Homo- and heterotetrameric architecture of the epithelial Ca2+ channels TRPV5 and TRPV6

Abstract

The molecular assembly of the epithelial Ca(2+) channels (TRPV5 and TRPV6) was investigated to determine the subunit stoichiometry and composition. Immunoblot analysis of Xenopus laevis oocytes expressing TRPV5 and TRPV6 revealed two specific bands of 75 and 85-100 kDa, corresponding to the core and glycosylated proteins, respectively, for each channel. Subsequently, membranes of these oocytes were sedimented on sucrose gradients. Immuno blotting revealed that TRPV5 and TRPV6 complexes migrate with a mol. wt of 400 kDa, in line with a tetrameric structure. The tetrameric stoichiometry was confirmed in an electrophysiological analysis of HEK293 cells co-expressing concatemeric channels together with a TRPV5 pore mutant that reduced Cd(2+) sensitivity and voltage-dependent gating. Immuno precipitations using membrane fractions from oocytes co-expressing TRPV5 and TRPV6 demonstrated that both channels can form heteromeric complexes. Expression of all possible heterotetrameric TRPV5/6 complexes in HEK293 cells resulted in Ca(2+) channels that varied with respect to Ca(2+)-dependent inactivation, Ba(2+) selectivity and pharmacological block. Thus, Ca(2+)-transporting epithelia co-expressing TRPV5 and TRPV6 can generate a pleiotropic set of functional heterotetrameric channels with different Ca(2+) transport kinetics.

Fig. 1. Immunoprecipitation of TRPV5 (upper) and TRPV6 (lower) proteins. Membranes of non- (ni), HA-TRPV5- or Flag-TRPV6-expressing oocytes were solubilized and subjected to endoF and endoH treatment. Glycosylated TRPV5 (gTRPV5) and TRPV6 (gTRPV6) proteins are indicated, and the protein bands labeled TRPV5 or TRPV6 represent the non-glycosylated core proteins.

Fig. 2. Determination of the TRPV5/6 oligomeric structure using chemical cross-linking. Lysates of (A) TRPV5- and (B) TRPV6-expressing oocytes incubated with sample buffer containing DTBP. Complexes were treated with DTT and loaded in the third lane.

Fig. 3. Immunoblot analyses of the oligomeric state of TRPV5 and TRPV6. Membranes from TRPV5- or TRPV6-expressing oocytes were solubilized in 0.5% (w/v) deoxycholate and subjected to sucrose gradient centrifugation. SDS indicates that 0.1% (w/v) SDS has been added to the sucrose gradient. The fractions with peak intensities of the marker proteins (phosphorylase B, 97 kDa; alcohol dehydrogenase, 150 kDa; catalase, 232 kDa; apoferritin, 442 kDa) are indicated.

Fig. 4. Co-localization of TRPV5 and TRPV6 in kidney. (A) Mouse kidney cortex sections were co-stained with antibodies against TRPV5 (left) and TRPV6 (right). (B) Immunoblotting of membrane preparations from oocytes expressing TRPV5 and TRPV6. To exclude cross-reactivity between the antibodies, the left blot was incubated with the TRPV5 antibody and the right blot was incubated with the TRPV6 antibody.

Fig. 5. Co-immunoprecipitation of TRPV5 and TRPV6. Copy RNA of HA-TRPV5 and/or Flag-TRPV6 was (co-)injected in oocytes and cell lysates were processed. (A) Immunoblot analysis demonstrated that both channel proteins are expressed and the applied antibodies do not cross-react. Co-immunoprecipitations were performed with the HA and Flag antibodies and subsequently immunoblots were probed using (B) the TRPV5 antibody and (C) the Flag antibody. Four oocytes expressing TRPV5 or TRPV6 were used for the immunoblot analysis depicted in (A), whereas 12 oocytes were processed for each condition in the co-immunoprecipitation experiments shown in (B) and (C). The total amount of the sample was loaded on the gel.

Fig. 6. Cd²⁺ sensitivity of TRPV5, TRPV6 and TRPV5^D542A mono- and multimers. (A–F) Current–voltage relationships obtained during voltage ramps in nominally divalent-free extracellular solutions in the absence and presence of 2, 20, 200 and 2000 µM CdCl₂ for cells transfected with (A) TRPV5555, (B) TRPV5^D542A, (C) TRPV55^D542A55, (D) a mixture of TRPV5 and TRPV5^D542A in a 3:1 ratio, (E) a mixture of TRPV6666 and TRPV5^D542A in a 1:1 ratio and (F) a mixture of TRPV666 and TRPV5^D542A in a 1:1 ratio. (G–I) Dose–response curves for the effect of Cd²⁺ measured at –100 mV. (G) Dose–response curves for TRPV5555 and TRPV5^D542A. From Hill functions fitted to the data (solid curves), we obtained values for K_D and n_Hill of 64 nM and 0.78, respectively, for TRPV5555 compared with 313 µM and 0.84 for TRPV5^D542A. Note that the Cd²⁺ sensitivity of the TRPV5555 concatemer was not significantly different from that of the TRPV5 monomers (K_D = 74 nM, n_Hill = 0.81; data not shown). (H) Dose–response curves for TRPV55^D542A55 and the mixture of TRPV5 and TRPV5^D542A. From a Hill function fitted to the TRPV55^D542A55 data (solid curve), we obtained values for K_D and n_Hill of 1.0 µM and 0.77, respectively. The data for the mixture of TRPV5 and TRPV5^D542A were not well fitted by a single Hill function, indicating a population of channels with distinct Cd²⁺ sensitivities. (I) Dose–response curves for TRPV6666 and for mixtures of TRPV5^D542A with TRPV6666 and TRPV666, respectively. From a Hill function fitted to the TRPV6666 data, values for K_D and n_Hill of 163 nM and 1.05, respectively, were obtained. Similar values were obtained for TRPV666 (K_D = 157 nM, n_Hill = 0.93) and for the TRPV6 monomer (K_D = 261 nM, n_Hill = 1.05). The dose–response curve for the mixture of TRPV6666 and TRPV5^D542A was well described by the weighted sum of the Hill functions for TRPV6666 and TRPV5^D542A (solid curve). In contrast, the dose–response curve for the mixture of TRPV666 and TRPV5^D542A was poorly fitted by the weighted sum of the Hill functions for TRPV6666 and TRPV5^D542A (dotted line).

Fig. 7. Dominant-negative effect of the TRPV5^D542A mutation on voltage-dependent gating of TRPV5/6 homo- and heterotetramers. (A) Voltage protocol. Voltage steps were delivered at a frequency of 0.5 Hz. Note that in these experiments the intracellular solution contained 3 mM MgCl₂ (calculated free intracellular Mg²⁺ = 127 µM) instead of the normal 1 mM to accentuate the voltage-dependent behavior of TRPV5/6. (B–F) Currents measured in divalent-free solution supplemented with 10 mM EDTA from cells expressing the indicated constructs or mixtures of constructs. (G and H) Voltage dependence of the apparent open probability for the constructs or mixtures of constructs indicated. The apparent open probability was determined as the current immediately upon stepping back to –100 mV normalized to the current at the end of the initial step to –100 mV.

Fig. 8. Expression and analysis of (hetero)tetrameric TRPV5/6 channels in HEK293 cells. (A) Currents at hyperpolarizing steps from the +20 mV holding potential to –100 mV. Extracellular Ca²⁺ and Ba²⁺ concentration was 30 mM. Current densities, expressed per unit membrane capacitance, were calculated from the current at –80 mV during the ramp protocols. (B) Normalized current block of heterotetrameric proteins by ruthenium red (1 µM). (C) Normalized I_Ba/I_Ca current ratio. (D) Inactivation kinetics of heterotetrameric proteins. Fast inactivation was assessed by the time for 10% decay (t_90%) of the current, and the slower run down by the time constant of a mono-exponential fit of the current during the last 1.5 s of the step.