Molecular insights into the structural and dynamical changes of calcium channel TRPV6 induced by its interaction with phosphatidylinositol 4,5-bisphosphate

Abstract

Transient receptor potential vanilloid subfamily member 6 (TRPV6) is an epithelial calcium channel that regulates the initial step of the transcellular calcium transport pathway. TRPV6 is expressed in the kidney, intestine, placenta, and other tissues, and the dysregulation of the channel is implicated in several human cancers. It has been reported that phosphatidylinositol 4,5-bisphosphate (PIP₂) activates TRPV6 and its close homologue TRPV5; however, the underlying molecular mechanism is less clear. Recently, a structure of rabbit TRPV5 in complex with dioctanoyl (diC₈) PIP₂, a soluble form of PIP₂, was determined by cryo-electron microscopy. Based on this structure, the structural model of human TRPV6 with PIP₂ was set up, and then molecular dynamics simulations were performed for TRPV6 with and without PIP₂. Simulation results show that the positively charged residues responsible for TRPV5 binding of diC₈ PIP₂ are conserved in the interactions between TRPV6 and PIP₂. The binding of PIP₂ to TRPV6 increases the distance between the diagonally opposed residues D542 in the selectivity filter and that between the diagonally opposed M578 residues in the lower gate of TRPV6. A secondary structural analysis reveals that residues of M578 in TRPV6 undergo structural and position changes during binding of PIP₂ with TRPV6. In addition, principal component analysis indicates that the binding of PIP₂ increases the dynamical motions of both the selectivity filter and the lower gate of TRPV6. These changes induced by PIP₂ favor the channel opening. Thus, this study provides a basis for understanding the mechanism underlying the PIP₂-induced TRPV6 channel activation.

1. Introduction

Transient receptor potential vanilloid subfamily member 6 (TRPV6) is a calcium-selective channel whose permeability ratio between calcium (Ca²⁺) and sodium is greater than 100 (1, 2). TRPV6 is mainly expressed in epithelial cells and regulates the initial step of the transcellular Ca²⁺ transport pathway that allows Ca²⁺ to enter into cells across the apical membrane (3). TRPV6 is responsible for the active Ca²⁺ absorption in the intestine and Ca²⁺ reabsorption in the kidney (4, 5), thus it plays an important role in calcium homeostasis. TRPV6 is also found in the placenta, pancreas, prostate, salivary gland, liver, testes, and other organs (6). Aberrant expression and mutations of TRPV6 are involved in several human malignancies, such as breast, ovarian, and colon cancers, and transient neonatal hyperparathyroidism (7, 8).

Recently, the structure of TRPV6 was determined by x-ray crystallography and cryo-electron microscopy (9–12). These structure models obtained by different approaches are very similar. TRPV6 is a 4‐fold symmetrical tetramer that mainly includes two segments: a wide intracellular skirt formed by ankyrin repeat domains (ARDs) and a transmembrane (TM) domain that contains a central ion conduction pathway. For each TRPV6 subunit, the TM domain includes six TM helices (TM1‐TM6), a pore helix (P‐helix) between TM5 and TM6, and a TRP helix. The channel pore is formed by TM5, P-helix, and TM6 from all four subunits. Along the ion permeation pathway, there are two constriction sites: a selectivity filter in the upper pore region and a lower gate in the lower pore of TRPV6. The selectivity filter is formed by four residues (⁵³⁹TIID⁵⁴²) located in the pore loop preceding the P-helix. The negatively charged residue D542 in the selectivity filter projects towards the center of the pore (13). This residue has been identified as a critical residue for Ca²⁺ selectivity, and the mutation of this residue (i.e. D542A) would result in a nonfunctional TRPV6 channel (14). The lower gate is defined by the narrowest residues at the intracellular end of TM6. When bound with 2-aminoethoxydiphenyl borate (2-APB), a TRPV6 inhibitor, the conserved residue M578 forms a narrow constriction (13), creating a hydrophobic seal on the intracellular end of TM6.

Phosphatidylinositol 4,5-bisphosphate (PIP₂) is a membrane phospholipid that regulates the activity of many ion channels including the TRP channels (15, 16). It has been reported that PIP₂ is required for the activity of TRPV6, and the depletion of PIP₂ leads to the inactivation of the channel (17, 18). Our group found that the function of TRPV6 is suppressed by oculocerebrorenal syndrome of Lowe protein (OCRL, an inositol polyphosphate 5-phosphatase) in part through its 5-phosphatase activity in decreasing the level of PIP₂ (19). PIP₂ also activates TRPV5 (20), the close homologue of TRPV6 which exhibits ~75% amino-acid sequence identity with TRPV6. Recently, a structure of rabbit TRPV5 in complex with dioctanoyl (diC₈) PIP₂, a soluble form of PIP₂, was determined by cryo-electron microscopy (21). TRPV5 was found in an open state when diC₈ PIP₂ is bound to the channel. The positively charged residues R302 in the linker loop, K484 in TM5, and R584 in TM6 of rabbit TRPV5 (corresponding to R302, K484, and R584 in human TRPV6) are involved in the interaction with the head group of PIP₂. To date, the structure of TRPV6 in complex with PIP₂ has been unavailable. However, the mutations of residues R302 and K484 (i.e. R302Q and K484Q) in TRPV6 significantly reduced the activity of the channel (21, 22), suggesting that these residues are involved in PIP₂ binding. Since TRPV6 and TRPV5 share a high degree of sequence identity, TRPV6 would have a PIP₂-binding site similar to TRPV5.

The structure of TRPV5 with diC₈ PIP₂ provides a molecular basis to understand how TRPV6 interacts with PIP₂. Nonetheless, the mechanism by which PIP₂ regulates the activity of TRPV6 is still unclear. To solve this problem, the structural model of human TRPV6 in complex with PIP₂ was set up based on the structure of TRPV5 with diC₈ PIP₂. Then, molecular dynamics simulations were performed for TRPV6 with and without the binding of PIP₂.

2. Materials and methods

2.1. Simulation system set up.

The structural model of human TRPV6 in complex with PIP₂ was set up based on the structure of rabbit TRPV5 in the presence of diC₈ PIP₂ (PDB ID: 6DMU) (21). Sequence alignment between human TRPV6 and rabbit TRPV5 (Figure S1) shows that the sequence identity and similarity of the two proteins are 82.5% and 92.2%, respectively. The structure of human TRPV6 (PDB ID: 6D7T) (13) was used as the initial structure for TRPV6. The reason why we choose this structure is that TRPV6 in this structure is in a closed state, and thus it can be used to investigate whether the binding of PIP₂ can activate TRPV6 and cause the opening of the channel. After superimposing the structure of human TRPV6 with that of rabbit TRPV5, the positions of four diC₈ PIP₂ were kept to model the PIP₂ lipids into the structure of human TRPV6. Then, the oleoyl and palmitoyl chain atoms of PIP₂ were modeled by CHARMM-GUI (23). Residue A467 in the original PDB structure of human TRPV6 was mutated back to Y467 in the simulation model, and the Ca²⁺ ions in the original PDB structure were retained in the model. In this study, the transmembrane region of human TRPV6, i.e. amino acids 291–610 that includes the linker helices (LH) LH1-LH2, pre-TM1, TM helices TM1-TM6, pore helix, and the TRP domain, was modeled. It is noted that the numbering of TRPV6 is based on the sequence of the annotated truncated form of TRPV6 (24). Compared to the endogenous human full-length TRPV6, it only consists of 725 amino acids and does not contain the N-terminal extended 40 amino acids. We used this numbering because it has been widely used to study the structural and functional properties of TRPV6.

Since the residues in ARD domains are not involved in the interaction with PIP₂, the ARD domains of TRPV6 were not included in the simulations to save the computational resources. Two systems of TRPV6 without and with PIP₂ binding were built, and these systems are denoted as TRPV6 and TRPV6-PIP₂, respectively. The membrane environment for TRPV6 was modeled by a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) lipid bilayer using CHARMM-GUI membrane builder (25). The two systems were then embedded in a lipid bilayer with TIP3P water molecules added to both sides of the bilayer (Figure S2). The simulation box dimensions were 128 Å × 128 Å × 120 Å along the x, y, and z axes, respectively. All systems were neutralized using NaCl with a concentration of 150 mM. The CHARMM36m all-atom force field (26) was assigned to the protein, ions, and water molecules, and the CHARMM36 lipid force field (27) was used for POPC and PIP₂ lipids.

2.2. Molecular dynamics simulations.

To investigate the structural and dynamical changes of TRPV6 caused by the binding of PIP₂, three independent molecular dynamics (MD) simulations were performed using the AMBER18 simulation package (28). The equilibrate simulation protocol is similar to our previous studies (29, 30). The simulation systems were first minimized totally 4,000 steps using steepest descent and subsequently conjugate gradient minimization methods. Then the whole system was equilibrated for 20 ps, while TRPV6, Ca²⁺ ions, and PIP₂ lipids were restrained in a constant number-pressure-temperature (NpT) ensemble with temperature at 50 K and pressure at 1 atm by applying a force constant of 100 kcal·mol⁻¹·Å⁻¹. After that, the system was gradually heated from 50 K to 300 K via six 100 ps in a constant number-volume-temperature (NVT) ensemble. During the heating simulations, the restraints on the proteins, Ca²⁺ ions, and PIP₂ are still maintained. The restraints were gradually reduced to zero in the subsequent 200 ps equilibration simulation at NpT of 300 K and 1 atm. Finally, 500 ns production simulations were carried out using Berendsen temperature and pressure coupling (31) without any restraint. The SHAKE constraints (32) were applied to all hydrogen-heavy atom bonds to permit a dynamics time step of 2 fs. Electrostatic interactions were calculated by the particle-mesh Ewald (PME) method (33, 34) and the cutoffs for PME and Lennard-Jones interactions were 10 Å, which is similar to the works of Ahamad et al. (35, 36).

2.3. Data analyses.

The root mean square deviation (RMSD) for the Cα atoms of TRPV6 was calculated to assess the equilibration of the simulation (Figure 1B). All the simulations reached plateau after 200 ns, especially for trajectory III of TRPV6-PIP2. To make sure that all the simulations got fully equilibrated, the last 200 ns simulations were used for analyses. The CPPTRAJ program of AMBER18 was used for the structural and dynamical analyses, including secondary structure analysis, root mean square fluctuation, distance calculation, density analysis, and principal component analysis. The DSSP method (37) was applied to determine whether an amino acid residue belonged to an α helix. To determine which residue of TRPV6 plays an important role in the binding of PIP₂, the interaction energy for each residue of TRPV6 was calculated using the molecular mechanics/generalized Born surface area (MM/GBSA) approach (38). To compare the interactions between TRPV6 and PIP₂, the representative structures for TRPV6 and TRPV6-PIP₂ were obtained by clustering analysis using MMTSB toolset (39). PDB structures for TRPV6 and TRPV6-PIP₂ were generated from the last 200 ns MD trajectories with a 100 ps interval. A centroid structure was obtained by averaging the PDB structures. Clustering analysis was performed using the K-means algorithm (40) based on the RMSD similarity of the structures. The structure that has the lowest RMSD from the centroid structure was obtained as the representative structure. The VMD software (41) was used for structure visualization. The average values and standard deviation for distances between the diagonally opposed residues were calculated and the significant differences for the variables were determined using the Student’s t-test with 95% confidence.

Figure 1. Modeling the interactions of TRPV6 with PIP₂.

A. Initial model of TRPV6 with phosphatidylinositol 4,5-bisphosphate (PIP₂) (labeled in purple). Residues D542 in the selectivity filter and M578 in the lower gate are present. For clarity, only two residues of D542 and M578 from the diagonally opposed subunits and one PIP₂ molecule are shown in the Figure. The detailed interactions between residues of TRPV6 and PIP₂ molecule are exhibited in the enlarged figure. The residues of TRPV6 that interact with the head group, the oleoyl chain, and the palmitoyl chain of PIP₂ molecule are labeled in blue, black, and orange, respectively. The residues of TRPV6 and PIP₂ molecule are respectively shown in Licorice and ball-and-stick models using VMD software. B. Root mean square deviation (RMSD) for the Cα atoms of TRPV6 without and with PIP₂. Three 500 ns independent simulations were performed for the two systems containing TRPV6 without and with PIP₂.

3. Results and discussion

3.1. Conserved amino-acid residues in TRPV5 and TRPV6 are involved in their interactions with PIP₂.

The structure of human TRPV6 with PIP₂ was modeled based on the structure of rabbit TRPV5 in complex with the soluble form of PIP₂ (PDB ID: 6DMU). The sequence alignment between human TRPV6 and rabbit TRPV5 is shown in Figure S1. The sequence identity between the two proteins is as high as 82.5%, suggesting that the two proteins may have a similar PIP₂-binding site. To validate our modeled structure of TRPV6 with PIP₂, the interactions between TRPV6 and PIP₂ were checked (Figure 1A). In the initial modeled structure, the head group of PIP₂ is coordinated with residues R302, R305, K484, and R584 of TRPV6, while the oleoyl chain of PIP₂ interacts with residues W583, D590, W593, and R594, and the palmitoyl chain of PIP₂ contacts with residues F329, C330, F472, Q473, M474, and F468 of TRPV6 (Figure 1A).

To determine whether the interactions between these residues and PIP₂ are maintained during the simulations, the interaction energy for each residue of TRPV6 was calculated and the averaged values are shown in Figure 2A. For the interaction energy, a negative value means the residue favors the interaction whereas a positive value indicates the residue disfavors the interaction. Residues K300, K301, R302, R305, K484, R492, and R584 exhibit large negative interaction energies, indicating these residues play important roles in the interactions with PIP₂. This is consistent with the structural data that the conserved residues R302, R305, K484, and R584 in TRPV5 are responsible for the binding of TRPV5 with diC₈ PIP₂ (21).

Figure 2. Interactions between TRPV6 and PIP₂.

A. Interaction energy for each residue of TRPV6. The residues that are favorable for the binding of the head group, the oleoyl chain, and the palmitoyl chain of PIP₂ are labeled in blue, black, and orange, respectively. The residues that are unfavorable for the binding of PIP₂ are labeled in red. The residues with the absolute energy value less than 1 kcal/mol are not labeled. B. Comparison of the conformational changes of the residues those are favorable for the binding of the head group of PIP₂. When PIP₂ is present, residues K300, R302, and R305 move towards R584 to attract the negatively charged head group of PIP₂. However, when PIP₂ is absent, these residues are away from R584 due to the repulsive forces. C. Details of how the residues those are unfavorable for the binding interact with PIP₂. The green line indicates there is a hydrogen bond between the two residues. D. Details of how TRPV6 interacts with the oleoyl and palmitoyl chains of PIP₂. The two monomers of TRPV6 are shown in yellow and pink, respectively. In Figures B, C, and D, the residues of TRPV6 and PIP₂ molecule are respectively shown in Licorice and ball-and-stick models using VMD software.

To examine how these positively charged residues interact with PIP₂, the representative structures showing the detailed interactions of these residues are exhibited (right figure of Figure 2B). To give a clear view, the interactions between TRPV6 and PIP₂ are also shown in the two-dimensional plot (Figure S3). For comparison, the conformation of these residues in TRPV6 without PIP₂ binding is present as well (left figure in Figure 2B). These residues are gathered together by the head group of PIP₂ since the negatively charged inositol group of PIP₂ draws these positively charged residues together. However, in the absence of PIP₂, residues K300 and R302 in the loop between LH1 and LH2, and R305 in LH2 are separated from residues K484 and R492 in TM5 and R584 in TM6 due to the repulsive electrostatic force. Distance between R305 in LH2 and R584 in TM6 is 14.79 ± 0.46 Å in TRPV6 whereas it is 10.28 ± 0.21 Å in TRPV6-PIP₂ (Figure 2B). Residue R305 moved 4.51 Å away from residue R584 when PIP₂ is absent, thus allowing LH2 to move away from the PIP₂-binding site of TRPV6.

The interaction energy result also shows that the negatively charged residues E303, D309, D489, D580, E588, D590, and E591 disfavor the interaction between TRPV6 and PIP₂ (Figure 2A). The detailed interactions of these residues are shown in Figure 2C. Most of these residues are closed to the positively charged residues that are involved in the interactions with the head group of PIP₂. It is noted that residue D590 that interacts with R589 is pointed away from the center of the positively charged residues where PIP₂ is located in position 1, however, these residues are pointed to the PIP₂ lipid in position 2. These aspartate and glutamate residues make an environment with negative electrostatic potential that would reduce the ability of TRPV6 binding to PIP₂.

Besides the positively charged residues that have large interaction energies, several residues favor the binding of TRPV6 with PIP₂ (Figure 2A). These residues are involved in the interactions with the tails of PIP₂. Among them, residues K314, V317, and W321 in pre-TM1 and R594 in TRP domain (labeled in black in Figure 2D) are involved in the interaction with the oleoyl chain of PIP₂; while residues F472, Q473, M474, and L475 in the loop between TM4 and TM5, and residues M491 and W495 in TM5 (labeled in orange in Figure 2D) are involved in the interaction with the palmitoyl chain of PIP₂. Compared to the initial modeled structure, the interactions between TRPV6 and the palmitoyl chain of PIP₂ were kept the same, whereas new interactions between the pre-TM1 helix of TRPV6 and the oleoyl chain of PIP₂ were identified.

In summary, the residues that contribute positively to the binding of the head group of PIP₂ are located in LH2, TM5, TM6 helices, and TRP domain. The residues that interact with the oleoyl chain of PIP₂ are located in pre-TM1 helix and TRP domain, while those that interact with the palmitoyl chain of PIP₂ are located in TM5 and loop between TM4 and TM5. Sequence alignments reveal that most of these residues are conserved between different species of TRPV5 and TRPV6 (Figure S4). It is noted that K300 in human TRPV6 is involved in the interaction with the negatively charged head group of PIP₂. However, the corresponding residues D300 in human TRPV5 is a negatively charged residue, which is unfavor for the binding of PIP₂. This interaction difference of residue at postion 300 between TRPV6 and TRPV5 may provide a novel strategy to design new antagonists to differentiate the two channels. For example, structure activity relationship analysis can first be used to screen small molecules and find the potential molecules for the binding of TRPV6 and TRPV5. Then, the specificity of the candidate molecules towards the two proteins can be determined by assessing the binding affinity of candidate antagonist with the two proteins. Amino-acid residues not conserved in the binding site may provide molecular basis for antagonist selectivity and specificity. Since the charge of the residue at position 300 is distinct between TRPV6 and TRPV5, the electrostatic interaction between the residue at position 300 and the candidate antagonist would be different.

3.2. PIP₂ causes structural and dynamical changes in TRPV6.

To investigate whether the binding of PIP₂ causes the structural change of TRPV6, secondary structure analyses were performed for each residue of TRPV6 and the results are averaged from the three independent simulations (Figure 3A). To compare the secondary structure of TRPV6 with and without PIP₂ binding, the occupancy for each helix was calculated by summing the occupancy of all the residues in the helix and then dividing it by the residue number. Results show that the occupancies of LH1, pre-TM1, TM1-TM6 helices, and TRP domain in TRPV6 are similar to those in TRPV6-PIP₂ (the detailed occupancies for these TM helices are shown in Table S1). However, the occupancy of LH2 helix in TRPV6-PIP₂ (73.85%) is much higher than that in TRPV6 (67.00%), suggesting that LH2 is more stable in TRPV6 upon PIP₂ binding. To further investigate the structural change caused by the binding of PIP₂, the radius of gyration (Rg) and solvent accessible surface area (SASA) for TRPV6 were calculated (Figure S5 and Figure S6). The average value of Rg is 34.99 ± 0.12 Å for TRPV6, whereas it is 34.89 ± 0.17 Å for TRPV6-PIP₂; the average SASA value is (8.18 ± 0.01) × 10⁴ Ų for TRPV6, whereas it is (8.28 ± 0.01) × 10⁴ Ų for TRPV6-PIP₂. These two variables are not altered by the binding of PIP₂, indicating the quaternary structure of TRPV6 is not significantly changed.

Figure 3. Structural and dynamical change of TRPV6 caused by the binding of PIP₂.

A. Secondary structure analysis was performed by calculating the helix occupancy for each residues of TRPV6. Transmembrane helices (TM) 1–6 are shaded in cyan. LH1, LH2, pre-TM1, pore helix, and TRP domain are shaded in gray. B. Root mean square fluctuation (RMSF) for each residue of TRPV6. The helix occupancy and RMSF values are calculated by averaging the values of TRPV6 tetramer from the three independent simulations. The regions corresponding to the secondary structure of TRPV6 are also shaded to be consistent with figure A.

The dynamical change of TRPV6 was studied by calculating the root mean square fluctuation (RMSF) of the Cα atoms that measures the fluctuation amplitude of each residue over the equilibrated trajectories (Figure 3B). To give a view of how RMSF was changed based on the secondary structure of TRPV6, the secondary structure elements labeled in Figure 3A were also shown in Figure 3B. Results show that the RMSF values for all the TM helices, pore helix, and TRP domain are low and are similar between TRPV6 and TRPV6-PIP₂, demonstrating these helices are dynamically stable during the simulations. However, the RMSF values for LH1, LH2, and pre-TM1 helices in TRPV6 are much larger than those in TRPV6-PIP₂, indicating these helices undergo large dynamical changes in the absence of PIP₂. This is consistent with the results of the distance calculation (Figure 2B) that LH2 moved away from the PIP₂-binding site when PIP₂ is absent.

3.3. PIP₂ increases the distances between the diagonally opposed residues in the selectivity filter and the lower gate of TRPV6.

Structural data show that the channel pore in TRPV5 with diC₈ PIP₂ is widened compared to that in TRPV5 without the binding of PIP₂ (21). To check whether the pore of TRPV6 is widened by the binding of PIP₂, the distances between the diagonally opposed residues representing the diameters of the selectivity filter and the lower gate were calculated (Figure 4). Residues D542 and M578 are respectively chosen to represent the diameters of the selectivity filter and the lower gate, since residue D542 in the selectivity filter is a critical residue for Ca²⁺ selectivity (14) and residue M578 in the lower gate forms a narrow constriction for the channel pore (13). Results of distance calculations show that the binding of PIP₂ increases the distance between the diagonally opposed residues D542 (3.87 ± 0.03 Å in TRPV6 vs. 4.48 ± 0.03 Å in TRPV6-PIP₂) and that between the diagonally opposed residues M578 (5.66 ± 0.10 Å in TRPV6 vs. 6.04 ± 0.10 Å in TRPV6-PIP₂), indicating the selectivity filter and the lower gate are enlarged when PIP₂ is bound to TRPV6.

Figure 4. The selectivity filter and the lower gate of TRPV6 are widened by the binding of PIP₂.

A. Comparison the pore region of TRPV6 without and with PIP₂. For clarity, only the two diagonally opposed pore regions of TRPV6 are shown. The distance between the diagonally opposed residues D542 in the selectivity filter is increased when TRPV6 binds with PIP₂ (3.87 ± 0.03 Å in TRPV6 vs. 4.48 ± 0.06 Å in TRPV6-PIP₂). B. Distribution of the dihedral angle for the side chain of residue D542. The dihedral angle is distributed around 70° with a small peak and around 290° with a large peak in TRPV6 without PIP₂ binding, whereas it is distributed around 60° with a large peak and around 185° with a small peak in TRPV6 with PIP₂. C. Intracellular view of the tetrameric residues M578 in the lower gate. The distance between the diagonally opposed residues M578 in the lower gate is increased when TRPV6 binds with PIP₂ (5.66 ± 0.10 Å in TRPV6 vs. 6.04 ± 0.10 Å in TRPV6-PIP₂). D. Distribution of the dihedral angle for the side chain of residue M578. The dihedral angle is distributed around 180° and around 290° in both TRPV6 with and without PIP₂ binding, however, the dihedral angle is distributed more around 180° in TRPV6 with PIP₂.

3.4. PIP₂ induces conformational changes of key residues in the selectivity filter and the lower gate of TRPV6.

To further check whether the channel pore is widened by the binding of PIP₂, the conformational changes of the key residues in the selectivity filter and the lower gate were investigated. The representative structure shows that the side chains of residues D542 point to each other and the center of the Ca²⁺ permeation pathway in TRPV6; whereas they point upwards to the extracellular side in TRPV6-PIP₂ (Figure 4A), which is consistent with the conformations of the residues in TRPV5 with the binding of PIP₂ (21). The calculations of the dihedral angle for the side chain of residue D542 also confirm that the distribution of the dihedral angle was altered in TRPV6 with PIP₂ (Figure 4B). The conformational changes of residue M578 in the lower gate were also studied by calculating the dihedral angle for the side chain of M578 (Figure 4D). The dihedral angle of M578 is mainly distributed around 180° and around 290° in both TRPV6 and TRPV6-PIP₂. However, when PIP₂ is bound to TRPV6, the distribution of the dihedral angle around 180° is increased while that of the dihedral angle around 290° is decreased.

Our secondary structure analysis reveals that the helix occupancy of residue M578 is decreased when PIP₂ is bound to TRPV6 (99.13% in TRPV6 vs. 86.18% in TRPV6-PIP₂) (Figure 3A). To further study whether residue M578 undergoes conformational changes, the mass density of residue M578 describing the location of the residue along the bilayer normal (z-axis) was calculated (Figure 5A). To give a clear view of the positions of the residue relative to the membrane surface, the final simulation structures are represented in Figure 5B. The black dashed lines in Figure 5B are the density peaks of the P atoms of POPC lipids, which indicate the surfaces of the membrane. The red dashed and dotted lines respectively represent the density peaks of residue 578 in TRPV6 and TRPV6-PIP₂. Compared to it in TRPV6, residue 578 in TRPV6-PIP₂ moved upwards to the cavity of the channel pore. Accompanied with the enlarged diameters of the selectivity filter and the lower gate, these conformational changes of residues D542 and M578 may facilitate the Ca²⁺ ions passing through the channel pore when PIP₂ is bound to TRPV6.

Figure 5. The location change of M578 in the lower gate of TRPV6.

A. Comparison of the mass density of M578 in TRPV6 without and with PIP₂. The black dotted lines indicate the density peaks of the phosphors (P) atoms (black lines) of POPC lipids. The red dashed and dotted lines indicate the density peaks of residue M578 (red lines) in TRPV6 and TRPV6-PIP₂, respectively. B. Structure showing the location of M578 (labeled in red) relative to membrane surface. The details of the conformations of M578 are presented in the enlarged figures.

3.5. PIP₂ increases the dynamical motions of the selectivity filter and the lower gate.

To further investigate the dynamical motion of the channel pore, principal component analysis (PCA) was performed (Figure 6). PCA is often used to extract protein’s large collective motions that reveal the physiological character of long trend dynamics (42–45). To clearly show the change of collective motion for the channel pore of TRPV6 by the binding of PIP₂, only the first eigenvectors that dominate the principal component of motions for the two diagonally opposed subunits of the channel are shown in Figure 6. PCA results show that in TRPV6 without PIP₂ binding, no significant motion was observed for the pore region, except that the TRP domains exhibit departure motion between the opposed subunits (left figure of Figure 6). In contrast, besides the TRP domains, the selectivity filter and the lower gate also exhibit large dynamical motions in TRPV6 with PIP₂ binding (right figure of Figure 6). These changes of dynamical motions in the selectivity filter and the lower gate may enlarge the channel pore of TRPV6, and thus regulate the function of the channel.

Figure 6. Principal component analysis showing PIP₂ binding causes different collective motions for the pore region of TRPV6.

The TRP domains (labeled in black dotted circles) show large dynamical motions in TRPV6 and TRPV6-PIP₂. However, residues in selectivity filter (labeled in red dotted circles) and lower gate (labeled in orange dotted circles) show large dynamical motions only in TRPV6-PIP₂.

4. Conclusions

In this study, the interaction between TRPV6 and PIP₂ was modeled based on the structure of TRPV5 with dioctanoyl PIP₂. Then the structural and dynamical changes of TRPV6 caused by the binding of PIP₂ were investigated by molecular dynamics simulations. Simulations results show that the positively charged residues K300, K301, R302, R305, K484, R492, and R584 of TRPV6 play important roles in the interaction with PIP₂. Accompanied with the conformational changes of residue D542 in the selectivity filter and residue M578 in the lower gate, the binding of PIP₂ to TRPV6 increases the diameters of both the selectivity filter and the lower gate of the channel. In addition, the binding of PIP₂ increases the dynamical motions of both the selectivity filter and the lower gate of TRPV6. These changes may ultimately result in the opening of the TRPV6 pore and the activation of the channel. The finding that PIP₂ activates TRPV6 by interacting with the positively charged residues may help to develop new therapeutic strategies through regulating the activation state of TRPV6. For example, new antagonist with negatively charged groups can be designed to compete the binding of TRPV6 with PIP₂, or novel activator can be created to fit into the binding site of PIP₂ to induce the opening of the channel.

Abbreviations

ARDs

ankyrin repeat domains

Ca²⁺

calcium

diC₈

dioctanoyl

LH

linker helices

MD

molecular dynamics

MM/GBSA

molecular mechanics/generalized Born surface area

NVT

number-volume-temperature ensemble

NpT

number-pressure-temperature ensemble

PCA

principal component analysis

PME

particle-mesh Ewald

PIP₂

phosphatidylinositol 4,5-bisphosphate

POPC

1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine

P-helix

pore helix

Rg

radius of gyration

RMSD

root mean square deviation

RMSF

root mean square fluctuation

SASA

solvent accessible surface area

TM

transmembrane

TRPV5/6

transient receptor potential vanilloid subfamily member 5/6

2-APB

2-aminoethoxydiphenyl borate

Data availability statement

All data and protocols are available upon request.