Structure-function analyses of human TRPV6 ancestral and derived haplotypes

SUMMARY

TRPV6 is a Ca²⁺ selective channel that mediates calcium uptake in the gut and contributes to the development and progression of human cancers. TRPV6 is represented by the ancestral and derived haplotypes that differ by three non-synonymous polymorphisms, located in the N-terminal ankyrin repeat domain (C157R), S1-S2 extracellular loop (M378V) and C-terminus (M681T). The ancestral and derived haplotypes were proposed to serve as genomic factors causing a different outcome for cancer patients of African ancestry. We solved cryo-EM structures of ancestral and derived TRPV6 in the open and calmodulin-bound inactivated states. Neither state shows substantial structural differences caused by the non-synonymous polymorphisms. Functional properties assessed by electrophysiological recordings and Ca²⁺ uptake measurements, and water and ion permeation evaluated by molecular modeling also appear similar between the haplotypes. Therefore, ancestral and derived TRPV6 have similar structure and function, implying that other factors are responsible for the differences in susceptibility to cancer.

Graphical Abstract

In brief (eTOC)

Neuberger et al. determined structures of ancestral and derived haplotypes of human calcium-selective channel TRPV6 in the open and inactivated states. Non-synonymous polymorphisms cause no structural differences between the states, changes in TRPV6-mediated currents, calcium uptake, or protein dynamics. They concluded that TRPV6 haplotypes have a similar structure and function.

Introduction

TRPV6, a representative of the vanilloid subfamily of transient receptor potential (TRP) channels, is a master regulator of organismal calcium homeostasis playing the key role in epithelial calcium transport of the intestine, placenta and other organs.^1–4 TRPV6 mutations and abnormal expression^5–9 have been linked to a range of human diseases associated with disturbed calcium homeostasis, including transient neonatal hyperparathyroidism, undermineralization and dysplasia of the human skeleton, hypercalciuria, chronic pancreatitis, various reproductive diseases, Pendred syndrome and Crohn’s-like disease.^5,10–21 Since calcium uptake is linked to cell proliferation and cancer progression, TRPV6 is also declared an oncochannel.^22–24 Indeed, TRPV6 was found to overexpress in some of the most severe human cancers, including leukemia, breast, prostate, colon, ovarian, thyroid and endometrial cancers,^6–8,22,25 and its overexpression correlates with the aggressiveness of the disease.^26,27

TRPV6 in humans is represented by a haplotype of the ancestral and derived variants, with the ancestral allele being minor and characterized by three nonsynonymous SNPs (C157R, M378V, M681T), which correspond to different regions in the protein, including the intracellular loop between the ankyrin repeats 3 and 4, the extracellular loop connecting transmembrane helices 1 and 2 and a structurally disordered region of the intracellular C-terminus between its two sections that bind calmodulin (CaM) (Fig. 1A–B). The ancestral variant of TRPV6 is more common in people of African descent (50–60% versus 1–11% in all other populations) and as such has emerged as a driver of higher incidence, higher mortality, and more aggressive forms of cancer, including breast, prostate, colon, stomach, and cervix cancer, in people of African descent.²⁸ Identification of molecular features distinct between the ancestral and derived TRPV6 might therefore help to devise patient-specific cancer treatment strategies and development of personalized medicine.

Figure 1. Functional assessment of hTRPV6_Anc and hTRPV6_Der.

A, Topology of hTRPV6 subunit, with three polymorphisms distinguishing the ancestral and derived haplotypes (C157R, M378V, M681T) labelled by red dots. B, Sequence alignment for ancestral and derived hTRPV6, with secondary structure elements and polymorphisms indicated. C, Mean current-voltage relationships for TRPV6_der and TRPV6_anc in response to voltage ramp applications. Maximum currents were used for the analysis. n is the number of measured cells. D-E, Dot-plots showing current amplitudes (mean ± SEM) obtained at −80 mV (D) and the ratio of current amplitudes (mean ± SEM) at +80 mV and −80 mV (E) calculated from measurements in C for TRPV6_der (n = 14) and TRPV6_anc (n = 11). The p values are shown for t-test with Welch’s correction. F, Representative measurements of intracellular Ca²⁺ concentrations ([Ca²⁺]_i) in HEK 293 cells expressing TRPV6_der, TRPV6_anc and TRPV6_der-D542A in response to external application of 5 mM Ca²⁺ in the absence and in the presence of 10 μM cis-22a. G, Representative measurements of [Ca²⁺]_i from HEK 293 cells expressing TRPV6_der and TRPV6_anc in response to increasing concentrations (0.2–5 mM) of extracellular [Ca²⁺]_o. H, Calcium concentration-dependences for TRPV6_der and TRPV6_anc calculated from experiments illustrated in G. Data are shown for the maximal [Ca²⁺]_i (mean ± SEM). n is the number of independent experiments. I-J Concentration-dependences for inhibition of TRPV6_der and TRPV6_anc by Cis-22a (I) and Gd³⁺ (J). Curves through the points (mean ± SEM) are the logistic equation 2 fits; n is the number of independent measurements.

While recent structural studies have revealed the TRPV6 molecular architecture^29,30 and the mechanisms of gating,³¹ inactivation by calcium-binding protein calmodulin (CaM)³² and inhibition by various natural and synthetic small molecules,^4,33–37 the structural basis of clinically relevant mutations and disease-prone genetic variants of this important ion channel remains largely unknown. In this study, we embark on comparative analysis of the structure and function of the ancestral and derived variants of human TRPV6 using a combination of single-particle cryo-electron microscopy (cryo-EM), electrophysiology, calcium imaging, molecular dynamics (MD) simulations and detailed mapping of the pore domain properties.

Results

Functional characterization of hTRPV6_Anc-Open and hTRPV6_Der-Open

We assessed the function of the two polymorphic channels that represent the ancestral (hTRPV6_Anc) and derived (hTRPV6_Der) variants of human TRPV6 using whole-cell patch-clamp electrophysiology and Ca²⁺ uptake measurements. Currents recorded from HEK293 cells expressing the ancestral or derived haplotypes in response to voltage ramps showed inward rectification, typical for TRPV6 (Fig. 1C).^38–40 The average current amplitude was modestly higher for hTRPV6_Anc compared to hTRPV6_Der (Fig. 1D), while the rectification index (the ratio of the current amplitudes at +80 mV and −80 mV) was indistinguishable (Fig. 1E). These results are in line with the previous electrophysiological examination of the derived and ancestral TRPV6 variants.⁴¹

We also measured Ca²⁺ influx mediated by hTRPV6_Anc and hTRPV6_Der expressed in HEK293 cells and found no remarkable changes (Fig. 1F–H). Importantly, TRPV6-mediated Ca²⁺ uptake was completely eliminated by the D542A mutation (Fig. 1F), which is known to abolish Ca²⁺ permeability of TRPV6,⁴² as well as by the TRPV6 inhibitor cis-22a³⁷ (Fig. 1F). For both hTRPV6_Anc and hTRPV6_Der, Ca²⁺ uptake was concentration-dependent (Fig. 1G). Logistic equation fitting of the maximal elevation of [Ca²⁺]_i in response to increased extracellular concentrations of Ca²⁺ ([Ca²⁺]_o) revealed the EC₅₀ of 1.32 ± 0.37 mM (n = 4) for hTRPV6_Anc and 1.36 ± 0.35 mM (n = 4) for hTRPV6_Der (Fig. 1H), supporting the idea that both TRPV6 variants display similar capability to uptake Ca²⁺.

We then tested the ability of the potent TRPV6 inhibitor Cis-22a³⁷ to suppress Ca²⁺ uptake through hTRPV6_Anc and hTRPV6_Der. Fitting of the concentration dependence of TRPV6 inhibition by Cis-22a (Fig. 1I) revealed similar potency for hTRPV6_Anc (IC₅₀ = 0.126 ± 0.016 μM, n_Hill = 0.81 ± 0.16, n = 3) and hTRPV6_Der (IC₅₀ = 0.127 ± 0.017 μM, n_Hill = 0.85 ± 0.18, n = 3). Likewise, the ion channel blocker Gd³⁺, which acts by binding to the main Ca²⁺ binding site formed by the side chains of D542,^29,30 showed similar potency when acting at hTRPV6_Anc (IC₅₀ = 10.7 ± 1.0 μM, n_Hill = 1.30 ± 0.13, n = 4) and hTRPV6_Der (IC₅₀ = 8.7 ± 1.1 μM, n_Hill = 0.90 ± 0.08, n = 4) (Fig. 1J). Hence, apart from the somewhat larger currents mediated by hTRPV6_Anc compared to hTRPV6_Der, both channels are very similar in terms of current rectification behavior, ability to uptake external Ca²⁺ and sensitivity to channel’s blockers, cis-22a and Gd³⁺.

Cryo-EM analysis of ancestral and derived human TRPV6

Given the similarity of functional characteristics of hTRPV6_Anc compared to hTRPV6_Der, we wondered if their different predicted role in cancer development can be explained by distinct structural features. Purified full-length ancestral and derived human TRPV6 proteins were subjected to single-particle cryo-EM analysis. For both haplotypes, cryo-EM data processing revealed two distinct populations of particles (Figs. S1–3, Table 1). The first ones yielded 4-fold symmetrical reconstructions (C4 symmetry), typical for the previously published structures of hTRPV6 in the apo open state (Fig. 2A,C). Particles in the second groups were not strictly symmetrical (C1 symmetry) due to the presence of an additional density located asymmetrically with respect to hTRPV6 (Fig. 2B,D). This density represents an endogenous calmodulin (CaM), apparently carried throughout protein purification with four Ca²⁺ ions bound (Fig. S4). The four resolved cryo-EM reconstructions also revealed densities for annular lipids surrounding the transmembrane region of TRPV6 as well as calcium ions bound to the ion channel selectivity filter (Fig. 2).

Table 1.

Cryo-EM data collection, refinement and validation statistics

Structure EMDB accession code PDB accession code	hTRPV6Anc-Open EMD-459339CUH	hTRPV6Anc-Inact EMD-459349CUI	hTRPV6Der-Open EMD-459359CUJ	hTRPV6Der-Inact EMD-459369CUK
Data collection / processing
Magnification	130,000x	130,000x	130,000x	130,000x
Voltage (kV)	300	300	300	300
Electron exposure (e−Å−2)	55	55	58	58
Defocus range (μm)	−0.8 to −2.5	−0.8 to −2.5	−0.8 to −2.0	−0.8 to −2.0
Reported pixel size (Å)	1.067	1.067	0.415	0.415
Exposures (no.)	18,006	18,006	12,516	12,516
Processing software
Motion correction	cryoSPARC v3.3	cryoSPARC v3.3	cryoSPARC v3.3	cryoSPARC v3.3
CTF estimation	cryoSPARC v3.3	cryoSPARC v3.3	cryoSPARC v3.3	cryoSPARC v3.3
Platform software for particle picking	cryoSPARC v3.3	cryoSPARC v3.3	cryoSPARC v3.3	cryoSPARC v3.3
Software for 2D/3D class. & Refinements	cryoSPARC v3.3	cryoSPARC v3.3	cryoSPARC v3.3	cryoSPARC v3.3
Symmetry imposed	C4	C1	C4	C1
Initial particle images (no.)	14,833,688	14,833,688	9,029,804	9,029,804
Final particle images (no.)	517,835	232,058	554,157	72,482
Map resolution (Å)	3.03	3.42	2.78	3.26
FSC 0.143
Refinement
Initial models used (PDB code)	7S89	6E2F	7S89	6E2F
Model resolution (Å)	3.03	3.42	2.78	3.26
FSC threshold	0.143	0.143	0.143	0.143
Map sharpening B factor (Å2)	−178.1	−151.1	−136.2	−93.2
Model composition
Non-hydrogen atoms	20,877	22713	21557	22531
Protein residues	2,376	2650	2440	2635
Ligands	49	62	65	62
Water	72	0	72	1
B factors (Å2)
Protein	61.00	73.94	74.30	80.32
Ligands	65.25	34.85	36.48	52.53
Water	45.08	n.a.	63.89	15.51
R.m.s. deviations
Bond lengths (A)	0.012	0.007	0.008	0.009
Bond angles (°)	1.502	1.387	1.408	1.406
Validation
MolProbity score	1.84	1.98	1.85	1.89
Clash score, all atoms	6.70	6.91	7.63	5.98
Poor rotamers (%)	1.16	0.39	1.51	0.44
Ramachandran plot
Favored (%)	93.52	88.17	95.72	89.67
Allowed (%)	6.31	11.30	4.28	10.03
Disallowed (%)	0.17	0.53	0.00	0.30

Figure 2. Cryo-EM characterization of ancestral and derived human TRPV6.

A-D, 3D cryo-EM density for hTRPV6_Anc-Open (A), hTRPV6_Anc-Inact (B), hTRPV6_Der-Open (C), and hTRPV6_Der-Inact (D), viewed intracellularly (left) and parallel to membrane (right) but cut off along the dashed lines shown in intracellular views. TRPV6 subunits are colored yellow, green, pink, and cyan, CaM dark green, lipids purple, and Ca²⁺ ions red. See also Figures S1–S4.

Structures in the open state

Building a molecular model for the 3D reconstruction of ancestral TRPV6 in the absence of CaM resulted in the hTRPV6_Anc-Open structure with the same general architecture as structures of derived TRPV6 solved previously^{29–31,33–37} (Fig. 3A–C). The hTRPV6_Anc-Open channel is assembled of four subunits with the central ion conducting pore in the middle of the transmembrane domain (TMD). The channel also includes an intracellular ‘skirt’ constructed from ankyrin repeat domains, linked by three-stranded β-sheets, N-terminal helices, and C-terminal hooks. The amphipathic TRP helices, a characteristic feature of the TRP channel family, connect the TMD to the C-terminal hook, running nearly parallel to the inner leaflet of the membrane and interacting with both the TMD and the skirt. The TMD consists of six transmembrane helices (S1–S6) and a re-entrant pore loop (P-loop) between S5 and S6. The first four transmembrane helices form the S1–S4 domain, which resembles the voltage-sensing domain in voltage-gated ion channels.⁴³ S5, P-loop, and S6 contribute to the pore domain, which leans against the S1–S4 domain of the adjacent subunit in a domain-swapped manner.³⁰

Figure 3. Structures of ancestral and derived human TRPV6 in the open state.

A-C, Structure of hTRPV6_Anc-Open viewed parallel to membrane (A), intracellularly (B) and extracellularly (C), with channel subunits colored yellow, light green, light pink, and cyan, residues R157 and V378 shown as space-filling models (dark blue) and Ca²⁺ ion in the selectivity filter as a red sphere. D, Superposition of hTRPV6_Anc-Open (green) and hTRPV6_Der-Open (pink), with side chains for residues at positions 157, 378 and 542 shown as sticks and Ca²⁺ ion in the selectivity filter as a red sphere. Only two of four subunits are shown, with the front and back subunits omitted for clarity. Boxes indicate regions expanded in E-H. E-H, Closeup view of the regions encompassing residues at positions 157 (E-F) and 378 (G-H) that are different in hTRPV6_Anc-Open (E,G) and hTRPV6_Der-Open (F,H). Cryo-EM density is shown as a blue mesh. See also Figures S5–S6.

For comparison, we also solved the structure of derived TRPV6 in the apo open state, hTRPV6_Der-Open. Superposition of hTRPV6_Anc-Open and hTRPV6_Der-Open made it obvious that the two structures are nearly identical (Fig. 3D). We took a close look at the residues that represent non-synonymous polymorphisms. Among the three of them, the first one represents a residue at the interface between the N-terminal helix and the loop connecting the third and fourth ankyrin repeat domains, R157 in hTRPV6_Anc-Open or C157 in hTRPV6_Der-Open, clearly defined in cryo-EM maps (Fig. 3E–F). Despite an apparent cation-π interaction between R157 and W29 in hTRPV6_Anc-Open (Fig. 3E), which is missing in hTRPV6_Der-Open (Fig. 3F), we observed no obvious change in the local (Fig. 3E–F) or global (Fig. 3D) fold of the TRPV6 protein.

Residues representing the second non-synonymous polymorphism are located in the S1-S2 extracellular loop, V378 in hTRPV6_Anc-Open or M378 in hTRPV6_Der-Open, also clearly defined in cryo-EM maps (Fig. 3G–H). Neither one of them appears to form specific interactions with other parts of the protein, nor alters the local (Fig. 3G–H) or global (Fig. 3D) fold of TRPV6. As residues C-terminal to Q637 are not supported by cryo-EM density in the hTRPV6_Anc-Open and hTRPV6_Der-Open structures, residues representing the third non-synonymous polymorphism, T681 in hTRPV6_Anc-Open or M681 in hTRPV6_Der-Open, are not resolved by our structures and presumably belong to the unfolded region of the C-terminus. Independent of their exact position, these residues do not appear to cause any significant change in the overall structure of the TRPV6 protein (Fig. 3D).

Open pore in hTRPV6_Anc-Open and hTRPV6_Der-Open

To see if any functional difference between ancestral and derived TRPV6 originates from structural characteristics of the ion-conducting pathway, we compared pores in hTRPV6_Anc-Open and hTRPV6_Der-Open (Fig. 4). The selectivity filter conformation does not show any difference (Fig. 4A–B) and appears to be the same as in all previously published TRPV6 structures,^29–37 emphasizing the conservative role of this region in cation selectivity and permeation.^29,30,44,45 The intracellular region of the pore is lined by S6 and exhibits dynamic behavior during TRPV6 activation and inactivation gating.^31,32 Nevertheless, this region in both hTRPV6_Anc-Open and hTRPV6_Der-Open looks similar, is characterized by the π-bulge in the middle of S6 and I575 making the pore narrow constriction (Fig. 4A–B), typical for the open or inactivated states of TRPV6.^31,32 In fact, measurements of the pore radius (Fig. 4C) unambiguously characterize the state of the pore in hTRPV6_Anc-Open and hTRPV6_Der-Open as open and essentially indistinguishable between these two structures.

Figure 4. Open pore of hTRPV6_Anc-Open and hTRPV6_Der-Open.

A-B, Pore-forming domain in hTRPV6_Anc-Open (A) and hTRPV6_Der-Open (B), with the residues contributing to pore lining shown as sticks. Only two of four subunits are shown, with the front and back subunits omitted for clarity. The pore profile is shown as a space-filling model (gray). The region that undergoes the α-to-π transition in S6 is highlighted in purple. C, Pore radius for hTRPV6_Anc-Open (green) and hTRPV6_Der-Open (pink) calculated using HOLE.⁵⁷ The vertical dashed line denotes the radius of a water molecule, 1.4 Å. D-E, Distribution of water (blue mesh) and Ca²⁺ ions (red density) averaged over MD simulations of hTRPV6_Anc-Open (D) and hTRPV6_Der-Open (E). F-G, Ca²⁺ ion (F) and water (G) density averaged over MD simulations of hTRPV6_Anc-Open (green) and hTRPV6_Der-Open (pink) measured along the channel pore. Thick lines represent the average values, while thin lines represent the mean ± σ. Ion binding sites R, P1-P4 are labeled in the panels. H, Ca²⁺ ion cumulative permeation through the pore at the membrane potential of −400 mV. Thin lines represent the number of ion permeation events along the three independent MD simulations of hTRPV6_Anc-Open (green) and hTRPV6_Der-Open (pink). Thick lines show the average Ca²⁺ permeation rates and labeled (mean ± σ). I, Distributions of Ca²⁺ conductance at −400 mV along the MD trajectories represented as boxplots. Green and pink circles represent conductance values calculated over 100 ns segments of MD trajectories; orange rhombs show the mean conductance ± SEM. The P value is for the t-test with Welch’s correction. See also Figures S5–S6.

To further compare pores in hTRPV6_Anc-Open and hTRPV6_Der-Open, we made projections of their landscapes and molecular hydrophobicity potential (MHP)⁴⁶ on the surface of a cylinder with the axis matching that of the ion channel pore (Fig. S5). It should be noted that such “dynamic molecular portraits” of the ion channel pore domains are very sensitive to small changes in the distribution of the physico-chemical properties of the surface of the pore walls⁴⁷. Such differences are difficult to detect only by analyzing spatial models of the corresponding protein regions. It was shown that the distribution of hydrophilic and hydrophobic areas appeared to be identical in hTRPV6_Anc-Open and hTRPV6_Der-Open, as well as nearly the same as in the open pore of C-terminally truncated derived hTRPV6 we analyzed previously⁴⁷. This further provides support to the idea that the pores in ancestral and derived TRPV6 have very similar characteristics.

Molecular dynamics simulations of hTRPV6_Anc-Open and hTRPV6_Der-Open

To assess possible differences in structure and permeation properties of ion-conducting pores in hTRPV6_Anc-Open and hTRPV6_Der-Open, we performed molecular dynamics (MD) simulations of these channels incorporated in lipid bilayers and surrounded by 250 mM water solution of CaCl₂ (see Methods). Both the ancestral and derived variants showed continuous distribution of water and ions throughout the entire pore length, strongly supporting their conducting conformations (Fig. 4D–G). Since TRPV6 is highly selective to calcium, we compared distribution of Ca²⁺ ions in the pore during MD simulations of hTRPV6_Anc-Open and hTRPV6_Der-Open (Fig. S6). In both channels Ca²⁺ ions bind to the same set of sites. At the extracellular vestibule, ions form a scattered spatial distribution at the recruitment sites²⁹ formed by negatively charged residues in the S5-P and P-S6 loops, which were rather flexible during MD simulations (labeled R in Fig. 4D–F). There are two sites in the selectivity filter, P1 that is formed by the carboxyl groups of D542, and P2, where Ca²⁺ ion is coordinated by T539. Analogous distribution of Ca²⁺ in the selectivity filter was observed in TRPV6 crystal structures using anomalous difference Fourier maps^29,30 and in MD simulations of TRPV6 conductance^45,48. The third Ca²⁺ binding site in the pore (P3) is located in the middle of the central cavity, near N572, and was also previously observed in TRPV6 crystal structures^29,30 and MD simulation studies of TRPV channels^45,49. The last Ca²⁺ binding site in the ion conducting pathway of TRPV6, P4, is at the intracellular entrance to the pore and formed by the carboxyl groups of D580.

Coming from the extracellular bulk solution where the water coordination number (CNw) is ~7, Ca²⁺ ions loose one to three water molecules at the selectivity filter entrance and then gain water back after passing P1 site on the way to the intracellular lumen. Binding to the P2-P4 sites occurs when Ca²⁺ ions are completely hydrated, by means of forming hydrogen bonds between the water molecules surrounding calcium and the pore-lining residues.

To obtain a more direct insight into the permeation of Ca²⁺ ions through the pores of the ancestral or derived haplotypes, we performed MD simulations of hTRPV6_Anc-Open and hTRPV6_Der-Open at the applied membrane potential of −400 mV. Such a high voltage is not physiological but is necessary to simulate enough permeation events to estimate channel conductance. Both haplotypes of TRPV6 exhibited fluctuating permeation rates along the MD trajectories. However, the resulting average permeation rates were found to be similar: 0.100 ± 0.010 and 0.075 ± 0.006 ions/ns for the ancestral and derived haplotypes, respectively (Fig. 4H). The corresponding Ca²⁺ conductance was also similar, G_Anc = 78 ± 7 pS and G_Der = 60 ± 8 pS (Fig. 4I).

TRPV6 pore represents a nanoscale transmembrane channel with heterogeneous distribution of spatio-chemical properties⁴⁵ (Fig. S5) and ions (Fig. 4D–E), which is necessary to provide effective ionic transport via cooperative “knock-on” or “knock-off” mechanisms determining high calcium selectivity and permeation in physiological conditions.^44,45,48 Based on the nearly identical distribution, coordination and permeation of Ca²⁺ ions through the pores of hTRPV6_Anc-Open and hTRPV6_Der-Open, it was concluded that both channels likely have similar calcium selectivity and conductance.

Structures in the calmodulin-bound inactivated state

The structure, function and dynamics of TRPV6 in the open conducting state showed no substantial differences between the ancestral and derived genetic variants (Figs. 1–5). While this is consistent with the previously reported similarity in biophysical properties of these channels, functionally relevant differences between ancestral and derived TRPV6 may originate from differences in their inactivation.⁴¹ We therefore compared the inactivated state structures of the ancestral and derived variants.

Figure 5. Structures of ancestral and derived human TRPV6 in the inactivated state.

A-C, Structure of hTRPV6_Anc-Inact viewed parallel to membrane (A), intracellularly (B) and parallel to membrane with only two of four subunits are shown and the other two (front and back subunits) omitted for clarity (C). Channel subunits are colored yellow, light green, light pink, and cyan, CaM dark green. Residues R157 and V378 are shown as space-filling models (dark blue) and Ca²⁺ ions as red spheres. D, Superposition of hTRPV6_Anc-Inact (blue) and hTRPV6_Der-Inact (yellow), with side chains for TRPV6 residues at positions 157, 378, 542, and 583 and K115 of CaM shown as sticks. Only two of four subunits are shown, with the front and back subunits omitted for clarity. Boxes indicate regions expanded in E-H. E-H, Closeup view of the regions encompassing residues at positions 157 (E-F) and 378 (G-H) that are different in hTRPV6_Anc-Inact (E,G) and hTRPV6_Der-Inact (F,H). Cryo-EM density is shown as a blue mesh. See also Figures S7.

The 3D reconstruction of ancestral TRPV6 with bound CaM (Fig. 2B) resulted in the inactivated state structure, hTRPV6_Anc-Inact (Fig. 5A–C). Consistent with the previously reported structure of the derived TRPV6 channel,³² CaM binds to hTRPV6_Anc-Inact with the stoichiometry of one Ca²⁺ binding protein per TRPV6 tetramer by docking into a ~50 Å × 50 Å cavity underneath the channel that is enclosed by the intracellular skirt, mainly composed of the ankyrin repeat domains. The N- and C-terminal lobes of CaM adopt a unique head-to-tail arrangement, with each lobe bound to two Ca²⁺ ions and an individual α-helical substrate: the proximal C-terminus of TRPV6 (residues 640–653) bound to the N-terminal lobe of CaM and the distal C-terminus of TRPV6 (residues 687–703) bound to the C-terminal lobe of CaM. Since 3D classes with empty Ca²⁺ binding sites have not been identified during cryo-EM data processing, we hypothesize that the occupancy of all four CaM binding sites (Fig. S4) is necessary for TRPV6 inactivation.

Most distinctly, lysine K115 of CaM sticks its side chain into the intracellular orifice of the ion channel pore formed by four tryptophans W583, one from each TRPV6 subunit at the S6 bundle crossing. These tryptophans trap the side chain of K115 by forming a tight cubic cage of their indole rings around the ε-amino group of K115. The tryptophan cage provides a unique environment for an atypically strong cation-π interaction between the π-system of four tryptophan indole rings and the positively charged ε-amino group of K115, which leads to breakage of four (one per subunit) open state-stabilizing salt bridges between Q473 in the S4-S5 linker and R589 in the TRP helix and narrowing of the intracellular pore due to tightening of the S6 bundle. The importance of W583 for TRPV6 and homologous TRPV5 function was confirmed by mutagenesis.^37,50–52

We also solved the structure of derived TRPV6 in the inactivated state, hTRPV6_Der-Inact, at higher resolution (3.26 Å) than before (3.98 Å).³² Superposition of hTRPV6_Anc-Inact and hTRPV6_Der-Inact clearly illustrates that the two structures are nearly identical (Fig. 5D). Similar to the apo (open) states, the residues that represent non-synonymous polymorphisms in hTRPV6_Anc-Inact and hTRPV6_Der-Inact show well-defined density in cryo-EM maps at the positions 157 (Fig. 5E–F) and 378 (Fig. 5G–H) and are not resolved at position 681 as the corresponding region of the structure appears disordered. Likewise, R157 in hTRPV6_Anc-Inact makes an apparent cation-π interaction with W29 (Fig. 5E), while C157 in hTRPV6_Der-Inact does not (Fig. 5F). Nevertheless, no obvious local (Fig. 5E–F) or global (Fig. 5D) changes in the TRPV6 protein structure are observed due to different residues at position 157. The same is true for position 378 represented by V378 in hTRPV6_Anc-Inact or M378 in hTRPV6_Der-Inact (Fig. 5D,G–H), strongly suggesting that neither one of the non-synonymous polymorphisms causes substantial changes in the inactivated state conformation.

Closed pore in hTRPV6_Anc-Inact and hTRPV6_Der-Inact

Since the ion-conducting pathway represents the main region of TRPV6 involved in Ca²⁺-induced inactivation by CaM (Fig. 5A–C), we also compared the pores in hTRPV6_Anc-Inact and hTRPV6_Der-Inact (Fig. 6). Similar to the open apo state structures, neither the selectivity filter nor the gate is different between the ancestral and derived TRPV6 (Fig. 6A–B). In both structures, there is a π-bulge in the middle of S6, typical for the inactivated or PCHPD-bound structures.^32,37 Needless to say, the pore radius in hTRPV6_Anc-Inact and hTRPV6_Der-Inact is nearly identical (Fig. 6C). While the pore radius of the selectivity filter in these inactivated-state structures is very similar to the open-state structures, emphasizing the dedicated role of this region in ion selectivity and permeation, the gate in hTRPV6_Anc-Inact and hTRPV6_Der-Inact is closed for ion or water permeation by the side chains of I575. The I575 side chains create a highly hydrophobic seal of the pore that is clearly seen in the MHP distributions for both ion channels (Fig. S7). All other hydrophilic and hydrophobic patterns are also nearly identical between the hTRPV6_Anc-Inact and hTRPV6_Der-Inact pores, further emphasizing the similarity of the inactivated states of the ancestral and derived haplotypes.

Figure 6. Closed pore of hTRPV6_Anc-Inact and hTRPV6_Der-Inact.

A-B, Pore-forming domain in hTRPV6_Anc-Inact (A) and hTRPV6_Der-Inact (B), with the residues contributing to pore lining shown as sticks. Only two of four subunits are shown, with the front and back subunits omitted for clarity. The pore profile is shown as a space-filling model (grey). The region that undergoes the α-to-π transition in S6 is highlighted in pink. C, Pore radius for hTRPV6_Anc-Inact (blue) and hTRPV6_Der-Inact (yellow) in comparison to hTRPV6_Anc-Open (green) and hTRPV6_Der-Open (pink) from Figure 4, calculated using HOLE.⁵⁷ The vertical dashed line denotes the radius of a water molecule, 1.4 Å. See also Figures S7.

Discussion

In the course of human evolution, the ancestral haplotype of TRPV6 has been largely replaced with a haplotype characterized by three non-synonymous SNPs (C157R, M378V, M681T) that are nearly fixed for the derived alleles in non-African populations.⁵³ Such selection likely occurred based on preexisting mutations in the ancestral African population, which subsequently became advantageous in a new environment and rose to high frequency.⁵⁴ It is well documented that the end of the last glacial period had triggered substantial migration events from Sub-Saharan Africa⁵⁵ which came along with the development of agricultural techniques and animal farming. The latter led to a drastic increase in milk production, which is associated with increased uptake of calcium. Intriguingly, the ancestral haplotype of the calcium-selective channel TRPV6 (R157, V378, T681) has been found to be more common in people of African ancestry (60% versus 1% to 11% in all other populations), putatively leading to higher incidence, higher mortality, and more aggressive forms of breast, prostate, colon, stomach, and cervix cancers among these populations.²⁸

Earlier studies have already questioned the functional link between the ancestral haplotype and cancer incidence. Thus, twice higher uptake of ⁴⁵Ca²⁺ has been reported for the ancestral compared to derived TRPV6 channels recombinantly expressed in Xenopus oocytes.⁵⁶ The corresponding increase in ⁴⁵Ca²⁺ influx was recapitulated by the single SNP M378V, while the other two SNPs (C157R/M681T) did not make an apparent contribution.⁵⁶ At the same time, another group performed patch-clamp experiments with the ancestral and derived TRPV6 channel variants and did not find statistically significant differences in biophysical characteristics.⁴¹

To get insights into the role of C157R, M378V and M681T in human TRPV6, we solved cryo-EM structures of ancestral and derived TRPV6 in the open and calmodulin-bound inactivated states. Compared to the closed state^29,30, the open states of the ancestral and derived haplotypes appear similar (Fig. 4A–C) and illustrate the same mechanism of TRPV6 activation³¹. Indeed, in both cases, the closed-to-open state conversion of this constitutively open channel is accompanied by an α-to-π transition in S6, which causes a ~100° rotation of its C-terminal portion and brings a completely different set of residues to face the pore in the gate region. This energetically unfavorable transformation of S6 is stabilized by the electrostatic interactions between Q473 in the S4-S5 linker and R589 in TRP helix as well as D489 in S5 and T581 in S6, which are missing in the closed state. The S6 transformation moves the gate-forming M578 side chains away from pore center and opens the channel for cation and water conductance^31,44. Given the similar pore dimensions (Fig. 4A–C), ion and water densities (Fig. 4D–G) and conductance (Fig. 4H–I), the gating process of activation and ion permeation are very much the same in the ancestral and derived haplotypes.

Likewise, we observed very similar conformations of ancestral and derived TRPV6 in the inactivated state (Fig. 6A–C). TRPV6 inactivation is proposed to occur via transformation of the open state upon binding of CaM³². When CaM binds to the open state, it inserts the K115 side chain into the ion channel intracellular entrance. K115 not only plugs the pore and prevents conductance, but strong cation-π interaction between the positively charged ε-amino group of lysine and the cubic cage of W583 indole rings also pulls the C-terminal S6 portions towards the pore center, creating an additional hydrophobic seal of the pore by I575 side chains. During inactivation, the strong K115-W583 interaction energetically compensates for the loss of the open state-stabilizing interaction between Q473 and R589, while the D489-T581 interaction remains intact. The pore geometry transformation and conformational rearrangements accompanying inactivation of ancestral and derived TRPV6 appear to be similar (Fig. 6A–C). Considering the similarity of the open states (Fig. 4), neither state nor gating function of TRPV6 showed substantial structural differences that can be linked to non-synonymous polymorphisms. Moreover, functional properties assessed by electrophysiological recordings and calcium uptake measurements as well as water and ion permeation evaluated by MD simulations also appeared very similar between the haplotypes. Given the modestly higher average current amplitude for hTRPV6_Anc compared to hTRPV6_Der (Fig. 1D), we conclude that other factors, like regulation of TRPV6 surface expression, rather than structure or function of the ancestral and derived variants are responsible for the differences in susceptibility to cancer.

STAR METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Cell Lines

For structural experiments, expression of ancestral and derived human TRPV6 was performed in HEK 293S cells lacking N-acetyl-glucosaminyltransferase I (GnTI⁻, mycoplasma test negative, ATCC #CRL-3022) that were maintained at 37°C and 6% CO₂ in Freestyle 293 medium (Thermo Fisher Scientific #12338–018) supplemented with 2% FBS. Baculovirus for infecting HEK 293S GnTI^– cells was produced in Sf9 cells (GIBCO) cultured in the Sf-900 III SFM media (GIBCO) at 27°C. For aequorin-based Ca²⁺ influx assay and patch-clamp experiments, TRPV6 channels were expressed in HEK 293T cells (mycoplasma test negative, ATCC #CRL3216) that were maintained at 37°C and 5% CO₂ in DMEM (Merck, #D6429) supplemented with 10% fetal calf serum (Thermo Fisher Scientific, #10270106), 100 μg/ml streptomycin and 100 U/ml penicillin (Merck, #P4333).

METHOD DETAILS

Constructs

Full-length ancestral and derived human TRPV6 used for cryo-EM was cloned into a pEG BacMam vector⁶⁹ with a C-terminal thrombin cleavage site followed by a streptavidin affinity tag (WSHPQFEK). Point mutations in wild-type human TRPV6 were introduced using the standard molecular biology techniques.^33,70

Expression and purification

hTRPV6 ancestral and derived wild-type variants were expressed and purified in identical manner and based on our previously established protocols.⁷¹ Bacmids and baculoviruses were produced using standard procedures.^{33–35,69,71–75} Baculovirus was made in Sf9 cells for ~72 hours (Thermo Fisher Scientific, mycoplasma test negative, GIBCO #12659017) and was added to suspension-adapted HEK 293 cells lacking N-acetyl-glucosaminyltransferase I (GnTI^–, mycoplasma test negative, ATCC #CRL-3022) that were maintained in Freestyle 293 media (Gibco-Life Technologies #12338–018) supplemented with 2% FBS at 37 °C and 5% CO₂. Twenty-four hours after transduction, 10 mM sodium butyrate was added to enhance protein expression, and the temperature was reduced to 30 °C. Seventy-two hours after transduction, cells were harvested by centrifugation at 5,471 × g for 15 min using a Sorvall Evolution RC centrifuge (Thermo Fisher Scientific), washed in phosphate-buffered saline pH 8.0, and pelleted by centrifugation at 3,202 × g for 10 min using an Eppendorf 5810 centrifuge. The cell pellet was solubilized under constant stirring for 2 hours at 4 °C in ice-cold lysis buffer containing 1% (w/v) n-dodecyl β-D-maltoside, 0.1% (w/v) CHS, 20 mM Tris-Cl pH 8.0, 150 mM NaCl, 0.8 μM aprotinin, 4.3 μM leupeptin, 2 μM pepstatin A, 1 mM phenylmethylsulfonyl fluoride, and 1 mM β-mercaptoethanol (βME). The non-solubilized material was pelleted in the Eppendorf 5810 centrifuge at 3,202 × g and 4 °C for 10 min. The supernatant was subjected to ultracentrifugation in a Beckman Coulter ultracentrifuge using a Beckman Coulter Type 45Ti rotor at 186,000 g and 4 °C for 1 hour to further clean up the solubilized protein. The supernatant was added to strep resin and rotated for 14–16 hours at 4 °C. The resin was washed with 10 column volumes of wash buffer containing 20 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM βME, 0.01% (w/v) GDN, and 0.001% (w/v) CHS, and the protein was eluted with the same buffer supplemented with 2.5 mM D-desthiobiotin. The eluted protein was concentrated using a 100 kDa NMWL centrifugal filter (MilliporeSigma Amicon) to 0.5 ml and then centrifuged in a Sorvall MTX 150 Micro-Ultracentrifuge (Thermo Fisher Scientific) using an S100AT4 rotor for 30 min at 66,000 × g and 4 °C before injection into a size-exclusion chromatography (SEC) column. hTRPV6 was further purified using a Superose^™ 6 10/300 GL SEC column attached to an AKTA FPLC (GE Healthcare) and equilibrated in 150 mM NaCl, 20 mM Tris-HCl pH 8.0, 1 mM βME, 0.01% GDN, and 0.001% CHS. The tetrameric peak fractions were pooled and concentrated using 100 kDa NMWL centrifugal filter to 3.5 mg/ml.

hTRPV6 was reconstituted in circularized NW11 nanodiscs (cNW11). cNW1 were prepared according to the standard protocol^75,76 as previously reported and stored at −80°C as ~2–3-mg/ml aliquots in the buffer containing 20 mM Tris pH 8.0 and 150 mM NaCl before usage. For nanodisc reconstitution, the purified protein was mixed with cNW11 nanodiscs and soybean lipids (Soy polar extract, Avanti Polar Lipids) at the molar ratio of 1:3:166 (hTRPV6:cNW11:lipid). The lipids were dissolved in the buffer containing 150 mM NaCl and 20 mM Tris pH 8.0 to reach the concentration of 100 mg/ml, and subjected to 5–10 cycles of freezing in liquid nitrogen and thawing in a water bath sonicator. The nanodisc mixture (500 μl) was rocked at room temperature for 1 hour. Subsequently, 40 mg of Bio-beads SM2 (Bio-Rad), pre-wet in the buffer containing 20 mM Tris pH 8.0 and 150 mM NaCl, were added to the nanodisc mixture, which was then rotated for one hour at 4°C. After adding 40 mg more of Bio-beads SM2, the resulting mixture was rotated at 4°C for another ~14 hours. The Bio-beads SM2 were then removed by pipetting. The sample was then centrifuged in a Sorvall MTX 150 Micro-Ultracentrifuge (Thermo Fisher Scientific) using a S100AT4 rotor for 30 min at 66,000 × g and 4 °C before injecting into a size-exclusion chromatography (SEC) column. Nanodisc-reconstituted hTRPV6 was then purified from empty nanodiscs using SEC with a Superose^™ 6 10/300 GL SEC column equilibrated with the buffer containing 150 mM NaCl, 20 mM Tris pH 8.0, and 1 mM βME. Fractions of nanodisc-reconstituted hTRPV6 were pooled and concentrated to 1.9 mg/ml using a 100-kDa NMWL centrifugal filter.

Cryo-EM sample preparation and data collection

UltrAuFoil R 1.2/1.3, Au 300 grids were used for plunge-freezing. Prior to sample application, grids were plasma treated in a PELCO easiGlow glow discharge cleaning system (0.39 mBar, 15 mA, “glow” 25 s, “hold” 10 s). A Mark IV Vitrobot (Thermo Fisher Scientific) set to 100% humidity at 4 °C was used to plunge-freeze the grids in liquid ethane after applying 3 μl of protein sample in SEC buffer (150 mM NaCl, 20 mM Tris pH 8.0, and 1 mM βME) to their gold-coated side using a blot time of 5 s, a blot force of 5, and a wait time of 15 s. The grids were stored in liquid nitrogen before imaging.

Images of frozen-hydrated particles of hTRPV6_ANC and hTRPV6_DER were collected on a Titan Krios transmission electron microscope (Thermo Fisher Scientific) operating at 300 kV and equipped with a post-column GIF Quantum energy filter and a Gatan K3 Summit direct electron detection camera (Gatan, Pleasanton, CA, USA).

For hTRPV6_ANC, 18,006 micrographs were collected in counting mode with raw image pixel size of 1.067 Å across the defocus range of −0.8 to −2.5 μm. The total dose of ~55 e⁻Å⁻² was attained by using the dose rate of ~20 e⁻pixel⁻¹s⁻¹ across 50 frames during the 2.5-s exposure time.

For hTRPV6_DER, 12,516 micrographs were collected in super-resolution mode with raw image pixel size of 0.415 Å across the defocus range of −0.8 to −2.0 μm. The total dose of ~58 e⁻Å⁻² was attained by using the dose rate of ~16 e⁻pixel⁻¹s⁻¹ across 50 frames during the 2.5-s exposure time.

Image processing and 3D reconstruction

Data were processed in cryoSPARC⁷⁷ (Figs. S1–S3). Movie frames were aligned using the patch motion correction. Contrast transfer function (CTF) estimation was performed on non-dose-weighted micrographs using the patch CTF estimation. Subsequent data processing was done on dose-weighted micrographs. Following CTF estimation, micrographs were manually inspected and those with outliers in defocus values, ice thickness, and astigmatism as well as micrographs with lower predicted CTF-correlated resolution (higher than 5 Å) were excluded from further processing (individually assessed for each parameter relative to the overall distribution). After several rounds of selection through 2D classification, particles were further 3D classified (heterogeneous refinement) into four classes. Particles representing the best class were re-extracted without binning (256-pixel box size) and further 3D classified.

Processing example (hTRPV6_Anc): The total number of 14,833,688 particles were picked using internally generated 3D templates and 13,499,268 of these particles were extracted with 8x binning (32-pixel box size). After several rounds of selection through 2D classification, 8,589,896 particles (representing 71 classes) were extracted with 4x binning (64-pixel box size) and further classified in several rounds of 2D (a subset selection is shown in Fig. S1). 5,673,939 particles representing 42 classes were selected, split into 4 equal batches, and further classified in 3D (heterogeneous refinement) into 4 classes for each batch. The best reconstruction of each of the four 3D classification rounds, representing 4,602,500 particles, were re-extracted without binning (256-pixel box size) and further classified in several rounds in 3D. The final set of 617,835 particles representing the best apo-state class and 232,058 particles representing the best calmodulin-bound state were subjected to homogenous, non-uniform, and finally CTF refinement.

The reported resolutions (Table 1) were estimated using the gold standard Fourier shell correlation (GSFSC). The local resolution was calculated with the resolution range estimated using the FSC = 0.143 criterion. Cryo-EM density visualization was done in UCSF Chimera⁷⁸ and UCSF ChimeraX.⁷⁹

Model building

All models of hTRPV6 were built in Coot,⁸⁰ using the previously published cryo-EM structure of hTRPV6 in the open state (PDB ID: 7MIO) for hTRPV6_Anc-Open and hTRPV6_Der-Open, and of calmodulin-bound hTRPV6 in the inactivated state (PDB ID: 6E2F) for hTRPV6_Anc-Inact and hTRPV6_Der-Inact, respectively, as a guides. The models were tested for overfitting by shifting their coordinates by 0.5 Å (using Shake) in Phenix,⁸¹ refining the shaken models against the corresponding unfiltered half maps, and generating densities from the resulting models in UCSF Chimera. Structures were visualized and figures were prepared in UCSF Chimera, UCSF ChimeraX, and Pymol.⁸² The pore radius was calculated using HOLE.⁸³

Aequorin-based Ca²⁺ influx assay

Measurements of intracellular Ca²⁺ concentrations [Ca²⁺]_i in TRPV6 expressing cells were performed as reported previously,⁸⁴ with several modifications. HEK 293T cells (mycoplasma test negative, ATCC #CRL3216) were maintained at 37°C and 5% CO₂ in DMEM (Merck, #D6429) supplemented with 10% fetal calf serum (Thermo Fisher Scientific, #10270106), 100 μg/ml streptomycin and 100 U/ml penicillin (Merck, #P4333). Cells cultured in 6-well plates (~60% confluence) were transfected with 1 μg/dish TRPV6 plasmid DNA³¹ and 0.1 μg/dish pG5A plasmid DNA encoding EGFP fused to Aequorea victoria aequorin, using Lipofectamine 2000 (Thermo Fisher Scientific, #11668019). Twenty-four hours after transfection, the cells were washed with HEPES-buffered saline (HBS) containing 150 mM NaCl, 5.4 mM KCl, 0.2 mM CaCl₂, 1 mM MgCl₂, 5 mM HEPES (pH 7.4) and 10 mM glucose, and mechanically resuspended in HBS. For reconstitution of aequorin, cell suspensions were incubated with 5 μg/ml coelenterazine (Carl Roth, #4094.3) in the HBS for 30 min at room temperature. Cells were washed twice by centrifugation at 2000 rpm for 5 min (Heraeus Pico 17 microcentrifuge, Thermo Fisher Scientific), resuspended in HBS and aliquoted into 96-well plates (1×10⁵ cells per well). Luminescence was detected at room temperature using a CLARIOstar microplate reader (BMG LABTECH GmbH). To monitor TRPV6-mediated Ca²⁺ influx, the extracellular concentration of Ca²⁺ ([Ca²⁺]_o) was increased to 5 mM by injecting the CaCl₂-containing HBS. In experiments illustrated in Fig. 1H, the [Ca²⁺]_o range was 0.2–5 mM. The experiments were terminated by lysing cells with 0.05% (v/v) Triton X-100 in HBS to record the total bioluminescence. The bioluminescence rates (counts/sec) were analyzed at 1-s intervals and calibrated as [Ca²⁺]_i values using the following equation:

$$p{\left[{\text{Ca}}^{2+}\right]}_{\text{i}}=0.332588\left(-\mathit{\text{log}}\left(\text{k}\right)\right)+5.5593$$ (1)

where k represents the rate of aequorin consumption, i.e., counts/s divided by the total number of counts.

To calculate the EC₅₀ values for the effects of [Ca²⁺]_o (Fig. 1H), we used GraphPad Prism 10.11 and the following equation:

$$E\left(c\right)=E_{\mathit{\text{min}}}+{\text{c}}^{*}\left(E_{\mathit{\text{max}}}-E_{\mathit{\text{min}}}\right)/\left(E{C}_{50}+\text{c}\right)$$ (2)

where $E\left(c\right)$ is [Ca²⁺]_i measured at the concentration $c$ of [Ca²⁺]_o, $E_{\mathit{\text{min}}}$ is the minimal [Ca²⁺]_i, $E_{\mathit{\text{max}}}$ is the maximal [Ca²⁺]_i, $E{C}_{50}$ is the half-maximal effective concentration.

The concentration dependencies for inhibitors Cis-22a and Gd³⁺ (Fig. 1I–J) were fitted using GraphPad Prism 10.11 and the following equation:

$$E\left(c\right)=E_{\mathit{\text{min}}}+\frac{E_{\mathit{\text{max}}}-E_{\mathit{\text{min}}}}{1+\frac{c^h}{{I{C}_{50}}^h}}$$ (3)

where $E\left(c\right)$ is TRPV6 activity at the concentration $c$ of the compound, $E_{\mathit{\text{min}}}$ is the minimal activity, $E_{\mathit{\text{max}}}$ is the maximal activity, $I{C}_{50}$ is the half-maximal inhibitory concentration, and $h$ is the Hill coefficient.

Patch-clamp measurements

Patch-clamp experiments with TRPV6 were performed as reported previously,^{39,40,84–86} with a few modifications. HEK 293T cells grown in 35-mm dishes to ~60% confluence were transiently transfected with 0.8 μg/dish TRPV6 plasmid DNA.³¹ Patch-clamp experiments were conducted 18–22 hours after transfection with cells displaying EGFP fluorescence. Whole-cell currents were recorded using an EPC10 patch-clamp amplifier and PatchMaster software (Version V2×69, Harvard Bioscience). Voltages were corrected for a liquid junction potential of 10 mV. Holding potential was +50 mV. Currents were elicited by voltage ramps from −100 mV to +100 mV over 200 ms applied every 2 seconds. The inward current amplitudes were measured at −80 mV and were normalized to the cell size as pA/pF. The capacitance was measured using the automated capacitance cancellation function of EPC10. The standard extracellular solution contained 140 mM NaCl, 2.8 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES-NaOH, and 11 mM glucose (all from Merck). Solutions were adjusted to pH 7.2 using an FE20 pH meter (Mettler Toledo) and to 290 mOsm using a Vapro 5520 osmometer (Wescor Inc). Patch pipettes were made of borosilicate glass (Science Products) and had a resistance of 2.0−3.7 MΩ when filled with the standard intracellular pipette solution containing 120 mM Cs-glutamate, 8 mM NaCl, 1 mM MgCl₂, 10 mM Cs-EGTA and 10 mM HEPES-CsOH. The intracellular solution was also adjusted to pH 7.2 and 290 mOsm. Data are presented as the means ± SEM. The results were tested for normal distribution using the Shapiro-Wilk or Kolmogorov-Smirnov tests followed by unpaired t-test with Welch’s correction using GraphPad Prism 10.11 as indicated in the figure legends. Significance was accepted at p ≤ 0.05.

MD simulations

Structural models of hTRPV6_Anc-Open and hTRPV6_Der-Open (residues 27–638) were inserted into a hydrated lipid bilayer composed of palmitoyloleoylphosphatidylcholine lipids (POPC) (about 600 lipids in the membrane, about 160×160×120 ų – simulation box size). Ca²⁺ and Cl⁻ ions were added to ensure zero net charge at 250 mM ionic concentration. To prevent the gate closure, intersubunit restraints were applied to the distances between Cα atoms of 572, 575, 578 residues. The distances were restrained between each residue of one subunit with all nine residues of other three subunits with the force constant of 10 kJ/(mol × Å²). To keep N572 side chains in the orientation like in cryo-EM structures (pointed to the pore axis), the dihedral angles C-Cα-Cβ-Cγ were additionally restrained with the force constant of 100 kJ/mol. Three replicas of the model system were constructed for both TRPV6 haplotypes by the protein rotation around the pore axis relative to a simulation box.

All replicas were first equilibrated in several stages: 5 × 10³ steps of steepest descent minimization followed by heating from 5 to 310 K during a 100-ps MD-run, then 10 ns of MD run with fixed positions of the protein, 10 ns of MD with fixed positions of the protein backbone, 20 ns of MD with fixed positions of the protein C_α atoms to permit membrane relaxation. Then, 500 ns MD-production runs were carried out for the hTRPV6_Anc-Open and hTRPV6_Der-Open systems (six simulations in total).

To model Ca²⁺ currents in hTRPV6_Anc-Open and hTRPV6_Der-Open, we used the similar procedure of model system construction and equilibration. Subsequently, an external electric field with a magnitude of E = −3.2 mV/Å was applied along the z-axis, resulting in a transmembrane potential of approximately −400 mV. 1000 ns MD-production runs were carried out for three independent replicas of the model system for both TRPV6 haplotypes.

MD simulations were performed using GROMACS 2021.4 package,⁸⁷ CHARMM36 force field (version july2020)^88–93 and the TIP3P water model.⁹⁴ Simulations were carried out with an integration time of 2 fs, constrained hydrogen-containing bond lengths by the LINCS algorithm,⁹⁵ imposed 3D periodic boundary conditions, constant temperature (310 K) and pressure (1 bar). Cutoff distance of 12 Å was used for evaluation of nonbonded interactions and the particle-mesh Ewald method⁹⁶ employed for treatment of long-range electrostatics. Recently developed multi-site Ca²⁺ model (CAM) was used for calcium ions modeling⁹⁷, which is optimized for Ca²⁺-protein interactions, and successfully used for MD simulations of Ca²⁺ conductance through the biomolecular pores, including TRPV channels.^45,48

Detailed mapping and visualization of hydrophobic organization of the pore domains was performed using the Molecular Hydrophobicity Potential (MHP) approach.⁴⁷ This method enables quantitative estimation of the spatial distribution of hydrophobic/hydrophilic properties on a molecular surface. The formalism of MHP assumes that each atom in the molecular system is attributed with its specific hydrophobicity constant and the sum of atomic contributions is calculated at the molecular surface.^46,98 In this study, we used the atomic hydrophobicity constants obtained by Wildman and Crippen.⁹⁸ The Molecular Surface Topography (MST) tool⁴⁷ available at https://model.nmr.ru/cell was used for calculating and mapping the landscape and MHP cylindrical projections.

The coordination number of water molecules (CN) in the hydration shell of Ca²⁺ ions was calculated as a number of water oxygen atoms within the 3.3-Å distance from the ion center, which corresponds to the first minimum on the ion-oxygen radial distribution function. Ca²⁺ conductance through the pore (G) at −400 mV was calculated as follows:

$$G=I/V_{TM}=\left(N\times q\right)/\left(t\times E\times H_Z\right)$$ (4)

where N is the number of ions permeated from the extracellular side of the pore (above P1 site) to its intracellular side (below P4 site) over the time $t$, $q$ is the Ca²⁺ charge equal to 3.2×10⁻¹⁹ С, $E$ is the electric field, $H_Z$ is the average length of the simulation box along z-axis.

QUANTIFICATION AND STATISTICAL ANALYSIS

Fitting of dose-response curves was performed using GraphPad Prism 10.11. The results were tested for normal distribution using the Shapiro-Wilk test followed by unpaired t-test with Welch’s correction, paired t-test or u-test (Mann-Whitney) using GraphPad Prism 10.11 as indicated in the figure legends. The significance was accepted at p ≤ 0.05 and data are presented as the means ± SEM. The statistical details of experiments can be found in figure legends (Figures 1 and 4) and in the Results. Cryo-EM data collection and refinement statistics are summarized in Table 1.

Highlights.

• Ancestral and derived haplotypes of the Ca²⁺ selective channel TRPV6 were characterized

• Cryo-EM structures of both haplotypes in open and inactivated states were determined

• Currents, Ca²⁺ uptake and molecular dynamics of two haplotypes were compared

• Ancestral and derived TRPV6 have a similar structure and function

Data and Code Availability

• The cryo-EM density maps were deposited to the Electron Microscopy Data Bank (EMDB) under the accession codes EMD-45933 for hTRPV6_Anc-Open, EMD-45934 for hTRPV6_Anc-Inact, EMD-45935 for hTRPV6_Der-Open, and EMD-45936 for hTRPV6_Der-Inact. The atomic coordinates have been deposited to the Protein Data Bank (PDB) under the accession codes 9CUH for hTRPV6_Anc-Open, 9CUI for hTRPV6_Anc-Inact, 9CUJ for hTRPV6_Der-Open, and 9CUK for hTRPV6_Der-Inact and are publicly available as of the date of publication. The accession codes are also listed in the Key Resources Table.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, Peptides, and Recombinant Proteins
Penicillin-Streptomycin	Merck	Cat# P4333
Coelenterazine	Carl Roth	Cat# 4094.3
Kanamycin	Fisher scientific	Cat# BP906–5
Tetracycline	Fisher scientific	Cat# BP912
IPTG	Zymo Research	Cat# I1001–5
Tris	Fisher scientific	Cat# BP152–1
NaCl	Fisher scientific	Cat# BP358–212
PMSF	Acros Organics	Cat# 215740500
2-Mercaptoethanol (βME)	Acros Organics	Cat# 125470100
Phosphate-Buffered Saline	Corning	Cat# 21–040-CV
Bio-beads SM-2	Bio-Rad	Cat# 152–8920
D-desthiobiotin	Sigma	Cat# D1411
Glyco-diosgenin (GDN)	Anatrace	Cat# GDN101
POPC:POPE:POPG	Anatrace	Cat# P516:P416:P616
Ruthenium red	Tocris	Cat# 11103–72-3
Sf-900 III SFM	Thermo Fisher Scientific	Cat# 12658027
Freestyle 293 expression medium	Thermo Fisher Scientific	Cat# 12338018
Sodium butyrate	ACROS Organics	Cat# 263191000
QuikChange II XL Site-Directed Mutagenesis Kit	Agilent	Cat# 200521
DMEM	Merck	Cat# D6429
Deposited Data
Cryo-EM density map for hTRPV6Anc-Open	This paper	EMD-45933
Cryo-EM density map for hTRPV6Anc-Inact	This paper	EMD-45934
Cryo-EM density map for hTRPV6Der-Open	This paper	EMD-45935
Cryo-EM density map for hTRPV6Der-Inact	This paper	EMD-45936
Coordinates for hTRPV6Anc-Open	This paper	9CUH
Coordinates for hTRPV6Anc-Inact	This paper	9CUI
Coordinates for hTRPV6Der-Open	This paper	9CUJ
Coordinates for hTRPV6Der-Inact	This paper	9CUK
Experimental Models: Cell Lines
HEK 293S GnTI−	ATCC	Cat# CRL-3022
Sf9	Gibco	Cat# 12659017
HEK 293T	ATCC	Cat# CRL3216
Recombinant DNA
Human derived Trpv6 cDNA (a.a. 1–725)	GenBank	AF365927
Oligonucleotides
Forward primer for C157R: 5’-CGCAACCTGATCTACTTTGGAG-3’	This paper	N/A
Reverse primer for C157R: 5’-AGGTGACCTTCTAAATGCGGTG-3’	This paper	N/A
Forward primer for M378V: 5’-GTGACACCTAAAGACGATATCAGAC-3’	This paper	N/A
Reverse primer for M378V: 5’-ATACGCCTCTTGAAGGAGTTTTTGC-3’	This paper	N/A
Forward primer for M681T: 5’-ACGCCCTCCGTGAGCCGC-3’	This paper	N/A
Reverse primer for M681T: 5’-AGGCAGACTCAGGTGTGGACTG-3’	This paper	N/A
Forward primer for D542A: 5’-CTGACTATTATTGCCGGGCC-3’	This paper	N/A
Reverse primer for D542A: 5’-AAT AATAGT CAGGAACAGCT CAAACG TG-3’	This paper	N/A
Software and Algorithms
cryoSPARC 4.3.0	Punjani et al.58	https://cryosparc.com/
UCSF Chimera	Pettersen et al.59	https://www.cgl.ucsf.edu/chimera/
UCSF ChimeraX	Pettersen et al.60	https://www.cgl.ucsf.edu/chimerax/
Pymol (Schrödinger)	DeLano Scientific61	http://www.pymol.org
PHENIX	Afonine et al.62	https://www.phenix-online.org/
COOT	Emsley et al.63	http://www2.mrc-lmb.cam.ac.uk/Personal/pemsley/coot
CHARMM-GUI	Jo et al.64, Wu et al.65	https://charmm-gui.org/
AmberTools20	AMBER 202066	https://ambermd.org/AmberTools.php
CPPTRAJ	Roe and Cheatham67	https://amberhub.chpc.utah.edu/cpptraj/
VMD 1.9.4	Humphrey et al.68	https://www.ks.uiuc.edu/Research/vmd/
PatchMaster V2x69	Harvard Bioscience	www.heka.com
GraphPad Prism 10.11	GraphPad Software	www.graphpad.com
Other
UltrAuFoil R1.2/1.3 -Au 300 mesh, Gold	Ted Pella (Quantifoil)	Cat# 688–300-AU-50
Size Exclusion Superose 10/300 column	GE Healthcare	Cat# 17–5172-01