Mechanism of Calcium Ion-Selective Channel Opening in the ChR2_L132C Mutant: A Molecular Dynamics Simulation

Abstract

Channelrhodopsin-2 (ChR2) is an important tool for optogenetics, and some of its mutants are Ca²⁺-selective channels. However, the mechanism for Ca²⁺-selective permeation is still unclear. In this study, molecular dynamic (MD) simulations for the Ca²⁺ permeation of the CatCh mutant were carried out to investigate the fundamental features of the selectivity of Ca²⁺. Research on the conformational changes in the key residues near the central gate (CG) of the channel suggested that E83, E90, and D253 play an important role in Ca²⁺ conductivity. The clustering analysis indicates that the above “EED triad” acts as a filter, and Ca²⁺ can only pass through if the EED is in a certain conformation. It was also found that hydrated Ca²⁺ can be coordinated with carboxyl groups, resulting in the loss of part of the water molecules in the hydrated shell and a reduction in ionic radius, which helps Ca²⁺ enter the channel.

1. Introduction

Optogenetics is a multidisciplinary technology that integrates optics, software control, gene manipulation technology, electrophysiology, etc., and the ChR2 [1], extracted from microorganisms, which is one of the most important tools in optogenetics. Channelrhodopsin-2 (ChR2) [1] is a non-selective cation channel that permits the passage of various cations, including H⁺, Na⁺, K⁺, and Ca²⁺. However, due to its low ionic conductivity and poor ion selectivity, ChR2 is not effective in some optogenetic fields. Therefore, designing and modifying ChRs with high conductance and ion selectivity is of great importance. Several ChR2 variants have been artificially engineered, including CatCh [2], which is a mutated form of ChR2 (ChR2-L132C) that has a six-fold increase in Ca²⁺ permeability compared to wild-type ChR2. This results in a 70-fold increase in photosensitivity and faster response kinetics.

The L132C mutation is a classical method used to increase Ca²⁺’s conductance of ChR2. Many subsequent experiments related to mutations have included the L132C mutant. For instance, Pan’s team conducted experiments in 2014 that showed increased photosensitivity in ChR2-L132C/T159C and ChR2-L132C/T159S mutants [3]. Similarly, Mager’s team reported in 2018 that the F219Y mutation near L132 in ChR2 also increases relative calcium ion permeability (PCa/PNa) [4]. CapChR1 (ChR2-S63D-L132C-T159C-N258E) and CapChR2 (CoChR-S43D-L132C-T139C-N238E) [5] were obtained by Hegemann’s group in 2022. Both variants contain the L132C mutation, and exhibit increased Ca²⁺ conductance. The MD simulation on CapChR2 found that extracellularly oriented D43 and E238 in the channel attract Ca²⁺ into the channel. The authors also suggested that the voltage dependence of Ca²⁺ conductance at CG could be reduced by mutating residues at CG of ChR2, such as S63D/N258E.

The crystal structure of the open state is crucial in explaining the mechanism of ion permeation. Although the crystal structures of the dark state of ChR2 and other ChRs have been found [6,7,8,9,10], and the structures of some intermediates in the photocycle have even been resolved [11], the explicit structure of the open state of ChR2 is still lacking. In addition, there are structural differences between the dark and open states of ChR2, as well as changes in the amino acid residues near the channel pores and the protonation state of the retinal with the gating state [12,13,14]. As a result, many existing mechanisms of ion permeation based on the dark-state crystal structure are insufficient.

Ca²⁺ plays a crucial role in various metabolic activities, including cell signaling, metabolic excitation, neuronal plasticity, muscle contraction, apoptosis, and cell growth. However, there have been few studies on the Ca²⁺-selective mechanism of ChRs. Further research on this mechanism will aid in understanding the ion-permeable mechanism of ChRs and facilitate the design of Ca²⁺-selectively enhanced ChRs. This study employs MD simulations to investigate the process of Ca²⁺ and Na⁺ permeation in CatCh, aiming to explain the reason for the increased Ca²⁺ conductance caused by the L132C mutation.

2. Materials and Methods

2.1. Modeling Transmembrane CatCh Systems

In our previous research [15], a conducting state of ChR2 was obtained through MD simulations, referred to as the ‘ChR2-P520-like’ structure. This structure is indicated as the most likely conductive state of ChR2 by the radius of the ion channel, the water distribution density of the channel, and the calculated absorption spectra of the retinal chromophore. The residues E83, E90, and D253 are deprotonated in this state. To create a mutant of L132C, the leucine residue at position 132 in the sequence was substituted with cysteine using the CHARMM-GUI 3.8 online tool [16]. The L132C mutant was then inserted into a pre-equilibrated lipid bilayer composed of 16:0/18:1c9 palmitoyloleylphosphatidylcholine (POPC) to create the transmembrane system of CatCh. Subsequently, the transmembrane CatCh was immersed in a 0.1 mol/L CaCl₂ solution environment, and a 300 ns MD simulation was conducted to obtain its open-state-like structure, which was designated as “CatCh-P520” (the pdb file is included in the Supplementary Materials) The CatCh system comprised 245 residues, 12145 water molecules, and 136 phospholipid molecules, totaling 58672 atoms. AmberTools v22 [17] was used to analyze the last 100 ns of equilibrium trajectories, and clustering analysis was employed to extract five representative conformations.

The CatCh-P520 structure obtained above was used for the construction of the CatCh-P520 transmembrane structure; using the Membrane Builder on the CHARMM-GUI 3.8 online website, the CatCh-P520 was embedded as an initial structure in a preequilibrated POPC lipid bilayer. Then, the composite structure of CatCh-P520 and the membrane were immersed into 0.15 mol/L of NaCl, a 0.1 mol/L CaCl₂ solution, and a mixed solution, respectively, to complete the construction of the transmembrane system of CatCh-P520.

2.2. MD Simulation Details

NAMD2.13 software [18] was used to perform the MD simulation of the CatCh transmembrane system. For all simulations, the CHARMM36m [19] force field was employed. The simulations used the Particle Mesh Ewald (PME) method [20] to calculate long-range electrostatic interactions. A truncation radius of 12 Å was set for short-range non-bonding interactions. The SHAKE algorithm [21] was used to limit covalent bond vibrations and all hydrogen atoms. The Langevin piston algorithm [22] was used to maintain a pressure of 1. The system temperature was controlled at 303.15 K using Langevin temperature. Harmonic restraints were used to restrain the partial dihedral and plane angles of the protein, which were gradually reduced in 6 pre-equilibration steps. MD simulations were performed in the NPT ensemble for 300 ns with a time step of 2 fs. Figure S2 shows the root mean square deviation (RMSD) of the CatCh.

MD simulations were performed on CatCh-P520 and ChR2-P520 transmembrane systems using NAMD 2.13 software. An electric field strength of 0.198 kcal/mol⁻¹ Å⁻¹ E⁻¹ was added in the direction of the z-axis of the whole system (perpendicular to the direction of the bilayer POPC membranes). The simulations were performed for 300 ns under the NPT ensemble. The permeation simulations’ conformations underwent cluster analysis using AmberTools. Free energy calculations were performed using Molecular Mechanics/the Poisson Boltzmann Surface Area (MM/PBSA) [23].

To investigate the permeation of Ca²⁺, CatCh-P520 was immersed in a 0.1 mol/L CaCl₂ solution after being embedded in a bilayer POPC membrane. A membrane voltage with an electric field strength of 0.198 kcal/mol⁻¹ Å⁻¹ E⁻¹ was applied along the z-axis, and another 300 ns MD simulation was performed. The ion permeation process typically takes a long time and, therefore, requires acceleration. In experiments, the photocurrents typically last for microseconds [24]. To accelerate the process of the MD simulation of ion permeation, a transmembrane voltage of approximately 900 mV [25] was used in this study. This voltage was about ten times higher than the membrane voltage in the physiological state (70–100 mV) and enabled the observation of the process of ion permeation in nanoseconds. Previous studies of CapChR2 [5] by Fernandez Lahore et al. also used voltages close to 1 V for the ion penetration of the membrane.

3. Results

3.1. Models of CatCh-P520 and Simulation of ion Permeation

A comparison of the structure of CatCh-P520 with the previously reported ChR2-P520-like open state (shown in Figure S1) indicates that the channel enlargement is mainly due to the partial displacement of TM2 of CatCh-P520 near the C132 site. The cavity containing amino acid position 132 of CatCh-P520 is larger than that of the ChR2-P520-like open state, and the movement of TM4 is the main cause of the increase in this cavity. Kleinlogel et al. [2] also reported that the increase in Ca²⁺ permeability may be due to the leucine mutation at position 132, which results in the formation of a cavity at the mutation site. This increases the flexibility of the helix, leading to the displacement of the helix and making the channel larger. The experiments were conducted with CatCh mutants.

The density map of Ca²⁺ in CatCh-P520 indicates the significant aggregation of Ca²⁺ near E101 on the extracellular side, suggesting that this site is prone to Ca²⁺ aggregation in the solution. In the equilibrium conformation of CatCh-P520, the carboxyl groups of E101 and E97 located in the extracellular access channel, and E90 and D253 at the central gate (Figure S3) are oriented towards the extracellular side, which attracts Ca²⁺ into the channel.

The duration of the Ca²⁺ residence at each residue position of the channel during permeation reflects the interaction between Ca²⁺ ions and the residues. Figure 1a shows that Ca²⁺ remained longer at E90 and E83 and shorter at E101 and E97, indicating a stronger interaction between Ca²⁺ and E90 and E83 during the permeation process. It is worth noting that when passing through selective channel proteins, such as KcsA, NaV, and CaV [26,27,28], the ions’ hydration shell was partially or completely removed. In contrast, non-selective ion channels allow ions to pass through without dehydration. Therefore, the hydration state is a crucial factor in describing ion permeation. To demonstrate the change in the hydration state of Ca²⁺, the variations in the coordination number of Ca²⁺ (including the coordination of Ca²⁺ to oxygen atoms in H₂O and residues) were counted, as shown in Figure 1b. At the beginning of the permeation process, Ca²⁺ was coordinated with an average total of 7.2 oxygen atoms, with the majority of coordinated oxygen atoms originating from H₂O molecules. Upon entering the channel for 4 ns and reaching E90 of the CG, hydrated Ca²⁺ released 1–2 H₂O molecules to form a coordination with the carboxyl oxygen. After an additional 10 ns, Ca²⁺ approached E83, and the H₂O molecule bound to Ca²⁺ and was restored to seven. When Ca²⁺ reaches E83 in the intracellular gate (ICG), the carboxylated oxygen of the glutamate replaces four H₂O molecules bound by Ca²⁺ due to the proximity of E83 to E82. At this point, both the H₂O and the carboxylate are coordinated with approximately four oxygen atoms to Ca²⁺. Finally, Ca²⁺ exits the channel to enter the cell. The calculation of Ca²⁺ binding energies to the carboxyl groups at CG (listed in Table S1) indicates that the substitution of water molecules by carboxyl oxygen at these sites is mainly due to the strong electrostatic attraction between the ion and the carboxyl group. This suggests that partial dehydration occurs only near the acidic amino acids during Ca²⁺ permeation. In contrast, hydrated Na⁺ was dehydrated to a much lesser extent as it permeated into the channel (as seen in Figure S4). According to the literature [29], KcsA, which exhibits high ion selectivity, tends to pass through the channel as a fully dehydrated ion. The weaker ion selectivity of CatCh is likely due to the change in hydration number, which prevents the strong dehydration-based selectivity mechanism from being utilized.

Figure 2 displays the one-dimensional potential mean force (PMF) of Ca²⁺ permeating into CatCh-P520. The PMF shows an energy barrier of 5 kcal/mol at the CG in the channel for Ca²⁺. Additionally, there are two relatively weak energy barriers (approximately 1 kcal/mol) before and after the CG, corresponding to E97 and E83 in the channel, respectively. The free energy barriers shown in the PMF correspond well with the binding sites observed in the simulations, suggesting that the energy barriers at these positions are the result of a significant interaction between several glutamic acids and Ca²⁺. These results also suggest that in CatCh, the Ca²⁺ conductance pathway is coupled to the position of GLU and ASP in the channel.

3.2. Effects of Amino Acids at CG

The figures show that Ca²⁺ interacts significantly with the channels in the CG and ICG, which may act as potential ‘selective filters’ for Ca²⁺. The residence time and PMF of Ca²⁺ permeation support this. It is important to note that the conformational changes in the amino acid residues E90, D253, and E83 play a key role in Ca²⁺ permeation.

As a divalent cation, Ca²⁺ strongly interacts with deprotonated acidic amino acids in the channel. Further analysis of the Ca²⁺ trajectories was conducted. The carboxyl positions of E83, E90, and D253 were used as a reference to perform the K-means clustering analysis of the trajectories with Ca²⁺ in the vicinity of CG. The clustering results are presented in

Table 1. Clustering results and corresponding energy decomposition.

Cluster	Percentage	Binding Free Energy (kcal/mol)
		E90	D253	E83
1	53.9%	−12.21	−16.65	−0.41
2	30.1%	−17.05	−1.29	−0.23
3	12.5%	−1.78	−0.40	−18.15

, and the top three representative conformations are shown in Figure 3. The results of the clustering analysis indicate that when Ca²⁺ is in close proximity to CG, the nearby residues predominantly adopt conformations corresponding to clustering results 1 and 2. In these states, Ca²⁺’s passage through CG was not observed. In clustering result 3’s state, the Ca²⁺ permeation at CG was observed, and clustering result 3 accounted for 12.5% of the total. The clustering results show that the percentage of structures capable of Ca²⁺ permeation is approximately 12.5%, indicating that even in the Ca²⁺-enhanced ChR2 variant of CatCh, Ca²⁺ permeation is low, which is similar to the results observed experimentally by Kleinlogel et al. [2].

The binding free energies of E83, E90, and D253 to Ca²⁺ were calculated in each of the three clusters (see Table 1 and Table S1) to represent the stabilizing effect of each of the three amino acid residues on Ca²⁺. In the first two clusters, E90 exhibited a high binding free energy for Ca²⁺ (−12.21 and −17.05 kcal/mol, respectively). Compared to cluster 3, D253 in cluster 1 had a binding free energy of −16.65 kcal/mol for Ca²⁺. This largely limits further Ca²⁺ penetration into CG. In cluster 2, D253 forms a hydrogen bond with water near R120. The effect of D253 on Ca²⁺ is weaker (−1.29 kcal/mol), but the attraction to Ca²⁺ is even weaker (only −0.23 kcal/mol) due to the carboxyl group of E83 not facing CG. However, in cluster 3, the carboxyl conformation of E83 towards CG allows Ca²⁺ to permeate more readily towards the intracellular side, resulting in the easier passage of Ca²⁺ through the channel.

3.3. Comparison of Ca²⁺ and Na⁺ Permeation Patterns

CatCh has a higher Ca²⁺ conductivity than ChR2, but like ChR2, its primary conductive ion is still Na⁺ [1,2]. To understand why CatCh’s Ca²⁺ conductivity is lower than that of Na⁺, an analysis of Na⁺ and Ca²⁺’s permeation is necessary. In the charged-field permeation simulation, we replaced the 0.1 mol/L CaCl₂ solution with a 0.15 mol/L NaCl solution and a mixed NaCl-CaCl₂ solution, respectively. We kept all other conditions unchanged to observe the process of corresponding ions.

Figure S3 illustrates that the F102 residue, situated at extracellular cavity 1 (EC1), blocks ion entry into the channel. The hydrophobic benzene ring, along with its spatial site blocking, prevents ions from entering the channel. Ca²⁺ primarily enters the channel through E101 on TM2 during permeation and then passes through E97 to the CG to interact with residues such as E90 at the CG. Compared to Ca²⁺, Na⁺ enters the channel more easily, and there are more pathways for Na⁺ to enter from the extracellular loop 1 of the channel. The Na⁺ passage was observed on both sides of F102 (Figure S3). The permeation mechanism of Na⁺ in the channel was simpler, and it remained in the central cavity for a shorter period of time than Ca²⁺. Na⁺ does not interact complexly with carboxylated residues in the central cavity. This is also evident from the change in the oxygen coordination of Na⁺ (as seen in Figure S4). The coordination number of Na⁺ for oxygen remains constant during permeation, and only two H₂O coordinations are removed, even at E83 and E82. This is one reason why Na⁺ conductance is generally greater than that of Ca²⁺ in ChRs.

Ca²⁺ blocks the permeation of other ions. When Ca²⁺ is located at CG in the channel, its two positive charges prevent other Ca²⁺ and Na⁺ ions from approaching CG before they pass through CG. However, the subsequent arrival of Ca²⁺ and Na⁺ ions may also promote further Ca²⁺ penetration at CG, resulting in successive ion penetration, similar to the ‘knock-on’ mechanism [5]. In contrast, Na⁺ can pass more densely through the channel.

4. Discussion

Due to the low Ca²⁺ conductance of most wild-type ChRs, we chose CatCh, a classical ChR2 Ca²⁺-enhancing mutant, to carry out the study on the permeation properties of Ca²⁺ in ChRs. And the structure of CatCh-P520 was obtained by mutation, and MD experiments using ChR2-P520 and ion-permeation simulations of CatCh-P520 were performed.

Previous discussions have established the structural differences between CatCh-P520 and ChR2-P520 (Figure S3). The permeation properties of ions are directly affected by the radius of their hydrated ions. The hydrated Ca²⁺ ion has a radius of approximately 3.8 Å, as the thickness of its first solvation shell is around 2.4 Å from the Ca²⁺, and the radius of a water molecule is usually considered to be 1.4 Å [30]. The minimum channel radius in the open state of ChR2 is approximately 3 Å [1]. The radius at the CG (E90 to D253 carboxylate distance) of CatCh-P520 obtained from the simulations in this paper is approximately 3.4 Å. Based on the radius, it appears challenging for hydrated Ca²⁺ to pass directly through the channels of ChR2 and CatCh-P520. This difference in radius may be a significant factor contributing to the greater Ca²⁺ permeability of CatCh compared to ChR2. This text suggests that hydrated Ca²⁺ ions must lose some water molecules at CG and ICG to reduce their radius and pass through the channel.

CatCh-P520 exhibits differences in the permeation processes of Ca²⁺ and Na⁺ due to their different ionic radii and carried charges. Since Ca²⁺ carries two positive charges, its permeation in CatCh-P520 is interconnected with GLU and ASP in the channel, i.e., its permeation path in CatCh-P520 follows the sequence E101, E97, E90/D253, and E83. This is similar to Ca²⁺ permeation in other channel proteins, such as in RyR [30], as Ca²⁺ passes near the selectivity filter along the sequence D4903, E4900, and D4899. Comparatively, Na⁺ penetrates more easily in CatCh-P520 due to its smaller size and smaller charge-carrying capacity and has shorter interaction time with GLU and ASP in the channel. The permeation characteristics mentioned are in line with the stronger Na⁺ conductance demonstrated by CatCh in the experiments.

Through the analysis of the Ca²⁺ permeation process in CatCh-P520, it was discovered that Ca²⁺ interacts more with residues in the channel near CG and ICG (refer to Figure 2). And after energy analysis, we focused the Ca²⁺ interaction near the CG and ICG on three conformations of residues E83, E90, and D253 (Figure 3). Within the channel, an ‘EED triad’ consisting of three carboxylic residues (E83, E90, and D253) facilitated Ca²⁺ binding. The reasons why Ca²⁺ penetration is more likely to occur in clustering result 3 are summarized as follows: D253 is ‘non-blocking’, E90 is ‘transferring’, and E83 is ‘attracting’. The clustering analysis reveals that when calcium ions are near CG in the channel, the ‘EED triad’ can adopt three conformations, as depicted in Figure 4. Clusters 1 and 2 are the dominant clusters prevalent. However, the strong force of D253 and E90 in cluster 1 creates a bilateral tethering structure with Ca²⁺, which restricts the penetration of Ca²⁺. In contrast, Ca²⁺ in cluster 2 does not experience the tethering effect of D253; at this point, D253 mostly forms a hydrogen bond with R120 or water in the vicinity of R120, which is in a different position to the channel. But it is difficult for Ca²⁺ to pass through the channel due to the lack of attraction to Ca²⁺ by E83. Ca²⁺ can only pass through when the residues in the channel are in the state under cluster 3, which satisfies the ‘non-blocking’ of D253, the ‘attracting’ of E83, and the ‘transferring’ of E90.

This text summarizes the structural features of CatCh-P520 and the ‘EED triad’ regulatory properties of Ca²⁺ in CatCh-P520 through simulations. Based on our findings, we propose the following ideas for enhancing the selectivity of ChRs for Ca²⁺: (i) Increase the radius of the open state of the channel to reduce the spatial site resistance of hydrated Ca²⁺ in the channel due to the large radius of hydrated Ca²⁺. For instance, the conductance of Ca²⁺ can be increased by the L132C mutation of ChR2, which increases the radius of the channel. (ii) Additionally, the enrichment of Ca²⁺ by carboxylate-bearing residues on the outside of EC1 (E41, E101) plays a crucial role in the permeation of Ca²⁺ due to the low concentration of Ca²⁺ in the cytosol. Increasing the number of carboxylate-bearing residues in the extracellular space would facilitate Ca²⁺ entry into the channel. (iii) The conformation of several residues, such as E83 of the ICG, influences Ca²⁺ permeation at the CG. The attraction of Ca²⁺ by the GLU of the ICG is favorable in helping Ca²⁺ to cross the energy barrier of the CG. Therefore, selectively mutating residues in the region between E83 and E90 to acidic amino acids may be an important method to increase Ca²⁺ permeability, similar to mimicking the EEEE motifs in other Ca²⁺-selective channel proteins [28,31,32].

5. Conclusions

In summary, our conclusions indicate that the reason for the increase in Ca²⁺ conductance in CatCh is due to the L132C mutation, which increases the channel’s radius. This makes the diameter of the partially dehydrated Ca²⁺ more compatible with the channel size, resulting in the higher permeability properties of Ca²⁺ in CatCh. Additionally, our findings suggest the reason for the generally lower conductance of Ca²⁺ in the various variants of ChR2, and we indicate the permeable properties of Ca²⁺ in CatCh. As the CatCh mutant only changes one amino acid residue far from the channel’s portion, the conclusions drawn from the simulations in CatCh can also be applied, to some extent, to ChR2 itself. Understanding these detailed mechanisms can aid in the design of subsequent Ca²⁺-selective enhancement variants of ChRs.