Conformational changes of pore helix coupled to gating of TRPV5 by protons

Abstract

The transient receptor potential channel TRPV5 constitutes the apical entry pathway for transepithelial Ca²⁺ transport. We showed that TRPV5 was inhibited by both physiological intra- and extracellular acid pH. Inhibition of TRPV5 by internal protons was enhanced by extracellular acidification. Similarly, inhibition by external protons was enhanced by intracellular acidification. Mutation of either an extra- or an intracellular pH sensor blunted the crossinhibition by internal and external protons. Both internal and external protons regulated the selectivity filter gate. Using the substituted cysteine accessibility method, we found that intracellular acidification of TRPV5 caused a conformational change of the pore helix consistent with clockwise rotation along its long axis. Thus, rotation of pore helix caused by internal protons facilitates closing of TRPV5 by external protons. This regulation by protons likely contributes to pathogenesis of disturbances of Ca²⁺ transport in many diseased states. Rotation of pore helix may be a common mechanism for crossregulation of ion channels by extra- and intracellular signals.

Introduction

Transient receptor potential (TRP) channels are widespread and play many important functions, ranging from store- and receptor-operated Ca²⁺ entry, thermal, tactile, taste, osmolar, and fluid flow sensing, embryonic development to epithelial Ca²⁺ and Mg²⁺ transport (Hoenderop et al, 2002; Montell et al, 2002; Clapham, 2003; Huang, 2004). TRP channels are classified into TRPC, TRPV, TRPM TRPP, TRPML, TRPN, and TRPA subfamilies. The TRPV subfamily is named after the first mammalian member of the subfamily, vanilloid receptor 1 (VR1), and contains six mammalian members, TRPV1–6 (Jordt et al, 2003). TRPV5 and TRPV6 are the only two highly Ca²⁺-selective TRP channels and mediate trans-epithelial Ca²⁺ transport in the kidney and intestine (Hoenderop et al, 2002). Structurally, all TRP channels have six predicted transmembrane (TM) segments and N- and C-terminal cytoplasmic tails similar to the topologies of voltage-gated K⁺, Na⁺, and Ca²⁺ channels, cyclic nucleotide-gated (CNG) channels, and hyperpolarization-activated channels (Montell et al, 2002; Clapham, 2003). The six-TM polypeptide subunits of TRP channels likely assemble as tetramers to form cation-permeable pores (Hoenderop et al, 2003).

The bacterial K⁺ channel KcsA contains two membrane-spanning segments and an intervening pore (P) loop (Doyle et al, 1998). The crystal structure of KcsA reveals an outer selectivity filter gate stabilized by pore helices and an inner gate formed by four inner helices of four identical subunits packed against each other as a bundle near the intracellular aspect of the membrane, giving an appearance of an inverted teepee (Doyle et al, 1998). This architecture is likely a general structure for all P-loop-containing channels including voltage-gated K⁺ channels, inward rectifier and Ca²⁺-activated K⁺ channels, and CNG cation channels (Flynn et al, 2001). How and which gate is activated by individual extra- and intracellular signals for each channel is under intense investigation. The location of the inner gate suggests its role as an activation gate for intracellular signals. However, signals acting from the intracellular aspect of the membrane also impact on the selectivity filter (Flynn and Zagotta, 2001).

TRPV5 is localized to the apical membrane of polarized epithelia and functions as a gatekeeper for transepithelial Ca²⁺ transport (Hoenderop et al, 2002). Transepithelial Ca²⁺ transport is inhibited in conditions associated with overproduction of acids (Sutton et al, 1979). How acid inhibits transepithelial Ca²⁺ transport is not known. One potential mechanism is by direct inhibition of the TRPV5 channel. Little is known about the structure and mechanism involved in the gating of TRP channels. Mutation of aspartate-542 of TRPV5 greatly reduces Ca²⁺ permeation, indicating that the aspartate and adjacent amino acids form the selectivity filter in the channel (Nilius et al, 2001). A stretch of ∼20 amino acids in the pre-selectivity filter region of TRPV5 and TRPV6 is highly homologous to that in KcsA and cysteine-substituted mutants of the region exhibit a cyclic pattern of reactivity to MTS reagents (Dodier et al, 2004; Voets et al, 2004). Thus, this region likely forms the pore helix of TRPV5 and TRPV6.

In this study, we show that the selectivity filter gate of TRPV5 is crossregulated by physiological extra- and intracellular pH (pH_i). Gating of TRPV5 by intracellular acidification is coupled with a clockwise rotation of the pore helix. This intracellular proton-induced conformational change of the pore helix facilitates gating by extracellular protons. These results provide a novel mechanism for crossregulation of ion channels by intra- and extracellular signals.

Results

Inhibition of TRPV5 by intracellular acidification

We examined pH_i regulation of TRPV5 by inside-out patch-clamp recording (Figure 1A). As shown, TRPV5 currents were recorded using voltage ramps (from −100 to 100 mV over 400 ms). In inside-out patches, TRPV5 exhibited inwardly rectifying currents in divalent-free (DVF) cytoplasmic solutions (see Figure 1B ramp I–V curve at pH_i 9.0: outward (at 100 mV) and inward currents (at –100 mV) were 197 and −412 pA, respectively). The Mg²⁺-independent inward rectification has been suggested due to intrinsic gating (Voets et al, 2003). Maximal TRPV5 currents occurred when inside-out membranes were exposed to solutions with pH 9.0–9.5 (Figure 1B and C). Decreasing pH from 9.0 to 4.9 caused a progressive reduction of TRPV5 currents, which reversed upon increasing the pH to 9.0 (Figure 1C). At an extracellular pH (pH_e) of 7.2 (illustrated in Figure 1A), the apparent pK_a for pH_i regulation was estimated at 7.31±0.12 (n=21) (Figure 1D). The Hill coefficient for pH_i regulation is ∼2.1, indicating that each channel is gated by more than one proton and with cooperativity. Thus, TRPV5 is steeply gated by internal protons at physiological and pathophysiological pH_i's.

Figure 1.

Regulation of TRPV5 by pH_i and its modulation by pH_e. (A) Inside-out recording and voltage ramp protocol (applied every 10 s). (B) Current–voltage (I–V) relationships of currents (in DVF solution) at different pH_i's. The degree of inward rectification in our study appears to be slightly less than that in some study (Voets et al, 2003). Mechanism(s) for the apparent difference awaits further study. (C) Inward currents (pA, at −100 mV) at different pH_i's over time. Currents through TRPV5 were inhibited by lanthanum (La³⁺, 10 mM). In mock-transfected cells, increasing pH_i from 4.9 to 9.4 caused a relatively insignificant increase of background inward currents (<50 pA, compared to the mean maximal inward TRPV5 currents at pH_i 9.4 (452±18 pA, n=21)). (D) Dose–response curve for inhibition of TRPV5 by intracellular acid. Relative currents (after subtraction of residual currents in 10 mM La³⁺, normalized to the maximal current, labeled as 100) at any given pH_i are shown.

Effects of internal protons on TRPV5 single-channel properties were further studied. As shown in the representative recording with only one active channel in the patch, single-channel open probability (P_o) was higher at pH_i 8.4 than at 7.0 (P_o: 0.86 versus 0.24; Supplementary Figure 1A). In multiple recordings, the average NP_o (number of channels in the excised patch [N] × P_o) at pH_i's 8.4 and 7.0 were 3.05±1.1 and 0.97±0.4, respectively (P<0.05; Supplementary Figure 1B). Single-channel conductance (measured between –25 and –75 mV) at pH_i's 8.4 and 7.0 were 78.2±4.6 and 60.4±2.6 pS, respectively (P<0.05; Supplementary Figure 1C). Thus, as for external protons (Yeh et al, 2003), intracellular protons reduce single-channel conductance as well as open probability.

Modulation of pH_i regulation of TRPV5 by pH_e

Extracellular protons also inhibit TRPV5 (Yeh et al, 2003). We investigated the potential modulation by pH_e of pH_i regulation of TRPV5. Effects of pH_i on TRPV5 at various extracellular (pipette) pH's were studied using inside-out recordings (Figure 2A). Increasing the pH_e from 7.2 (indicated by the open triangle curve in Figure 2A and second bar from right in Figure 2B) to 8.4 (black curve and black bar in Figure 2) caused a shift in the pK_a for pH_i regulation toward more acidic values (see Supplementary Figure 2A and B for original traces). Conversely, decreasing pH_e from 7.2 to 6.8 caused a shift in pK_a for pH_i regulation toward more alkaline pH (Figure 2A and B; Supplementary Figure 2C and D). Thus, opening and closing of the channel by extracellular alkalinization and acidification, respectively, facilitates opening and closing by intracellular alkalinization and acidification. Intra- and pH_e could crossregulate TRPV5 by acting on the same gate or by regulating different activation gates. The pH_e-induced shift in pK_a for pH_i regulation based on normalized relative current, nevertheless, suggests that both pH's crossregulate the same gate.

Figure 2.

Modulation of pH_i regulation of TRPV5 by pH_e. (A) Dose–response curves for pH_i regulation (pH_i from 9.5 to 4.9) at different pH_e's. Experimental paradigm is as in Figure 1D. (B) pK_a for pH_i regulation (Y-axis) at different pH_e's (X-axis).

Modulation of pH_e regulation of TRPV5 by pH_i

If both pH's regulate the same gate, pH_i would also modulate pH_e regulation of the channel. Using whole-cell recordings (Figure 3A), decreasing intracellular (pipette) pH's from 8.4 to 6.4 caused a shift in pK_a for pH_e regulation from 5.1±0.1 to 7.3±0.1 (Figure 3A and B), indicating that intracellular acidification increases the sensitivity to extracellular protons. Glutamate-522 in the pre-pore helix region of TRPV5 is an pH_e sensor (Yeh et al, 2003). We found that modulation by pH_i or pH_e regulation was blunted in the E522Q mutant (Figure 3C and D), supporting that pH_i's and pH_e's crossregulate the same gate.

Figure 3.

Modulation of pH_e regulation of TRPV5 by pH_i. (A) Dose–response curves for inhibition of WT TRPV5 by extracellular acidification (pH_e from 9.0 to 4.4) at different pH_i. Relative currents (normalized to the maximal current, labeled as 100) at any given pH_e are shown. (B) pK_a for pH_e regulation (Y-axis) at different pH_i (X-axis) for WT TRPV5. (C) Same as panel A, except for E522Q TRPV5 mutant. The mean inward current densities at pH_i 7.2 and pH_e 7.4 were 438±29 pA/pF for E522Q mutant and 531±42 pA/pF for WT. (D) Same as panel B, except for E522Q TRPV5 mutant. pK_a for WT (from panel B) is shown in circles connected by line for comparison.

Both intra- and extracellular acidification decrease the estimated diameter of TRPV5 pore

To investigate whether internal and external protons crossregulate the selectivity filter gate, we first examined the effects of pH_i and pH_e on the estimated diameter of ion permeation pore. The pore diameter was estimated by measuring the relative permeability of organic monovalent cations of increasing size (Sabovcik et al, 1995; Voets et al, 2004).

Like many Ca²⁺-permeating channels, TRPV5 conducts monovalent cations in the absence of Ca²⁺ ions (Nilius et al, 2001; Yeh et al, 2003). With Na⁺ as the sole cation in the extra- and intracellular solution, TRPV5 currents reversed at 0 mV (−0.30±0.52 mV, n=18; Figure 4A). Replacing Na⁺ in the extracellular solution with the same concentration of monomethylammonium (MA⁺; 3.6 Å), dimethylammonium (DMA⁺; 4.6 Å), or N-methyl-D-glucamine (NMDG⁺; 6.8 Å) resulted in reductions of inward currents and shifts of reversal potentials toward negative membrane potentials. The permeability ratios of MA⁺, DMA⁺, and NMDG⁺ relative to Na⁺ (P_X/P_Na) were determined from the biionic reversal potential (Hille, 2001). The diameters (Å) of ion permeation pores were estimated (using a modified excluded volume equation (P_X/P_Na)^1/2=dP−dX/dP−dNa, where dP, dX, and dNa are diameters of the pore, cation X⁺, and Na⁺, respectively; see Sabovcik et al, 1995) at ∼10.3, 7.5, and 5.6 for pH_i's 9.0, 7.4, and 6.0, respectively (Figure 4B and C). These estimated pore diameter values agree with those using another excluded volume equation (Dwyer et al, 1980; Voets et al, 2004) (Supplementary Figure 3). Thus, intracellular acidification reduces the estimated TRPV5 pore diameter. Similarly, extracellular acidification reduces the estimated pore diameter. As shown in Figure 4D and E, the pore diameters (Å) were estimated at ∼11.8, 7.5, and 6.3 for pH_e's 9.0, 7.4, and 6.0, respectively.

Figure 4.

Effects of pH_i and pH_e on the estimated pore size. (A) Relative permeability of MA⁺, DMA⁺, and NMDG⁺ to Na⁺ at pH_i and pH_e 7.4. Whole-cell currents were first recorded in a bath solution containing 130 mM NaAsp, which was subsequently replaced by solutions containing 130 mM X-Asp. X is MA, DMA or NMDG. Shown are I–V curves at these different extracellular cations. The differential rectification in different extracellular cations and the reduction of outward currents in extracellular NMDG⁺ may be due to pore block by partially permeant cations. (B) Relationships of the relative permeability ratios of permeating ions over Na⁺ versus diameter of permeating ions at different pH_i's. Experiments were performed as in panel A, with pH_i at 9.0, 7.4, or 6.0. The relative permeability ratios (P_X/P_Na) were calculated according to the equation P_X/P_Na=exp ΔE_rev × F/RT. ΔE_rev is the shift of reversal potential upon changing the bath solution from NaAsp to X-Asp. X is MA, DMA, or NMDG with diameters of 3.6, 4.6, and 6.8 Å, respectively. (C) Estimated pore diameter (Å) at pH_i's 9.0, 7.4, and 6.0. (D) Relationships of the relative permeability versus diameter of permeating ions at different pH_e's. (E) Estimated pore diameter at pH_e 9.0, 7.4, and 6.0.

Effects of pH_i and pH_e on the monovalent cation selectivity of TRPV5

The effects of pH on the selectivity filter of TRPV5 were further investigated by examining the monovalent cation selectivity. Permeability ratios of Li⁺, K⁺, and Cs⁺ relative to Na⁺ were determined by replacing extracellular Na⁺ with the respective ions. At pH_i 7.4, the relative permeability ratios of Li⁺, K⁺, and Cs⁺ to Na⁺ were 0.85±0.04, 0.76±0.04, and 0.65±0.05, respectively (Figure 5A and Table I). This selectivity sequence (Na⁺>Li⁺>K⁺>Cs⁺) follows Eisenman selectivity sequence X, indicating a strong-field-strength-binding site (Hille, 2001; Nilius et al, 2001). An increase in the diameter of the selectivity filter would decrease the electrostatic field strength of the ion-binding site. Indeed, we found that increasing pH_i from 7.4 to 9.0 changed the selectivity sequence to Eisenman III or IV, indicating a weak-field-strength-binding site (Figure 5B and Table I). These results support the idea that intracellular protons regulate the selectivity filter gate. Sequence X (at pH_i 7.4) represents the second strongest-field-strength-binding sequence among the 11 Eisenman sequences. Accordingly, decreasing pH_i from 7.4 to 6.0 did not change the Eisenman sequence (Table I).

Figure 5.

Relative permeability of inorganic monovalent cations to Na⁺ at pH_i 7.4 (A) or 9.0 (B). Whole-cell experiments were performed with extracellular at pH 7.4 and pH_i at 7.4 (panel A) or 9.0 (panel B). Currents were first recorded in a bath solution containing 130 mM NaAsp, which was subsequently replaced by solutions containing 130 mM X-Asp. X is Li, Cs, or K. Shown are I–V curves at these different extracellular cations. The two left panels show I–V curves in the full range. The right panels show I–V curves of the boxed regions in an expanded scale.

Table 1.

Effects of pH's on relative permeability of inorganic monovalent cations to Na⁺

		K+	Li+	Cs+	Permeability sequence	Eisenman sequence
pHi (at fixed pHe 7.4)	6.0	0.80±0.09	0.92±0.03	0.71±0.05	Na>Li>K>Cs	X
pHi (at fixed pHe 7.4)	7.4	0.76±0.04	0.85±0.04	0.65±0.05	Na>Li>K>Cs	X
pHi (at fixed pHe 7.4)	9.0	1.15±0.13	1.01±0.11	1.09±0.10	K>Cs>Na=Li	III or IV
pHe (at fixed pHi 7.4)	7.4	0.79±0.04	0.91±0.04	0.68±0.05	Na>Li>K>Cs	X
pHe (at fixed pHe 7.4)	9.0	1.10±0.07	0.88±0.11	0.98±0.11	K>Na=Cs>Li	IV or V
Data shown are the relative permeabilities of the respective ions to Na+ (mean±s.e.m.).

We further examined the effects of pH_e on TRPV5 selectivity profile. Consistent with the results of increasing the pore diameter, extracellular alkalinization from pH 7.4 to 9.0 (at a fixed pH_i 7.4) altered the selectivity sequence from Eisenman sequence X to a weaker-field-strength-binding sequence (sequence IV or V; Table I). These results, together with the effects of pH on the pore diameter, indicate that both internal and external protons regulate the selectivity filter gate.

Intracellular acidification causes clockwise rotation of the pore helix

Amino acids 527–538 of TRPV5, likely from the pore helix of TRPV5 (Dodier et al, 2004; Voets et al, 2004), are similar to the pore helix of KcsA (Figure 6A). The selectivity filter of TRPV5 contains aspartate-542 critical for Ca²⁺ permeation (Nilius et al, 2001). The conformational state of pore helix is critical for the open conformation of selectivity filter (Doyle et al, 1998). We therefore hypothesized that internal and external protons crossregulate the selectivity filter gate via conformational changes in the pore helix.

Figure 6.

Accessibility of the pore helix to MTSET. (A) Alignment of the pore helix and selectivity filter of TRPV5 versus KcsA. Amino acids 527–543 of TRPV5 are shown. D542 of TRPV5 is critical for permeation of Ca²⁺ ions. Amino acids A529–F537 were individually mutated to cysteine. (B) Illustration of externally applied MTSET (1 mM) in whole-cell recordings at pH_i 8.4 or 6.8. (C–F) Effects of MTSET on A529C (C), L530C (D), F531C (E) or S532C (F) at pH_i 8.4 (○) or 6.8 (•). Inward currents (at −100 mV) from voltage ramp applied every second. Cells were transfected with varying amounts of DNA (0.5–5 μg) to produce similar current density for different mutants (ranging from 354±42 to 554±57 pA/pF at −100 mV and pH_e 8.4). Currents were normalized to the maximal current prior to application of MTSET. (G) Percent inhibition by MTSET (after 3 min) of cysteine-substituted mutants (from A529 to F537) at pH_i 8.4 (white bar) or 6.8 (black bar). (H) Helical wheel projection of amino acids of the pore helix of TRPV5. MTSET accessibility area of the pore helix is shown in gray. E535 and S532 are not accessible at pH_i 8.4, but become accessible at 6.8. In contrast, L530 and F537 are accessible at pH_i 8.4, but become inaccessible at 6.8. Amino acids L536, A529, and T533 are accessible at both pH_i's. Amino acids F531 and F534 are inaccessible at either pH_i.

To test this hypothesis, we introduced cysteines into amino acids 529–537 of TRPV5 and examined reactivity of the mutants to extracellular membrane-impermeant cysteine-modifying reagent MTSET at either pH_i 8.4 or 6.8 (Figure 6B). Application of MTSET (1 mM) caused a fast inhibition of A529C at both pH_i's 6.8 and 8.4 (Figure 6C). The inhibition was not reversible by washout (not shown), indicating covalent modification of cysteine residues by MTSET. The second-order rate constants for inhibition were ∼200 and ∼75 M⁻¹ s⁻¹ at pH_i's 6.8 and 8.4, respectively. In contrast, L530C was inhibited by MTSET at pH_i 8.4, but barely inhibited after 3 min application of MTSET at pH_i 6.8 (Figure 6D). F531C was not inhibited at either pH_i (Figure 6E); S532C was inhibited at pH_i 6.8, but not at pH_i 8.4 (Figure 6F). Figure 6G summarizes the effects of MTSET on cysteine-substituted amino acids from A529C to F537C. A529C, L530C, T533C, L536C, and F537 were >75% inhibited by MTSET at pH_i 8.4 in 3 min, whereas F531C, S532C, F534C, and E535C were not. As in helical wheel projection, amino acids sensitive to MTSET are located in the same face of the helix (Figure 6H). In contrast, the MTSET-resistant amino acids are located in a separate face. Assuming that lack of irreversible inhibition reflects inaccessibility to MTSET, these results indicate that the face of pore helix bound by L536 and L530 is exposed to solvents at pH_i 8.4. Interestingly, the solvent-exposed face at pH_i 6.8 is different from that at pH_i 8.4. As shown in Figure 6G and H, E535C and S532C were not accessible at pH_i 8.4, but accessible at pH_i 6.8. In contrast, F537C and L530C, two of the residues accessible at pH_i 8.4, were not accessible at pH_i 6.8 (Figure 6G and H). TRPV5 is mostly open at pH_i 8.4 (∼90% of the maximal current) and predominantly closed at pH_i 6.8 (∼30% of the maximal current). Thus, closing of TRPV5 by internal protons is associated with a clockwise rotation of the pore helix.

Cysteine substitution of pore helix does not affect gating by protons

To exclude that cysteine substitution affects gating by protons, we examined three representative cysteine-substituted mutants (L530C, F531C, and S532C) located in separate faces of the pore helix (see Figure 6H). As shown in Figure 6, L530C is accessible at pH_i 8.4, but not at 6.8, whereas F531C is not accessible at either pH_i. S532C is accessible at pH_i 6.8, but not at 8.4. Here, we found that the sensitivity to pH_i for each of L530C, F531C, and S532C mutants was not different from that of wild-type (WT) TRPV5 (Figure 7A). Also, the sensitivity to pH_e and its modulation by pH_i for the mutants were essentially the same as for the WT (Figure 7B–F). These results indicate that cysteine substitution for L530, F531, and S532 does not affect gating by pH, and support our conclusion that changes in MTSET reactivity for amino acids in the pore helix are due to rotation along its along axis.

Figure 7.

No effects of cysteine substitution on pH gating. (A) pH_i regulation of L530C, F531C, and S532C mutants versus WT. Relative currents (maximal current=1) versus pH_i's are shown. (B–E) Dose–response curves for pH_e regulation at pH_i 8.4 or 6.8 for WT (B), L530C (C), F531C (D), and S532C (E). (F) pK_a for pH_e regulation at pH_i 8.4 and 6.8 for WT, L530C, F531C, and S532C. pK_a values are from panels B, C, D, and F, respectively.

Amino acid aspartate-542 is important for intracellular acidification-induced changes of the pore diameter

Mutation of aspartate-542 of TRPV5 to alanine (D542A) increases the estimated pore diameter (Voets et al, 2004), suggesting that the side chain of aspartate-542 lines the narrowest part of the ion permeation pathway, forming the molecular sieve for permeating ions. We found that, unlike for WT TRPV5, the relative permeability of MA⁺, DMA⁺, and NMDG⁺ to Na⁺ for D542A mutant was not altered by intracellular acidification (Figure 8A; see Figure 4B for comparison). The estimated pore diameters (Å) at pH_i's 9.0, 7.4, and 6.0 for D542A mutant were ∼10, 9.7, and 9.2, respectively (Figure 8B). For comparison, the estimated pore diameter (Å) at pH_i's 9.0, 7.4, and 6.0 for WT TRPV5 were ∼10.3, 7.5, and 5.6, respectively (see Figure 4C). These results support the conclusion that intracellular protons regulate the selectivity gate of TRPV5 formed by aspartate-542.

Figure 8.

Role of aspartate-542 in pH_i-mediated changes of the pore diameter. (A) The relative permeability ratios of permeating ions over Na⁺ versus diameter of permeating ions at different pH_i's for D542A mutant. Experimental paradigm is as in Figure 4B. (B) Estimated pore diameter (Å) at pH_i's 9.0, 7.4, and 6.0 for D542A mutant. Experimental paradigm is as in Figure 4C. (C) Solutions and diagram of whole-cell experiments of the relative permeability ratio of Ca²⁺ versus Na⁺. (D) I–V curves of WT TRPV5 currents recorded at 140 mM NaAsp bath solution (labeled as Na⁺) and upon changing to 130 mM NaAsp+10 mM CaCl₂ (labeled as Ca²⁺). Upward arrow indicates reversal potentials in 130 NaAsp+10 CaCl₂-containing bath solution. Currents inactivated rapidly in four sweeps (numbered 1, 2, 3, and 4; one voltage ramp sweep per second). (E) Effects of intracellular acidification on Na⁺ versus Ca²⁺ currents. Experimental paradigm is as shown in panels C and D. Na⁺ currents (‘Ina', gray bar) are inward currents (−80 mV) at 130 mM NaAsp-containing bath solution. Ca²⁺ currents (‘Ica', white bar) are inward currents (at −80 mV) from the first sweep after changing to Ca²⁺-containing bath solution. Currents at pH_i's 7.4 and 6.0 were normalized to pH_i 9.0. Mean±s.e.m., n=12–15 for each group. (F) Relative permeability ratio of Ca²⁺ versus Na⁺ (P_Ca/P_Na) for WT TRPV5 (gray bar) and D542A mutant (white bar) at different pH_i's (9.0, 7.4, 6.0). The permeability ratio of Ca²⁺ versus Na⁺ was calculated according to the equation, P_Ca/P_Na=([Na⁺]_i/4[Ca²⁺]_o) exp(ΔE_rev F/RT) {(1+exp(ΔE_rev F/RT)}. ΔE_rev's (shift of reversal potential from Na⁺ to Ca²⁺ bath solution) were 7.6±2.5, 22.3±4.5, and 13.5±2.2 mV for pH_i's 9.0, 7.4, and 6.0, respectively (see Supplementary Figure 4). Mean±s.e.m., n=8–12 for each group.

The physiological function of TRPV5 is to mediate transepithelial Ca²⁺ transport. pH_e affects TRPV5-mediated Ca²⁺ entry (Yeh et al, 2003). Here, we further examined the effects of pH_i on Ca²⁺ permeation and selectivity. Whole-cell Na⁺ current (at pH_i 9.0, 7.4, or 6.0) was first recorded in a bath solution containing (in mM) 140 NaAsp (Figure 8C). Subsequently, current was recorded in a bath containing (in mM) 130 NaAsp plus 10 CaCl₂. Owing to the anomalous mole fraction effect, inward currents would be predominantly carried by Ca²⁺ ions in the 10 mM CaCl₂-containing solution (Vennekens et al, 2001). The intracellular solution contains 140 mM Na⁺. As shown in a representative recording at pH_i 7.4, currents reversed at ∼0 mV in the 140 mM Na⁺-containing bath solution, indicating Na⁺ currents (Figure 8D; labeled ‘Na⁺'). In the 10 mM Ca²⁺-containing bath solution, currents reversed at a positive membrane potential (indicated by the upward arrow), indicating inward Ca²⁺ currents (Figure 8D; labeled ‘Ca²⁺'). Ca²⁺ entry through TRPV5 raises the local intracellular Ca²⁺ concentration and causes feedback inhibition of the channel (Vennekens et al, 2000). As shown, TRPV5 Ca²⁺ currents decreased over 4 s (Figure 8D; labeled ‘1–4'). The magnitude of the first Ca²⁺ current (at −100 mV) recorded immediately after changing to the Ca²⁺ solution (labeled ‘1' in Figure 8D) was rather consistent (about 50–60% relative to that of the Na⁺ current in all experiments) and thus used for examining effects of pH_i on TRPV5-mediated Ca²⁺ current. Similar to the Na⁺ current (labeled ‘Ina'), intracellular acidification and alkalinization decreased and increased Ca²⁺ current, respectively (labeled ‘Ica'; Figure 8E). Thus, the regulation by pH_i is likely important for the physiological function of TRPV5 in Ca²⁺ transport.

The effect of pH_i on Ca²⁺ selectivity was examined by measuring the relative permeability ratio of Ca²⁺ to Na⁺ (P_Ca/P_Na). As reported by others (Nilius et al, 2001), we found that TRPV5 was highly Ca²⁺-selective and D542A mutation markedly decreased the selectivity (P_Ca/P_Na at pH_i 7.4: 65±7 and 3.2±0.4 for WT TRPV5 and D542A mutants, respectively; Figure 8F). Interestingly, the relative permeability ratio of Ca²⁺ to Na⁺ was decreased by either alkalinization to pH_i 9.0 or acidification to pH_i 6.0 (P<0.05 versus pH_i 7.4; Figure 8F and Supplementary Figure 4). These results are consistent with the idea that the side chains of aspartate-542 of selectivity filter are oriented to provide the most stable interaction with Ca²⁺ ions at physiological pH_i ∼7.4. Displacement of the side chains caused by rotation of the pore helix (either clockwise rotation at pH_i 6.0 or counter-clockwise rotation at pH_i 9.0) destabilizes the interaction between side chains and Ca²⁺ ions and decreases the selectivity for Ca²⁺. The decrease in the relative permeability of Ca²⁺ to Na⁺ by intracellular acidification, together with the decrease in single-channel conductance and open probability, leads to reduction of calcium entry. Intracellular alkalinization, however, causes an increase in single-channel conductance and open probability besides its effect on the relative permeability of Ca²⁺ to Na⁺. The net effect of alkalinization on TRPV5 is to increase the calcium entry (see Figure 8E above).

Amino acid lysine-607 is important for pH_i sensing and crossregulation by intra- and extracellular protons

We further searched for pH_i sensor(s). There are more than 140 titratable amino acids in the intracellular portion of TRPV5. We focused on 17 amino acids surrounding the putative intracellular entrance of the pore (i.e., from the fourth TM domain to the proximal C-terminus; see Figure 9A). These include lysine-484 in the intracellular loop between the fourth and fifth TM domains (IC loop; Figure 9A), aspartate-489 and arginine-492 in the fifth TM domain, glutamate-535 and aspartate-542 in the pore region, tyrosine-595 in the sixth TM domain, and aspartate-580, histidine-582, arginine-584, glutamate-588, aspartate-590, glutamate-591, and lysine-607 in the proximal carboxyl-terminus of the channel. We mutated each of these amino acids to a nontitratable amino acid and found that only mutation of lysine-607 to asparagine (K607N) decreased the sensitivity of the channel to inhibition by intracellular acidification. The pK_a for inhibition by intracellular acidification were 7.41±0.15 and 6.17±0.13 for WT TRPV5 and K607N mutant, respectively (P<0.05, n=5 each; Figure 9B). We further examined a role of lysine-607 in the crosstalk between pH_i and pH_e. As in Figure 7F, intracellular acidification from pH_i 8.4 to 6.8 increased the pK_a for pH_e regulation of WT TRPV5 from 5.1±0.07 to 6.8±0.08 (Figure 9C). The increase in the pK_a for pH_e regulation by intracellular acidification was blunted for K607N mutant (pK_a: 4.9±0.06 and 5.6±0.06 at pH_i's 8.4 and 6.8, respectively). Thus, lysine-607 is an pH_i sensor that contributes to crossregulation by pH_i and pH_e.

Figure 9.

Amino acid lysine-607 is important for pH_i sensing and crossregulation by pH_i and pH_e. (A) Location of lysine-607 (K607) in the intracellular proximal carboxyl terminus (Ct). Illustrated is side-view of two subunits of a tetrameric channel. Only the fourth, fifth, and sixth TM (labeled ‘4', ‘5', and ‘6') and pore helices are shown. (B) Mutation of lysine-607 to asparagine (K607N) decreases intracellular acidification-mediated inhibition. Experimental paradigm is as in Figure 1. WT and K607N mutant channels in inside-out patches were exposed to pH_i ranging from pH 9 to 5. (C) pK_a for pH_e regulation at pH_i's 8.4 and 6.8 for WT (•) and K607N (○). Experimental paradigm is as in Figure 7F. pK_a's for pH_e regulation were studied using whole-cell recording at pH_i 8.4 or 6.8. (D–E) Effects of MTSET on double mutants S532C/K607N (D) and L530C/K607N (E) at pH_i 8.4 (○) or 6.8 (•). Experimental paradigm is as in Figure 6. (F) Percent inhibition by MTSET (after 3 min) of cysteine-substituted and K607N double mutants (L530C/K607N, S532C/K607N, E535C/K607N, and F537N/K607N) at pH_i 8.4 (white bar) or 6.8 (black bar). Experimental paradigm is as in Figure 6. ^* Indicates P<0.05 by unpaired t-test, double mutants versus single cysteine-substituted mutants at pH_i 6.8 taken from Figure 6G (L530C, S532C, E535C, and F537C, respectively).

S532C and E535C were not accessible to MTSET at pH_i 8.4, but accessible at pH_i 6.8. In contrast, L530C and F537C were accessible at pH_i 8.4, but not accessible at pH_i 6.8 (see Figure 6). These changes in accessibility are most consistent with the interpretation that intracellular acidification causes a clockwise rotation of the pore helix. We found that double mutants S532C/K607N (Figure 9D and F) and E535C/K607N (Figure 9F) were less sensitive to MTSET at pH_i 6.8 than single-cysteine mutants S532C and E535C, respectively (see Figure 6F and G). In contrast, double mutants L530C/K607N (Figure 9E and F) and F537C/K607N (Figure 9F) were more sensitive at pH_i 6.8 than single mutant L503C and F537C, respectively. Thus, K607 contributes to the rotation of pore helix associated with gating by internal protons. As controls, the effects of MTSET on double mutants at pH_i 8.4 (Figure 9D–F) were not significantly different from that on single mutants (Figure 6G). Also, the pK_a for pH_i regulation for the double mutants (6.21±0.15, 6.09±0.13, 6.26±0.17, and 6.11±0.14 for L530C/K607, S532C/K607N, E535C/K607N, and F537C/K607N, respectively; n=3–4 for each) were not significantly different from that for K607N mutant (Figure 9B), confirming that cysteine substitution does not affect pH_i gating. These results support the conclusion that binding of protons to lysine-607 leads to rotation of the pore helix. Mutation of lysine-607 blunts the rotation caused by internal protons.

Discussion

Using the substituted cysteine accessibility method (SCAM), we report in this study that intracellular acidification causes conformational change(s) of the pore helix of TRPV5. One of the changes likely involves clockwise rotation of the pore helix along its long axis. This conclusion is based on the following assumptions. First, only cysteine residues at a water-accessible surface of the protein will react with MTSET. Second, reaction with MTSET will inhibit TRPV5. These two assumptions are likely valid in our present study given that MTSET is positively charged and membrane-impermeant (Holmgren et al, 1996) and that MTSET is large in size and the pore helix is in close proximity and critical for ion conduction (Doyle et al, 1998). Finally, intracellular acidification could, in theory, alter the local electrostatic environment of the cysteine residues and therefore affect their reactivity to MTSET without causing rotation of the pore helix (Wilson et al, 2000). However, the localization of MTSET-reactive and nonreactive residues to two separate, opposite faces of a predicted α-helix and the pattern of reactivity at pH_i 8.4 versus 6.8 (see Figure 6H) provide compelling evidence for a clockwise rotation of the pore helix caused by intracellular acidification. Other types of movements/conformational changes of the pore helix such as translation and/or tilting may also occur in intracellular acidification.

The pore helix of ion channels is critical for ion conduction. Aromatic amino acids of the pore helices of KcsA make contacts with aromatic amino acids in the selectivity filter to hold the filter open (Doyle et al, 1998). Moreover, pore helices point their carboxyl-termini to a water-filled cavity at the membrane center. The partial negative charge of the oriented pore helices provides a favorable electrostatic force to stabilize K⁺ ions at the membrane center (Roux and MacKinnon, 1999). In the crystal structure of another bacterial K⁺ channel KirBac1.1 in the closed state, pore helices no longer point directly at the center of the channel cavity (Kuo et al, 2003). It is believed that this misalignment of pore helices contributes to the closed state. For TRPV5, rotation with or without translation and/or tilting of the pore helix caused by intracellular acidification may cause a similar misalignment of the helix dipoles and close the channel. Rotation of the pore helix can also alter conformation of the selectivity filter.

Binding of extracellular protons to glutamate-522 in the pre-pore helix region of TRPV5 causes conformational changes and channel closure (Yeh et al, 2003). We find that pH_i modulates pK_a for pH_e regulation and vice versa. Mutation of glutamate-522 blunts the modulation by pH_i of pH_e regulation. Together, these results indicate that extra- and pH_i crossregulate the same activation gate. Where is this activation gate? The crystal structure of KcsA suggests two potential activation gates: an outer gate formed by the selectivity filter and pore helix and an inner gate formed by bundle crossing of four inner helices near the intracellular aspect of the membrane (Doyle et al, 1998; Flynn et al, 2001). Gating of voltage-dependent K⁺ channels is associated with structural rearrangements of the inner helices and changes in their accessibility to intracellular Ag⁺ (del Camino and Yellen, 2001), suggesting that bundle crossing is the activation gate in response to changes in TM voltage. Bundle crossing also serves as activation gate for intracellular signals. Phillips et al (2003) showed that gating of Kir6.2 channels by intracellular ATP involves movement of inner helices at the bundle crossing. Based on the crystal structure, Jiang et al (2002) proposed that opening of the MthK channel by intracellular Ca²⁺ occurs via bending of the inner helices at a hinge located deep in the membrane to splay open the bundle.

In contrast, studies of CNG channels revealed that, though the bundle crossing undergoes conformational changes associated with gating by intracellular cyclic nucleotide, it does not form a physical barrier for permeating ions in the closed state (Flynn and Zagotta, 2001). The pore helix and selectivity filter of the CNG channels both undergo conformational changes in association with gating by intracellular cyclic nucleotide (Becchetti et al, 1999; Liu and Siegelbaum, 2000). Gating-associated movement of pore helix has also been found for other channels including voltage-dependent K⁺ channels (Yang et al, 1997) and KcsA (Perozo et al, 1999). Whether the fifth and/or sixth TM helices of TRPV5 form an inner gate and, if so, whether they serve as an activation gate for internal protons is not known. Our results, nevertheless, suggest that the pore helix and selectivity filter of TRPV5 is at least one of the activation gates for intracellular protons (see the model in Figure 10). Gating by intracellular acidification occurs, at least partly, at the selectivity filter via rotation of the pore helix.

Figure 10.

Model for crossregulation of TRPV5 by pH_i and pH_e. Conformational changes caused by intra- and extracellular acidification are illustrated. See text for details.

How does the pore helix senses intracellular protons? The inner helices of KcsA make direct contacts with pore helices (Doyle et al, 1998). It is believed that rotation of the pore helix of CNG channel in response to cGMP binding to the intracellular binding site occurs indirectly through conformational changes of the sixth TM inner helix (Flynn and Zagotta, 2001; Flynn et al, 2001). We propose that binding of protons to intracellular proton sensor(s) such as lysine-607 causes conformational changes of the sixth (and/or the fifth) TM helix (model in Figure 10). Conformational changes of the sixth TM helix then cause rotation of the pore helix of TRPV5 similarly to that in the CNG channel. Extracellular protons cause conformational changes of the channel by binding to a pH sensor (glutamate-522) located immediately upstream of the pore helix (Yeh et al, 2003). We suggest that both extra- and intracellular protons cause pore helices to rotate in the clockwise direction, leading to closing (narrowing) of the selectivity filter gate. Rotation of the pore helices by intracellular acidification will therefore facilitate closing of the channel by extracellular acidification and vice versa. This model of gating by protons at the level of pore helix and selectivity filter is supported by the results that protons decrease selectivity for Ca²⁺, pore diameter, and the single-channel conductance, as well as reduction in open probability.

TRPV5 is the apical entry pathway for transepithelial Ca²⁺ transport in many tissues, including kidney, intestine, placenta, and prostate (Hoenderop et al, 2002). Overproduction of acid, as in metabolic acidosis or high dietary animal protein intake, inhibits renal transepithelial Ca²⁺ transport, causing increased urinary Ca²⁺ excretion and kidney stone diseases (Sutton et al, 1979; Huang, 2004). Most conditions of acid overproduction leading to kidney stone diseases cause both extra- and intracellular acidosis (such as high dietary animal protein intake) (Amanzadeh et al, 2003; Nicoletta and Schwartz, 2004). We report in this study that internal and external protons crossinhibit TRPV5-mediated Ca²⁺ entry. We suggest that inhibition of TRPV5-mediated renal Ca²⁺ transport by internal and/or external protons contribute to the pathogenesis of disturbances of Ca²⁺ transport in these diseases. Moreover, crossregulation by internal and external protons will allow the effects of extracellular acidosis on TRPV5 be modulated by the pH_i and vice versa. For example, extracellular acidification from pH_e 7.4 to 6.4 inhibits TRPV5 by 18% at pH_i 7.6, but by 24% at pH_i 6.8 (see Figure 3A). Thus, the combined effects of extra- and intracellular acidosis on calcium transport may be amplified or blunted depending on existing pH_e and pH_i.

Gating of ion channels in response to external and internal signals is a crucial cell function. How signals regulating ion channels are integrated across the cell membrane is poorly understood. The pore helix of ion channels is pivotal for ion conduction through the selectivity filter and accessible to both external and internal signals (Flynn et al, 2001). Thus, conformational changes of the pore helix may be a common mechanism for integrating extra- and intracellular signals in the regulation of ion channels.

Materials and methods

Molecular biology and cell culture

Mutant rabbit TRPV5's in pCDNA3 mammalian expression vector were constructed as described (Yeh et al, 2003). Chinese hamster ovary (CHO) cells (at ∼50% confluence) were cotransfected with cDNA for pEGFP (0.5 μg) plus cDNAs for WT or mutant TRPV5 (0.5–5 μg) using lipofectamine-plus transfection kits (Gibco) as described (Yeh et al, 2003). About 24–48 h after transfection, cells were dissociated by limited trypsin treatment and placed in a chamber for recording. Transfected cells were identified using epifluorescent microscopy for recordings.

Electrophysiological recordings

Inside-out and whole-cell recordings were performed using an Axopatch 200B patch-clamp amplifier (Axon Instruments) (Yeh et al, 2003). For whole-cell recordings, the pipette and bath solution contained (in mM) 140 Na-Asp (sodium aspartate), 10 NaCl, 10 EDTA, 10 HEPES (titrated to pH as specified) and 140 Na-Asp, 10 NaCl, 1 EDTA, and 10 HEPES (pH as specified), respectively. For inside-out recordings, membrane patches were excised into a bath solution containing (in mM) 140 NaAsp, 10 NaCl, 10 EDTA, and 10 HEPES (pH as specified).

To determine the relative permeabilities of MA⁺, DMA⁺, and NMDG⁺ to Na⁺ in whole-cell recordings, the initial bath and pipette solution contain (in mM) 130 Na-Asp, 10 NaCl, 1 EDTA, 10 HEPES (pH as specified) and 130 Na-Asp, 10 NaCl, 10 EDTA, 10 HEPES (pH as specified), respectively. Na-Asp-containing bath solution was then replaced by one that contains 130 X-Asp, 10 X-Cl, 10 HEPES, and 1 EDTA (X is MA, DMA, or NMDG). X-Asp solutions were prepared by mixing methylamine (Fluka Chemika), dimethylamine (Aldrich), or NMDG (Sigma) (preweighted to a final concentration of 140 mM) with aspartic acid (preweighted to a final concentration 130 mM) and HCl (preweighted to a final concentration 10 mM) and adjusted pH using MA, DMA, or NMDG, respectively. For determination of the relative permeabilities of K⁺, Li⁺, and Cs⁺ to Na⁺, K-Asp, Li-Asp, and Cs-Asp solutions were prepared similarly.

To test the accessibility of substituted cysteines, MTSET (Toronto Research Chemicals) was added to the bath solution from 100 mM stocks. Stock solutions were prepared daily and kept at 4°C until use. Voltage protocols for each experiment were described in the individual figure. Currents were low-pass filtered at 1 kHz using an eight-pole Bessel filter, sampled every 0.1 ms (10 kHz) with Digidata-1300 interface and stored directly onto a computer hard disk using pCLAMP9 software. Data were transferred to compact discs for long-term storage.

Data analysis

To analyze the sensitivity of the channel to inhibition by extracellular or intracellular protons, relative currents at different pH values were fitted with modified Hill equation using the Sigma-Plot program (Yeh et al, 2003). To calculate the apparent second-order rate constant for inhibition of channels by MTSET, time constant was obtained by fitting the time course of inhibition of channels by MTSET with a single exponential. Rate constant was calculated by dividing the reciprocal of the time constant by concentration of reagent (Yeh et al, 2003). The permeability ratios of different monovalent cations versus Na⁺ were calculated based on the shifts of reversal potential (ΔE_rev) on exchanging the extracellular solutions, P_X/P_Na=exp(ΔE_rev × F/RT), where R, T, and F have the usual meaning (Hille, 2001). MA⁺, DMA⁺, and NMDG⁺ (pK_a 9.25, 10.62, and 9.5, respectively) dissociate significantly at pH 9.0. Thus, concentrations of ionized MA⁺, DMA⁺, and NMDG⁺ at pH 9.0 were calculated using the respective pK_a. P_X/P_Na was calculated as exp(ΔE_rev × F/RT) × ([Na⁺]/[ionized X⁺]). The permeability ratio of Ca²⁺ versus Na⁺ was calculated according to the equation, P_Ca/P_Na=([Na⁺]_i/4[Ca²⁺]_o) exp(ΔE_revF/RT){(1+exp(ΔE_rev F/RT)}. The activity coefficients for Na⁺ and Ca²⁺ were assumed 0.75 and 0.33, respectively. All potentials were corrected for liquid junction potentials. Liquid junction potentials were calculated and verified experimentally (Neher, 1992; Barry, 1994). The diameter of ion permeation pore was estimated by fitting P_X/P_Na versus diameter of permeating ions using a modified excluded volume equation: (P_X/P_Na)^1/2=dP−dX/dP−dNa, where dP, dX, and dNa are diameters of the pore, cation X⁺, and Na⁺, respectively (Sabovcik et al, 1995). Data are shown as mean±s.e.m. of the number of observations. Statistical comparison was made using unpaired Student's t-test.

Supplementary Material

Supplementary data