Structural and Dynamic Insights into SARS-CoV-2 Spike-Protein-Montmorillonite Interactions.

The spike (S) protein of SARS-CoV-2 has been found to play a decisive role in the cell entry mechanism of the virus and has been the prime target of most vaccine development efforts. Although numerous vaccines are already in use and more than half of the world population has been fully vaccinated, the emergence of new variants of the virus poses a challenge to the existing vaccines. Hence, developing an effective drug therapy is a crucial step in ending the pandemic. Nanoparticles can play a crucial role as a drug or a drug carrier and help tackle the pandemic effectively. Here, we performed explicit all-atom molecular dynamics simulations to probe interactions between S protein and Montmorillonite (MMT) nano clay surface. We built two systems with different counterions (Na+ and Ca2+), namely Na-MMT and Ca-MMT, to investigate the effect of different ions on S protein-MMT interaction. Structural modification of S protein was observed in the presence of MMT surface, particularly the loss of helical content of S protein. We revealed that electrostatic and hydrophobic interactions synergistically govern the S protein-MMT interactions. However, hydrophobic interactions were more pronounced in the Na-MMT system than in Ca-MMT. We also revealed residues and glycans of S protein closely interacting with the MMT surface. Interestingly, N165 and N343, which we found to be closely interacting with MMT in our simulations, also have a critical role in cell entry and in thwarting the cell's immune response in recent studies. Overall, our work provides atomistic insights into S protein-MMT interaction and enriches our understanding of the nanoparticle-S protein interaction mechanism, which will help develop advanced therapeutic techniques in the future.

Introduction

The year 2020 was one of a remarkable era in human history, although not a memorable one. The coronavirus disease 2019 (COVID-19) pandemic posed an unprecedented global challenge for public health and the economy. The pandemic, since its emergence, has claimed millions of lives and pushed billions into economic and social jeopardy. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which is responsible for the spread of COVID-19, is highly contagious in comparison to its close relative SARS-CoV,[1] which was linked to an epidemic in the year 2002.[2]

The SARS-CoV-2’s spike (S) glycoprotein plays a crucial role in the viral invasion of the host cell. Human angiotensin-converting enzyme 2 (hACE2) is an enzyme present in the cells of the lungs, heart, and kidney, which helps regulate the blood pressure.[3,4] The hACE2 also acts as a host receptor for the SARS-CoV-2 virus, and the interaction between S glycoprotein and hACE2 is found to be the mechanism for the cell entry[5−7] of the virus. The receptor binding domain (RBD) in the ”up” conformation[7] among many of the functional domains of S protein is the one that binds to the hACE2. The critical role of S protein-hACE interactions in the viral hijacking makes it the primary target for many therapeutic and vaccine approaches.[5,8,9]

Nanoparticles (NPs) are revolutionizing almost all aspects of our lives, including health care.[10] NPs show great potential in the development of therapeutic materials and targeted drug delivery. The efficacy of NPs as an efficient drug delivery medium or any other biomedical application can be attributed to their high surface-to-volume ratio, which facilitates adsorption of biomolecules and a variety of other chemical species. Hence, owing to the excellent physicochemical properties offered by the NPs, they have attracted the attention of the biomedical research community for as long as they have been discovered. In fact, NP-biomolecular interactions are studied extensively, both experimentally[11−13] and computationally.[14−16] Zhang et al.[17] used a nanosized graphene oxide sheet as a carrier of anticancer drugs for targeted delivery to MCF-7 cells and human breast cancer cells. They found that the nanocarriers delivered the drug to the target with high specificity and reduced toxicity. Nano clays, particularly montmorillonite are also a class of NPs which are extensively studied for many potential applications such as drug delivery,[18−20] MMT–organic interaction[21] and interactions with virus[22] and protein.[23−25] MMT surface belongs to the smectite class of minerals, a 2:1 phyllosilicate; an octahedral alumina layer is sandwiched between two tetrahedral silica layers. The MMT surface also undergoes isomorphic substitution of Al3+ atoms present in the central alumina layer, with a lower charge species such as Mg2+, which creates a net negative charge on the surface. Hence, the MMT surface can attract the positively charged species in the environment. Hence, in its natural state the MMT’s chemical formula is (Na,Ca)0.33(Al,Mg)2(Si4O10)(OH)2·nH2O. Block et al.[22] studied the interaction between ϕ6 virus and MMT clay via transmission electron microscopy. They noted severe disassembly of virus and loss of envelope, which was attributed to the strong electrostatic and van der Waals forces between MMT and virus. Anderson et al. investigated the interaction between model protein Gb1 and different clay surfaces, including MMT, via MD simulations. They observed marginal influence of the presence of mineral surfaces on the protein structure, except for birnessite surface, which showed significant perturbation of protein’s coformation.[23] Köhler et al.[14] performed molecular dynamics (MD) simulations for studying fibrinogen (Fg)–(mica, graphite) interactions. They observed a weaker Fg–mica interaction than Fg–graphite. The Fg–mica interaction was driven by electrostatic effects between Fg and mica surface and the solvated counterions present on the mica surface. Kubaik et al.[15] noted in the MD simulations that the adsorption of lysozyme on silica was mainly driven by electrostatics and supported by weaker hydrophobic forces. Simulations also showed minor variation in lysozyme’s conformation due to adsorption, while experiments reveal the effects of pH on the adsorption. Lecot et al. built atomistic models of different silane monolayers functionalized on silica(SiO2) surfaces and extensively studied their surface properties.[26] Additionally, they studied adsorption of ACE2(present as ACE2–RBD complex)[27] and streptavidin(present as streptavidin–biotin complex) onto these silane monolayer surfaces via MD simulations. Lecot et al. reported that adsorption of ACE2 on silane monolayers leads to a substantial increase in the binding energy between ACE2 and RBD, but the conformation of ACE2 is not affected much; hence, the bioactivity of ACE2 remains conserved.[27] NPs are also contributing to tackling the COVID-19 pandemic in every sense possible. Silver NPs have been hypothesized to be used against SARS CoV-2,[28] as it is found to have antiviral activity in a previous study,[29] and hence can help in blocking the viral entry to the host cell. Computational investigation of the interaction between S protein and model surfaces (gold and silica)[30] have shown the exposure of RBD and have some interesting findings.

In this study, we performed explicit molecular dynamics (MD) simulations to investigate the interaction between the montmorillonite (MMT) clay model surface and SARS-CoV’s S protein. Specifically, the study aims to unravel the nature of S protein–MMT surface interaction and its effect on S protein’s structure. Also, the effect of two different counterions (Na+, Ca2+) on protein–surface interactions was investigated.

Methods

In total, three simulation systems were prepared: (1) S protein without any surface (control), (2) S protein-Na-MMT, (3) S protein-Ca-MMT. All simulations were performed with NAMD[31] simulation package. The equilibration runs were performed in the NPT ensemble, while the production runs were executed in the NVT ensemble. The production simulations for Na-MMT and Ca-MMT were run for 120 ns, while the control run was simulated for 60 ns. Since, only time-averaged properties were evaluated for control run, which were found to be well converged within the simulated time scale. The montmorillonite surface and S protein were modeled using CLAYFF[32] and CHARMM36[33] force-field, while TIP3P[34] model was used to model water molecule. Periodic boundary conditions were applied in all directions for all the systems. The simulation box length is kept such that the minimum distance between protein and the box’s edge in all directions is more than 10 Å. A time step of 1 fs was used to integrate the equations of motion. van der Waals interactions were cutoff at 12 Å with a switching distance of 10 Å. The long-range electrostatic interactions were evaluated using particle mesh ewald[35] (PME) summation. The bonds involving hydrogen atoms were constrained with SHAKE[36] algorithm.

Protein–Surface Setup

The fully glycosylated structure of S trimer obtained from CHARMM-GUI’s[37] COVID-19 archive was used as the starting structure for the simulations. The S trimer structure prepared in CHARMM-GUI is a head-only model with 1146 residues and was modeled using the Cryo-EM structure[38] (PDB ID: 6VXX). The titratable residues of protein were assigned correct protonation states assuming a pH of 7.0. The protein was then solvated in a water box of size 185 Å × 91.5 Å × 199 Å with SOLVATE plugin in visual molecular dynamics[39] (VMD). The glycosylated S protein trimer contains a charge of −15e; hence an equal number of counterions (Na+) were added to the solvation box, which totaled the system to 670461 atoms. The solvated and ionized system is then subjected to the minimization of 10,000 steps with the conjugate gradient method. The step was followed by a run of 200 ps, in which the temperature is increased slowly in steps from 0 to 310 K in an NVT ensemble. The backbone atoms of protein were frozen, while the side chain and glycan atoms were harmonically constrained with force constant of 0.5 kcal/mol during the above two steps. In the next step, all constraints were released, further the system was minimized for 5000 steps and was thermalized incrementally from 0 to 310 K in an NPT ensemble for 5 ns. The system is then submitted for the production run in an NVT ensemble as a control system.

The atomistic model of montmorillonite (MMT) surface was designed with Atomistic Topology Operations in MATLAB’s atom[40] package. A unit cell with a basic formula as Al2Si4O10(OH)2 is replicated in X and Y directions. Then about two-thirds of the central aluminum atoms were replaced with Mg2+ atoms in the obtained crystal lattice to mimic the isomorphic substitution of an MMT clay surface(Figure a,b). The constructed MMT surface has an XY dimension of 196 Å × 196 Å and a charge of −628e.

Figure 1

(a) Top view and (b) side view of atomistic model of MMT surface represented in licorice representation. Yellow, red, pink, and cyan colors represent silicon, oxygen, aluminum and magnesium, respectively. (c) Atomic model of simulation system with S protein and MMT surface in new cartoon and quick surf representation, respectively, while RBD of S protein is highlighted in Surf drawing.

To create S protein-MMT systems, the coordinates of equilibrated S protein from the control equilibration run (described in the first part of this section) were taken as the initial structure of S protein. The protein was then solvated in a water box of size 196 Å × 196 Å × 188 Å. The system was then ionized with counterions, taking the protein and MMT charges into account. In this step, two systems were created, namely Ca-MMT and Na-MMT, with Ca2+ and Na+ as counterion species, respectively. The total number of atoms in Ca-MMT and Na-MMT were 736 518 and 735 876, respectively. The boxes were then placed on top of the MMT surface (Figure ). The protein is positioned so that the RBD is facing the MMT surface (Figure ), and the minimum protein to surface distance is 8 Å. Supporting Information (SI) Figure S9 shows the electric dipole moment vector for RBDs of all three chains in both(Na-MMT, Ca-MMT) the systems at the end of the simulation. The electric dipole moments are in agreement with a similar work dealing with the adsorption of a negatively charged protein on a negatively charged surface.[41] The merged system is then minimized for 2500 steps, followed by an NVT run for 200 ps with protein and surface atoms fixed. Further, the system is subjected to an unconstrained minimization of 2500 steps and stepwise temperature increment in an NPT ensemble for 2 ns. Further, the two systems (Ca-MMT, Na-MMT) were submitted for the production runs.

Analysis

All the analyses was performed using tools present in VMD and NAMD and in-house Tcl scripts. Root-mean-square fluctuation(RMSF) for residues of S protein was calculated to quantify the protein’s conformational changes using the ”measure rmsf” command in the VMD. The RMSF was calculated, using the initial structure as the reference and taking into account all the frames of the trajectory. The interaction energy between S protein and MMT surface is calculated to understand the nature of interactions using the ”pair-interaction” utility in NAMD. Secondary structure calculation on S protein was performed using STRIDE algorithm[42] in VMD. Contact analysis between MMT-ions, MMT-residues, and MMT glycans was performed using the ”measure contact” utility in VMD. The distance cutoff used for calculating the contacts was 5 Å for all the cases. In the case of MMT-ions, only those ions were considered where the ion comes in contact with MMT for at least 40% frames of the whole trajectory. Electrostatic potential maps were calculated using PMEpot[43] module in VMD.

Results and Discussion

Effect on Residue Flexibility

Root mean-square fluctuation (RMSF) for each chain of the S trimer (chains A, B, and C) were calculated for all the three systems (control, Ca-MMT, Na-MMT). Figure shows the distribution plots for RMSF values of amino acid residues of all the three chains A, B, and C. SI Figure S1 shows the deviation of RMSF the deviation of RMSF values of residues in MMT systems from the values in the control system. The deviation is represented by ΔRMSF = (RMSF)control – (RMSF)(Na–MMT/Ca–MMT). An offset of 5 and −5 was added to ΔRMSF values of Ca-MMT and Na-MMT, respectively. The offset is added to separate the curves of Ca-MMT and Na-MMT and bring more clarity to the visualization. Hence, the ΔRMSF values above the baseline indicate a “decrease” in the RMSF value of the residue in comparison to the control system and vice versa for the value below the baseline. In general, it can be observed that due to the presence of the surface, the flexibility of the residues has reduced, indicated by the decrease in RMSF values of many residues. As can be observed from the Figure a, control shows highest number of residues in higher RMSF value range, among all the three systems. The distribution for chain A (Figure a) shows that the curve for control system is broader and spread across higher RMSF values, while the distributions for Na-MMT and Ca-MMT are narrower and are shifted toward smaller values of RMSF. Also, the control distribution has a peak at around 1.5 Å while the other two systems have a peak at lower RMSF values. This implies there are some residues that have an increase in RMSF value due to the presence of the surface. This can also be observed from SI Figure S1a, the N terminal residue and residues in the C terminal domain have high RMSF values for the control system. Chain B seems to be the most affected by the presence of MMT surface, as the distribution (Figure b) shows a drastic increase in the number of residues with smaller RMSF values for Na-MMT and Ca-MMT. The Ca-MMT system has the highest peak for the lowest RMSF value at around 1.2 Å. The control’s distribution for chain B is again broader and lacks any significant peak compared to the other two systems. Notably, the residues from 250 to 600 have got affected the most (SI Figure S1b), which is the RBD region (331–528) of chain B. Chain C shows moderate changes in the presence of MMT surface(Figure c), only a few residues (Pro 251, Gly 252, Asp 253, Arg 683, Ala 684, and Arg 682) showed significant changes (SI Figure S1c).

Figure 2

Distribution of RMSF values for (a) chain A (b), chain B, and (c) chain C in control, Na-MMT, and Ca-MMT systems.

Overall, the RMSF analysis shows that the MMT surface considerably affects the flexibility of all the three chains of the S protein, particularly chain B. It indicates a strong interaction between the MMT surface and the S protein. Moreover, the RBD region of chain B indicated the strongest interaction with surface among all the chains, which is reflected by reduction in RMSF values of most residues in the presence of MMT surface (both Ca-MMT and Na-MMT). The Ca-MMT system has a stronger effect on residue flexibility than Na-MMT. Further, we characterize the nature of interaction and effects on other structural features of the protein.

Influence on Secondary Structure of S Protein

Protein’s structure is intimately connected to its function. Hence, changes in secondary structure of a protein can give critical insights about its interaction with the surrounding. Therefore, we evaluated the changes in the secondary structure of S protein due to its interaction with the MMT surface. Figure shows the average probability of secondary structure contents (coil, turns, beta-sheet, and helix) for all the three systems. It can be observed that there is a heavy loss of helical content in the presence of the MMT surface as the average probability for helix reduced from 21% in the control system to 6.4% and 6.3% in Ca-MMT and Na-MMT system, respectively. Conversely, the average probability for turns increased in Na-MMT and Ca-MMT to 39.7% for both the systems from 23.7% in the control system. The turns must have increased at the cost of helical content, that is, the residues which lost the helical content must be the ones that got converted to turns. All other structure contents (coils and beta sheets) showed minor variations. As shown in SI Figure S2, changes in the secondary structure of all the three chains were also similar to the changes observed in overall protein structure. Another important point to be noticed here is that the changes in secondary structure that occurred are irrespective of the counterions, that is, changes are almost identical in both Na-MMT and Ca-MMT. The loss of helical content of the S protein must be due to strong, attractive interaction with the MMT surface. A similar observation was also made by Kubiak and Mulheran,[44] where they found loss of helical content of lysozyme protein in the presence of charged Mica surface.

Figure 3

Secondary structure probability of S protein in control, Na-MMT and Ca-MMT systems.

Nature of Protein–Surface Interaction and Influence of Counterions

To understand the nature of the interaction between the spike protein and MMT surface, time evolution of interaction energy is calculated between all three chains and MMT surface for both Na-MMT and Ca-MMT (Figure ). The interaction analysis revealed electrostatics to be the dominant mode of interaction between S protein and MMT surface, which is in line with previous studies.[23,44] However, as found in the subsequent section on ”contact analysis”, the protein can interact with the surface via hydrophobic contacts. As can be observed from Figure a,b, all chains except chain A in the Ca-MMT system have negative interaction energy indicating attraction with the MMT surface. The average interaction energy in Ca-MMT for chains A, B, and C was 773, −1483, and −157 kJ/mol, respectively, and for Na-MMT, it is −269, −1047, and −241 kJ/mol for A, B, and C, respectively. Although the average interaction energies show attractive interaction between chains and surface, only chain B shows consistently strong attraction and is more pronounced for the Ca-MMT system. Other chains show relatively weaker interaction with the surface. The highest interaction energy between chain B and surface explains the maximum structural deviation for chain B in RMSF analysis. Since chain B is the most flexible among all three chains, which lets it interact with the MMT surface more readily, we observe the strongest interaction between chain B and the MMT surface.

Figure 4

Time evolution of interation energy between chains of S protein and MMT surface in (a) Ca-MMT and (b) Na-MMT system. (c) Distribution of contact frequency of ions (Na+, Ca2+) on MMT surface in Na-MMT and Ca-MMT systems.

MMT surface’s high negative charge (−628e) is balanced by the presence of positively charged ions (Na+, Ca2+). The presence of ionic species is known to influence protein–surface interaction.[23,41,45] To quantify the extent of ion-surface interaction, we evaluated the contact frequency of ions with the surface. Figure c shows the number of ions and the corresponding percentage of contact frequency with the MMT surface for Na-MMT and Ca-MMT systems. It can be observed from Figure c that MMT surface in the Ca-MMT system has a very strong interaction with Ca2+ ions, as indicated by a very high number of Ca2+ ions with 100% contact frequency. Na+ ions in the Na-MMT system show weaker interaction with MMT surface, as can be observed from significantly less number of Na+ ions in the high contact frequency range. However, there are a substantial number of Na+ ions in the 40–60% range of contact frequency, but total number of Na+ ions in contact with MMT surface are still very less in comparison to the number of Ca2+ ions. The binding of Ca2+ ions firmly to the MMT surface than Na+ ions was also found in a previous study.[23]SI Figure S5 shows electrostatic maps of systems. It can be observed that the Ca-MMT system (SI Figure S5b) shows higher positive potential density at the MMT surface, indicated by the blue color, than in the case of the Na-MMT system (SI Figure S5a). However, it can be observed that the region around protein in potential maps is white for both the systems and the color is much denser in the case of Na-MMT than in the Ca-MMT system. The denser white color in Na-MMT could be because the Na+ ions readily diffuse from the surface to the bulk and interact with negatively charged polar residues of the protein, creating a net neutral environment around the protein. However, divalent Ca2+ ions hold on firmly to the MMT surface and are not readily available for protein to directly interact with therefore, a lighter white color is observed in the potential map of the Ca-MMT system. This is in agreement with Anderson et al.’s study,[23] where they proposed the surrounding of Gb1 protein by Na+ ions, screening the protein–surface interactions and consequently preserving the protein’s conformation in the Na-MMT system. While for the Ca-MMT system, they found that the Gb1 protein’s conformation was disturbed since ca2+ ions were tightly bound to the MMT surface, and the protein is more exposed than in Na-MMT. Hence, to examine the effect of ions on S protein’s conformation, we compared the root-mean-square deviation(RMSD)(SI Figure S3) of all three chains of S protein in Na-MMT and Ca-MMT systems. However, as seen in SI Figure S3, we do not find any significant difference in the time evolution of the RMSD between the Na-MMT and Ca-MMT. Although the RMSD for chain A (SI Figure S3a) showed slight differences, they are not remarkable enough. To understand which chain is most affected due to the presence of the MMT surface, we compared the RMSD evolution of each chain within a system for Na-MMT and Ca-MMT. We can observe from Figure that chain B has the highest RMSD, and the RMSD curve for chain B is unstable compared to other chains in both systems. The RMSD analysis indicates that chain B conformation is the most affected in the presence of MMT surface. Moreover, structural representation colored with electrostatic potential shows the distribution of potential for the systems (SI Figure S6). It can be observed that the surface in the Ca-MMT system (SI Figure S6b) shows more part of it covered with blue color than Na-MMT system (S6a), which indicate positive potential and must be because of stronger adsorption of Ca2+ ions than Na+ ions on the MMT surface. Also, a difference can be observed in the distribution of potential on protein’s structure between both systems. The colors of some regions of protein’s structure in the Ca-MMT system (SI Figure S6b) are bright red, while the same region in the Na-MMT system (SI Figure S6a) are paler in color. This is again due to Na+ encompassing the protein, while in the Ca-MMT system, most of Ca2+ ions are strongly adsorbed on the MMT surface.

Interacting Residues and Glycans

Contact analysis is performed to identify the critical residues and glycans of S protein closely involved in the interaction with the MMT surface. We searched for all the pairs of atoms, with one atom from S and the other from MMT, which are within 5 Å of each other.

SI Figure S7 shows the time evolution of the number of contacts between protein atoms and MMT surface atoms. It can be observed from SI Figure S7a that protein in the Na-MMT system shows much higher and sustained contact with MMT surface than in Ca-MMT. SI Figure S7b,c shows contact number for each protein chains in Na-MMT and Ca-MMT respectively. It can be noted that the maximum contribution for protein-MMT surface contact is from chain B in the case of Na-MMT. At the same time, in the case of Ca-MMT also, chain B shows the maximum number of contacts, but chain A also shows a substantial contribution. We calculated the residue-wise number of contacts to know the type of residues that play a dominant role in the close interaction between S protein and MMT surface and consequently the nature of interactions. In our simulations, we found residues from the N terminal region and some residues from RBD of S protein in close contact with the MMT surface. Figure a,b shows the number of contacts of interacting residues for chain A and chain B, respectively. It can be noted from Figure a that for chain A, VAL445, which belongs to the RBD region of S protein, has the highest number of contacts in both the systems (Na-MMT, Ca-MMT). The GLY446 is another residue that shows contact with the MMT surface in chain A for both systems. LYS444 residue of chain A in the Na-MMT system also shows a significant number of contacts. The interacting residues of chain B show much higher contacting frequency (almost ten times) (Figure b) than Chain A. Moreover, the number of interacting residues in chain B is more than in chain A. In the case of Na-MMT, all interacting residues of chain B belong to the N terminal. Residues with the highest number of contacts in chain B for the Na-MMT system are MET1, PHE2, PHE4, and LEU5, while VAL3 and LEU7 also show a significant number of contacts. However, chain B in Ca-MMT has only MET1 showing noticeable contact; however, VAL 445 also shows some degree of contact toward the surface(SI Table S1b). Figure c,d show the representative snapshot of Na-MMT and Ca-MMT system’s trajectory. SI Table S1 shows the list of all interacting residues for Na-MMT (SI Table S1a) and Ca-MMT (SI Table S1b). It can be observed from the contact analysis of S protein and MMT surface that most of the interacting residues are hydrophobic, such as methionine, valine, phenylalanine, and leucine. It implies the presence of hydrophobic interaction between these residues and the MMT surface. Moreover, the presence of phenylalanine, which contains an aromatic side chain, indicates a possibility of cation-π interaction with ions (Na+, Ca2+) adsorbed on top of the MMT surface. Also, S protein residues show much higher contact in the Na-MMT system than Ca-MMT system. The reason must be less interference of surface cations in the case of Na-MMT since, as discussed in the earlier section, the Ca2+ ions are strongly adsorbed and have broader coverage of MMT surface than Na+ in the Na-MMT system. Hence, residues in the Na-MMT system have more surface available for interaction.

Figure 5

Number of contacts between key interacting residues and MMT surface in Na-MMT and Ca-MMT system for (a) chain A and (b) chain B. Representative snapshot of interacting residues and MMT surface in (c) Na-MMT and (d) Ca-MMT system. The chains and interacting residues are shown in new cartoon and licorice and MMT surface is shown in quick surf representation.

The S protein is extensively glycosylated, and the glycans are known to play a crucial role in viral cell entry.[46−48] Since S protein is densely covered with glycans, it can play an important role in the interaction between S protein and the MMT surface. Hence, we calculated the number of contacts and identified glycans with high contact frequency with the MMT surface. SI Figure S8a shows the time evolution of the total number of contacts between glycans and MMT surface for Ca-MMT and Na-MMT. The time evolution of the number of contacts for glycans shows similar behavior to that of protein chains, as the glycans in Na-MMT show higher and more sustained contacts than in Ca-MMT. We also examined the chain-wise time evolution of contacts of glycans. It can be observed that glycans attached to chain B have the maximum number of contacts in both Na-MMT (SI Figure S8b) and Ca-MMT (SI Figure S8c), as observed similarly in protein residue-MMT contacts. Next, to identify the key interacting glycans, we computed the number of contacts for each glycan. Figure a,b shows the number of contacts of the key interacting glycans for chain B and chain C, respectively. We found N17 from N terminal domain (NTD), N165, and N343 from RBD in the case of chain B to be highly interactive with MMT surface (Figure a) for both the systems. However, N17 in Na-MMT shows much higher contacts than in Ca-MMT, while N165 exhibits more contacts in Na-MMT than in Ca-MMT. Similarly, N149 and N165 of chain C show high contacting affinity for the Ca-MMT system, while chain C in Na-MMT also have these two glycans coming in contact with the MMT surface, but the frequency is much lesser (SI Table S2a). However, chain A has only one glycan having contact with the surface per system (SI Table S2), with only N165 in Na-MMT showing high contact frequency (SI Table S2a). In particular, N165 is a key interacting glycan across all the chains in both systems. N165 was also found to play an important role in the interaction of S protein with ACE2 and antibodies in previous studies.[48,49] Also, N343, which is found to be a key interacting glycan, was found to play a critical role in the binding of various antibodies to S protein.[49]Figure c,d show representative snapshots of the interacting glycans along with the MMT surface for the Na-MMT and Ca-MMT system, respectively.

Figure 6

Number of contacts between key interacting glycans and MMT surface in Na-MMT and Ca-MMT system for (a) chain B and (b) chain C. Representative snapshot of interacting glycans and MMT surface in (c) Na-MMT and (d) Ca-MMT system. The S protein, interacting glycans and MMT surface are shown in new cartoon, licorice and quick surf representation.

Conclusions

In this work, via MD simulations, we investigated fully glycosylated S protein-MMT surface interaction and the effect of two different counterions, Na+ and Ca2+, on them. The structural analysis revealed a reduction in flexibility of many residues and heavy loss of helical content in the S protein in the presence of the MMT surface. Chain B of all three chains of S protein was found to have the strongest interaction with the MMT surface. The nature of the interaction between S protein and MMT surface was found to be predominantly electrostatic. The interaction energy analysis revealed stronger electrostatic interaction in the case of the Ca-MMT system. However, we also found many neutral hydrophobic residues, such as valine, methionine, leucine, and phenylalanine, closely interacting with the MMT surface, indicating the presence of hydrophobic interaction between S protein and MMT surface. These hydrophobic interactions were found to be more pronounced in the Na-MMT system. Also, The Ca2+ ions showed stronger adsorption than Na+ on the MMT surface, which impacts the mechanism of protein–surface interaction in Na-MMT and Ca-MMT. Additionally, we found many N-glycans closely interacting with the MMT surface. Particularly, N165 and N343 among the interacting residues were found to have a prominent role in S protein-antibody and S protein-ACE2 interactions in recent studies. Overall, our work provides detailed atomistic information about the nature of S protein-MMT interaction, which can lead to the design of better therapeutic agents in the future.