigher rates of DAPI uptake in the presence of controlexternal solutions (containing 1 mM Ca2? and 1 mM Mg2?)than cells expressing wild-type Cx46 (Fig. 8c). The dyeuptake in cells expressing either human Cx46 or T19M wascompletely blocked by application of 200 lM La3? (Fig. 8c).Thus, the T19M mutation disrupted gap junctional plaque formation and extracellular divalent cation-dependentregulation of hemi-channel gating regardless of the speciesof origin of the Cx46.spectrum of alterations of connexin behavior. This residueis important for both cell biological and physiologicalfunctions, since Cx46T19M exhibits impaired gap junctionassembly and abnormal gap junctional channel and hemichannel activities.The inability of T19M to support intercellular communication is likely due to its poor formation of gap junctionplaques. The absence of gap junctions could potentiallyresult from altered trafficking of the connexin or reducedassembly of connexons into immunodetectable gap junction plaques. The detection of T19M hemi-channel activityby electrophysiology and dye uptake implies that themutant forms oligomers that traffic properly to the plasmamembrane. Therefore, the most likely explanation is thatthe mutant protein is inefficient at assembling into gapjunction plaques. Several other connexin mutants thatcause cataracts (including Cx50D47N and Cx50R23T)(Arora et al. 2008; Thomas et al. 2008) or other diseases(Di et al. 2002; Marziano et al. 2003; VanSlyke et al. 2000)lack gap junction function because of inefficient formationof gap junction plaques ascribed to impaired trafficking ofthe connexin. In contrast, T19M has impaired formation ofgap junction plaques, but it can traffic to the plasmamembrane.In our studies, T19M did not cause significant inhibition of co-expressed wild-type Cx46 or Cx50 gap junctionDiscussionIn this paper, we have shown that mutation of the conserved threonine-19 to methionine causes a unique123

152J.-J. Tong et al.: The Connexin46 MutantabT19MWT50 pA1s50 pA1sPeak Tail Current e (mV)(3)*(10)-40Fig. 6 T19M hemi-channels show alterations in voltage gating.Ensemble averaged current traces recorded from cells expressingwild-type rat Cx46 (a) or T19M (b) in response to a 2-s voltage clampstep to 80 mV followed by a hyperpolarizing step to -60 mV. Theholding potential was -60 mV. Dashed line indicates zero currentlevel. c Averaged peak tail currents at -60 mV. The number of cellstested is indicated within parentheses. d Bar graph summarizes thet1/2’s of deactivation of peak tail currents for wild-type rat Cx46(hatched bars) and T19M (black bars) at -80, -60, and -40 mV.Data are graphed as mean SEM. *p \ 0.01 (Student’s t test orMann–Whitney rank sum test compared with T19M-transfectedcells). The number of cells analyzed is indicated within parenthesesactivity. This loss-of-function behavior for gap junctionalintercellular communication is similar to that reported forsome other cataract-linked connexin mutants (e.g.,Cx50D47N, Cx46N63S, Cx46fs380) (Arora et al. 2008;Pal et al. 2000), but contrasts with the dominant-negativeeffect induced by other mutants (e.g., Cx50P88S,Cx50P88Q) (Arora et al. 2006; Pal et al. 1999). Thebehavior of T19M suggests that either it does not cooligomerize with wild-type Cx46 or Cx50 or that itspresence in mixed gap junction channels has no significant effect on function.123T19M also showed enhanced hemi-channel activity inthe presence of physiological concentrations of divalentcations. A ‘‘gain of hemi-channel function’’ has beenreported for some cataract-associated connexin mutantsincluding Cx50G46V (Minogue et al. 2009; Tong et al.2013) and Cx46G143R (Ren et al. 2013); however,unlike Cx46T19M, these mutants form gap junctionplaques and support high levels of intercellular communication. A few Cx43 mutants that form non-functionalgap junction plaques have been reported to form functional hemi-channels based on ATP release (Dobrowolskiet al. 2007). Some Cx26 mutants associated with keratitis-ichthyosis-deafness

Axioplan 2 microscope (Carl Zeiss Inc., Mu nchen, Ger-many) equipped with a mercury lamp, and images were acquired with a Zeiss AxioCam digital camera and Zeiss AxioVision software (Carl Zeiss Inc.). Figures were assembled using Adobe Photoshop CS3 Extended (Adobe Systems Inc., San Jose,